Abstract
Rickettsia felis is associated with fever, headache, myalgia, and macular rash in some infected humans and has been detected in the cat flea (Ctenocephalides felis) in many countries around the world. While some naturally exposed cats have been assessed for antibodies against R felis, to our knowledge, no one has reported use of polymerase chain reaction (PCR) to attempt to amplify R felis DNA from client-owned cats and the fleas collected from them. In this study, we assayed 92 pairs of cat blood and flea extracts from Alabama, Maryland and Texas, using PCR assays that amplify a region of the citrate synthase gene (gltA) and the outer membrane protein B gene (ompB). Of the 92 pairs, 62 of 92 (67.4%) flea extracts and none of the cat blood samples were positive for R felis DNA.
Rickettsia species are obligate intracellular Gram-negative bacteria that are divided into two distinct groups, the spotted fever group (SFG) and the typhus group (TG). The SFG group includes one of the oldest known vector-borne diseases, Rocky Mountain Spotted Fever and Mediterranean Fever, caused by R rickettsii and R conorii, respectively (Watt and Parola 2003). All of the SFG are tick-borne, except R akari (louse-borne) and R felis (flea-borne). Ranging from mild to severe, clinical manifestations of SFG-associated diseases in people are fairly non-specific with acute fever, headache, lymphadenopathy, and a macular rash (with a characteristic inoculation eschar at the bite site) being recognized most frequently (Watt and Parola 2003). Rickettsia felis was originally detected in a commercial cat flea (Ctenocephalides felis) colony and was initially designated the ELB agent (Adams et al 1990). At that time, the organism was believed to be a TG rickettsia because it induces antibodies that are serologically cross-reactive to R typhi antigens (Azad et al 1992). Further study by use of genetic sequencing and bacteriological techniques has shown the organism to belong to the SFG (Higgins et al 1996, Azad et al 1997, Bouyer et al 2001, La Scola et al 2002, Blair et al 2004).
Fever, headache, myalgia, and macular rash in humans have been attributed to R felis infection in several countries around the world (Schriefer et al 1994, Zavala-Velazquez et al 2000, Raoult et al 2001, Boostrom et al 2002, Oliveira et al 2002, Richter et al 2002, Parola et al 2003, Perez-Arellano et al 2005, Bernbeu-Wittel et al 2006). In addition, one person in Mexico developed neurological symptoms following R felis infection, suggesting that the organism may be the cause of severe debilitating disease in some people (Galvao et al 2004). Rickettsia felis has been detected in C felis, C canis, and Pulex irritans; these fleas have a worldwide distribution (Azad et al 1992, Williams et al 1992, Radulovic et al 1995, Azad et al 1997, Raoult et al 2001, Oliveira et al 2002, Marquez et al 2002, Zavala-Velazquez et al 2000, 2002, Richter et al 2002, Boostrom et al 2002, La Scola et al 2002, Parola et al 2003, Rolain et al 2003, Blair et al 2004, Kelly et al 2004, Shaw et al 2004, Rolain et al 2005, Bauer et al 2006, Bernbeu-Wittel et al 2006, Bitam et al 2006, Marie et al 2006, Zavala-Velazquez et al 2006). Ctenocephalides felis is a biological vector for R felis; the organism can be transmitted transovarially and transstadially within the flea (Wedincamp and Foil 2002). Cats are commonly infested by C felis and as they are increasing in popularity as pets around the world, there has been increasing concern regarding flea-associated, zoonotic organisms like R felis and Bartonella species (Brown et al 2003). Prevalence rates of R felis DNA have been determined in C felis collected from cats in the United Kingdom (Kenny et al 2003, Shaw et al 2004), France (Rolain et al 2003), Israel (Bauer et al 2006), New Zealand (Kelly et al 2005), Australia (Schloderer et al 2006), and Thailand (Parola et al 2003). However, very few cats have been proven to have active R felis infection. Research colony cats exposed to fleas known to be carriers of R felis became seropositive by the 4th month post-exposure and two of the 12 cats had R felis DNA detected in blood by polymerase chain reaction (PCR) assay (Wedincamp and Foil 2000). In studies of naturally exposed cats (Sorvillo et al 1993, Breitschwerdt 2005, Case et al 2006), antibodies to R felis or related organisms were detected, but the cats were not assessed for current infection by PCR assay or culture. To our knowledge, R felis prevalence rates in C felis and the cats from which they were collected have not been determined in the United States.
In a previous study, our laboratory reported the prevalence rates of Bartonella species, Mycoplasma haemofelis, ‘Candidatus M haemominutum’, Ehrlichia species, Neorickettsia risticii, and Anaplasma phagocytophilum in blood of cats and the C felis collected from the cats concurrently (Lappin et al 2006). The purpose of this study is to report the prevalence rates of R felis DNA in extracts from the cat blood and fleas from that study.
Cat blood and their corresponding flea samples (92 pairs) were collected by veterinarians in Alabama (59 flea groups), Maryland (25 flea groups) and Texas (eight flea groups) as previously described (Lappin et al 2006). After DNA extraction and assessment by other PCR assays, the DNA samples were stored at −20°C until used in this study. For the cat samples, DNA had been extracted from 200 μl of blood in EDTA. For the flea samples, DNA had been extracted from between 1 and 14 fleas per cat, as previously described (Lappin et al 2006). After thawing the samples at 20°C, R felis PCR assays were performed using an adaptation of previously described protocols (Shaw et al 2004). Both the citrate synthase gene (gltA) and outer membrane protein B (ompB) specific oligonucleotide primers were utilized for all samples (Shaw et al 2004). The PCR assays were performed as previously described, except the final primer concentration was decreased from 0.5 μM to 0.4 μM of each primer per reaction to allow for optimization for use with the cat whole blood samples. Endpoint sensitivity testing was performed on a log dilution series of R felis genomic DNA for each primer set to ensure that no sensitivity was lost on either the cat whole blood or flea DNA samples. The stock concentration for gltA gene (5.32×102 μg/μl) and ompB gene (5.45×102 μg/μl) was determined by a spectrophotometer. For the individual assays, 1.10 fg DNA and 0.110 fg DNA (gltA and ompB, respectively) were detected. In both PCR assays, a negative control, consisting of PCR reagents and PCR water as template (no DNA template added), was utilized on each run. In addition, on each run in both PCR assays, purified R felis DNA and R rickettsii DNA were assessed as positive controls. DNA from positive samples in either PCR assay were purified using the Qiagen Gel Purification kit (Qiagen, Valencia, CA) according to manufacturer's protocol and submitted for sequencing at a commercially available laboratory (Macromolecular Research Core Laboratory, Colorado State University, Fort Collins, CO). Resultant sequences were then analyzed by comparison to sequences in GenBank using the BLAST program on the National Institutes of Health (NCBI) website (http://www.ncbi.nlm.nih.gov).
An amplicon consistent with that of R felis (gltA amplicon of 443 bp or ompB amplicon of 662 bp) was detected in 62 of 92 flea groups (67.4%). Of the positive samples, 62 of 62 (100%) were positive by use of the gltA primers and 59 of 62 (95.2%) were positive by use of the ompB primers. Both genes of interest were amplified from flea DNA extracts from Alabama and Maryland only. The Texas flea DNA extracts were negative on both PCR assays. Ample DNA for genetic sequencing was available for 52 samples positive with gltA primers and 49 samples positive with ompB primers. Results of sequencing showed that all samples were most homologous with R felis when compared to GenBank Accession number CP000053.1 (Rickettsia felis URRWXCal2, complete genome). The percentage homology ranged from 92% to 99% and from 88% to 100% for gltA and ompB, respectively. Rickettsia felis DNA was not amplified from any of the 92 samples of cat whole blood using either of the PCR assays. The overall prevalence rates of DNA of the infectious agents in the blood of the cats and their fleas are shown in Table 1.
Distribution of Rickettsia felis, Bartonella species, and hemoplasma species PCR assay results from cat blood and flea digests in the United States
Hemoplasma and Bartonella species results were initially published in Lappin et al 2006. Rickettsia felis results are those that were positive in either or both of the PCR assays.
Rickettsia felis DNA was amplified from 67.4% of the flea groups from client-owned cats in this study. While R felis DNA has been amplified from fleas collected from cats in multiple studies in different countries, to our knowledge, this is the highest prevalence rate reported to date. Appropriate controls were included on each PCR assay run and the amplicons that were sequenced were most homologous with R felis and so we believe that the positive PCR assay results document infection by R felis or a closely related organism. The differences between this study and those previously reported may simply relate to the geographical areas studied. It is also possible that differences between studies merely relate to the small sample sets available for evaluation. For example, while we failed to amplify R felis DNA from any of the eight flea groups from Texas, the prevalence was 4% in a previous study (Schriefer et al 1994). It is also possible that the differences in prevalence rates relate to differences in the sensitivity of the PCR assays utilized in the different studies.
DNA of R felis, Bartonella species, and hemoplasma species were amplified alone and in combination in fleas in this study (Table 1) which is similar to results obtained in a study in the United Kingdom (Shaw et al 2004). However, prevalence rates of the different agents varied between the two studies and as previously discussed may simply reflect climatic differences between the study areas. For example, the overall prevalence rates of R felis, Bartonella species, and hemoplasma species DNA in flea pools were 80.4% and 50% in the studies performed in the United States and United Kingdom, respectively. The R felis prevalence rates were 67.4% and 21.0% in the studies performed in the United States and the United Kingdom, respectively. In addition, Bartonella clarridgeiae DNA was amplified from 45.6% of flea groups in this study (Lappin et al 2006) but none of the fleas collected from cats in the United Kingdom (Shaw et al 2004).
While 67.4% of the flea groups assayed in this study were positive for R felis DNA, positive results were not obtained from any of the 92 cat blood samples. Rickettsia felis can be maintained within populations of fleas by vertical transmission and is transmitted between hosts by fleabites (Wedincamp and Foil 2002). As the fleas assessed in this study were collected directly from the cats it can be assumed that the cats were exposed to the organism by fleabites. Previous studies of naturally infected or experimentally exposed cats have documented that cats develop serum antibody responses to R felis or related organisms suggesting that at least transient infection occurs (Sorvillo et al 1993, Wedincamp and Foil 2000, Breitschwerdt 2005, Case et al 2006). The failure to amplify R felis DNA from the blood of cats described here can potentially be explained in several ways. First, it is possible that R felis DNA was present in the blood samples tested but at a copy number below the detectable limit of the assays used. While we believe that this hypothesis is unlikely based on the analytical sensitivities of the assays, future studies should also assess potentially more sensitive assays such as real-time PCR. It is also possible that the cats were infected with R felis but the organism was sequestered in tissues other than blood such as endothelial cells or dermal tissues; in a previous study of opossum, R typhii DNA was amplified from the spleen but not the liver by PCR assay (Williams et al 1992). Lastly, only a single blood sample was tested in the cats; it is possible that bacteremia was intermittent and missed during the sampling period.
Based on the results of the study described herein, it appears unlikely that domestic cats are effective reservoir hosts for R felis. It is possible that opossum or other animals serve as reservoir hosts for R felis (Williams et al 1992, Boostrom et al 2002). However, because the organism can be maintained within fleas by vertical transmission, a reservoir is not required.
Rickettsia felis, Bartonella species, or hemoplasma species DNA was amplified from 80.4% of the flea groups assessed in this study (Lappin et al 2006). We believe these results support the recommendation that flea control be maintained on cats in an attempt to lessen exposure of cats or people to these agents (Brown et al 2003). However, additional studies will be required to determine whether flea control lessens the prevalence of these infections in cats or humans.