Ecology of Rickettsia felis: A Review

Abstract

It has been two decades since the first description of Rickettsia felis, and although a nearly cosmopolitan distribution is now apparent, much of the ecology of this unique microorganism remains unresolved. The cat flea, Ctenocephalides felis, is currently the only known biological vector of R. felis; however, molecular evidence of R. felis in other species of fleas as well as in ticks and mites suggests a variety of arthropod hosts. Studies examining the transmission of R. felis using colonized cat fleas have shown stable vertical transmission but not horizontal transmission. Likewise, serological and molecular tools have been used to detect R. felis in a number of vertebrate hosts, including humans, in the absence of a clear mechanism of horizontal transmission. Considered an emerging flea-borne rickettsiosis, clinical manifestation of R. felis infection in humans, including, fever, rash, and headache is similar to other rickettsial diseases. Recent advances toward further understanding the ecology of R. felis have been facilitated by stable R. felis-infected cat flea colonies, several primary flea isolates and sustained maintenance of R. felis in cell culture systems, and highly sensitive quantitative molecular assays. Here, we provide a synopsis of R. felis including the known distribution and arthropods infected; transmission mechanisms; current understanding of vertebrate infection and human disease; and the tools available to further examine R. felis.

Although a number of bacteria are associated with blood-feeding arthropods, members of the genus Rickettsia are recognized for their unique relationship with the vector. One such agent is Rickettsia felis, a gram-negative bacterium predominately described in the cat flea, Ctenocephalides felis. First identified in cat fleas in 1990 (Adams et al. 1990), R. felis has since been identified in a variety of fleas. Within the last 20 yr, there also have been a growing number of reports implicating R. felis as a human pathogen. Despite the increasing appearance of R. felis in the literature, the basic transmission mechanisms of R. felis in nature is unknown. Research progression toward an understanding of the epidemiology and pathogenesis of R. felis has been hampered by the limited knowledge of basic R. felis biology. However, the persistence of R. felis in colonized cat fleas and more recently the cultivation of several R. felis isolates from arthropod hosts have aided the characterization of R. felis transmission.

Phylogenetically, rickettsiae belong to the α-sudivision of Proteobacteria, with arthropod-borne Ehrlichia and Anaplasma as close relatives (Dumler et al. 2001). Within the genus Rickettsia, the organization of the disease-causing organisms were traditionally grouped into the typhus group (TG) and spotted fever group (SFG). Recent genome analyses have determined that additional groups exist within the genus; R. felis is currently placed in the transitional group (TRG) (Gillespie et al. 2007, Weinert et al. 2009); however, existence and placement of R. felis in this group has been debated (Fournier et al. 2008).

Structurally, rickettsiae are coccobacilli (rod-shaped) bacteria that average 0.7–2.0 μm by 0.3–0.5 μm in size. The physical composition of Rickettsia includes a typical trilaminar, gram-negative cell wall structure consisting of inner and outer membranes separated by a peptidoglycan layer (Pang and Winkler 1994). A sizable body of literature describing polymorphic rickettsiae exists; however, the significance of the morphologically distinct forms to rickettsial biology is unknown (Gulevskaia et al. 1975, Philip et al. 1983, Sunyakumthorn et al. 2008).

Among the more recently recognized emerging rickettsial infections, R. felis has proved challenging to characterize. A rickettsia-like organism, first observed by electron microscopy in the midgut epithelial cells of colonized adult cat fleas, was designated “ELB agent” after the source of the fleas, the Elward Laboratory cat flea colony (El Soquel, CA) (Adams et al. 1990). Amplification of rickettsial genes encoding citrate synthase (gltA) and the genus-specific 17-kDa antigen genes confirmed the presence of rickettsiae in the Elward Laboratory flea colony (Azad et al. 1992), and the name R. felis was proposed (Higgins et al. 1996). Subsequent amplification of rickettsial genes encoding the 17-kDa antigen and the 190-kDa antigen (ompA) genes from a colony of cat fleas maintained at Louisiana State University (LSU; Baton Rouge, LA) confirmed the original molecular characterization and further classified R. felis as an SFG Rickettsia (Bouyer et al. 2001). Although initial attempts to produce a sustained culture of either the ELB or the LSU strains of R. felis failed (Radulovic et al. 1995, Bouyer et al. 2001), later attempts to isolate and propagate R. felis from cat fleas maintained by Flea Data (Freeville, NY) proved successful (Raoult et al. 2001). The Flea Data isolate was designated strain Marseille-URRWXCal₂ (also reported as strain California 2; Ogata et al. 2005) and is the current type strain for R. felis (LaScola et al. 2002). Additional strains of R. felis have been isolated from colonized and wild-caught cat fleas and propagated using various arthropod cells lines (Horta et al. 2006b, Pornwiroon et al. 2006). The genome and proteome of R. felis (Marseille-URRWXCal₂) recently have been described (Ogata et al. 2005, Ogawa et al. 2007), and unique characteristics include the presence of conjugative plasmids pRF and pRFδ and the expression of several proposed virulence factors, respectively. To further exemplify this truly dynamic organism, it seems that plasmid content can vary among isolates of R. felis and among Rickettsia species (Baldridge et al. 2008, Fournier et al. 2008).

The recognition of R. felis as an emerging pathogen, in conjunction with its intriguing biology, provides exciting avenues for research. As detection of R. felis in arthropods and vertebrate hosts becomes more widespread, the distribution of human flea-borne rickettsiosis cases also expands (Fig. 1). Here, we summarize the current knowledge regarding the ecology of R. felis, including the expanding range of invertebrate and vertebrate hosts; the potential routes of transmission; the tools and techniques used to diagnose R. felis infection; and the characteristics and dynamics of R. felis rickettsioses. We will also present topics that require further study.

Fig. 1.

Reported distribution of R. felis. R. felis–positive arthropods have been identified in the following countries (gray): Afghanistan, Algeria, Argentina, Australia, Canada, Chile, Croatia, Cyprus, Democratic Republic of Congo, Ethiopia, Gabon, Indonesia, Israel, Japan, New Zealand, Peru, Portugal, United Kingdom, and Uruguay. R. felis–positive arthropods and cases of flea-borne spotted fever have been identified in the following countries (black): Brazil, Egypt,* France, Germany, Laos, Mexico, South Korea, Spain, Taiwan, Thailand, Tunisia,* and the United States (*not reported from arthropods in this country).

Rickettsia felis Infection of Arthropods

The cat flea serves as the primary vector and reservoir of R. felis. Consequently, most of our current understanding of R. felis infection in arthropods comes from the R. felis/C. felis model. The distribution of R. felis within fleas has been examined using microscopic (transmission electron microscopy [TEM]) and molecular methods (polymerase chain reaction [PCR]). specific tissues in which R. felis has been identified in adult cat fleas include the midgut epithelial cells (also identified in larval fleas), muscle cells, fat body, tracheal matrix, ovaries, epithelial sheath of testes, and salivary glands (Adams et al. 1990, Bouyer et al. 2001, Macaluso et al. 2008). Within the epithelial cells of the salivary glands, R. felis were typically free in the cytosol, often surrounded by a “halo” (cleared cytoplasmic contents) consistent with previous ultrastructural descriptions of R. felis in tick-derived cell culture (Pornwiroon et al. 2006) and in other flea tissues (Adams et al. 1990). The wide dissemination of R. felis within the flea host suggests a close association between fleas and the bacteria; however, a correlation between rickettsial distribution in flea tissues and the distinct transmission route has not been determined.

Quantitative real-time PCR (qPCR) assays have been developed for use as a diagnostic tool and to quantitatively assess rickettsial load within fleas. Both SYBR Green and probe-based qPCR assays have been described as methods for detecting and quantifying R. felis load in fleas. In an antibiotic susceptibility study, replication of multiple isolates of R. felis were measured in XTC-2 cells treated with various antibiotics; however, assay details and R. felis load calculation methods in this study were not clear (Rolain et al. 2002). In another study, a probe-based method targeting a portion of the ompB gene, which encodes rickettsial outer membrane protein B, was able to differentiate between R. felis– and R. typhi–infected fleas and had an assay sensitivity of 1 copy/μl (determined using serial dilutions of a plasmid containing a portion the ompB gene) (Henry et al. 2007). Comparison of crude (boiled flea lysate) versus kit-based DNA extraction methods determined crude extraction of DNA was insensitive, resulting in limited detection of Rickettsia (Henry et al. 2007). A recently developed SYBR Green assay was used to describe the kinetics of R. felis infection in cat fleas during the metabolically active periods of flea bloodmeal acquisition and oogenesis (Reif et al. 2008). In this study, a mean rickettsial load of 3.9 × 10⁶ rickettsiae per flea with a range of 1 × 10³–1.6 × 10⁷ rickettsiae per flea was identified. R. felis infection density was also determined by comparing the ratio of rickettsiae to flea genes. During9d of flea bloodmeal feeding, rickettsial load remained relatively stable; larger-sized, female fleas had greater rickettsial burdens than male fleas (Reif et al. 2008). Tissue-specific R. felis infection burden has not been examined in fleas but may provide clues to the specific tissues involved in R. felis transmission. Additional studies examining molecular interactions between R. felis and the arthropod host will provide valuable information to decipher R. felis biology and epidemiology.

The presence of R. felis in commercial and institutional flea colonies has been extensively studied (Table 1). After the initial identification of R. felis in the Elward Laboratory cat flea colony, a 1999 survey examining the prevalence of R. felis infection in commercial and institutional colonies showed that R. felis infection was widespread in controlled cat flea colonies (Higgins et al. 1994). In addition to cat fleas, R. felis has been detected molecularly in a panoply of blood-feeding arthropods worldwide, in 28 countries spanning five continents (Table 2). As the only currently defined biological vector, the infection of cat fleas by R. felis is concerning from a public health perspective because of the cat fleas’ habit of indiscriminate host selection (Azad et al. 1992). In contrast to other described rickettsial species, R. felis is unique because it may potentially infect both insect and acarine hosts and has been detected in both ticks and mites (Ishikura et al. 2003, Choi et al. 2007, Tsui et al. 2007, Oliveira et al. 2008). The prevalence of R. felis infection in surveyed wild-caught arthropods ranges from 0.8 to 100%, depending on species and geographic location, but typically is <25% (Table 2). Although R. felis has been detected molecularly in numerous arthropod species, there exists the potential for arthropods that have just consumed an R. felis-infected bloodmeal to appear positive for infection, despite being a noncompetent vector. Therefore, vector competence must be assessed, and additional studies will be required to discern the biological significance of R. felis infection in these various arthropod hosts.

Table 1.

Prevalence of R. felis in institutional and commercial flea colonies

Institution	Year assessed	R. felis prevalence	Reference
Ag Research Consultantsa	1994	70%	Higgins et al. 1994
American Biological Supply	1990	0%	Adams et al. 1990
	1992	0%	Azad et al. 1992
Elward Labs (EL)	1992	100%	Azad et al. 1992
	1994	83%	Higgins et al. 1994
	2007	0%	Pornwiroon et al. 2007
Flea Data	2001	20–100%b	Raoult et al. 2001
Heska	1998	0%	Noden et al. 1998
	2008	0%	Macaluso et al. 2008
Louisiana State University (LSU)a	1994	86%	Higgins et al. 1994
	2000	65%	Wedincamp and Foil 2000
	2002	65%	Wedincamp and Foil 2002
	2007	94%	Pornwiroon et al. 2007
	2007	100%	Henry et al. 2007
	2008	96%, 67%, 35%	Reif et al. 2008
Mercka	1994	66%	Higgins et al. 1994
Nu-Era Research Farms	1994	83%	Higgins et al. 1994
Professional Laboratory Research Service (PLRS)a	1994	50%	Higgins et al. 1994
	2007	16%	Pornwiroon et al. 2007
Texas A&M Universityb	1994	43%	Higgins et al. 1994
University of California	1990	0%	Adams et al. 1990
University of Maryland at Baltimore (UMAB)	1992	0%	Azad et al. 1992
USDA, MAVERLb	1994	93%	Higgins et al. 1994

^aCreated from or supplemented with originally R. felis-infected EL Lab fleas.

^bPotentially contaminated with originally R. felis-infected EL Lab fleas.

Table 2.

Geographic distribution of R. felis in wild-caught arthropods

Country	Surveyed vector + for R. felis	Prevalence of R. felis-infection	Surveyed mammalian hosts with + R. felis vector	Reference
Algeria	Archaeopsylla erinacei and C. canis	100%	Rodents and hedgehogs	Bitam et al. 2006
Afghanistan	rodent fleas	9%	Gerbils	Marie et al. 2006
Argentina	C. felis	22.6%	Dogs	Nava et al. 2008
Australia	C. felis and Echidnophaga gallinacea	ND	Dogs and cats	Schloderer et al. 2006
Brazil	Ctenocephalides spp.	ND	Dogs and cats	Oliveira et al. 2002
	C. felis	28–80%	Dogs	Horta et al. 2005
	Ctenocephalides sp. Rhipicephalus sanguineus and Amblyomma cajennense	ND	Dogs and horses	Cardoso et al. 2006
	C. felis and C. canis	36% C.f., 19.1% C.c.	Dogs	Horta et al. 2006a
	C. felis and Polygenis atopus	41% C.f., 3–8% P.a.	Opossums, dogs and cats	Horta et al. 2007
	R. sanguineus and C. felis	ND	Dogs and horses	Oliveira et al. 2008
Canada	C. felis	18%	Cats	Kamrani et al. 2008
Chile	C. felis	70%	Cats	Labruna et al. 2007b
Croatia	Haemaphysalis sulcata	20–26%	Sheep and goats	Duh et al. 2006
Cyprus	C. felis	5.6%	Rats	Psaroulaki et al. 2006
Democratic Republic of Congo	Pulex irritans, E. gallinacea, Xenopsylla brasiliensis, Tunga penetrans	10.7%	Collected off host	Sackal et al. 2008
Ethiopia	Fleas	ND	ND	Raoult et al. 2001 a
France	C. felis	8.1%	Cats	Rolain et al. 2003
	C. felis, C. canis, A. erinacei	17% C.f., 27% C.c., 100% A.e.	Dogs and cats	Gilles et al. 2008a
Gabon	C. felis	ND	Monkey	Rolain et al. 2005
Germany	C. felis, A. erinacei	100% A.e., 9% C.f.	Dogs and cats	Gilles et al. 2008b
Indonesia	Xenopsylla cheopis	ND	Rodents and shrews	Jiang et al. 2006
Israel	C. felis	1–13%	Cats	Bauer et al. 2006
Japan	Ixodes ovatus, Haemaphysalis flava, Haemaphysalis kitasatoe	ND	None collected by flagging	Ishikura et al. 2003
Mexico	C. felis	ND	Dogs	Zavala-Velazquez et al. 2002
New Zealand	C. felis	15%	Dogs and cats	Kelly et al. 2004
Peru	C. felis	ND	Dogs	Blair et al. 2004
Portugal	A. erinacei and Ctenophtalmus sp.	3.5%	Rodent and hedgehog	DeSousa et al. 2006
South Korea	Chigger mites	ND	Wild rodents	Choi et al. 2007
Spain	ND	40%	Dog	Perez-Arellano et al. 2005
	C. felis	ND	Dogs and cats	Marquez et al. 2002
	C. felis	ND	Dogs and cats	Marquez et al. 2006
	C. felis and C. canis	28.4%	Dogs and cats	Blanco et al. 2006
Taiwan	Mesostigmata mite	ND	Rodents	Tsui et al. 2007
	C. felis	18.8%		Tsai et al. 2009
Thailand	C. felis and C. canis	4.4%	Dogs and ferret-badger	Parola et al. 2003
United Kingdom	C. felis	6–12%	Dogs and cats	Kenny et al. 2003
	C. felis	9–21%	Dogs and cats	Shaw et al. 2004
United States	C. felis	1.8%	Opossums	Williams et al. 1992
	C. felis	3.7%	Opossums	Schriefer et al. 1994b
	P. irritans	ND	Dog	Azad et al. 1997
	C. felis	3.8% 1993, 2.1% 1998	Opossums	Boostrom et al. 2002
	Anomiopsyllus nudata	0.8%	Rodents	Stevenson et al. 2005
	Carios capensis	1.5%	Brown pelican	Reeves et al. 2006
	C. felis	ND	Cats	Hawley et al. 2007
	X. cheopis	24.8%	Mice	Eremeeva et al. 2008
Uruguay	C. felis and C. canis	41%	Dogs and cats	Venzal et al. 2006

^aReported as unpublished data in this reference.

ND, not determined.

Transmission Routes of R. felis

Global dissemination of C. felis is contributing to the vast distribution of R. felis. Although R. felis has been identified molecularly in a number of arthropod species, the cat flea is currently the only arthropod associated with biological transmission of R. felis. Maintenance of R. felis in the environment is most likely a function of stable vertical transmission, through transstadial and transovarial transmission within cat flea populations (Azad et al. 1992, Wedincamp and Foil 2002). Mechanisms of possible R. felis transmission routes were postulated with the discovery of R. felis in flea reproductive tissue (Azad et al. 1992), because other rickettsial species had been shown to be vertically transmitted (Farhang-Azad et al. 1985, Azad and Beard 1998). Initial reports describing R. felis vertical transmission within flea populations used PCR to detect R. felis in freshly deposited cat flea eggs, showing that R. felis could be transovarially transmitted, a finding that correlated with the high prevalence (≈90%) in the Elward Laboratory flea colony (Azad et al. 1992). A subsequent study examined the prevalence of R. felis in newly emerged unfed adult cat fleas from eight colonies in the United States and identified R. felis infection in all colonies, with prevalence ranging from 43 to 93% (Higgins et al. 1994). The authors attributed high R. felis infection prevalence in colonized fleas to efficient vertical transmission among flea cohorts and horizontal transmission between fleas through co-feeding. Because newly emerged, unfed adult fleas were examined, acquisition of R. felis through vertical transmission is the most reasonable conclusion, because horizontal routes of transmission by fleas co-feeding or acquisition at immature stages were not definitively shown.

The most comprehensive R. felis vertical transmission study was performed by Wedincamp and Foil (2002), in which the efficiency of R. felis vertical transmission in cat fleas over 12 generations without the aid of an infectious bloodmeal or vertebrate host was described. It was reported that R. felis infection prevalence waned from 65 to 2.5% by the 12th generation in fleas fed bovine blood using an in vitro feeding system compared with the steady 65% infection prevalence for cat-fed fleas. Although it is plausible that, in the cat-fed colony, R. felis infection prevalence was boosted by occasional rickettsemias, uninfected fleas fed on these same cats did not acquire infection. R. felis transmission also was not observed by copulation or direct contact between infected and uninfected fleas. With the exception of vertical transmission among colonized cat fleas, biological transmission of R. felis by other arthropods has not been described; R. typhi and R. felis are the only pathogenic rickettsial species transovarially transmitted in arthropods other than ticks or mites (Azad et al. 1992).

Horizontal transmission of viable R. felis to juveniles by adult fleas has not been shown; however, the potential for horizontal transmission between arthropod and vertebrate or arthropod and arthropod is likely. Observed by ultrastructural studies in the salivary glands of cat fleas (Macaluso et al. 2008), saliva produced during blood feeding may provide one route of R. felis transmission to a susceptible vertebrate host. Evidence for rickettsiae transmission from flea to host through salivary secretion is supported by PCR amplification of R. felis DNA in the blood of laboratory cats exposed to R. felis-infected cat fleas and their subsequent seroconversion (Wedincamp and Foil 2000). The containment of the fleas in a feeding capsule ensured that cat hosts did not become exposed to infection during grooming procedures (e.g., ingesting infected fleas or flea feces), strongly suggesting horizontal transmission through flea feeding. Additional studies examining veterinary clinic cats infested with R. felis-infected fleas were unable to amplify R. felis DNA but could detect antibodies to R. felis, indicating possible past infection or exposure (Hawley et al. 2007, Bayliss et al. 2009). Although DNA and serological analyses do not directly address the viability of R. felis in the mammalian host, large amounts of dead rickettsiae would be required for a mammalian host to generate a strong antibody response. Transmission of rickettsiae in saliva during blood feeding has been shown with other rickettsial species (reviewed in Azad and Beard 1998) and likely is a mechanism of R. felis transmission.

Co-feeding between R. felis-infected and uninfected fleas or other susceptible arthropods poses another possible horizontal transmission route of R. felis. In this scenario, an infected flea would be able to transmit R. felis in saliva or by regurgitation of rickettsiae to an uninfected flea feeding nearby. Pathogen transmission through co-feeding has been described for numerous arthropods including Ixodes scapularis in the transmission of Borrelia burgdorferi (Patrican 1997) and in Culex species with the transmission of the West Nile virus (McGee et al. 2007). Rickettsial species, including Rickettsia massiliae in Rhipicephalus turanicus and Rickettsia rickettsii in Dermacentor andersoni, have also been documented to use co-feeding as a transmission strategy (Philip 1959, Matsumoto et al. 2005). In one study, Rickettsia-free fleas did not acquire R. felis infection after feeding with R. felis-infected fleas for 5 d on an artificial feeding system (Wedincamp and Foil 2002). Neither the presence of rickettsiae in the shared meal or the susceptibility of the uninfected fleas to infection were determined. Although R. felis transmission through co-feeding was not observed, additional studies are needed before this potential transmission route is ruled out.

Larval feeding may be another potential route for R. felis horizontal transmission. Cat flea larvae feed on adult flea feces as their main source of nutrition because 80–90% of total blood proteins are retained in adult flea feces (Silverman and Appel 1994). In addition to their regular diet, cat flea larvae lead cannibalistic lives preying on eggs and other larvae (Silverman and Appel 1994, Lawrence and Foil 2000, Hsu et al. 2002), all of which can lead to a greater success during larval and adult maturation (Lawrence and Foil 2000). Feeding on R. felis-infected feces and/or cannibalism of R. felis-infected eggs/larvae may facilitate R. felis transmission to uninfected fleas. In one study, uninfected larvae allowed to feed on PCR-positive R. felis-infected feces, eggs, and earlier-instar larvae were unsuccessful in acquiring R. felis; however, R. felis viability and persistence in flea feces and infective dose had not been assessed (Wedincamp and Foil 2002). Additional studies examining R. felis viability, persistence, and infectivity in adult flea feces are needed.

Transmission of rickettsiae to vertebrates through infected feces is a common transmission strategy used by the TG Rickettsia. The predominant mode of R. typhi transmission is through fleas feeding on rickettsemic vertebrate hosts and the subsequent shedding up to 10 d of infectious R. typhi in their feces (Azad and Beard 1998). In addition to R. typhi being horizontally transmitted to a vertebrate host through infected flea feces, vertical transmission within flea populations has also been described, although at a lower rate (Farhang-Azad et al. 1985). The ability of fleas to transmit R. typhi both horizontally and vertically suggests similar mechanisms of transmission for R. felis. Although the ability of fleas to acquire R. felis during blood feeding and then shed viable R. felis in their feces has not been fully examined, in vitro studies have shown an infectious extracellular state of R. felis (Sunyakumthorn et al. 2008). The increased availability of more sensitive detection methods will allow for detection of low-level transmission by other postulated transmission routes.

Along with Rickettsia species, numerous bacterial endosymbionts have been described in domestic and wild-caught fleas (Gorham et al. 2003, Murrell et al. 2003, Pornwiroon et al. 2007, Jones et al. 2008). Abundant endosymbionts in arthropod populations may regulate the ability of R. felis to colonize through either similar tissue tropism or nutrient competition. In ticks, the interference phenomenon, the establishment of one species of Rickettsia inhibiting the transovarial transmission of a second rickettsial species (Burgdorfer 1988, Macaluso et al. 2002), and niche competition have been described. In fleas, an examination of R. felis and R. typhi co-infection prevalence, where R. felis-infected cat fleas were fed blood containing R. typhi for 9 d, showed the ability of cat fleas to acquire dual infection; however, infection rates were at a lower prevalence than in either single infection, indicating that in cat fleas previously infection with R. felis may inhibit R. typhi infection (Noden et al. 1998). Complete inhibition of R. typhi infection in R. felis-infected fleas was not observed in all fleas, and whether or not R. felis is able to inhibit vertical transmission of R. typhi to flea progeny or inhibit transmission to a susceptible vertebrate host is not known. During the initial description of R. felis, Wolbachia were described in the ovaries of cat fleas (Adams et al. 1990); subsequently, Wolbachia species have been described in several species of fleas (Gorham et al. 2003, Rolain et al. 2003, Dittmar and Whiting 2004, Luchetti et al. 2004). The possibility of R. felis and Wolbachia interactions are especially interesting because both organisms occupy many of the same host cells (niches) and possibly compete for similar host resources. In insects, the influence of microbial interactions on Wolbachia abundance has been shown (Bordenstein et al. 2006, Goto et al. 2006). specifically in fleas, assessment of microbiota in the R. felis–free Elward Laboratory cat flea colony showed Wolbachia to be the predominant bacteria compared with decreased detection of Wolbachia in the LSU colony with an R. felis prevalence of ≈94% (Pornwiroon et al. 2007). In addition to Wolbachia, R. felis-infected arthropods have also been infected with either Bartonella clarridgeiae (Rolain et al. 2003), Bartonella henselae (Shaw et al. 2004), Bartonella quintana (Rolain et al. 2003, Marie et al. 2006), and Hemoplasma sp. (Shaw et al. 2004) or Spiroplasma, Stenotrophomonas, and Staphylococcus (Pornwiroon et al. 2007). Occupying the same cells or organs in the arthropod host suggests R. felis must be able to contend with other vertically maintained endosymbionts and microbiota acquired during feeding (Pornwiroon et al. 2007). For example, Spiroplasma negatively affected the population of Wolbachia during co-infection of Drosophila melanogaster (Goto et al. 2006). The biological impact of the interspecific relationship of co-infecting rickettsial and other bacterial species in the flea host requires further examination.

Bacterial species, such as Wolbachia and Spiroplasma, are readily able to manipulate their host’s biology, e.g., cytoplasmic incompatibility, male-killing, feminization, to facilitate their transmission (Werren 1997, Dobson et al. 1999, Fry et al. 2004, Montenegro et al. 2006, Duron et al. 2008). The fitness cost associated with R. felis infection of fleas is not clear and needs to be examined because other pathogenic Rickettsia, such as R. rickettsii and R. prowazekii, have been shown to negatively impact the fitness of their host arthropod (Snyder and White 1945, Burgdorfer and Brinton 1975, Niebylski et al. 1999). Like Wolbachia and Spiroplasma, other insect-specific Rickettsia species can negatively impact the fitness of their arthropod hosts, such as the rickettsial species in the twospotted lady beetle, Adalia bipunctata, which is associated with male embryo mortality (Werren et al. 1994, Perlman et al. 2006). The ability of R. felis to influence the biology of the arthropod vector may delineate which arthropod species are competent vectors (Pornwiroon et al. 2007).

Regulation of R. felis transmission within a vector population has also been hypothesized. One of the most used R. felis–infected cat flea colonies is at LSU, with studies reporting R. felis infection prevalence ranging from 35 to 100% (Higgins et al. 1994, Wedincamp and Foil 2000, Henry et al. 2007, Pornwiroon et al. 2007, Reif et al. 2008). In a recent study, the prevalence of R. felis infection in a cat flea colony was inversely correlated to R. felis infection loads in individual fleas (Reif et al. 2008). In three trials, flea- and Rickettsia-naïve cats were exposed to ≈100 cat fleas, with varying prevalence of R. felis infection, for 9 d, and daily R. felis infection load and infection density were assessed in individual flea lysates. When colony R. felis infection prevalence was greatest, R. felis infection load in individual fleas was low. However, as colony R. felis infection prevalence decreased, individual flea R. felis infection loads increased by as much as 4.75-fold, suggesting that R. felis infection is regulated in the vector population. Regulating infection prevalence and load in a vector population may represent a maintenance strategy for R. felis infection in LSU colonized cat fleas, whereby the prevalence of waning infection signals greater infection burdens in individual fleas in an effort to facilitate transmission in a higher percent of progeny or in a susceptible mammalian host. Further studies examining the population dynamics of R. felis are required to identify the mechanisms of R. felis infection at the vector population level.

Rickettsia felis Infection of Mammals and Humans

Serological-based studies have tried to define the prevalence and incidence of R. felis infection in specific populations of domestic and wild animals. A definitive mammalian host(s) has not been identified in the epidemiology of R. felis and may vary by geographic location and distribution of arthropod vectors. Several peri-domestic species associated with the cat flea vector have been implicated, including cats (Higgins et al. 1996, Boostrom et al. 2002, Case et al. 2006, Labruna et al. 2007b, Bayliss et al. 2009); dogs (Richter et al. 2002, Oteo et al. 2006); and opossums (Williams et al. 1992, Schriefer et al. 1994b, Boostrom et al. 2002), all of which have been seropositive for R. felis infection outside of laboratory experiments (Table 3). Serological testing of domestic animals in areas with R. felis-infected fleas showed that infestation of local animals with R. felis–infected fleas did not always correlate to seropositive animals (Williams et al. 1992, Horta et al. 2005, Pinter et al. 2008). Examination of cat serum from veterinary clinics by indirect immunofluorescent assay (IFA) found five febrile cats and two control cats reactive against R. felis antigen (Bayliss et al. 2009); however, in a similar study, R. felis DNA could not be amplified from cats naturally infested with R. felis-infected cat fleas (Hawley et al. 2007). A study in Chile examined fleas from local cats by PCR and cats by serology for R. felis infection and reported that ≈70% of fleas were positive for R. felis infection and that 70% of cats had a titer ≥1:64 for R. felis. Under laboratory conditions, previously flea- and Rickettsia-naïve cats were infested with R. felis-infected cat fleas and most cats either continuously or intermittently exposed to R. felis-infected cat fleas seroconverted by 4 mo after infestation (Wedincamp and Foil 2000). Although these studies showed horizontal transmission from fleas to cats, transmission from mammal to arthropod has not been shown, and doubts have been raised about whether the latter occurs (Weinert et al. 2009). However, only occasional horizontal transmission from mammal to arthropod may be needed to enhance or maintain R. felis in nature (Weinert et al. 2009). The lack of a description of a definitive mammalian host impedes epidemiological studies of R. felis.

Table 3.

Detection of R. felis infection in mammals

Country	Mammalian hosts positive for R. felis (% positive)	Tissue tested	Assay	Reference
Chile	Cat 73%a	Serum	IFA	Labruna et al. 2007a
Germany	Dog NA	Serum	MIF, Western blot	Richter et al. 2002
Peru	Rat NA	Blood	ND	Blair et al. 2004 b
Spain	Dog NA	Serum	PCR gltA, ompA, ompB	Oteo et al. 2006
Taiwan	Cat¥	Serum	IFA	Tsai et al. 2009
United States	Opossums 30%a	Serum	DFA	Williams et al. 1992
	Opossums 33%a	Blood	PCR 17 kDa, gltA, IFA	Schriefer et al. 1994b
	Cat 8%	Serum	ND	Higgins et al. 1996 b
	Cat 81a,c	Serum	Serology, PCR	Wedincamp and Foil 2000
	Cat 100c	Serum	IFA	Wedincamp and Foil 2002
	Opossums 33% 1993 and 22% in 1998a	Serum	IFA	Boostrom et al. 2002
	Cat <15%	Serum	IFA	Boostrom et al. 2002 b
	Cat 11%	Serum	IFA	Case et al. 2006
	Cat NAa	Blood	PCR gltA, ompB	Hawley et al. 2007
	Cat 3.9%	Serum	IFA	Bayliss et al. 2009

^aInfested with R. felis-infected C. felis.

^bReported as unpublished data in this reference.

^cInfected under experimental conditions.

NA, not available.

Human cases of R. felis infection have been reported in 12 countries worldwide (Fig. 1). The first human case of R. felis infection was reported from Texas in 1994 (Schriefer et al. 1994a); since then, additional human cases have been reported from Spain (Perez-Arellano et al. 2005, Bernabeu-Wittel et al. 2006, Nogueras et al. 2006, Oteo et al. 2006), Germany (Richter et al. 2002), France (Raoult et al. 2001), Egypt (Parker et al. 2007), Brazil (Raoult et al. 2001, Galvao et al. 2006), Mexico (Zavala-Velazquez et al. 2000, Galvao et al. 2006, Zavala-Velazquez et al. 2006), Thailand (Parola et al. 2003), Taiwan (Tsai et al. 2008), South Korea (Choi et al. 2005), Laos (Phongmany et al. 2006), and Tunisia (Znazen et al. 2006) (Table 4). Interestingly, only in Tunisia and Egypt were there reported human R. felis infection cases and no additional reports of R. felis–positive arthropods. Serosurveys to detect R. felis infection in humans have also been conducted. In Spain, serum samples tested by IFA to survey two populations of people for past exposure to R. felis reported up to 7.1% of people had antibodies to R. felis (Bernabeu-Wittel et al. 2006, Nogueras et al. 2006). Misdiagnosis of R. felis infection as another rickettsial infection, such as murine typhus or Mediterranean spotted fever, is also common (Schriefer et al. 1994a, Raoult et al. 2001), implying a higher prevalence of R. felis infection in areas where multiple rickettsial species overlap.

Table 4.

Human cases and associated clinical manifestations of flea-borne rickettsiosis caused by R. felis

Country	No. cases	Rash	Fatigue	Acute				Other	Reference
				Fever	Headache	Arthralgia	Myalgia
Brazil	2	2/2	ND	2/2	ND	ND	ND	Stupor 2/2, coma 1/2, thrombocytopenia 1/2, vomiting 2/2, elevated liver enzymes 1/2	Raoult et al. 2001
Egypt	1	ND	ND	1/1	ND	ND	ND	ND	Parker et al. 2007
France	2	2/2	ND	2/2	ND	ND	ND	ND	Raoult et al. 2001
Germany	1	1/1	1/1	1/1	1/1	ND	ND	Splenomegaly, elevated liver enzymes	Richter et al. 2002
Laos	1	ND	ND	1/1	ND	ND	ND	ND	Phongmany et al. 2006
Mexico	3	3/3	3/3	2/3	2/3	1/3	2/3	Nuchal pain 1/3, leucopenia 1/3, leucocytosis 1/3, anemia 1/3, thrombocytosis 1/3, abdominal pain 2/3, nausea 1/3, vomiting 1/3, diarrhea 1/3, photophobia 2/3, hearing loss 1/3, conjunctivitis 1/3	Zavala-Velazquez et al. 2000
	1	0/1	ND	1/1	ND	ND	ND	Skin lesions, cough, pulmonary infiltrates	Zavala-Velazquez et al. 2006
South Korea	3	ND	ND	3/3	ND	ND	ND	ND	Choi et al. 2005
Spain	5	0/5	ND	5/5	4/5	4/5	4/5	Dry cough 3/5, abdominal pain 1/5, elevated liver enzymes 5/5, conjunctivitis 1/5	Perez-Arellano et al. 2005
	2	1/2	ND	2/2	ND	2/2	ND	Malaise 2/2, elevated liver enzymes 2/2	Oteo et al. 2006
Taiwan	1	0/1	1/1	1/1	1/1	0/1	ND	Chills, pyuria, acute polyneuropathy	Tsai et al. 2008
Thailand	1	0/1	ND	1/1	1/1	ND	ND	Chills, hepato-splenomegaly, leukopenia, vomiting	Parola et al. 2003
Tunisia	8	8/8	ND	8/8	ND	ND	ND	Peripheral adenopathy 2/8, interstitial pneumopathy 1/8	Znazen et al. 2006
USA	1	0/1	ND	1/1	ND	ND	ND	ND	Schriefer et al. 1994a

ND, not determined.

The clinical disease caused by R. felis infection has been designated several names including flea-borne spotted fever, cat flea typhus, and R. felis rickettsiosis. Clinical manifestations of R. felis infection are similar to symptoms caused by other rickettsial organisms and can range in severity (Table 4). Typical symptoms can include fever, rash, headache, myalgia, and eschar at the bite site (Brouqui et al. 2007). More severe symptoms can result from visceral (abdominal pain, nausea, vomiting, and diarrhea) and neurologic (photophobia and hearing loss) involvement (Zavala-Velazquez et al. 2000, Galvao et al. 2006). The variable presentation of clinical disease can make diagnosis difficult (Azad and Radulovic 2003), and refinement of the full spectrum of clinical disease associated with R. felis infection will expedite accurate diagnoses.

Detection and Diagnosis of R. felis Infection in Vertebrate and Invertebrate Hosts

Detection of R. felis infection can be accomplished by serological or molecular diagnosis (Tables 3 and 5). Serological assays, such as direct immunofluorescent assay (DFA), can be used to detect the presence of rickettsiae in humans and animals; however, periods of rickettsemia may quickly subside, making detection difficult. Rickettsial infection is routinely diagnosed using serological methods that use the human or animals’ antibody response to rickettsial antigens. The most common methods of serological diagnosis include microimmunofluorescence (MIF), IFA, enzyme-linked immunosorbent assay (ELISA), Western blotting, and cross-adsorption assays. Despite many serological assays being rickettsial group specific, multiple serologic tests are often required to confirm diagnosis of the specific infecting rickettsiae because cross-reactivity between rickettsial species is common. Also, because these assays rely on development of specific antibodies that appear at the earliest 7 d after the onset of symptoms, these methods are more commonly used to confirm suspected rickettsial diagnosis. R. felis has been detected serologically in humans and domestic animals; however, diagnosis through these methods can be a challenge because of lag-time in diagnosis until specific antibody development; cross-reactivity with other rickettsial species; variable cut-off titer values; and reagent availability. Further compounding the issues of accurate diagnosis is the clinical similarity of many rickettsial diseases. In areas where R. felis-infected arthropods (specifically infected fleas) were identified, several retrospective studies, using serological diagnosis, have been conducted to determine R. felis infection prevalence among local human and animal populations (Bernabeu-Wittel et al. 2006, Nogueras et al. 2006, Pinter et al. 2008). Ideally, serological results should be verified by molecular tests, although this is not always possible, especially when looking for evidence of past infections.

Table 5.

Assays used to diagnose human cases of flea-borne rickettsiosis

Country	No. cases	Diagnostic method	Tissue tested	Reference
Brazil	2	MIF, nested PCR gltA	Serum	Raoult et al. 2001
	1	Diagnostic methods not provided	NA	Galvao et al. 2006
Egypt	1	PCR 17 kDa	Blood	Parker et al. 2007
France	2	MIF	Serum	Raoult et al. 2001
Germany	1	MIF, Western blot, nested PCR PS120 protein gene	Serum	Richter et al. 2002
Laos	1	MIF, Western blot, cross-adsorption	Serum	Phongmany et al. 2006
Mexico	3	Skin biopsy, PCR 17 kDa, serological test	Serum, skin	Zavala-Velazquez et al. 2000
	1	IFA, PCR 17 kDa	Whole blood	Zavala-Velazquez et al. 2006
	2	Diagnostic methods not provided	NA	Galvao et al. 2006
Spain	5	MIF, Western blot, cross adsorption	Serum	Perez-Arellano et al. 2005
	2	PCR gltA, ompA, ompB	Serum	Oteo et al. 2006
	33	IFA	Serum	Bernabeu-Wittel et al. 2006
	7	IFA	Serum	Nogueras et al. 2006
South Korea	3	Multiplex-nested PCR rompB, gltA, sequencing and RFLP	Serum	Choi et al. 2005
Taiwan		qPCR groEL, 17 kDa, ompB, IFA	Whole blood	Tsai et al. 2008
Thailand	1	IFA, Western blot, cross adsorption	Serum	Parola et al. 2003
Tunisia	8	IFA, Western blot, cross adsorption	Serum	Znazen et al. 2006
United States	1	PCR 17 kDa, RFLP, Southern blot	Whole blood	Schriefer et al. 1994a
	3	ELISA	Serum	Wiggers et al. 2005

NA, not available.

In general, molecular diagnosis of rickettsial infection offers greater sensitivity and specificity and can also be made at the time of illness (e.g., biopsy an eschar). Molecular detection usually involves PCR assays to first determine rickettsial infection, followed by restriction fragment-length polymorphism (RFLP) analysis, dot-blot hybridization, or sequencing of rickettsial genes to identify the specific infecting rickettsial species. Genes commonly used for rickettsial detection include the Rickettsia genus specific 17-kDa antigen gene, the 16S rRNA gene, the citrate synthase gene (gltA), ompB, and ompA (this gene is truncated in R. felis). Molecular detection is common in arthropod surveys of R. felis infection, and assessment of individual arthropods can help elucidate the prevalence of R. felis in specific areas. For example, in a survey of fleas collected from rodents and a hedgehog in Portugal, R. felis was detected and differentiated from R. typhi in Archaeopsylla erinacei and Ctenophtalmus sp. by amplification and sequencing of gltA and ompB (DeSousa et al. 2006). Likewise, in Israel, multiple genotypes of R. felis were identified in cat fleas by detection and sequencing of the 17-kDa antigen gene, gltA, ompA, and ompB (Bitam et al. 2006), which shows the specificity of molecular assays in differentiating not only rickettsial species but also in differentiating between genotypes of a single rickettsial species.

Molecular detection assays have also been used to describe R. felis infection in mammals including humans. In a survey of opossums infested with R. felis-infected cat fleas, blood samples were collected, and sequencing of amplified gltA and 17-kDa gene fragments showed R. felis infection in the opossum (Schriefer et al. 1994b). In Germany, R. felis infection was diagnosed in a person and their dog by amplification from serum samples and sequencing of the R. felis PS120 protein gene (Richter et al. 2002). More sensitive molecular tools have been used to identify specific rickettsial infections. Development of qPCR assays able to detect as little as 1–10 specific rickettsial gene copies allow for detection of low-level R. felis infections (Henry et al. 2007, Reif et al. 2008). The development of species-specific probes allows for the immediate differentiation and diagnosis of specific rickettsial infection (Rozmajzl et al. 2006, Henry et al. 2007). Although molecular assays may be more sensitive and specific, limits in machine and reagent availability are common. Molecular assays also only detect current rickettsial infections, whereas serological-based assays can diagnose evidence of past infections. Currently, there is no standard protocol for physicians to diagnose R. felis infection in patients, and despite circumstantial evidence of infection, R. felis has not been isolated from a human case of infection.

Identification of arthropods and mammals infected or able to be infected with R. felis can serve as a platform for developing methods to study R. felis-host interactions in more detail. R. felis–host interactions can be examined using both in vitro and in vivo systems. Several cell culture systems for in vitro examination of R. felis have been established and provide a valuable tool for examining R. felis biology (Table 6). R. felis can be successfully cultivated in a variety of vertebrate- and invertebrate-derived cell lines, maintained at or below 32°C. Vertebrate-derived cell lines that support R. felis growth include Vero (Radulovic et al. 1995, LaScola et al. 2002), L929 (Radulovic et al. 1995), and XTC-2 (Raoult et al. 2001, LaScola et al. 2002) cells. Arthropod cell lines capable of supporting R. felis growth include ISE6 (Pornwiroon et al. 2006), C6/36 (Horta et al. 2006b), Aa23 (Sakamoto and Azad 2007), and Sua5B (Sakamoto and Azad 2007) cells.

Table 6.

Cell culture systems for R. felis studies

	Cell line	Type of host cell	Temperature	R. felis strain	Reference
Vertebrate	Vero	African green monkey kidney	34°C	NA	Radulovic et al. 1995
			28–32°C	Marseille-URRWFXCal2	Raoult et al. 2001
	L929	Murine fibroblast	34°C	NA	Radulovic et al. 1995
			34°C	LSU	Sakamoto and Azad 2007
	XTC-2	Toad tadpole Xenopus laevis	28°C	Marseille-URRWFXCal2	Raoult et al. 2001
Invertebrate	ISE6	Tick Ixodes scapularis	32°C	LSU	Pornwiroon et al. 2006
	C6/36	Mosquito Aedes albopictus	25°C	Pedreira	Horta et al. 2006b
	Aa23	Mosquito Aedes albopictus	RT	LSU	Sakamoto and Azad 2007
	Sua5B	Mosquito Anopheles gambiae	RT	LSU	Sakamoto and Azad 2007

NA, not available; RT, room temperature.

The development of bioassays using arthropods and mammals susceptible to R. felis are crucial to study R. felis biology in vivo. Established cat flea colonies infected with R. felis, such as the LSU C. felis colony, provide a valuable tool to study, in vivo, the biology (infection kinetics, tissue tropism, etc.) of R. felis within its primary host and vector. Within the flea host, the interspecific relationship of R. felis with other flea bacterial endosymbionts that may affect the establishment or transmissibility of R. felis, and vector competence of individual fleas, can also be addressed. As of yet, no in vivo mammalian model of R. felis infection has been established. Establishment of an animal model will facilitate a more accurate understanding of R. felis, defining pathogenesis in both animals and humans and delineating the transmission of R. felis from arthropod to animal, and vice versa. Identification of key molecules in both the arthropod and mammalian hosts that aid in establishment and maintenance of R. felis infection and subsequent transmission could aid in the development of transmission control strategies within vector populations and vaccine targets for human or animal infection.

In summary, R. felis has been detected molecularly in numerous arthropod species around the world; the cat flea is currently the only defined biological vector and reservoir. Within the flea, R. felis infection is disseminated, having been identified in the midgut, ovaries, and salivary glands. Under laboratory conditions, maintenance of R. felis in flea cohorts is primarily by vertical (transovarial and transstadial) transmission. Seroconversion and detection of R. felis DNA in cats infested with R. felis-infected fleas provides evidence for horizontal transmission; however, definitive transmission of viable organisms between mammals and arthropods has not been shown. As the etiologic agent of flea-borne rickettsiosis, human cases of R. felis infection have been documented in 12 countries around the world, with patients presenting with typical symptoms of rickettsial infection including: fever, rash, and myaliga. Several serological, molecular, and biological assays are available to study the biology and ecology of R. felis.

Despite being transmitted by a cosmopolitan vector and listed as an emerging infectious disease, our current understanding of R. felis biology is incomplete, with several issues that need to be addressed. Primarily, the transmission mechanisms of R. felis, specifically identification of competent arthropod vectors and vertebrate reservoirs, defining the alternate routes of R. felis transmission, and developing accurate detection methodology and a definition of disease, should receive further attention. Every year there are reports of arthropod, animal, and human cases of R. felis infection from additional countries. Continued development and implementation of molecular tools to delineate the specific roles of the arthropod and vertebrate host in the R. felis transmission cycle will guide accurate risk assessment and controls measures.