Transstadial Transmission of Francisella tularensis holarctica in Mosquitoes, Sweden

In Sweden, human cases of tularemia caused by Francisella tularensis holarctica are assumed to be transmitted by mosquitoes, but how mosquito vectors acquire and transmit the bacterium is not clear. To determine how transmission of this bacterium occurs, mosquito larvae were collected in an area where tularemia is endemic, brought to the laboratory, and reared to adults in their original pond water. Screening of adult mosquitoes by real-time PCR demonstrated F. tularensis lpnA sequences in 14 of the 48 mosquito pools tested; lpnA sequences were demonstrated in 6 of 9 identified mosquito species. Further analysis confirmed the presence of F. tularensis holarctica–specific 30-bp deletion region sequences (FtM19inDel) in water from breeding containers and in 3 mosquito species (Aedes sticticus, Ae. vexans, and Ae. punctor) known to take blood from humans. Our results suggest that the mosquitoes that transmit F. tularensis holarctica during tularemia outbreaks acquire the bacterium already as larvae.

O utbreaks of tularemia are caused by the bacterium Francisella tularensis holarctica throughout the Northern Hemisphere and by F. tularensis tularensis in North America only. Routes of infection include transmission from blood-sucking arthropods and through contact with infected dead or live animals, as well as from aerosols, dust, and water (1). Two primary disease manifestations, ulceroglandular and glandular tularemia, are associated with vector-borne transmission of the bacterium (1). Traditionally, mosquitoes are considered the primary vectors of F. tularensis holarctica to humans in Russia and Scandinavia (2)(3)(4)5). Moreover, mosquito-borne transmission of tularemia may be becoming more common in central Europe; evidence shows that this infection has reemerged during the past decade (6,7).
The ulceroglandular form of tularemia is by far the most common in Sweden; most human cases occur in late summer and early fall and are assumed to be transmitted by mosquitoes (4,8). A total of 5,754 human cases of tularemia were reported during 1931-1993, and the incidence of infection varies greatly among these years, ranging from a few cases in some years to >2,700 cases during 1967 (8). In the Örebro area of central Sweden, widespread mosquitoassociated tularemia outbreaks fi rst occurred during 2000 and 2003 (4,5), after which human cases have continued to occur in this new area where tularemia is endemic (www. smi.se/statistik/harpest). However, how vector mosquitoes acquire the bacterium is still not clear.
The demonstrated ability of F. tularensis holarctica strains to survive in association with protozoa indicates that ubiquitous aquatic protozoa might be an important environmental reservoir for the bacterium (9-11). Moreover, mosquito larvae, mainly the species A. sticticus and other fl oodwater mosquitoes, exert a predatory effect on aquatic protozoan populations (12). These factors indicate that mosquito larvae may be exposed to F. tularensis in their natural aquatic environment. We investigated the natural occurrence of F. tularensis in mosquitoes hatched from larvae collected in an area where tularemia was endemic. Because of unknown mechanisms, the bacterium F. tularensis is extremely diffi cult to isolate directly from environmental samples. Thus, our study focuses entirely on molecular techniques.

Sample Collection
Mosquito larvae were sampled on August 28, 2008, in Örebro, an area where tularemia is endemic. Nine human tularemia cases (3.24/100,000 persons) were reported from Örebro County in June-September 2008 (www. smi.se/statistik/harpest). Two sampling locations were selected, Ormesta (WGS84; 59°16′12′′N, 15°16′48′′E) and Vattenparken (WGS84; 59°16′55′′N, 15°15′00′′E), on the basis of a geographic distribution study of human tularemia cases in the area (13). Both locations are situated at Lake Hjälmaren near the city of Örebro and are characterized by lush vegetation of reed belts and deciduous trees and bushes. Using a standard dipper, we collected mosquito larvae from shallow temporary water bodies in the transition zone between reed and willow bush habitats (Ormesta 1), in the deciduous forest (Ormesta 2), and in a ditch covered by bush and grass (Vattenparken).
Mosquito larvae from each water body were reared to adults in their original pond water (Ormesta 1, containers A and B; Ormesta 2, containers C, D, E, F, and G; and Vattenparken, container H). At the start of this study, timezero samples of the original water from each container were collected and stored at -20°C. Emerging adult mosquitoes were collected by a mechanical aspirator, killed by freezing, and stored at -80ºC until species was identifi ed. During identifi cation, adult mosquitoes were kept cold on a chill table, illuminated by a cold light lamp, and identifi ed to species based on morphologic features. Identifi ed mosquitoes were sorted by area, species, and sex in pools of 1-50 specimens and returned to the -80ºC freezer.

DNA Extraction from Mosquitoes and Water Samples
For DNA extraction, 10 μL of 2.8 M NH 4 OH solution and 450 mg each of 1.0-mm and 0.1-mm silica beads were added to each pooled mosquito sample. Samples were homogenized for 60 s (BeadBeater FastPrep; BioSpec Products, Inc., Bartlesville, OK, USA). The homogenized samples were incubated at room temperature for 15 min, and 60 μL of sterile water was added. DNA extraction was performed by using SoilMaster DNA Extraction Kit (Epicentre Biotechnologies, Madison, WI, USA) according to the manufacturer's instructions. The resulting DNA pellet was resuspended in 60 μL Tris EDTA buffer.
DNA extraction from water samples was performed as previously described (14). Two milliliters of each water sample was centrifuged at 16,000 × g for 1 h, 1.9 mL of the resulting supernatant was discarded, and DNA was extracted from the remaining volume by using a SoilMaster DNA Extraction Kit (Epicentre Biotechnologies).
Mosquitoes and water samples then underwent a F. tularensis holarctica-specifi c PCR, based on the 30bp-deletion region FtM19 and using the FtM19InDelF/R primer pair, and modifi ed from PCR (14). Each reaction consisted of 1-3 μL templates, 1x SsoFast EvaGreen Supermix (Bio-Rad, Hercules, CA, USA) 400 nmol/L for each of the primers FtM19Indel F/R (5′-GAATTACATAAAGTTCATGGTCCAGTAC-3′ and 5′-GTTTCAGAATTCATTTTTGTCCGTAA-3′) and Milli-Q (Millipore) water to give a fi nal volume of 20 μL. An initial denaturation at 98°C for 2 min was followed by 50 cycles at 98°C for 5 s, 60°C for 5 s, and a melt curve 65°C-95°C on a Bio-Rad CFX96. Positive control mixtures, using DNA from F. tularensis holarctica and negative control mixtures without a template, were included in each PCR run.

F. tularensis in Mosquitoes Hatched from Field-collected Larvae
The 334 adult mosquitoes of 9 species hatched from mosquito larvae collected in the tularemia-endemic Örebro area were analyzed in 48 pools; 14 pools (29%) were positive for the F. tularensis lpnA gene ( Table 1). Eleven of the 14 lpnA-positive samples were possible to sequence ( Figure 1). All obtained sequences showed high sequence similarity (>97%) with F. tularensis in alignment with published sequences from representatives of subspecies of F. tularensis and their closest known relatives (i.e., Francisella-like endosymbionts).

F. tularensis in Water
Water samples from 5 of the 8 water containers used for rearing were positive for the F. tularensis lpnA gene ( Table 2). Four of these containers yielded mosquitoes that were positive for the F. tularensis lpnA gene. However, 3 water containers used for rearing that were negative for the F. tularensis lpnA gene, all yielded adult mosquitoes positive for the F. tularensis lpnA gene. Thus, there was no correlation between detection of F. tularensis in water from a specifi c container used for rearing and detection of F. tularensis in adult mosquitoes hatched from the container ( Table 2).

Discussion
We detected F. tularensis holarctica DNA in adult mosquitoes hatched from fi eld-collected larvae sampled in an area in Sweden endemic for tularemia. This fi nding suggests that mosquitoes came in contact with the causative agent of the disease, F. tularensis holarctica, in the aquatic habitat of the mosquito larvae. We have previously shown that F. tularensis holarctica persists in natural aquatic environments between outbreaks (14) and in association with protozoa (10,11). Mosquito larvae of fl oodwater mosquitoes (i.e., A. sticticus) are predators on protozoa in temporary wetland environments (12). Our results suggest natural transstadial transmission of F. tularensis holarctica from its water reservoir via female mosquitoes to their vertebrate blood-meal hosts, including humans. The observation of water containers negative for F. tularensis that yielded mosquitoes positive for the bacterium and vice versa suggests that mosquitoes were truly positive for F. tularensis and not cross-contaminated with water that tested positive; however, we cannot exclude varying sensitivity of the real-time PCR analysis for water and mosquito samples. Further studies of the tissue tropism of the bacterium within the mosquito body are needed to confi rm how F. tularensis holarctica is transmitted by mosquito vector.
Transstadial transmission by mosquitoes after ingestion of pathogenic microorganisms as larvae has previously been shown for Rift Valley fever virus (16 Aedes communis Ae. punctor Ae. cinereus Ae. sticticus 4 (17) 2 ND 2 (12) 1 1 2b Ae. vexans Culiseta alaskaensis Cs. annulata Culex pipiens/torrentium 1 (5) ND ND 1 (4) fever virus to hamsters (16). Transstadial transmission of F. tularensis subspecies has also been reported in several species of ticks (3,17). In a recent study, transmission of F. tularensis novicida was tested in laboratory strains of the tropical mosquitoes Anopheles gambiae and Ae. aegypti (18). However, the bacterium was not transmitted transstadially to adult mosquitoes, and female mosquitoes exposed to F. tularensis novicida in a blood meal were not able to transmit the bacterium to mice. Results of this study, along with our results, contribute to the growing body of data that indicate differences in the ecology, including vectors and reservoirs, of Francisella species, subspecies, and even populations (3,14,19,20).
We detected F. tularensis DNA in 29% of the pooled samples of adult mosquitoes hatched from fi eld-collected larvae, indicating that transmission of the bacterium from water can generate a relatively high proportion of infected adult mosquitoes in an area endemic for tularemia. In line with our results, the F. tularensis fopA gene was detected in 30% of mosquito pools sampled in Alaska (18). However, further studies of host-seeking female mosquitoes in areas where tularemia is endemic are required to identify the range of mosquito species naturally infected with F. tularensis holarctica and the temporal distribution of the bacterium in these potential vector species in relation to the onset of outbreaks.
We detected F. tularensis holarctica in the fl oodwater mosquito species Ae. sticticus and Ae. vexans, the snowpool mosquito species Ae. punctor, and a mixture of Cx. pipiens and Cx. torrentium mosquitoes. The 3 Aedes spp. mosquitoes feed primarily on mammals and commonly   (13), a period when fl oodwater mosquito species are dominating the Swedish mosquito fauna (21). Notably, the detection rate of F. tularensis was higher in fl oodwater pools of female mosquitoes (46%) than in snow-pool pools of female mosquitoes (17%). The observation that F. tularensis holarctica occur in the fl oodwater mosquito Ae. sticticus is especially noteworthy because this nuisance species is now increasing its geographic range within Sweden (22). We suggest that the transmission of the bacterium F. tularenis holarctica via blood-feeding mosquitoes to humans in areas of Sweden where tularemia is endemic originates from the aquatic habitat of mosquito larvae. However, further studies are needed to confi rm transmission of the bacterium from its aquatic reservoir by blood-feeding female mosquitoes to their vertebrate hosts. The fi nding of F. tularensis holarctica DNA in adult mosquitoes, hatched from larvae collected in an area where tularemia is endemic, indicates that disease transmission in outbreaks originates in the pond habitats of the mosquito larvae.