Fleas as parasites of the family Canidae

Plague

Fleas are mainly the vectors of bacteria. Historically, the most important bacterial agent which is transmitted by fleas is the plague bacterium, Yersinia (Y.) pestis. Human disease caused by Y. pestis has been historically associated with rats, mainly with Rattus (Ra.) norvegicus and Ra. rattus. The origin of Y. pestis is now thought to be the steppe regions of Central Asia. In nature, the plague agent is transmitted by fleas among the rodent populations. Mainly rodent fleas are involved in the natural transmission, and more than 80 flea species belonging to different genera were found to transmit Y. pestis in nature [17, 18]. These flea species mainly parasitize rodents, but they may be occasionally found on pets and also on dogs which can be shown by the occasional detection of fleas of wild rodents or other wild animals on dogs. The transmission of plague to humans via flea bites is almost exclusively through the oriental rat flea (X. cheopis) which played the major role as bridging vector between the rodent and rat populations and humans.

While the role of cats for the transmission of plague has been established for a long time [19], dogs are considered to be less susceptible to plague and their role for the transmission of Y. pestis has not been established equally well. Dogs seem to develop some clinical illness after infection with Y. pestis, including symptoms like fever, lethargy and bubos [20]. Epidemiological data show that plague patients are significantly more likely reported for sleeping in the same bed as a dog [21]. These data indicate that probably a long and close contact with dogs and their fleas is necessary for the transmission from dogs to humans [22]. These epidemiological data are strengthened by experimental data, showing that the flea species found most commonly on dogs, the cat flea and the dog flea, are considered to be poor vectors of Y. pestis due to their poor ability to become blocked [23]. They may possess a small although not essential importance by their ability to become pestiferous. The mouth parts may be contaminated with Y. pestis during blood sucking of bacteriaemic blood. The transmission may occur mechanically or by a newly discovered unblocked transmission mechanism recently shown to play a role in the transmission of plague in the flea species O. montana [24]. However, a role of P. irritans, which is found frequently in low prevalence rates on dogs (see above), is under discussion and has not yet been resolved [25]. The role of another flea, occassionally detected on dogs, Echidnophaga (E.) gallinacea, in the transmission of plague is unclear, although it is known to become infected with Y. pestis [26]. However, it is also thought to be a poor vector for the agent due to its stick-tight behaviour [23, 26]. The role of Tunga (T.) penetrans as a plague vector is unknown [25]. Possibly the male fleas, which are free hematophagous ectoparasites, may act as pestiferous vectors while the females of T. penetrans are not capable to do so, because they are embedded in the host epidermis [27]. No data are available on Archeopsylla erinacei and its potential role as a vector of Y. pestis. The ability of Rh. lutzi to transmit the plague bacterium is unknown. However, a non-identified Rh. sp. is listed as vector in an established wild rodent plague focus in South America [28]. Fleas of the genus Megabothris were also involved in the natural transmission of Y. pestis in natural foci in the northwestern U.S.A. [28], while information on species of the genera Ceratopsyllus, Paraceras, Cediopsylla and Ceratophyllus to our best knowledge is not available.

Rickettsioses

There are mainly two species of rickettsiae which are naturally transmitted by fleas, Rickettsia (R.) typhi, the pathogen of murine typhus and R. felis, a recently discovered Rickettsia species causing flea-borne spotted fever.

Murine typhus is a zoonosis which is maintained in nature mainly by a flea-rat-flea transmission cycle [29]. Murine typhus is one of the few rickettsioses which are distributed globally except in the Antarctic [29]. Rats of the genus Rattus, mainly R. rattus and R. norvegicus are of major importance for the primary transmission cycle as vertebrate hosts. Besides these two species, many other wild and domestic animals may serve as additional vertebrate hosts. Among them, dogs were also found to become infected with R. typhi in different areas of the world. In Spain, between 10 to 12% of dogs showed antibodies against R. typhi, while during an outbreak of murine typhus in Austin, Texas, U.S.A., 44% of dogs tested exhibited antibodies agains R. typhi [30–33]. Because of these data, the potential role of pet and stray dogs during urban outbreaks of murine typhus has to be readdressed.

R. typhi has deen detected in at least ten species of fleas of the genera Ctenocephalides., Echidnophaga, Leptopsylla, Monopsyllus, Nosopsylla, Pulex and Xenopsylla [29]. Among them is C. felis which is frequently found on dogs all over the world and the most prevalent flea species on dogs in many areas (Table 1). Several other flea species are known to play a role as actual or potential vectors of R. typhi, among them are E. gallinacea, P. irritans and X. cheopis. They are found occasionally on dogs and therefore may support an urban dog-flea-dog transmission cycle (Table 1). X. astia, X. bantorum, X. brasiliensis, Leptopsylla (L.) segnis and Nosopsylla (N.) fasciatus are primary rodent fleas which may accidentally feed on humans or on dogs [1].

R. felis is the pathogen of flea-borne spotted fever. Interestingly, together with R. typhi, it is the only other Rickettsia species which seems to be globally distributed. Both of these rickettsiae are transmitted by fleas. While the agent of murine typhus has been known for many decades, the agent of flea-borne spotted fever was only detected recently. In 1990, a rickettsia-like organism in an adult flea colony in the Elward Laboratory, Maryland was detected and named "ELB agent" [34]. Later, the further characterization of this agent resulted in the classification as a spotted fever group rickettsiae [35]. In recent years, the widespread geographical distribution of R. felis and of the disease caused by this agent increased, indicating that R. felis is the most widely distributed Rickettsia species.

The cat flea (C. felis) was identified as the primary vector and reservoir of R. felis. So far, this Rickettsia species has been detected in more than 30 countries on five continents except Antarctica [36–41]. The reason for this almost universal distribution may be seen in the global distribution of C. felis which finally transported its ectoparasites and their endoparasites into all parts of the world. The high infection rates of colonized cat flea populations from 40 up to 90% [42], together with experimental data, imply that R. felis is well-adapted to its vector including transstadial and transovarial transmission leading to the high infection rates of fleas [43, 44]. Although the cat flea is thought to be most important as a vector, R. felis was detected in at least eleven other flea species [36]. Among them are P. irritans, Archaeopsylla (A.) erinacei, E. gallinacea, C. canis, T. penetrans and X. cheopis. These species are found either sporadically or more frequently also on dogs.

Although a number of serological studies have been conducted to identify the mammalian host(s) of R. felis, no definitive mammalian host has been identified so far. Among the peri-domestic mammals which exhibit antibodies against R. felis, cats, dogs and opossums are found [36]. Cats are believed to be the most important hosts for the cat flea and for the rickettsiae transmitted by cat fleas. In different areas of the world antibody seroprevalence rates of 4 to 100% of cats were reported in several studies (summarized in [36]). However, the definitive role of the cat as a mammalian host supporting the life cycle of R. felis, has not been established. While the transmission of fleas to cats had been proven either by serology or by molecular biology [44], the horizontal transmission of R. felis from mammals to fleas so far has not been shown [45]. Therefore, the role of cats for the life cycle of R. felis remains unclear. Even more unclear is the role of domestic dogs for the maintainance of the natural life cycle of R. felis. Only few seroprevalence studies in dogs are available. In Spain, 16% of dogs showed specific antibodies against R. felis [31]. Two other reports indicate an association of PCR-positive dogs to the occurance of sporadic human cases of flea-borne spotted fever. In Germany and in Spain, R. felis-infected dogs were present in families with human cases [46, 47]. The dog seems to play the role as a transport vehicle for C. felis and also for its parasite, R. felis to humans. So far, no wild animal reservoir of R. felis has been detected. There is, however, some evidence, that the opossum (Didelphis virginiana) may play a role [48, 49].

Another Rickettsia species was recently found in fleas. In the USA, a sylvatic transmission cycle of R. prowazekii was detected [50]. When exploring this transmission cycle, R. prowazekii could be detected in lice and in squirrel fleas (Or. howardi). The exact transmission between the squirrels and from squirrels to humans remains unclear. However, it is hypothesized that R. prowazekii is excreted by the fleas (and lice) with their feces. After drying of feces, the pathogen may be inhaled as aerosol and then may cause typhus. A role for dogs within this so-called sylvatic transmission has not been postulated so far. However, dogs may serve as mechanical carriers of fleas from outside to the home and therefore serve as a bridging host.

Bartonellosis

Besides these two pathogens of the genus Bartonella, dogs and canids seem to play a role as putative natural hosts for B. vinsonii subsp. berkhoffii [54–56]. Various data show that stray dogs in the tropics exhibit high antibody prevalence rates against B. vinsonii subsp. berkhoffii. In sub-Saharan Africa and in Asia 26% to 65% of domestic dogs tested were antibody positive against this species [57, 58]. In contrast, dogs in non-tropical regions (Europa, U.S.A., northern Africa) were seropositive only in < 5% [52, 59, 60]. However, in California, B. vinsonii subsp. berkhoffii was detected by serology in up to 35% and in 28% by PCR in coyotes (Canis latrans). This high prevalence and the apparent chronic bacteriaemia imply a role of the coyote as natural host for B. vinsonii subsp. berkhoffii [59]. Possibly, the frequent contacts of stray dogs with wild canids cause the high prevalence rates while the contacts of domestic dogs with coyotes in California seem to be limited. B. vinsonii subsp. berkhoffii has also been incriminated as a cause of human heart disease [61]. The infection of humans by flea-borne bartonellae mainly occurs via cat scratches. An infection of cats via infected fleas could be experimentally demonstrated, although this way is thought not to play a major role in the transmission and epidemiology of flea-borne Bartonella. The infection of humans via fleas seems possible, however this mode of infection also does not seem to play a major role for the epidemiology of Bartonella. Sporadic cases of transmission of bartonellae by dog scratches or dog bites document the potential risk of transmission from dog to human and therefore may also involve dogs as bridging hosts from animals to humans [62, 63].

Miscellaneous

A number of other bacterial pathogens have been isolated from fleas. These are Coxiella burnetii, Francisella tularensis, Staphylococcus aureus, Salmonella enteritidis, Borrelia burgdorferi, Borrelia duttoni, Listeria monocytogenes, Y. pseudotuberculosis, Erysipelothrix rhusiopathiae, Burkholderia (Bu.) mallei, Bu. pseudomallei, and Brucella abortus [1]. Probably, most of these pathogens were detected by chance after a bacteriaemic bloodmeal. The potential to serve as vectors for the agents listed, although unknown and never tested experimentally, seems to be minimal.