Rickettsial Infections among Ctenocephalides felis and Host Animals during a Flea-Borne Rickettsioses Outbreak in Orange County, California

Due to a resurgence of flea-borne rickettsioses in Orange County, California, we investigated the etiologies of rickettsial infections of Ctenocephalides felis, the predominant fleas species obtained from opossums (Didelphis virginiana) and domestic cats (Felis catus), collected from case exposure sites and other areas in Orange County. In addition, we assessed the prevalence of IgG antibodies against spotted fever group (SFGR) and typhus group (TGR) rickettsiae in opossum sera. Of the 597 flea specimens collected from opossums and cats, 37.2% tested positive for Rickettsia. PCR and sequencing of rickettsial genes obtained from C. felis flea DNA preparations revealed the presence of R. typhi (1.3%), R. felis (28.0%) and R. felis-like organisms (7.5%). Sera from opossums contained TGR-specific (40.84%), but not SFGR-specific antibodies. The detection of R. felis and R. typhi in the C. felis fleas in Orange County highlights the potential risk for human infection with either of these pathogens, and underscores the need for further investigations incorporating specimens from humans, animal hosts, and invertebrate vectors in endemic areas. Such studies will be essential for establishing a link in the ongoing flea-borne rickettsioses outbreaks.


Introduction
Murine typhus (also known as endemic typhus) caused by Rickettsia typhi has been endemic in California since at least 1915 when the first case was identified [1]. It was recognized as a major public health concern in both Texas and California in the 1940s. This was followed by a marked decline in the number of reported cases in humans, possibly due to the intensive rat and vector control programs that were instituted [2]. In the U.S., human cases of flea-borne rickettsial diseases have been increasing since the 1990s, occurring primarily in southcentral Texas, Hawaii, and southern California [2][3][4]. Recently an outbreak of rickettsial disease(s) referred to as flea-borne rickettsioses in areas of California traditionally known to be endemic for murine typhus has been reported. Unfortunately the etiologic agent(s) for this outbreak has not been identified [5]. In California during 2013, 105 cases of flea-borne rickettsioses were reported to the California Department of Public Health (CDPH) and were classified as confirmed (n = 67) or probable (n = 38). Overall, 98% of these cases were reported in Orange and Los Angeles counties [3]. Under Title 17, California Code of Regulations (CCR) §2500, Rocky Mountain spotted fever and other rickettsial diseases (non-Rocky Mountain spotted fever), including typhus and typhus-like illnesses, are classified as reportable communicable diseases which must be reported within seven days of identification [6]. It has been noted that in these flea-borne rickettsioses outbreaks, the classical murine typhus urban transmission cycle involving rats (Rattus norvegicus and Rattus rattus) and the Oriental rat flea (Xenopsylla cheopis) has shifted to a suburban cycle that involves cats (Felis catus), Virginia opossums (Didelphis virginiana), and cat fleas (Ctenocephalides felis) [5,7]. This change in vector species for murine typhus is not unheard of as other flea species, lice, and mites have also been implicated as potential vectors and reservoirs for R. typhi [8,9].
Another possible etiology for the California flea-borne rickettsioses outbreak could be infections with the flea-borne pathogen, R. felis, which is associated with the prevalent cat flea [5]. R. felis was first recognized as a human pathogen and the causative agent for flea-borne spotted fever in 1991 in a patient/case in Texas [10], after its detection in a colonized cat flea in 1990 [11]. C. felis has historically been the only defined biological vector for R. felis, where the agent is maintained through transovarial and transstadial transmission [12]. However ticks, mites, mosquitoes, and more than 20 species of fleas have been found to be naturally infected with R. felis [13], including X. cheopis, the main vector of R. typhi [4,[14][15][16] Both R. typhi and R. felis have worldwide distribution [17,18] and their areas of endemicity overlap, especially in urban centers and port cities [4,14]. Co-infections with the two species in single cat and rat fleas has been documented to occur both under both natural [4,19] and experimental conditions [20]. Similarly, a previous study conducted in California has shown this sympatric relationship between R. typhi and R. felis in cat fleas collected from opossums and cats [5,21]. As a result of their overlapping endemicity, coupled with the sharing of flea vectors, and the similar presentation of clinical illness in humans [17], the actual estimation of the burden and distribution of murine typhus and flea-borne spotted fever is complex [4].
Epidemiologic studies conducted in California and Texas have showed a positive correlation between flea-borne human typhus cases and the geographic distribution of rickettsia seropositive opossums and rickettsia infected fleas [5,22], but no correlation between rickettsia infected fleas and seropositive opossums [5]. Although flea-borne rickettsial diseases due to R. typhi and R. felis are well-described individually, significant gaps exist in our knowledge and understanding of their concurrent epidemiology. Here we report the identification of R. felis, R. typhi, Candidatus Rickettsia senegalensis, and Candidatus Rickettsia asemboensis in C. felis, as well as high prevalence of antibodies against typhus group rickettsiae (TGR) in opossums during the 2011-2013 flea-borne rickettsioses outbreak in Orange County, CA.

Material and Methods
The study was conducted utilizing three groups of fleas and blood samples from animals collected as part of the comprehensive flea-borne rickettsioses prevention program in Orange County, California, by the Orange County Mosquito and Vector Control District (OCMVCD): 1) opossums caught in small mammal traps (Tomahawk Live Trap1, Tomahawk, WI) set in response to reports of human cases of flea-borne rickettsioses and pet cats owned by humans diagnosed with flea-borne rickettsioses from 2011-2012, 2), opossums collected by OCMVCD as part of routine flea-borne typhus surveillance, but not geographically associated with cases of flea-borne rickettsioses (collections for 2013), and 3) historical biobanked specimens from the county collected in 1969-1988 from opossums and in the environment using "white sock" method [23]. Opossums were euthanized in a CO 2 gas chamber according to American Veterinary Medical Association (AVMA) protocols [24]. Blood and flea samples from pet cats were taken by local veterinarians. Blood samples and fleas were collected, processed, and stored at -80°C for testing. Fleas were individually washed in molecular-grade water and mechanically disrupted using disposable pellet pestles (Fisher Scientific) or pipette tips. Genomic DNA was extracted from the fleas using Prepman Ultra sample preparation kits (Applied Biosystems, Foster City, CA) and from the opossum sera using DNeasy blood and tissue kits (QIAGEN, Valencia, CA) according to the manufacturers' instructions, using a final elution volume of 100 μl. A qPCR assay (Rick17b) that amplifies and detects a 115-bp segment of the 17-kDa antigen gene [25] was used to screen flea specimens for the presence of rickettsial DNA. Flea DNA preparations positive for rickettsiae (Rick17b positive) were subsequently tested using species-specific qPCR assay (RFelB) for R. felis [26] and confirmed with a second species-specific assay (RFel_Phosp_MB) targeting the membrane phosphatase gene from R. felis [27]. Samples that tested positive with Rick17b but negative for RfelB were further tested using two qPCR assays namely: the Ca. R. asemboensis-specific assay (Rasem) that amplifies and detects a 112-bp of the ompB gene [15] and the R. typhi-specific assay (Rtyph) that targets a 122-bp fragment of the ompB gene of R. typhi [28]. PCR amplification and sequencing of gltA, ompB, rrs, sca4, and R. felis plasmid genes pRF and pRFδ were attempted on a subset of flea DNA samples as previously described [15]. Sequencing reactions were performed utilizing both DNA strands with Big Dye Terminator v3.1 Ready Reaction Cycle Sequencing Kit (Life Technologies, Foster City, CA), according to the manufacturer's instructions on an ABI 3500 genetic analyzer (Applied Biosystems). Sequences were assembled using CodonCode Aligner version 5.0.1 (CodonCode Corporation, Centerville, MA) and exported to MEGA version 6 software (CEMI, Tempe, AZ) where they were compared with other historical strains sequences available in GenBank. Evolutionary analyses were conducted in MEGA version 6 based upon single-locus sequence typing of rrs, gltA, ompB, sca4, and R. felis plasmid (pRF) genes [29]. Chisquared tests were used to compare the infection rates across the three time periods using the GraphPad Software Inc. (2013). A p-value of 0.05 or less was considered significant.
To assess opossums for evidence of previous infection with rickettsiae, serum samples (n = 262) were screened in duplicate for spotted fever group rickettsiae (SFGR)-and typhus group rickettsiae (TGR)-specific immunoglobulin G (IgG) at a dilution of 1:100 using enzyme linked immune-sorbent assays (ELISAs) as previously described by Graf et al. [30], except that R. conorii str. Morocco was used as the SFGR ELISA antigen. An anti-opossum conjugate was made using unlabeled anti-opossum antibody from Bethyl Labs, Inc. (Montgomery, TX) and KPL Sure-Link HRP conjugation kit (Gaithersburg, MD). Positive controls were opossum serum pools positive for R. typhi and cat serum from cats fed upon by C. felis infected with R. felis [31]. To confirm the results of the SFGR and TGR ELISAs, 40 random opossum sera were sent to Naval Medical Research Center, Silver Spring, Maryland for quality control and were retested in parallel using both the in-house conjugate and a commercially available anti-opossum IgG (Alpha Diagnostics Intl. Woodlake Center, San Antonio, TX) and anti-cat IgG (KPL, Gaithersburg, MD) labeled with horseradish peroxidase (HRP). An assessment of cross-reactivity of antibodies against R. felis was conducted using R. felis immune sera from two cats and two mice with SFGR-and TGR-ELISAs as described above, except that anti-mouse IgG and anti-cat IgG conjugates were used (KPL). The same sera had previously been shown to be reactive to R. felis str. LSU by IFA as previously described [32].
GenBank Accession numbers: The sequences for the Candidatus Rickettsia senegalensis (previously described as Rickettsia sp. PU01-02) obtained from C. felis have been deposited in the GenBank with accession numbers KT304218, KT304219, KT304220 for rrs, ompB, and sca4, respectively.
Ethical approval: The Orange County Mosquito and Vector Control District (OCMVCD) do not have a formal Institutional Animal Care and Use Committee (IACUC) since it is not considered a research institution. However, it does follow the protocols for animal handling for disease surveillance purposes as outlined by the California Department of Public Health, Vector Borne Diseases Section, and adhered to American Veterinary Medical Association (2013) guidelines for animal euthanasia. Scientific permits were obtained from the California Department of Fish and Wildlife (CDFW) to collect and sample small mammals.

Flea results
DNA preparations were obtained from 597 individual C. felis of which 587 (98.3%) were removed from opossums and 10 (1.75%) were removed from cats. Of the 597 flea DNA extracts, 222 (37.2%) were positive for rickettsial DNA (opossums n = 219 of 587 (37.3%) and cats n = 3 of 10 (30%)) using the Rick17b qPCR genus-specific assay. The results for the genus-, group-and species-specific qPCR assays are shown in Table 1  To confirm the identity of rickettsiae identified by the various qPCR assays, PCR amplification and sequencing was performed on either one or a combination of the following genes: gltA, rrs, ompB, sca4 and/ or pRF genes. For RfelB and RFel_Phos_MB positive DNA extracts (n = 11), PCR amplicons were generated and sequences determined for ompB. A 1444-bp ompB sequence was generated for all 11 DNA preparations, and they were all 100% identical to R. felis URRWXCal2 (Accession no. CP000053). R. felis plasmid pRF was detected in 3/11 samples and the 402-bp sequence generated for all three was 100% identical to R. felis URRWXCal2 plasmid pRF (CP000054).
A subset of three flea DNA samples that were RfelB positive and /or RFel_Phosp_MB negative were further evaluated to determine their identity. The sequences for rrs, sca4, and ompB genes were determined and compared with those available in GenBank. The three sequences of 1,009-bp fragment of the rrs gene were 100% identical with those of Ca. R. senegalensis (KF666476). From a 1,046-bp sequence of sca4 gene, the three DNA preparations had a 100% match to Ca. R. senegalensis (KFF666474). From the same 1,046-bp sequence, the similarity of the three preps to that of R. felis URRWXCal2 (CP000053) was only (98.4%). Phylogenetic  2); the closest validated Rickettsia species is R. felis (94.78% match to R. felis URRWXCal2). A 1431-bp fragment of the ompB sequence from the three flea DNA samples was 100% identical to Rickettsia sp. DS_018 (KP398500) detected in cat fleas from California. A PCR amplicon for the plasmid pRF was not produced for the three samples. From one of the two flea DNA samples positive by the R. asemboensis-specific qPCR assay (Rasemb), an amplicon was obtained with the ompB PCR. A 1,401-bp ompB sequence was determined to be 100% identical to Rickettsia sp. DS_006 (KP398499) detected in a cat flea from California and was 99.86% (1399/1401) similar to Candidatus R. asemboensis (JN315972). An additional 734-bp sequence of the ompB gene was found to be 100% (734/734) identical to R. sp. cf9 (DQ379483), 99.86% (733/734) to Ca. R. asemboensis (JN315972), 99.73% (732/734) similar to R. sp. RF2125 (JX183538), and R. sp. R14 (HM370113). Thus, this R. felis like organism (RFLO) agent was found to be similar to Ca. R. asemboensis.
A PCR amplicon (381-bp) for the gltA gene was produced in 5 out of the 8 Rtyph positive qPCR assay samples assessed. The gltA sequences generated for the five samples assessed were all 100% identical to R. typhi str. B9991CWPP (CP003398).
Twenty-seven (53%) out of 51 fleas assessed from 1969-88 were positive by Rick17b and RfelB qPCR assays but only 5 (9.8%) were positive by Rfel_Phos_MB qPCR assay. PCR amplification for the ompB, gltA, and sca4 genes attempted in 11 DNA extracts from these positive flea DNA preparations yielded no amplicons. A shorter fragment (599-bp) of ompB produced amplicons in 4/11 samples. These four samples were 100% identical to R. felis.

Rickettsia typhi positive opossum sera
A total of 262 opossum sera were assessed for the presence of rickettsia DNA. All the sera were negative for rickettsia DNA using the genus-specific Rick17b qPCR assay. Of the sera tested for the presence of SFGR-and TGR-specific IgG, 107/262 (40.84%) had IgG antibodies reactive against TGR antigens with endpoint titers ranging from 100 to 6400 (Fig 1). None of the opossum sera were positive for antibodies against SFGR antigens (S1 Fig). Positive control serum samples, previously shown to be reactive to the R. felis str. LSU IFA slides, from two cats fed on by R. felis str. LSU infected fleas and two mice infected with R. felis were evaluated by the SFGR-and TGR-specific ELISAs. Serum samples from the two cats and two mice exposed to R. felis, but not negative control sera, reacted to the SFGR-specific ELISA, but these same samples did not react to the TGR-specific ELISA (S2 Fig). Therefore, the opossum TGR-ELISA positive sera were most likely the result of a previous infection with R. typhi and not due to a R. felis infection.

Discussion
The investigation presented herein follows upon the recent outbreak of flea-borne rickettsial disease in Orange County, California, from 2011-2013 [3,5]. It compares samples collected from opossums caught in small mammal traps around case homes and away from the reported cases during the ecologic investigations of flea-borne rickettsioses (2011-2012), samples from opossums captured throughout the county in 2013, and historical biobanked specimens from the county from 1969-1988. Overall, rickettsial infections were detected in 37.2% of the cat fleas analyzed from opossums and cats. Although R. felis URRWXCal2 was the predominant rickettsia in the cat fleas (28%), Ca. R. senegalensis (7.2%), Ca. R. asemboensis (0.3%) and R. typhi (1.3%) were also detected in the fleas. While flea-borne rickettsioses outbreaks in southern California and Texas have previously been attributed to murine typhus [2,33,34], in relation to R. felis, R. typhi was not very common (1.3%) in this investigation in C. felis fleas removed from opossums and cats. These rates of R. typhi infections are similar to those reported previously in southern California [5,21].
The RfelB assay [26] previously thought to be specific for R. felis was found to detect both R. felis and the newly described Ca. R. senegalensis. As such the RfelB assay should no longer be considered a R. felis-specific assay. However, the recently developed R. felis-specific assay-RFel_Phosp_MB did not detect Ca. R. senegalensis or Ca. R. asemboensis, but only detected R. felis [27]. Thus, we recommend the use of the latter assay in assessing specifically for the presence of R. felis in arthropod vectors and vertebrate hosts.
PCR and sequencing of either one, or a combination, of gltA, ompB, rrs, sca4, and pRF genes resulted in the identification of four species within fleas assessed from Orange County. These rickettsiae included: R. typhi, R. felis, and the RFLOs Ca R. senegalensis, and Ca. R. asemboensis. Candidatus R. senegalensis was recently isolated in pure culture in XTC-2 cell line from C. felis fleas removed from cats from Senegal [35]. Like other RFLOs, the pathogenic role of Ca. R. senegalensis is unknown. A very close relationship in the gltA gene sequence (99.9%) between Rickettsia sp. RF31 and Ca. R. senegalensis has been noted [35]. The former, R. sp. RF31, was first described in C. felis obtained from the Thai-Myanmar border, however limited gene sequence information was reported [36]. In the United States R. sp. RF31 was first reported in cat fleas in South Carolina in 2005 [37], and belonging to this group also is DS-018 (KP398500) detected in California [38]. This report confirms existence of Ca. R. senegalensis (or a very closely related agent) in California, United States.
We report in addition to R. typhi, R. felis, and Ca. R. senegalensis, the presence of Ca. R. asemboensis among C. felis fleas. We speculate that Rickettsia sp. DS_006 (KP398499) reported in California [38], and other closely related genotypes reported worldwide [15,26,36,39,40] are the same as Ca. R. asemboensis. Within this genogroup, only Ca. R. asemboensis has been isolated in cell culture and has the full genome sequence available in GenBank [41]. Recently, a Ca. R. asemboensis-genotype was detected in the blood of a monkey (Macaca fascicularis) in Malaysia [42]. Although the Ca. R. asemboensis genogroup appears to have a wide distribution, it has not been associated with any clinical illness.
During this study, DNA from archived cat flea samples collected from 1969 through 1988 were also extracted and tested. Rickettsia DNA was detected in these samples using the pan-rickettsial qPCR assay targeting the 17kDa (Rick17b) and species-specific assays (RFelB and RFel_Phosp_ MB) indicating that R. felis was circulating in C. felis collected in Orange County. Unfortunately, it was not possible to produce the longer PCR amplicons necessary for sequencing gene fragments of ompB, ompA, and sca4. This may be due to DNA degradation resulting from prolonged storage. We were unable to confirm the rickettsia present in samples collected in 1969-1988. Comparing results from 2011-2012 and 2013, RFLOs only made up 0-15% of all RfelB positives. It is therefore more likely that a majority of positive fleas contained R. felis DNA and very little if any RFLOs. Interestingly, we detected R. felis DNA in fleas collected 1969-1988, suggesting that R. felis was circulating in the flea population prior to its first description in 1990.
Rickettsial DNA was not found in the blood of opossums in this study, which is consistent with the results from previous investigations in which rickettsial DNA and/or evidence of rickettsemia was not detectable in the blood of opossums [5,43], peridomestic small mammals [33], and the Norway rat (Rattus norvegicus) [39]. Additionally, in an experimental study, intraperitoneal inoculation of Didelphis aurita (black-eared opossum) with R. felis did not produce rickettsemia [44]. Although the present study and other previous studies did not find rickettsemia in the opossums' blood, R. felis DNA has been detected in tissues from opossums and Rattus norvegicus [5,45], while R. typhi DNA was detected in the spleen of the opossum [46].
Our study reports a high prevalence (40.84%) of antibodies against R. typhi, however antibodies against SFGR were not detected in these same samples. This is especially intriguing since the opossums were infested with fleas (average flea index of 53.5, range 0-501 fleas per animal), of which 29.3% were infected with the rickettsial pathogens: R. typhi (TGR) (1.3%) and R. felis (SFGR) (28%). This observation is in agreement with previous studies, which reported anti-R. typhi antibody prevalence of 42-71% in opossums trapped within the vicinity of human flea-borne rickettsioses cases in Los Angeles County and Austin, Texas [22,33], and an 11% prevalence in Orange County [7]. However, a recent study conducted in Orange County, and coinciding with a flea-borne rickettsioses outbreak, did not detect any anti-R. typhi antibodies in the opossums by IFA [5]. Although the classic transmission cycle for murine typhus involves infected rats and rat fleas (X. cheopis), the latter are rarely reported in these flea-borne rickettsial disease endemic areas [22,39]. While X. cheopis is known as the biological vector for R. typhi, our study found no X. cheopis on opossums and cats, and 1.3% of cat fleas infesting opossums were infected with R. typhi. These findings align with studies conducted from 1916 through 1948 where an association between flea-borne rickettsial disease with opossums and cat fleas had already been made [47,48]. The previous reports, together with the high seropositivity for R. typhi antibodies in the opossums suggest that the opossums and cat fleas may play an important role in the epidemiology of murine typhus [8].
Although opossums have been shown to be susceptible to R. felis infection through experimental studies [44] and natural infection [2], the present study did not detect antibodies against SFGR in the opossums assessed. Contrastingly, high seroprevalence (22%) to anti-R. felis has been noted in opossum populations in Texas [2], utilizing R. typhi str. Wilmington and R. felisinfected cat flea midguts on IFA slides. Boostrom et al. [2] noted that 8% of opossum sera were reactive to R. typhi, but their IFA could not distinguish R. typhi from R. felis in 6 samples, denoting cross-reactivity or dual infections. In the present study, our ELISA, utilizing serum from mice infected with, and cats fed upon by cat fleas infected with, R. felis str. LSU, reacted with R. conorii str. Morocco but not with R. typhi str. Wilmington ELISA antigens. Fang and Raoult [49] found that IgG antibodies against R. felis cross-reacted more often with antigens from SFGR and less with antigens from TGR using microimmunofluorescence (MIF). Though some studies based only on IFA suggest that flea-borne spotted fever induced antibodies react well to both R. typhi and R. felis IFA antigens [10,50], the ELISAs used in this study did not show such crossreactivity.
Our study provides two important findings, first that cat fleas were predominantly infected with R. felis, but to a lesser extent with R. typhi and RFLOs. Second, in contrast to the small proportion of R. typhi infected fleas in relation to R. felis infected fleas, opossums had high prevalence of anti-R. typhi IgG antibodies and none to SFGR. The epidemiology of flea-borne rickettsial diseases warrants further investigations to incorporate specimens from humans, animal hosts, and invertebrate vectors in endemic areas to establish a link in the ongoing fleaborne rickettsioses outbreaks.
Supporting Information S1 Fig. Aphotograph of screening plates showing both positive (dark green) and negative (colorless circles) for R. typhispecific IgG antibodies in opossum sera (plates A and B) and negative for R. conorii-specific IgG antibodies (plates C and D). Note the positive control (red arrows) and the negative controls (blue arrows). Serum from a cat infected with R. felis was evaluated with the SFGR and TGR ELISAs. Cat serum with anti-R. felis reacted with R. conorii antigen of the SFGR ELISA (plate C, red arrows). (TIF) S2 Fig. Orange County titer data for the opossum sera (n = 23) positive for anti-R. typhi antibodies during screening (Plates A,B and C). Note that for plate A, the positive control (column 1), 3 negative controls (column 2-4) and cat serum infected with R. felis was included as a control for R. typhi antigen (column 5, yellow arrow). The cat serum with anti-R. felis did not react with R. typhi antigen of the TGR ELISA (panel yellow arrow). (TIF)