Rickettsia felis in cat fleas, Ctenocephalides felis parasitizing opossums, San Bernardino County, California

Abstract

Los Angeles and Orange Counties are known endemic areas for murine typhus in California; however, no recent reports of flea-borne rickettsioses are known from adjacent San Bernardino County. Sixty-five opossums (Didelphis virginiana) were trapped in the suburban residential and industrial zones of the southwestern part of San Bernardino County in 2007. Sixty out of 65 opossums were infested with fleas, primarily cat fleas, Ctenocephalides felis (Bouché, 1835). The flea minimum infection rate with Rickettsia felis was 13.3% in pooled samples and the prevalence was 23.7% in single fleas, with two gltA genotypes detected. In spite of historic records of murine typhus in this area, no evidence for circulation of R. typhi in fleas was found during the present study. Factors contributing to the absence of R. typhi in these cat fleas in contrast to its presence in cat fleas from Orange and Los Angeles Counties are unknown and need to be investigated further in San Bernardino County.

Introduction

Rickettsia felis is an obligate intracellular bacterium that causes cat flea rickettsiosis (Reif & Macaluso, 2009). While most commonly associated with the cat flea, Ctenocephalides felis (Bouché), R. felis has been detected in other flea species, ticks, mites and even a non-hematophagous arthropod, the book louse, Liposcelis bostrychophila (Adams et al., 1990; Reif & Macaluso, 2009; Thepparit et al., 2011). Whether arthropods other than C. felis are competent vectors for R. felis or the PCR detection just reflects rickettsial DNA passively acquired with a bloodmeal from a rickettsiemic host is unknown. Infected fleas are most frequently found on cats, dogs and peridomestic opossums (Didelphis spp.); however, fleas containing R. felis DNA have also been collected from rodents, hedgehogs, horses, sheep, goats, gerbils and monkeys (Perez-Osorio et al., 2008; Reif & Macaluso, 2009). Rickettsia felis maintains a persistent infection in C. felis through transovarial and transstadial transmission (Azad et al., 1992; Hirunkanokpun et al., 2011). Cats exposed to R. felis-positive C. felis could both seroconvert and may exhibit PCR positive blood, suggesting that horizontal transmission can also occur (Wedincamp & Foil, 2000), but whether cats or other animals serve as a reservoir of R. felis needs investigation.

At the time of its original discovery, R. felis was thought to be another pathogen, similar to Rickettsia typhi, which causes murine typhus, and circulates in cat fleas and is associated with opossums (Adams et al., 1970, 1990; Sorvillo et al., 1993). However, data based on molecular detection and identification of R. felis worldwide indicate that its circulation does not necessarily coincide with that of R. typhi, which is found far less frequently (Eremeeva et al., 2008; Perez-Osorio et al., 2008; Sackal et al., 2008; Reif & Macaluso, 2009; Richards et al., 2010). The first and only US association of R. felis with human illness was documented in 1991 when R. felis DNA was detected in the blood of a patient from Texas suspected to have murine typhus (Schriefer et al., 1994). As of 2011, other human cases as a consequence of R. felis infection were reported from 15 other countries (Perez-Osorio et al., 2008; Richards et al., 2010; Socolovschi et al., 2010). These cases manifested with non-specific symptoms of varying incidences of rash, fatigue, fever, headache, arthralgia and myalgia, indicating that disease as a result of R. felis may be significantly under-recognized and under-diagnosed in the U.S.A., as is true of other rickettsioses.

In the present study, we describe extensive surveillance for flea-borne pathogens in cat fleas infesting opossums in San Bernardino County, CA, and discuss the public health relevance of these findings.

Materials and methods

Site description and sampling method

The study was conducted in the southwestern corner of San Bernardino County, CA. This mostly suburban residential and industrial area is bounded by Riverside, Los Angeles and Orange Counties to the south, west and southwest, respectively; and by the San Bernardino Mountains to the north. Twenty-seven sites were selected based on past reports of opossum presence. Two to three large live-capture wire traps per site were baited with cat food and fruit, and set in late afternoon every other week from March to October 2007. Traps were checked the following morning for animal capture, with non-target species released immediately. Trapped opossums were anaesthetized and the ectoparasites were removed, identified to species using standard taxonomic keys, frozen at − 70 °C and shipped on dry ice to the Centers for Disease Control and Prevention (CDC, Atlanta, GA, U.S.A,) for testing.

Preparation of DNA and testing for flea-borne rickettsiae

Fleas were processed individually if 10 or fewer fleas were found or in pools of 2 fleas if 11 or more fleas per opossum were collected. A maximum of 50 fleas per animal were tested for the latter group. Fleas were surface disinfected, frozen in liquid nitrogen and pulverized using Konte pestles (Kimble-Kontes, Vineland, NJ, U.S.A.) (Abramowicz et al., 2011). The flea powder was re-suspended in 160 μL Nuclei lysis solution (Promega, Madison, WI, U.S.A.) supplemented with 40 μL 0.5 _M EDTA and 20 μL of 20 mg/mL Proteinase K (Qiagen, Valencia, CA, U.S.A.), and incubated overnight at 56 °C. DNA was extracted using the Wizard SV 96 Genomic DNA purification system (Promega) and a Biomek 2000 Robotic Workstation (Beckman Coulter, Fullerton, CA, U.S.A.), and was eluted with 120 μL of sterile nuclease-free water and stored at 4 °C.

Flea DNA was tested for the presence of R. typhi and R. felis DNA using the Brilliant QPCR reagent kit (Stratagene, Cedar Creek, TX, U.S.A.) and a species-specific TaqMan for the citrate synthase (gltA) gene of Rickettsia as described previously (Karpathy et al., 2009). Positive control plasmids were constructed by cloning target fragments from R. typhi strain Wilmington and R. felis strain LSU into pCR2.1 (Invitrogen, Carlsbad, CA, U.S.A.), according to the manufacturer’s instructions. All reactions were run with a non-template control in which sterile water was substituted for the DNA sample. Amplicons were sequenced using an Applied Biosystems Big Dye Terminator v3.1 Cycle Sequencing Kit and purified using the ABI BigDye XTerminator kit (Applied Biosystems, Carlsbad, CA, U.S.A.). Assembly and sequence analysis was performed using the Sequencher 4.8 software (Gene Codes, Ann Arbor, MI, U.S.A.) and analysed using MEGA4 (Tamura et al., 2007).

Data analysis

The flea index was calculated as an arithmetic average of flea number per opossum. Prevalence was estimated as percentage of the total number of samples in which R. felis was detected. Minimum infection rates (MIR) were calculated as percentage of a ratio between the total number of flea pools positive for R. felis and the total number of fleas tested. Statistical assessment was done using Student’s t-test. Confidence intervals (CI) for MIR and prevalence rates were calculated using the Wilson score method without continuity correction (Newcombe, 1998).

Results

During the study period, 65 opossums were trapped at the 27 sites (Fig. 1A). Among them, 34 (26 adults and 8 subadults) were females and 31 (8 adults and 23 subadults) were males. The largest collections of 17 and 12 opossums were trapped in May and September, respectively. Four to eight animals were trapped at 4 sites, whereas one to three opossums were collected from the other 23 sites during the same trapping period (Fig. 1A).

Fig. 1.

Geographical map of the area studied. (A) Distribution of the trapping sites, opossum collection data and total number of fleas collected at each site. Triangles indicate collection sites where four to eight animals were trapped per site; one to three animals were trapped in other locations. Inset illustrates the location of the studied area relative to Los Angeles and Orange Counties of California; (B) the distribution of fleas and prevalence and/or average minimum infection rates of R. felis in cat fleas per animal collected at each site with fleas. Arrows indicate sites of fleas infected with R. felis genotype RF2125.

A total of 1870 fleas were collected from 60 (92.3%) of the 65 opossums trapped (1–289 fleas per animal, flea index 31.2). Male opossums had a higher although not statistically significant flea index of 34.4 than females (flea index 24.3, t-test P = 0.56). Among the fleas collected, 1867 were C. felis, 2 were human flea, Pulex irritans Linnaeus, 1758 and 1 was sticktight flea, Echidnophaga gallinacea Westwood, 1875. In total, 1258 fleas were tested, including 570 pools of 2 fleas from 33 opossums, and 118 samples of 1 flea from 27 opossums. One hundred and fifty-two out of 570 pools (minimum infection rate 13.3%, 95% CI 11.48–15.43) and 28 out of 118 single flea samples (prevalence 23.7%, 95% CI 16.96–32.16) were tested positive for R. felis DNA by gltA TaqMan. The distribution of tested fleas with their average MIR or prevalence is displayed in Fig. 1B. Rickettsia felis-infected fleas were present on 44 out of 60 flea-infested opossums (73%); C. felis from 13 of the 27 opossums infested with < 10 fleas were not infected whereas only 4 of the 33 heavily infested animals did not have infected flea pools. The E. gallinacea and one of the two P. irritans also tested positive for R. felis DNA. None of the fleas tested positive for R. typhi DNA.

All PCR amplicons were sequenced and two genotypes of R. felis were identified based on their gltA nucleotide sequences. One was identical to that of R. felis URRWXCal2 (NC_007109), whereas the other type had four mutations at the nucleotide positions 26(C/T), 83(T/G), 170(C/T) and 179(G/A) and was identical to RF2125 (AF516333). Genotype URRWXCal2 was dominant and found in 93.4% of 180 samples sequenced. The RF2125 genotype was found in 12 C. felis from 9 opossums trapped in 8 sites from 5 different administrative affiliations (Fig. 1B). Fleas carrying both genotypes of R. felis were collected from the same animal, including two adult females, three adult males and four young males.

Discussion

Murine typhus is endemic in the state of California with human cases primarily reported from Los Angeles and Orange Counties (Civen & Ngo, 2008). San Bernardino County is of particular interest as it is directly adjacent to these counties. It has similar ecological characteristics; however, no recent reports of human exposure are known despite historic records of both murine typhus and R. typhi in rats from San Bernardino (Beck et al., 1944; Beck & Van Allen, 1950). We evaluated the presence and distribution of opossums and their associated ectoparasites, and determined the prevalence of the flea-borne rickettsiae, R. typhi and R. felis in San Bernardino County. Sixty-five D. virginiana were trapped in the southwestern portion of the county where they live in relatively close contact with humans in a predominantly urban and suburban environment in the most densely inhabited section of the largest county in the U.S.

The prevalence of R. felis in cat fleas ranges from 1% to 45.8% and has been reported as high as 100% in small batches of samples tested (Reif & Macaluso, 2009). The prevalence determined in the present study is within the same range; however, we observed a significant difference between MIR of pooled fleas and prevalence determined by testing individual samples. Therefore, it may be that the MIR underestimates the prevalence of R. felis in pools, possibly because the greater amount of DNA in pools or other contaminants may inhibit the TaqMan assay to some degree (Demeke & Jenkins, 2010). If one corrects for a possible underestimation of prevalence of infection in flea pools from the 9 animals with greater than 40% of their pools infected (both fleas are actually infected rather than only one) and the remaining 24 animals with less than 33% infected pools (one of two fleas actually infected), the prevalence of R. felis in pooled fleas is 20.2%. This is within the CI estimated from fleas that were tested individually.

It is also unknown if the determined prevalence and infection rate of R. felis in cat fleas parasitizing opossums in San Bernardino County is sufficient for efficient transmission to humans. Clearly, opossums which are highly flea infested are more likely to carry infected fleas than poorly infested animals. Furthermore, the significance of co-circulation of two genetic types of R. felis, URRWXCal2 (or Cal2) and RF2125 reported here for the first time in cat fleas from California is not well understood. The RF2125 type was first found in 2 dog flea, C. canis (prevalence 5.9%, n = 34) from Thai-Myanmar border and then was detected in other flea species from Europe, Northern Africa, Uruguay and Eastern U.S. (Parola et al., 2003; Bitam et al., 2006; Loftis et al., 2006; Venzal et al., 2006; Nelder et al., 2009). Beside cat and dog fleas, the RF2125 genotype of R. felis was detected in E. gallinacea (100%, n = 12), hedgehog flea, Archeopsylla erinacei (100%, n = 4) and P. irritans (Bitam et al., 2006; Loftis et al., 2006; Hornok et al., 2010).

In our previous studies R. felis was shown to co-circulate with R. typhi (Eremeeva et al., 2008; Karpathy et al., 2009; Abramowicz et al., 2011), similarly to when it was first discovered in its sylvatic cycle in CA (Williams et al., 1992). However, in those sites R. felis infection was not easily recognized as a separate flea-borne cause of human illness as current diagnosis relies on serological testing with cross-reactive R. typhi antigen, and is therefore reported as murine typhus (Green et al., 2011). Whether R. typhi also circulates in a classic rat flea, Rattus norvegicus–Xenopsylla cheopis cycle in San Bernardino County is unknown as only opossums and their ectoparasites were available for this study. The roof rat, Rattus rattus, appears to be more common in this region; however, they are most frequently infested with mouse flea, Leptosylla segnis (Schoenherr) and Northern rat flea, Nosopsylla fasciatus (Bosc, 1800) with seasonal variations (Schwan et al., 1985). Although L. segnis can be infected with R. typhi, it is unlikely to cause human infection as it rarely bites people; the potential of N. fasciatus for transmission of R. typhi needs further evaluation (Traub et al., 1978). Further study focusing on rats and rat fleas may provide more insight into the dynamics of rickettsial circulation and the occurrence of rickettsial diseases in San Bernardino County. By contrast, suggestions that opossums and associated cat fleas can serve as an alternative animal host for R. typhi and sylvatic reservoir of murine typhus (Civen & Ngo, 2008; Adjemian et al., 2010) may not be completely justified as reliable findings of those associations in the absence of active rodent cycles of R. typhi are infrequent (Williams et al., 1992; Boostrom et al., 2002; Karpathy et al., 2009) and need further evaluation.

Ctenocephalides felis is a very common flea in pets, and cats can easily acquire them from opossums entering urban and suburban properties. As about 32.4% of U.S. households have one or more cats (AVMA, 2007) awareness of the potential risk of flea-borne rickettsial pathogens should be an important component in the disease prevention programme.