Abstract
Rickettsia conorii is the causative agent of Mediterranean spotted fever (MSF) and Israeli spotted fever (ISF) transmitted by the brown dog tick Rhipicephalus sanguineus. In areas where MSF or ISF are prevalent, dogs have high prevalence of R. conorii -neutralizing antibodies. However, the true role of dogs in the persistence of the R. conorii transmission cycle is unknown, and their reservoir competence for this pathogen has remained untested. We assessed the ability of dogs infected with R. conorii to transmit the pathogen to previously uninfected Rh. sanguineus ticks. Dogs were infected either via needle-inoculation of cultured rickettsiae or naturally via infected tick bite. Dogs were monitored for clinical signs of infection, for rickettsemia by PCR, and for seroconversion and were subjected to infestation with uninfected ticks at different time points. Rh. sanguineus larvae and nymphs successfully acquired the agent from both needle-inoculated and tick-infected dogs and transmitted it transtadially. Tick-infected dogs remained infectious to ticks for at least a month postinfection. The molted ticks were, in turn, infectious to naïve dogs. These results demonstrate that dogs are capable of acquiring R. conorii from infected Rh. sanguineus ticks and transmitting infection to cohorts of uninfected ticks, thus confirming for the first time that dogs are indeed competent reservoirs for R. conorii. In addition, dogs with different genetic backgrounds appear to differ in their susceptibility to R. conorii infection.
Introduction
Unfortunately, these different opinions are exclusively based on indirect or circumstantial evidence. R. conorii has been isolated from Rh. sanguineus ticks removed from dogs. Serological studies conducted in various countries endemic for MSF have demonstrated a correlation between human disease and R. conorii antibody prevalence in canines (Tringali et al. 1986, Keysary et al. 1988, Herrero et al. 1992). Moreover, dog owners have higher odds of being infected with R. conorii than persons without pets (Mumcuoglu et al. 1993, Bacellar et al. 1995). In addition, several studies have reported the presence of either viable rickettsiae or rickettsial DNA in the dogs' blood (Durand 1930, Senneville et al. 1991, Estrada-Pena and Venzal Bianchi 2006, Solano-Gallego et al. 2006). In one controlled laboratory experiment, Kelly et al. (1992) repeatedly reisolated the Zimbabwean strain of R. conorii from the blood of artificially inoculated dogs and suggested that dogs may be important in the epidemiology of this pathogen by acting as domestic reservoirs of infection for ticks, albeit for short periods. However, this supposition has remained untested, as the authors did not attempt to infect ticks by feeding them on infected dogs.
To date, there is no direct evidence of R. conorii transmission from infected dogs to uninfected ticks. Questions of whether domestic dogs can amplify the infection in the population of vectors, or merely feed infected ticks and bring them into contact with humans, remain unanswered. Therefore, the goal of this study was to evaluate the reservoir competence of dogs for R. conorii by assessing their ability to acquire the pathogen and transmit it to previously uninfected tick vectors. The first phase was to determine the ability of dogs infected with R. conorii via needle inoculation to transmit the agent to uninfected Rh. sanguineus ticks. This first phase also served to supply ticks for the R. conorii-infected colony. In the second phase of the study, we assessed the ability of the dogs exposed to R. conorii via a tick bite to transmit the naturally acquired infection to uninfected vectors.
Materials and Methods
Ticks and model animals
Rh. sanguineus ticks in this study were derived from our colony of North American origin and maintained in our laboratory as previously described (Troughton and Levin 2007; Levin et al. 2009). Specific pathogen free New Zealand white rabbits (Oryctolagus cuniculus) were used as hosts for feeding all developmental stages of uninfected ticks. Between feedings, ticks were kept in environmental incubators at 24°C±1°C and 90% relative humidity. For the transmission experiments, ticks were placed inside feeding bags glued to a shaven area on a dog's back. After engorgement, xenodiagnostic ticks were allowed to molt to the nymphal stage and individually tested for the presence of rickettsial DNA.
For this study, we used 18–24-month-old male dogs–either mixed breeds or purpose-bred beagles. All experimental procedures were performed in accordance with Institutional Animal Care and Use Committee approved protocols. All dogs were housed indoors, in a climate-controlled animal facility that precluded an unintended exposure to any arthropod-borne agent including rickettsiae. The absence of antibodies to spotted fever group rickettsiae in the blood of each dog was confirmed before inoculation or infestation by the indirect fluorescent antibody (IFA) as described next. The appetite, behavior, and level of activity of each dog were monitored daily throughout the study. The weight and body temperature of infected animals were measured thrice per week. Venous blood and serum samples were collected twice weekly after infection and tested for the presence of rickettsial DNA and anti-R. conorii antibodies.
Dogs needle-inoculated with R. conorii
R. conorii isolates of African (R. conorii conorii strain Malish-ATCC VR614) and Mediterranean (R. conorii israelensis strain T487) origin were grown in Vero E6 cells at 32°C in antibiotic-free minimal essential medium supplemented with 2% fetal calf serum and 2 mg/mL L-glutamine (CDC Core Facility). Rickettsiae were purified by Renografin density gradient centrifugation, frozen at −70°C in SPG with 5 mM MgCl2 and 1% Renografin76, and titered by plaque assay on E6 cells as previously described (Eremeeva et al. 2003).
A mixed-breed dog A2 and a beagle A1 were intravenously inoculated with 1×106 R. conorii conorii. Mixed-breed dogs B1 and B2, and a beagle B3 were intravenously inoculated with 1×106 R. conorii israelensis (ISTT). Rh. sanguineus larvae were placed on dogs A1, B1, B2, and B3 on the day of inoculation, and uninfected nymphs were placed on dogs A2, B2, and B3 one day after inoculation (Table 1). Engorged larvae and nymphs were collected daily, allowed to molt, and used to establish R. conorii-infected colonies, or for the subsequent experiments involving tick-borne infectious challenge. Representative samples of freshly molted acquisition-fed ticks were tested by PCR for the presence of rickettsial DNA.Dogs were monitored daily for 4 weeks for changes in temperature, weight, and behavior. Blood and serum samples were obtained from all dogs three/week for 5–6 months for PCR and serologic detection of rickettsial infection.
Dogs infected with R. conorii israelensis via tick bite
Ticks fed on dogs needle inoculated with R. conorii israelensis were placed on mixed-breed dog C1 together with uninfected Rh. sanguineus nymphs for proliferation of R. conorii infection in the tick colony by cofeeding as previously described (Zemtsova et al. 2010). Adult ticks derived from these nymphs were used as the source of infection for the subsequent experiments. The dog C1 itself was not subjected to follow-up xenodiagnostic infestations but was monitored for clinical signs of infection as well as by blood-PCR and serology.
Ten male and 10 female Rh. sanguineus infected with R. conorii israelensis by means of cofeeding on the dog C1 were placed on naïve beagles C2 and C3 (day 0) and allowed to feed for 8 days. After removal of all adult ticks, dogs were fitted with new feeding bags, and cohorts of uninfected Rh. sanguineus larvae were placed on days 9, 16, 23, and 30 postinfection. To avoid inflammatory reactions at the feeding site, the first and third groups of larvae were placed on dog C2, and the second and fourth groups of larvae were placed on dog C3 (Table 2).
Diminished survival of ticks exposed to the strain Malish (Levin et al. 2009) precluded implementation of this second phase using R. conorii conorii.
PCR and serology
Whole blood (200 μL) and serum (500 μL) samples from dogs were aseptically collected into either 1.7 mL microcentrifuge tubes (Corning, Inc., Lowell, MA) containing 5 μL of 2% EDTA (Sigma Aldrich, St. Louis, MO) or Microtainer serum separator tubes (Becton Dickinson and Co., Franklin Lakes, NJ). Blood samples were stored frozen at −20°C until tested by PCR for the presence of R. conorii DNA, and sera were refrigerated for serologic testing.
DNA extraction and PCR procedures were carried out in separate facilities. DNA was extracted from ticks and blood samples by using the Qiagen DNEasy Blood and Tissue kit (Qiagen, Inc., Valencia, CA) according to manufacturer's protocols. The presence of rickettsial DNA was detected by PCR using primers RR190-547F and RR190-701R to amplify a 154-bp fragment of the rompA gene of Rickettsia as described by Eremeeva et al. (Eremeeva et al. 2003). Blood samples were tested in duplicates with negative samples included in all extraction and PCR rounds. Water was used as a no-template negative control. PCR was interpreted as positive when both duplicate reactions had positive results.
IFA was performed on dog sera, as previously described (Lennette et al. 1995), using FITC labeled goat anti-dog IgG (γ) conjugate diluted as per manufacturer's recommendations (KPL, Inc., Gaithersburg, MD). Slides were spotted with Rickettsia conorii Morocco (VR141) antigen, air-dried, and fixed in acetone. Serum samples were initially screened at 1/16 and 1/256 dilutions, and positive samples were titered to the endpoint in a twofold dilution series. Serologic data are reported as the reciprocal of the last dilution showing positive fluorescence. Titers ≥16 were considered positive.
Results
Transmission of R. conorii conorii and R. conorii israelensis from needle-inoculated dogs to Rh. sanguineus ticks
The dogs A1 (beagle) and A2 (mixed-breed) responded to needle inoculation with R. conorii conorii with mild fever, loss of appetite for 2–4 days, and transient lethargy during the first week after inoculation. All symptoms spontaneously resolved without intervention by the day 8 postinoculation. Rickettsial DNA was continuously detectable in the blood of dog A1 from day 2 to 8 postinoculation, whereas dog A2 was blood-PCR positive from day 3 though day 7 and again on day 24. Dogs A1 and A2 seroconverted by day 10 and 7, and the immune response peaked between 14 and 21 days postinoculation with antibody titers reaching 16,384 and 8192 respectively. Both dogs remained blood-PCR negative and seropositive for the remainder of observation with antibody titers declining to 512 by day 152 postinoculation.
Rh. sanguineus larvae and nymphs fed on both R. conorii conorii–inoculated dogs successfully acquired the infection and transmitted it transstadially (Table 1). The resulting prevalence of infection in nymphal ticks that had fed as larvae on beagle A1 was 17%, and the prevalence of infection in adult ticks that had fed as nymphs on the mixed-breed dog A2 was 93%.
Needle inoculation of dogs with R. conorii israelensis resulted in variable clinical presentation. Dog B1 was febrile for the first 5 days postinoculation, anorexic on days 2–4, and lethargic on day 4. It was blood-PCR positive only on days 1 through 8 postinoculation, seroconverted by day 15, and exhibited a peak antibody titer of 16,384. The antibody titers declined to 512 by day 90 after inoculation and to 256 by day 156. Dog B2 experienced a decreased appetite but no fever, and had detectable rickettsemia on days 4, 7, and 21 postinoculation. It seroconverted by day 7 and exhibited a peak antibody titer of 4096 on days 14–18. The titers declined to 512 by day 92, and the dog became seronegative within 6 months postinoculation. Conversely, the beagle-dog B3 did not exhibit any symptoms after the inoculation; it ate and behaved normally, and retained a normal body temperature. Blood samples from B3 were all PCR negative for R. conorii. This dog seroconverted by day 7, but the antibody titers reached only 132, and the dog became seronegative within only 3 months postinoculation.
Rh. sanguineus larvae and nymphs fed on all three R. conorii israelensis—inoculated dogs acquired the infection and transmitted it transstadially (Table 1). The resultant prevalence of infection in freshly molted nymphs ranged from 12% to 37%, and in adult ticks—from 10% to 15%. It is noteworthy that dog B3 successfully transmitted infection to both xenodiagnostic larvae and nymphs, despite being PCR-negative throughout the experiment.
Transmission of R. conorii israelensis from dogs infected via tick bite to uninfected Rh. sanguineus
When mixed-breed-dog C1 was infested with ticks that had previously fed on an R. conorii israelensis–inoculated dog, it did not exhibit any symptoms of infection. However, rickettsial DNA was detected in the dog's blood on two occasions (days 25 and 56 postinfestation), and antirickettsial antibodies developed by day 10. Antibody titers peaked at 8192 on days 14–21, declined to 512 by day 52, and the dog became seronegative at 227 days postinfestation. Blood samples for PCR were collected thrice/week throughout the 7.5-month long observation period, and only two samples were positive. It is noteworthy that the immune response to rickettsial infection had already passed its peak by the time the pathogen was first detected by PCR, and antibody titers were on the decline when rickettsial DNA was detected in the blood the second time.
Beagle dog C2 exhibited decreased appetite on day 5 postinfestation and elevated temperature on days 4–6. Blood was PCR positive on days 43, 45, and 48 postinfestation. Seroconversion occurred by day 13, and antibody titer peaked at 2048 on days 17–22. After that, antibody titers gradually declined, and the dog became seronegative by day 214 postinfestation. Here again, the antibody response was already on the decline by the time of rickettsial DNA detection in the dog's blood. Beagle dog C3 showed no clinical signs of R. conorii infection. Its appetite, temperature, and behavior remained normal, and it failed to seroconvert. However, it had a PCR-positive blood sample on day 43 postinfestation.
Uninfected Rh. sanguineus larvae were able to acquire R. conorii from both dogs C2 and C3 and transmitted the agent transstadially (Table 2). This transmission occurred at the time when rickettsial DNA was not detectable in the blood by PCR, and despite extremely high antibody titers in dog C2. The dogs remained infectious to uninfected larvae for at least a month after infected ticks had completed their feeding, although the prevalence of infection varied between cohorts of recipient ticks (Table 2). Larger proportions of ticks acquired the pathogen if placed on dogs at two and 3 weeks after placement of infected adult ticks than if placed at 1 or 4 weeks. The difference in prevalence of infection was statistically significant between the second or third cohorts and the fourth cohort of ticks (P χ2=0.004).
Discussion
Results of this study demonstrate for the first time that dogs infected with R. conorii conorii and R. conorii israelensis are indeed infectious to Rh. sanguineus ticks. Moreover, we show that dogs acquiring the R. conorii israelensis infection naturally, via tick bite, can be infectious to ticks for at least 4 weeks after initial exposure to the agent regardless of the presence or severity of clinical symptoms. Successful transmission of R. conorii israelensis from dogs to ticks was observed while rickettsemia was undetectable by PCR, and in the presence of high antibody titers.
In 1967, Burdgdorfer and Varma suggested that ticks serve as both vectors and reservoirs for most of the spotted fever group rickettsiae (Burgdorfer and Varma 1967). General acceptance of this hypothesis may simply be due to its frequent reiteration, or stem from an assumption that vertebrate hosts of ticks once exposed to a pathogen become immune for the rest of their lives and, thus, are excluded from maintenance of the natural transmission cycle. However, maintenance of SFG rickettsiae with ticks as the only reservoir would require efficient transovarial and transstadial transmission with bacteria causing no deleterious effect on the reproductive fitness or viability of the tick host. This does not always appear to be the case, as at least some strains of Rickettsia rickettsii and R. conorii have been shown to negatively affect fecundity and survival of infected ticks (Niebylski et al. 1999, Levin et al. 2009, Socolovschi et al. 2009). This suggests that these agents could not be maintained indefinitely without regular horizontal proliferation among ticks through susceptible vertebrate hosts.
Since the principal vector of R. conorii—Rh. sanguineus—is a common parasite of dogs worldwide, these hosts are routinely exposed to the pathogen and have a potential for affecting its maintenance cycle—whether by increasing or decreasing prevalence of infection in ticks. A number of studies have demonstrated a correlation between human disease and R. conorii antibody prevalence in canines (Tringali et al. 1986, Keysary et al. 1988, Herrero et al. 1992). Also, dog owners may have higher odds of being infected with R. conorii than persons without pets (Mumcuoglu et al. 1993, Bacellar et al. 1995).
In our study, all but one dog seroconverted within 2 weeks postinfection, regardless of the mode of inoculation and developed high antibody titers. These titers, however, declined significantly within the next 3 months. Two of our dogs that seroconverted after exposure to infected ticks became negative after 214 and 227 days postinfection. Evidence for the relatively short life of anti-R. conorii immunity in naturally infected dogs can be found in a number of studies where both the prevalence and titers of antibodies declined rapidly in winter when ticks were inactive but rose again the next spring (Espejo-Arenas et al. 1990, Espejo et al. 1993, Delgado and Carmenes 1995, Ortuno et al. 2009). This is in contrast to the serologic response in humans, who maintain IgG titers for several years after contracting boutonneuse fever (Mansueto et al. 1985). At least two surveys found no significant differences in seroprevalence to R. conorii between young (<1 year old) and old dogs (Delgado and Carmenes 1995, Harrus et al. 2007). This again confirms that canine antibodies lasted less than a year.
Despite frequent exposure to R. conorii, dogs rarely become clinically ill when infected by either infected ticks or inoculation of infected blood (Durand 1930, Mumcuoglu et al. 1993, Ortuno et al. 2009), or these symptoms are so mild and of such short duration that they are not noticed. One report described a febrile illness associated with R. conorii infection in three Yorkshire terriers, which exhibited anorexia and lethargy for 2–3 days (Solano-Gallego et al. 2006). In our study, most of the dogs had similar clinical symptoms of infection, although the severity of symptoms varied depending on the bacterial strain, the mode of infection, and on the genetic background of the animals. Needle inoculation with the isolate Malish caused mild fever, anorexia, and lethargy. These symptoms appeared more pronounced in a mixed-breed dog than in a pure-bred beagle. Needle inoculation with an identical dose of R. conorii israelensis caused yet milder symptoms in mixed-breed dogs and none at all in a beagle. When naïve dogs were subjected to infestation with R. conorii israelensis-infected ticks, two out of three exhibited no adverse reaction at all. Although a description of clinical symptoms after R. conorii infection in dogs was not a subject of this study, our records show that the propensity for clinical illness was lower in beagles than in mixed-breed dogs.
Regardless of the symptoms, rickettsial DNA was repeatedly detected in the blood of most dogs whether needle inoculated or bitten by infected ticks. Interestingly, the dogs that exhibited more pronounced clinical symptoms (A1, A2, and B1) had only short periods of detectable rickettsemia lasting less than 10 days postinfection. On the other hand, blood-PCR was positive much later—between 25 and 56 days postinfection—in dogs with milder or no symptoms. It is also noteworthy that in dogs infected with R. conorii israelensis via tick bite, rickettsial DNA was detected much later than in needle-inoculated dogs. This dissimilarity is likely due to the difference in the infectious dose received by animals, but may also be related to the inoculation routes. Similarly to our results, Estrada-Pena and Bianchi detected persistence of R. conorii DNA in dog blood for at least 3 months (Estrada-Pena and Venzal Bianchi 2006).
Uninfected ticks feeding on one of the naturally infected dogs at the peak of antibody response successfully acquired R. conorii and transmitted it transstadially. Further, in two of the dogs, rickettsial DNA was detected well after this peak, whereas antibody titers were on a decline. These episodes of detectable rickettsemia did not trigger a boost in antibody titers. Thus, the presence of antibodies per se did not prevent dogs from being infectious to ticks, or from recurring rickettsemia. The transmission of R. conorii despite the presence of high antibody titers suggests that the antibodies may not be protective, and indicates a possibility that even seropositive dogs may serve as competent reservoirs upon reinfection.
Together, results of this study conclusively demonstrate for the first time that the domestic dog (C. familiaris) is a competent reservoir of R. conorii. Dogs infected with R. conorii, whether by artificial or natural means, were able to transmit the infection to previously uninfected Rh. sanguineus ticks. Dogs were infectious to ticks for at least a month after exposure to infected ticks, and the transmission success was not dependant on the presence of clinical symptoms. Further, the highest rates of transmission were seen in the absence of detectable rickettsemia in the blood, and in the presence of high antibody titers. We subjected dogs to infestations with uninfected ticks for only up to 4 weeks after the initial exposure. It remains to be studied whether the period of infectivity for ticks in naturally infected dogs can last longer than 1 month postinfection.
The mildness of a tick-borne infection in reservoir animals normally would allow vertebrate hosts to remain active, which will increase exposure of the animal to uninfected ticks, and spreading the infected ones. Therefore, absence of severe clinical illness in dogs infected with R. conorii via the natural route of a tick bite indirectly confirms ecological coadaptation between this agent and the domestic dog as a natural reservoir.
Relative ecological importance of the systemic transmission (due to rickettsemia) in dogs vs. transmission between cofeeding ticks remains to be assessed. Prevalence of infection in ticks acquiring R. conorii israelensis from naturally infected dogs in this study was lower than when donor and recipient ticks were simultaneously fed (Zemtsova et al. 2010). However, the longevity of the systemic transmission may well compensate for its lower efficiency and allow the infectious dogs to produce more infected ticks and spread them over a larger territory.
Although Rh. sanguineus in peridomestic environment feeds primarily on dogs, it can also be found in a diverse range of wild and domestic animals, including rodents, lagomorphs, ungulates, carnivores, and humans (Mumcuoglu et al. 1993, Walker et al. 2000, Kahn and Line 2010). Unsurprisingly, all these host species are also exposed to R. conorii, and some of them may also serve as reservoirs alongside or in competition with dogs. Antibodies against R. conorii have been also detected in a variety of domestic animals including pigs, donkeys, cattle, goats, sheep, rabbits, mules, and horses, potentially suggesting a role in maintenance of MSF (Herrero-Herrero et al. 1989). The European rabbit (O. cuniculus) has been suggested to play a role in transmitting R. conorii along the French Mediterranean coast, because a spectacular reduction in MSF cases was seen during an outbreak of myxomatosis in 1952, which destroyed the wild rabbit population (reviewed by Rovery and Raoult 2008). In addition, small ruminants have been suspected as potential reservoirs of R. conorii based on the fact that in Cyprus the prevalence of infection in Rh. sanguineus ticks collected from sheep and goats was significantly higher than in those from dogs (Psaroulaki et al. 1999). The ability of these or any other vertebrate hosts for transmission of R. conorii to ticks following natural infection remains to be studied. Only then will their relative importance as reservoirs in the maintenance of this agent be able to be compared with that of the domestic dog.
Footnotes
Acknowledgments
The authors thank Davidia N. Grant for her invaluable help with the maintenance of infected and uninfected tick colonies. They gratefully acknowledge Dr. William L. Nicholson and Aubree J. Roche for their contributions to this study.
Disclosure Statement
No competing financial interests exist.
