Detection of Rickettsia Species in Fleas Collected from Cats in Regions Endemic and Nonendemic for Flea-Borne Rickettsioses in California

Abstract

Rickettsia typhi, transmitted by rat fleas, causes most human flea-borne rickettsioses worldwide. Another rickettsia, Rickettsia felis, found in cat fleas, Ctenocephalides felis, has also been implicated as a potential human pathogen. In the continental United States, human cases of flea-borne rickettsioses are reported primarily from the southern regions of Texas and California where the cat flea is considered the principal vector. In California, more than 90% of locally acquired human cases are reported from suburban communities within Los Angeles and Orange counties despite the almost ubiquitous presence of cat fleas and their hosts throughout the state. The objective of this study is to assess the presence and infection rate of Rickettsia species in cat fleas from selected endemic and nonendemic regions of California. Cat fleas were collected from cats in Los Angeles County (endemic region) and Sacramento and Contra Costa counties (nonendemic region). Sequencing of 17 amplicons confirmed the presence of R. felis in both the endemic and non-endemic regions with a calculated maximum likelihood estimation of 131 and 234 per 1000 fleas, respectively. R. typhi was not detected in any flea pools. Two R. felis-like genotypes were also detected in fleas from Los Angeles County; Genotype 1 was detected in 1 flea pool and Genotype 2 was found in 10 flea pools. Genotype 1 was also detected in a single flea pool from Sacramento County. Results from this study show that R. felis is widespread in cat flea populations in both flea-borne rickettsioses endemic and nonendemic regions of California, suggesting that a high prevalence of this bacterium in cat fleas does not predispose to increased risk of human infection. Further studies are needed to elucidate the role of R. felis and the two R. felis-like organisms as etiologic agents of human flea-borne rickettsioses in California.

Introduction

R ickettsia species (Class: Alphaproteobacteria, Order: Rickettsiales) are gram-negative, obligate intracellular bacteria capable of invading both the midgut epithelial–cells of arthropods and the endothelial cells lining the blood vessels of vertebrates (Eisen and Gage 2012). Replication of the bacteria in vertebrate host cells leads to cell lysis, releasing bacteria into the bloodstream, where they can be acquired by hematophagous arthropods. Rickettsia typhi, the causative agent of murine typhus in humans, was the first species isolated and linked to human cases of flea-borne rickettsioses worldwide. R. typhi is transmitted by the Oriental rat flea, Xenopsylla cheopis, through flea bite or by inoculation of infected flea feces into the bite site or abraded skin and maintained in a rat–flea–rat cycle involving commensal rats (Rattus norvegicus and Rattus rattus) (Eisen and Gage 2012). Most human infections occur in port cities worldwide where rats and fleas are abundant (Civen and Ngo 2008). In 1990, Rickettsia felis was described from a laboratory colony of cat fleas, Ctenocephalides felis (Adams et al. 1990, Higgins et al. 1996) and soon thereafter implicated as the causative agent of human disease in a Texas patient (Schriefer et al. 1994). Since its discovery, human cases due to R. felis, referred in the literature as flea-borne spotted fever or cat-flea typhus (Abdad et al. 2011), were reported from Africa, Australia, Europe, the Middle East, North and South America, and Southeast Asia (Raoult et al. 2001, Pérez-Arellano et al. 2005, Zavala-Velazquez et al. 2006, Znazen et al. 2006, Ben-Zvi et al. 2010, Parola 2011, Williams et al. 2011, and Edouard et al. 2014). Laboratory studies provided evidence that R. felis could be transmitted to cats through the bite of infected cat fleas; however, the role of cats or other animals as possible reservoirs of these Rickettsia remains unknown (Wedincamp and Foil 2000, Reif and Macaluso 2009, Eisen and Gage 2012).

In the continental United States, human cases of flea-borne rickettsioses are reported from the southern portions of Texas and California where cat fleas and opossums have been implicated as the principal vector and potential reservoirs, respectively (Adams et al. 1970, Civen and Ngo 2008). A limited number of epidemiological studies have been performed to assess the overall prevalence of R. felis and R. typhi in cat fleas in Southern California. Rickettsial DNA was identified by polymerase chain reaction (PCR) in 52 (31.1%) of 167 cat fleas collected from opossums in Los Angeles County, 12 of which were further characterized as R. typhi (9 fleas) and R. felis (3 fleas) (Williams et al. 1992). Eremeeva et al. (2012) detected R. felis DNA in 882 (47.1%), R. typhi DNA in 25 (1.3%), and DNA of both organisms in 32 (1.7%) of 1873 cat flea pools recovered from cats, roof rats, and opossums trapped within the vicinity of households of human cases in Los Angeles and Orange counties. Rickettsia DNA was not recovered from the blood of any of the host animals nor were animals seropositive for Rickettsia species. R. felis was detected, however, in other tissues of four opossums. In the neighboring San Bernardino County, only R. felis DNA was detected in 152 (26.7%) of 570 cat flea pools and in 26 (22.6%) of 115 individual fleas recovered from opossums (Abramowicz et al. 2012).

In California, human cases of flea-borne rickettsioses are reported to public health authorities as “Rickettsial Diseases” (non-Rocky Mountain spotted fever), including typhus or typhus-like illness. Between 2010 and 2014, an average of 80 such cases was reported annually. Cat fleas and their hosts are ubiquitous in California, yet locally acquired human cases of flea-borne rickettsioses are reported almost exclusively from suburban communities within Los Angeles and Orange counties (Civen and Ngo 2008, Eremeeva et al. 2012). The reasons for the apparent regional endemicity of disease in California are not known, but could be due to differences in the presence, infection rate, and/or species of rickettsia in populations of cat fleas from other areas of the state. Surveillance bias due to physician awareness of the disease could also be a contributing factor. Documenting the distribution of R. typhi and R. felis in the presumed cat flea vector in California is necessary to elucidate the epidemiology of flea-borne rickettsioses and the apparent regional distribution of human disease. This study provides an initial comparison of the presence and species composition of rickettsiae in cat fleas from three counties of the state, comprising both endemic and nonendemic regions.

Materials and Methods

Flea collection

Cat fleas were collected from domestic cats in three counties of California. Collection sites included two animal shelters in Los Angeles County (flea-borne rickettsioses endemic region) and two trap/neuter/release (TNR) programs, one in Sacramento County and one in Contra Costa County (flea-borne rickettsioses nonendemic region). One to ten fleas were collected from each cat using a fine-tooth comb and were placed directly into a microcentrifuge tube (one per cat) containing 70% ethanol (Carolina Biological Supply Co.). All fleas in each tube made up a flea pool.

Molecular analyses

Pooled fleas from each cat were prepared and underwent molecular analyses. Fleas were washed in molecular-grade water to remove alcohol and opened midsection using disposable 18G to 20G needles. Flea tissues were vortex mixed with 180 μL of lysis buffer and 20 μL of proteinase K and DNA extracted following the manufacturer's instructions (QIAmp DNA Mini Kit; Qiagen, Inc.). DNA was eluted in 100 μL of elution buffer. A sample containing only elution buffer was extracted with each batch of flea samples and used as a negative extraction control.

The absence of PCR inhibitors was determined by the amplification of a fragment of ∼200 bp of the mitochondrial 16S rDNA using primers Sen_mt16S: 5′-TACATAACACGAGAAGACC-3′ and Rev_mt16S: 5′-GTGATTGCGCTGTTATCC-3′, as previously described (Vobis et al. 2004), with modifications. Briefly, all assays were performed in 15 μL of total volume using 1 μL of extracted DNA as template. The reaction mixture contained 7.5 pmol of each primer and a 1× “ready to use” premix (SYBR Premix Ex Taq Tli RNAse H Plus, Takara Bio) containing an optimized buffer, nucleotides, a heat-activated DNA polymerase, and SYBR Green I for detection of amplifications using an intercalator-based real-time PCR technology. A 1× of ROX reference dye was used to standardize fluorescence detection across wells. All reactions were run on a StepOne Plus Real-Time PCR System (Life Technologies). Amplifications were performed under the following conditions: one hot-start cycle at 95°C for 1 min followed by 45 cycles of 94°C for 10 s, 56°C for 20 s, and 72°C for 20 s. Amplification was detected by SYBR Green I fluorescence with emission at 522 nm. Melting curves were obtained for each amplification using temperature increments of 0.5°C/s. Positive amplification were characterized by a melting temperature peak between 80.9°C and 82.5°C. Negative extraction controls were used in each batch tested. Flea DNA was amplified from all samples.

Rickettsia genus-specific primers (ompB-F1: 5′-CAGGATTGGTAACTGCTTC-3′; ompB-R1: 5′-ACCGTTACCTATATCACCGG-3′) were chosen to amplify a 1560 bp region spanning approximately the first one-third of the ompB gene, based on the R. felis Marseille-URRWXCal2 strain (GenBank ID NC_007109). PCRs were carried out in a 50 μL volume reaction mix (0.30 μM of each primer, 0.3 mM dNTPs, 1 mM MgCl₂, 1× PCR buffer, and 0.05 U/μL Taq DNA polymerase) with 100 ng of DNA. Reactions were performed in a Veriti™ 96-well thermal cycler (Applied Biosystems) using the following conditions: 95°C for 2 min, 40 cycles of 95°C for 30 s, 54.8°C for 45 s, and 72°C for 30 s, followed by a final extension of 7 min at 72°C. Positive (Rickettsia rickettsii or Rickettsia amblyommii) and negative controls were used in each run. Amplicons were visualized on 1.5% agarose gels stained with ethidium bromide.

All positive PCR products were restriction enzyme digested with PstI to distinguish between R. felis (expected fragment sizes: 885, 597, 78 bp) and R. typhi (expected fragment sizes: 957, 300, 294 bp). Seventeen samples with the expected R. felis restriction pattern were chosen for nucleotide sequence analysis. In addition, restriction enzyme digestion of PCR products producing fragments distinctly different from R. felis or R. typhi was further characterized. Amplicons from PCR-positive samples were spin column purified (QIAquick PCR Purification Kit; Qiagen) and quantified photometrically (NanoVue Spectrophotometer; Fisher Scientific). The purified amplicons were sequenced bidirectionally by a commercial facility (Eurofins MWG Operon), using the same primer pairs designed for amplification plus two internal primers, ompB-F2 (5′-GCTGCGCCTTCTACATTA-3′) and ompB-R2 (5′-CAACAAATCCTGCAGCTC-3′). Sample sequences were compared to the R. felis Marseille-URRWXCal2 strain (GenBank ID NC_007109) and R. typhi Wilmington strain (GenBank ID NC_006142) reference sequences using CLC Main Workbench 7.0.3 (CLC Bio) and NCBI/BLAST (http://blast.st-va.ncbi.nlm.nih.gov/Blast.cgi).

Statistical analysis

The pooled infection rate of fleas with Rickettsia species was calculated using maximum likelihood estimation (MLE) methods for unequal pool sizes. The bias-corrected MLE method for point estimations and the skewness-corrected function to compute asymptotic 95% confidence limits were calculated using PooledInfRate version 4.0 (Brad Biggerstaff, CDC; available from www.cdc.gov/westnile/resourcepages/mosqsurvsoft.html). Fisher's exact test was used for comparison of the Rickettsia infection rates of fleas between Northern and Southern California. Significance was defined as p < 0.05.

Results

All fleas in the study were identified as C. felis using a standard taxonomic key (Lewis et al. 1998). A total of 130 cats were sampled and therefore 130 flea pools collected: 76 from the flea-borne rickettsiosis endemic region of Los Angeles County [Animal Shelter 1 (n = 41) and Animal Shelter 2 (n = 35)] and 54 from the nonendemic region of Sacramento County (n = 29) and Contra Costa County (n = 25). A total of 520 fleas were collected, with a median of five fleas per cat. Rickettsia DNA was detected in 77 flea pools by PCR analysis (Table 1); the Rickettsia species in three samples from Southern California could not be determined due to insufficient DNA available for restriction enzyme digestion or sequencing. These samples were considered as positives, however, when analyzing the overall infection rate. The pooled infection rate by MLE of Rickettsia species was not significantly higher in fleas tested from the endemic region (42 pools out of 281 fleas tested; pooled infection rate of 213 per 1000 fleas) when compared to the nonendemic region (35 pools out of 239 fleas tested; pooled infection rate of 234 per 1000 fleas) (p = 0.283). When comparing the pooled infection rate of R. felis only, there was a significant difference between endemic and nonendemic areas: 29 pools out of 277 fleas tested (pooled infection rate of 131 per 1000 fleas) were R. felis positive from the endemic region and 35 pools out of 239 fleas tested (pooled infection rate of 234 per 1000 fleas) were R. felis positive from the nonendemic region (p < 0.0001).

Table 1.

Presence and Species of Rickettsia in Cat Flea Pools from Regions Endemic and Nonendemic for Flea-Borne Rickettsioses in California

Collection site	Pools positive for Rickettsia felis, n	Pools positive for Rickettsia Genotype 1, n	Pools positive for Rickettsia Genotype 2, n	Coinfected pools, n	Nonspeciated positive pools	Total positive pools, n	Total negative pools, n	Total pools tested, n
Northern California/nonendemic region
Sacramento County TNR^a	17	—	—	1^b		18	11	29
Contra Costa County TNR	17	—	—	—		17	8	25
Southern California/endemic region
Los Angeles County shelter 1	16	1	3	1^c	1	22	19	41
Los Angeles County shelter 2	12	—	6	—	2	20	15	35
Total	62	1	9	2	3	77	53	130

TNR—trap, neuter, release program for stray and feral cats.

Coinfection of R. felis with Rickettsia Genotype 1.

Coinfection of R. felis and Rickettsia Genotype 2.

All amplicons but one from fleas collected in the nonendemic region had restriction patterns consistent with R. felis; none had restriction patterns consistent with R. typhi. Two restriction patterns, distinct from R. typhi and R. felis, were seen in amplicons from fleas collected in the endemic region and in a single flea pool from the nonendemic region. Genotype 1 (GenBank ID KP398499) was identified in one flea pool from Animal Shelter 1 in the endemic region. In addition, one flea pool from the nonendemic region (Sacramento County TNR program) presented with restriction patterns, suggesting a coinfection of Genotype 1 with R. felis. Genotype 2 (GenBank ID KP398500) was identified in 10 pools from the endemic region; 3 flea pools from Animal Shelter 1, 6 flea pools from Animal Shelter 2, and 1 flea pool from Animal Shelter 1, which had restriction patterns suggesting the presence of a coinfection of Genotype 2 with R. felis. Although the Rickettsia species could not be determined for three samples, the remaining 28 infected cat flea pools from the endemic region were shown to harbor R. felis DNA only (Table 1).

Sequence analysis of 17 amplicons (7 from Los Angeles County, 5 from Sacramento County, and 5 from Contra Costa County) having the expected restriction pattern for R. felis confirmed a 99–100% identity to the R. felis Marseille-URRWXCal2 strain (GenBank ID NC_007109). The nucleotide sequence of Genotype 1 was 99% identical to a proposed new species, Candidatus Rickettsia asemboensis (GenBank ID JN315972.2), and 94% identical to R. felis. Sequencing analysis of a flea pool from the nonendemic region demonstrated a coinfection of R. felis and Genotype 1. Genotype 2 was 99% identical to the newly reported Candidatus R. senegalensis identified in fleas from Senegal (GenBank ID KF666470.1) and shared 96% identity with R. felis. A phylogenetic tree (Fig. 1) was generated based on 1431-bp sequences of the ompB of the two R. felis-like genotypes from the flea pools and closely related organisms by using the maximum-likelihood method on the basis of the Kimura 2-parameter model (Kimura 1980), with bootstrap support based on 1000 replicates (Molecular Evolutionary Genetics Analysis [MEGA] software, version 6) (Tamura et al. 2013).

FIG. 1.

Phylogenetic tree based on 1431 bp from the outer membrane protein B (ompB) sequences of Rickettsia felis-like genotypes obtained from flea pools in Southern California (in boldface) and closely related organisms constructed by using the maximum-likelihood method on the basis of the Kimura 2-parameter model. Each bacterial name is followed by geographic origin data and the GenBank accession number is provided in parentheses. The numbers at the nodes indicate percentages of bootstrap support based on 1000 replicates. Percentages corresponding to partitions reproduced in fewer than 50% of bootstrap replicates are collapsed. The scale bar indicates 0.02 substitutions per nucleotide position.

Sequencing primers were designed to differentiate between R. felis and Genotype 2 when sequencing an amplicon suspected of containing both sequences, based on the restriction digest pattern. Primer ompB-15bp-R (5′-TACTGCATTAGCAGGACC-3′) binds to the wild-type R. felis sequence in a region that contains a 15 bp deletion in Genotype 2. Primer ompB-15bpdel-R (5′- CCATTATTAGTACCTATTACTGC-3′) binds to the Genotype 2 sequence but not to R. felis. The presence of a mixed infection of R. felis and Genotype 2 was confirmed in one flea pool collected from a cat at Los Angeles County Animal Shelter 1.

Discussion

With the exception of R. typhi, the global epidemiology and disease ecology of flea-borne Rickettsia species are still poorly understood. Such a conundrum exists in California where the majority of reported human cases (over 90%) occur in residents of Los Angeles and Orange counties, but rarely elsewhere in the state despite statewide prevalence of the implicated cat flea vector and its potential hosts. In the 1970s, a cycle involving R. typhi, cat fleas, and opossums was proposed to explain this Southern California phenomenon (Adams et al. 1970), and subsequent studies suggested that, rather than R. typhi, a different Rickettsia species named R. felis might be the etiologic agent of human illness in this region (Reif and Macaluso 2009, Eremeeva et al. 2012). The etiologic agent causing human flea-borne rickettsioses in California, however, is yet to be confirmed by molecular methods such as the detection of rickettsial DNA using PCR analysis or isolation of the infectious agent from clinical specimens. Rickettsiae are often found sequestered within endothelial cells reducing the number of organisms circulating in the bloodstream (Labruna and Walker 2014). Because of this, it can be extremely difficult to detect or isolate these organisms in a patient's blood sample even before antibiotic therapy has begun. In addition, the antibody-based testing protocols currently used for diagnosis use R. typhi antigens, which may result in cross-reactivity between antigenically related rickettsiae (Civen and Ngo 2008, Reif and Macaluso 2009) and thus prevent confirmation of the Rickettsia species involved.

In this study, the infection rate in cat fleas with R. felis and the absence of R. typhi were consistent with findings from past studies in Southern California, which demonstrated a minimum infection prevalence of 13–32% R. felis and 0–2% R. typhi (Abramowicz et al. 2012, Eremeeva et al. 2012). R. felis has also been reported from regions of Texas and Mexico endemic for flea-borne rickettsioses. In Corpus Christi, Texas, 14 (2.6%) of 529 cat fleas recovered from opossums were PCR positive for the presence of Rickettsia DNA and sequencing analysis demonstrated R. felis and R. typhi presence in 11 and 3 cat flea specimens, respectively (Boostrom et al. 2002). In Yucatán, Mexico, 11 (20%) of 54 pools of cat fleas collected from domestic dogs were positive for Rickettsia DNA and sequences from 2 positive flea pools were 100% identical to R. felis (Zavala-Velázquez et al. 2002). The wide distribution of R. felis in cat fleas from flea-borne rickettsioses endemic and nonendemic regions of North America suggests that a high prevalence of this bacterium in cat fleas does not predispose to increase risk of human infection. Human flea-borne rickettsioses cases are uncommon in California with a total of 217 confirmed cases reported between 2010 and 2014 (www.cdph.ca.gov/HealthInfo/discond/Pages/Typhus.aspx). Having limited numbers of R. typhi-infected fleas in the environment may decrease transmission risk and may explain the low number of human cases. R. typhi is yet to be detected in cat fleas outside of the endemic region in California.

A comparison of cat flea pools positive for R. felis among the three California counties sampled in this study revealed that the infection rate of this rickettsia species in regions considered flea-borne rickettsioses nonendemic (i.e., Sacramento and Contra Costa counties) was higher than in Los Angeles County, where ∼70% of the human cases are reported. Thus, differential risk of acquiring flea-borne rickettsioses between endemic and nonendemic areas of California cannot be ascribed to differential exposure to R. felis from infected cat fleas. Other factors may be responsible for the nonendemic status of regions in California such as lack of identification of the disease agent by physicians when patients present with potentially nonspecific symptoms of flea-borne rickettsioses (Civen and Ngo 2008) or involvement of other hematophagous arthropods in R. felis transmission. R. felis DNA has been reported previously in other flea species, ticks, and mosquitoes, suggesting the role of other ectoparasites as potential vectors of R. felis (Socolovschi et al. 2012, Abarca et al. 2013, Ramírez-Hernández et al. 2013, Dieme et al. 2015, Peniche-Lara et al. 2015).

The identification of two R. felis-like genotypes in cat fleas raises questions as to whether other Rickettsia species may be causative agents of flea-borne rickettsioses. Genotype 1, detected in two flea pools, is closely related to Candidatus R. asemboensis, a newly recognized rickettsia detected in five species of fleas (Ctenocephalides canis, C. felis, Echidnophaga gallinacea, X. cheopis, and Pulex irritans) from Asembo, Kenya (Jiang et al. 2013). In that study, four C. canis and 71 (94.7%) of 75 C. felis pools were infected with Candidatus R. asemboensis, whereas R. felis was detected in only three C. felis pools. One cat flea pool was coinfected with both agents (Jiang et al. 2013). Genotype 2, detected in 10 flea pools in this study, was most similar to Candidatus R. senegalensis recently detected in cat fleas from Senegal (Mediannikov et al. 2014). The distribution of these rickettsiae in cat flea populations and their potential to cause human disease should be considered for further investigation.

Why human cases of flea-borne rickettsioses cases in California are primarily reported from Los Angeles and Orange counties remains unclear. The results from this study demonstrate that R. felis is present in cat fleas in both endemic and nonendemic regions of the state and suggest that at least two additional R. felis-like genotypes may be circulating in populations of these fleas. However, because only a limited number of fleas and locations were examined in this study, a more comprehensive survey might shed more light on the prevalence of these Rickettsia species and R. typhi. Other considerations in future research should include flea pool size and PCR targets. Comparisons across uneven numbers of pooled fleas, combined with a PCR targeting a relatively large DNA fragment, have the potential to produce a dilution effect that may decrease the efficiency of rickettsia detection. At the same time, the current study design required amplification of a large DNA fragment, when used in conjunction with the restriction enzyme digestion, to adequately differentiate between R. felis and R. typhi. Many questions regarding the basic biology and ecology of flea-borne rickettsioses in California are yet to be answered, including confirmation of the etiologic agent(s).

Footnotes

Acknowledgments

The authors thank Crystal Perreira and Leslie Foss for their assistance with flea collections from the Contra Costa Co. TNR program, Jessica Chinison and Dr. Dominique Griffon for lab support, and Dr. Mark Novak for his assistance with flea collections from the Sacramento Co. TNR program. They thank the veterinarians and staff from each collection site for providing the opportunity to collect flea specimens. The authors acknowledge Dr. Vicki Kramer and Dr. Charsey Cole Porse for providing critical review of the article.

Author Disclosure Statement

No competing financial interests exist.

References

Abarca

, López

, Acosta-Jamett

, Martínez-Valdebenito

. Rickettsia felis in Rhipicephalus sanguineus from two distant Chilean cities. Vector Borne Zoonotic Dis, 2013; 13:607–609.

Abdad

, Stenos

, Graves

. Rickettsia felis, an emerging flea-transmitted human pathogen. Emerg Health Threats J, 2011; 4:7168.

Abramowicz

, Wekesa

, Nwadike

, Zambrano

, et al. Rickettsia felis in cat fleas, Ctenocephalides felis parasitizing opossums, San Bernardino County, California. Med Vet Entomol, 2012; 26:458–462.

Adams

, Schmidtmann

, Azad

. Infection of colonized cat fleas, Ctenocephalides felis (Bouché), with a rickettsia-like microorganism. Am J Trop Med Hyg, 1990; 43:400–409.

Adams

, Emmons

, Brooks

. The changing ecology of murine (endemic) typhus in Southern California. Am J Trop Med Hyg, 1970; 19:311–318.

Ben-Zvi

, Meltzer

, Nogueras

, Segura

, et al. First detection of human infection with Rickettsia felis in Israel. Am J Med Sci, 2010; 340:343.

Boostrom

, Beier

, Macaluso

, et al. Geographic association of Rickettsia felis-infected opossums with human murine typhus, Texas. Emerg Infect Dis, 2002; 8:549–554.

Civen

, Ngo

. Murine typhus: An unrecognized suburban vectorborne disease. Clin Infect Dis, 2008; 46:913–918.

Dieme

, Bechah

, Socolovschi

, Audoly

, et al. Transmission potential of Rickettsia felis infection by Anopheles gambiae mosquitoes. Proc Natl Acad Sci U S A, 2015; 112:8088–8093.

10.

Edouard

, Bhengsri

, Dowell

, Watt

, et al. Two human cases of Rickettsia felis infection, Thailand. Emerg Infect Dis, 2014; 20:1780–1781.

11.

Eisen

, Gage

. Transmission of flea-borne zoonotic agents. Annu Rev Entomol, 2012; 57:61–82.

12.

Eremeeva

, Karpathy

, Krueger

, Hayes

, et al. Two pathogens and one disease: Detection and identification of flea-borne Rickettsiae in areas endemic for murine typhus in California. J Med Entomol, 2012; 49:1485–1494.

13.

Higgins

, Radulovic

, Schriefer

, Azad

. Rickettsia felis: A new species of pathogenic rickettsia isolated from cat fleas. J Clin Microbiol, 1996; 34:671–674.

14.

Jiang

, Maina

, Knobel

, Cleaveland

, et al. Molecular detection of Rickettsia felis and Candidatus Rickettsia asemboensis in fleas from human habitats, Asembo, Kenya. Vector Borne Zoonotic Dis, 2013; 13:550–558.

15.

Kimura

. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences. J Mol Evol, 1980; 16:111–120.

16.

Labruna

, Walker

. Rickettsia felis and changing paradigms about pathogenic rickettsiae. Emerg Infect Dis, 2014; 20:1768–1769.

17.

Lewis

, Lewis

, Maser

. Genus Ctenocephalides Stiles and Collins. The Fleas of the Pacific Northwest. Oregon: Oregon State University Press, 1988:33–36.

18.

Mediannikov

, Aubadie-Ladrix

, Raoult

. Candidatus ‘Rickettsia senegalensis’ in cat fleas in Senegal. New Microbes New Infect, 2014; 3:24–28.

19.

Parola

. Rickettsia felis: From a rare disease in the USA to a common cause of fever in sub-Saharan Africa. Clin Microbiol Infect, 2011; 17:996–1000.

20.

Peniche-Lara

, Dzul-Rosado

, Pérez-Osorio

, Zavala-Castro

. Rickettsia typhi in rodents and R. felis in fleas in Yucatán as a possible causal agent of undefined febrile cases. Rev Inst Med Trop Sao Paulo, 2015; 57:129–132.

21.

Pérez-Arellano

, Fenollar

, Angel-Moreno

, Bolaños

, et al. Human Rickettsia felis infection, Canary Islands, Spain. Emerg Infect Dis, 2005; 11:1961–1964.

22.

Ramírez-Hernández

, Montoya

, Martínez

, Pérez

, et al. Molecular detection of Rickettsia felis in different flea species from Caldas, Colombia. Am J Trop Med Hyg, 2013; 89:453–459.

23.

Raoult

, La Scola

, Enea

, Fournier

, et al. A flea-associated Rickettsia pathogenic for humans. Emerg Infect Dis, 2001; 7:73–81.

24.

Reif

, Macaluso

. Ecology of Rickettsia felis: A review. J Med Entomol, 2009; 46:723–736.

25.

Schriefer

, Sacci

Jr , Dumler

, Bullen

, et al. Identification of a novel rickettsial infection in a patient diagnosed with murine typhus. J Clin Microbiol, 1994; 32:949–954.

26.

Socolovschi

, Pagés

, Raoult

. Rickettsia felis in Aedes albopictus mosquitoes, Libreville, Gabon. Emerg Infect Dis, 2012; 18:1687–1689.

27.

Tamura

, Stecher

, Peterson

, Filipski

, et al. MEGA6: Molecular Evolutionary Genetic Analysis version 6.0. Mol Biol Evol, 2013; 30:2725–2729.

28.

Vobis

, D'Haese

, Mehlhorn

, Mencke

, et al. Molecular phylogeny of isolates of Ctenocephalides felis and related species based on analysis of ITS1, ITS2, and mitochondrial 16S rDNA sequences and random binding primers. Parasitol Res, 2004; 94:219–226.

29.

Wedincamp

Jr , Foil

. Infection and seroconversion of cats exposed to cat fleas (Ctenocephalides felis Bouché) infected with Rickettsia felis . J Vector Ecol, 2000; 25:123–126.

30.

Williams

, Izzard

, Graves

, Stenos

, et al. First probable Australian cases of human infection with Rickettsia felis (cat-flea typhus). Med J Aust, 2011; 194:41–43.

31.

Williams

, Sacci

Jr , Schriefer

, Andersen

, et al. Typhus and typhus-like rickettsiae associated with opossums and their fleas in Los Angeles County, California. J Clin Microbiol, 1992; 30:1758–1762.

32.

Zavala-Velazquez

, Laviada-Molina

, Zavala-Castro

, Perez-Osorio

, et al. Rickettsia felis, the agent of an emerging infectious disease: Report of a new case in Mexico. Arch Med Res, 2006; 37:419–422.

33.

Zavala-Velázquez

, Zavala-Castro

, Vado-Solís, Ruiz-Sosa

, et al. Identification of Ctenocephalides felis fleas as a host of Rickettsia felis, the agent of a spotted fever rickettsiosis in Yucatán, Mexico. Vector Borne Zoonotic Dis, 2002; 2:69–75.

34.

Znazen

, Rolain

, Hammami

, et al. Rickettsia felis infection, Tunisia. Emerg Infect Dis, 2006; 12:138–140.