Genetic Variants of Anaplasma phagocytophilum Infecting Dogs in Western Washington State

JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 2005, p. 796 801 Vol. 43, No. 2 0095-1137/05/$08.00 0 doi:10.1128/jcm.43.2.796 801.2005 Copyright 2005, American Society for Microbiology. All Rights Reserved. Genetic Variants of Anaplasma phagocytophilum Infecting Dogs in Western Washington State Florence M. Poitout, 1 * Joanne K. Shinozaki, 1 Patrick J. Stockwell, 2 Cynthia J. Holland, 3 and Sanjay K. Shukla 2 Phoenix Central Laboratory for Veterinarians, Everett, Washington 1 ; Clinical Research Center, Marshfield Clinical Research Foundation, Marshfield, Wisconsin 2 ; and Protatek Reference Laboratory, Chandler, Arizona 3 Received 7 July 2004/Returned for modification 7 August 2004/Accepted 17 October 2004 Eight dogs from western Washington State suspected of being infected with Anaplasma phagocytophilum because of the finding of morulae in peripheral blood neutrophils were studied for determination of the etiologic agent of disease. All cases were diagnosed between April 2003 and April 2004. Six of the eight dogs had no travel history during the 6 months prior to presentation. Two dogs had traveled within the Northwest United States and Canada. Fever, lethargy, and anorexia were the most common clinical signs in the dogs. Lymphopenia, thrombocytopenia, and an elevated activity of alkaline phosphatase in the serum were the most common laboratory findings. All dogs tested during the acute phase of clinical signs were seropositive for A. phagocytophilum antibodies but negative for Ehrlichia canis antibodies. PCR amplification and direct sequencing of portions of the 16S rrna gene from the whole blood of all seven dogs that were tested yielded A. phagocytophilum after a comparison to bacterial sequences available in the GenBank database. Five genetic variants were identified based on one or two nucleotide differences in the 16S rrna gene sequences at nucleotide positions 54, 84, 86, and 120. Individual dogs were infected with more than one variant. Treatment with doxycycline or tetracycline resulted in a rapid resolution of clinical signs. The occurrence of canine granulocytic anaplasmosis in western Washington State suggests that A. phagocytophilum infection should be considered in differential diagnoses of dogs presenting with lethargy, anorexia, fever, and lameness, particularly in the context of lymphopenia, thrombocytopenia, and increased serum alkaline phosphatase. The zoonotic importance of A. phagocytophilum should support an increase in surveillance for horses and people residing in this area. * Corresponding author. Mailing address: Phoenix Central Laboratory, 11620 Airport Rd., Everett, WA 98204-3742. Phone: (425) 355-8787. Fax: (425) 355 3676. E-mail: poitout@msn.com. Anaplasma phagocytophilum is a newly designated tick-borne pathogen that joins together three previously described species, Ehrlichia equi, Ehrlichia phagocytophilum, and human granulocytic agent (HGE agent), after a reorganization of the families Rickettsiaceae and Anaplasmataceae in the order Rickettsiales (10). The disease caused by A. phagocytophilum was formerly designated granulocytic ehrlichiosis, but it is now reported as granulocytic anaplasmosis. This new classification was justified by Dumler et al. (10) on the basis of a high degree ( 99.8%) of similarity in the 16S rrna genes of these three previous Ehrlichia species. However, disparities in seroprevalence, incidence, clinical severity, and disease manifestation in people, horses, and ruminants in North America and Europe suggest that there are still undiscovered differences among these bacteria (1, 9, 23). Ehrlichia ewingii, another etiologic agent of canine granulocytic inclusions, is predominantly found in the Southern and Southeast United States. Genetic analyses revealed that A. phagocytophilum and E. ewingii belong to distinct groups of bacteria (10). A. phagocytophilum is also distinct from Ehrlichia canis, the etiologic agent of canine monocytic inclusions. A. phagocytophilum causes granulocytic inclusions in domestic dogs, horses, cattle, sheep, goats, llamas, cats, and people (3, 6, 20, 27, 29, 35). Ruminants in the United States other than llamas have not been reported to be infected by A. phagocytophilum, unlike their European counterparts. In dogs, A. phagocytophilum infection can manifest as acute lethargy, anorexia, and fever. A. phagocytophilum infection has been described in California, the upper Midwest and Northeast United States, British Columbia, and Europe. It is a tick-borne disease transmitted by Ixodes pacificus in the Western United States and Ixodes scapularis in the Midwest and Eastern United States (23, 30). In the Western United States and Canada, equine granulocytic anaplasmosis (EGA) was reported as early as 1968 in California (22) and in 1996 in British Columbia (5). In California, A. phagocytophilum infections were reported for dogs in 1982 (21) and for people in 1994 (26). EGA was first reported in Washington State in 1987 (22), but to our knowledge, cases of canine and human anaplasmosis have not been previously reported in Washington State. The aim of this report was to describe clinical, clinicopathologic, and molecular findings for eight dogs that were naturally infected with A. phagocytophilum in western Washington State. MATERIALS AND METHODS Case selection. Eight dogs that were diagnosed with granulocytic inclusions or morula-like structures from laboratory submissions to Phoenix Central Laboratory (Everett, Wash.) between April 2003 and April 2004 were studied. The diagnosis of granulocytic ehrlichiosis was made by finding ehrlichial morulae in neutrophils in Wright-stained peripheral blood smears examined by light microscopy. Data collection. For each dog, the following information was recorded when available: signalment, historical complaints, the dog s travel history for 6 months prior to diagnosis, potential tick exposures, clinical signs, results of a physical 796

VOL. 43, 2005 A. PHAGOCYTOPHILUM VARIANTS IN DOG INFECTIONS 797 TABLE 1. Relevant characteristics of diagnosis and place of residence for eight dogs from western Washington State with A. phagocytophilum infection Dog no. Signalment a Mo and yr of diagnosis City of residence b Historical complaints and clinical signs 1 Beagle, FS, 4 yrs April 2003 Gig Harbor Lethargy, anorexia, occasional coughing, fever, dehydration 2 American Eskimo, MN, 9 yrs May Gig Harbor Lethargy, anorexia, occasional coughing, fever, dehydration 3 Beagle cross, FS, 6 yrs June Tumwater Lethargy, inappetence, left forelimb lameness, fever 4 Yorkshire terrier, FS, 6 yrs July Gig Harbor Lethargy, anorexia, left hindlimb lameness, fever 5 Cocker spaniel, MN, 13 yrs July Blaine Lethargy, inappetence, occasional coughing, fever, pale mucous membranes 6 Briard, FS, 12 yrs October Renton Right hindlimb lameness, right stifle pain, slight weight loss, possible PU/PD c 7 Cocker, M, 6 yrs March 2004 Poulsbo Lethargy, anorexia, fever, left rear lameness, slight weight loss, exaggerated breathing 8 Bichon frise, M, 8 yrs April 2004 Port Orchard Lethargy, anorexia, left rear lameness, fever a MN, neutered male; M, intact male; FS, spayed female. b All cities are located in Washington State. c PU/PD, polyuria-polydipsia. examination, clinicopathologic findings, treatment, and response to treatment. Clinicopathologic evaluations comprised complete blood counts and serum chemistry analyses. Blood smears were reviewed by a board-certified veterinary clinical pathologist. The percentage of infected neutrophils was also evaluated, and the platelet count was estimated when an accurate platelet count could not be obtained by a hematology analyzer (Advia 120; Bayer Diagnostic Division, Tarrytown, N.Y.) because of platelet clumping. For platelet count estimation, the mean platelet number counted in 10 fields at a magnification of 100 was multiplied by 20,000/ l. Serology. For six of the eight infected dogs, serum samples were titrated for antibodies to E. canis and A. phagocytophilum (Protatek Reference Laboratory, Chandler, Ariz.) by indirect fluorescent antibody testing during the acute phase of the disease. For two of the eight dogs, recovery-phase serum samples were titrated for antibodies to A. phagocytophilum 7 to 12 months after the initial infection. 16S rrna gene PCR and DNA sequencing. For seven of the eight infected dogs, EDTA-anticoagulated whole blood was collected and stored at 20 C until samples were sent frozen to the laboratory for further analysis. DNAs were extracted from 200 l of EDTA-anticoagulated whole blood by use of a QIAamp DNA mini extraction kit (QIAGEN Inc., Valencia, Calif.) according to the manufacturer s protocol. The DNAs were eluted into 200 l of elution buffer. PCRs were performed in a PE 7000 thermocycler by use of a QIAGEN HotStar kit. Each 30- l reaction mix contained 15 l of HotStar mix, 20 pmol of each primer (HGE1 [5 TCC TGG CTC AGA ACG AAC 3 ] and 706R [5 TCC TGT TTG CTC CCC ACG CTT TC 3 ]), 2.5 l of template DNA, and 8 l ofh 2 O. The amplification products were resolved in an ethidium bromide-containing 1.5% agarose gel and photographed by use of the GelDoc 2000 imaging system (Bio-Rad Inc., Hercules, Calif.). Reactions were considered positive if they showed an 771-bp amplicon for the primer set HGE1-706R. Double-distilled, filtered (0.45- m pore size), autoclaved sterile water was used as a negative control, and a previously identified A. phagocytophilum DNA from a dog blood sample was used as a positive control. Five of seven dog blood samples were also independently tested by PCRs with the A. phagocytophilum-specific primers HGE521 and HGE790. In each case, the single amplicon of the expected size was column purified by use of a QIAGEN purification kit (QIAGEN Inc.) and then sequenced. The DNA sequences were aligned by use of the Seqman program (DNASTAR Inc., Madison, Wis.). The consensus sequence was compared to bacterial sequences available in the GenBank database by use of the BLAST 2.0 program (National Center for Biotechnology Information, Bethesda, Md.). Genetic variants were confirmed by comparing the A. phagocytophilum 16S rrna gene sequences from the seven infected dogs with that of the prototype A. phagocytophilum human strain (GenBank accession number U02521) (8). Nucleotide sequence accession numbers. The rrna gene sequences for all of the variants determined for this study have been deposited in GenBank under accession numbers AY741095 to AY41099 for Washington (WA) variants 1 to 5, respectively. RESULTS Dogs. The mean age of infected dogs was 8 years, and the mean weight was 17.9 kg (39.7 lb). Four dogs were females, and four were males (Table 1). Six dogs had been spayed or neutered, and two were intact males. No breed was overrepresented (Table 1). Seven of the eight dogs were diagnosed between April and July, and one dog was diagnosed in October (Table 1). All dogs were residents of western Washington State (Table 1). There was no history of travel for six of the eight dogs during the 6 months prior to presentation. One dog had traveled to Oregon, Idaho, and Montana, and one dog had traveled to Idaho, Montana, and Canada (Alberta and Saskatchewan). None of the dogs owners had observed any tick attachment. There was no history of concurrent disease reported. None of the dogs were receiving medications at the time of diagnosis. Historical complaints included lethargy and inappetence or anorexia for up to 5 days prior to presentation for seven of the dogs (87%). Lameness was reported for five dogs (62%). Owners also reported occasional coughing for three dogs (50%). Possible polyuria and polydipsia were reported for one dog (Table 1). Physical examination. Physical examination abnormalities consisted of fever in eight dogs (100%), mild to moderate dehydration in three dogs (37%), slight weight loss in two dogs (25%), and pale mucous membranes in one dog (12%) (Table 1). Lameness was evident during physical exams for five of the eight dogs, as reported by the clients (62%). Orthopedic exams revealed one dog with pain upon palpation of the right stifle joint, but they were normal for the other dogs. Joint effusion was not detected in any dog. Likewise, no neurological deficits were identified in any of the dogs. Coughing was observed in three dogs, with one of the three suspected of having episodic tracheal collapse. One dog was reported to have slightly exaggerated breathing. No cardiac abnormalities were detected at the time of the exams. Abdominal palpation was within normal limits for all eight dogs. No lymphadenopathy was found in any dog. Laboratory findings. Complete blood counts and serum chemistry profiles were available for all eight dogs (Table 2). Urinalysis results were not available for any of the dogs. Hematologic abnormalities included leukopenia in five dogs (62%), with lymphopenia in eight dogs (100%) and mild to moderate neutropenia in three dogs (37%). Thrombocytopenia was observed in seven of the eight dogs (87%), although some platelet clumps were present in all dogs, and microscopic evaluations estimated decreased platelet numbers in these seven dogs (Table 2). A mild, normocytic, normochromic, nonregenerative anemia was identified in three dogs (37%). Blood smear examinations revealed granulocytic inclusions or moru-

798 POITOUT ET AL. J. CLIN. MICROBIOL. TABLE 2. Hematologic findings for eight dogs from western Washington State with A. phagocytophilum infection Hematologic parameter Value for indicated dog 1 2 3 4 5 6 7 8 Reference range Hematocrit (%) 43.8 40.6 43.0 39.4 29.3 35.6 42.5 36.4 37.0 55.0 Reticulocutes/ l 12,810 4,500 10,200 White blood cells (10 3 / l) 3.7 9.3 4.5 6.2 4.2 4.2 7.7 4.8 6.0 17.0 Bands / l 37 0 0 0 0 0 0 0 0 300 Neutrophils / l 3,441 8,463 3,530 5,456 3,906 3,540 6,660 3,744 3,600 11,500 Lymphocytes / l 185 372 730 496 252 480 690 336 1,000 4,800 Monocytes / l 37 465 160 248 42 150 290 672 150 1,350 Eosinophils / l 0 0 20 0 0 10 20 48 0 1,250 Basophils / l 0 0 20 0 0 20 20 0 0 500 Platelets (10 3 / l) 40 10 80 150 150 100 250 36 200 500 % of neutrophils containing organisms 21 9 12 32 29 15 30 15 la-like structures from all eight dogs (100%). Nine to thirty-two percent of neutrophils had inclusions (Table 2). Infected neutrophils generally contained one inclusion, although two or more were occasionally found (Fig. 1). Inclusions were not identified in leukocytes other than neutrophils. Serum biochemical abnormalities were usually mild and nonspecific for all eight dogs. Increased alkaline phosphatase values were noted for six of eight dogs (75%), hypophosphatemia occurred in five of eight dogs (62%), decreased venous CO 2 was noted for three of seven dogs (42%), and low total T 4 and hyperlipasemia were each seen in two of seven dogs (28%). Fecal exams were performed on two dogs, and one dog had evidence of roundworms. Arthrocentesis was not performed on any of the dogs. Serology. Five of the six dogs for which serological data were available were seropositive for A. phagocytophilum (83%) (Table 3). All five dogs tested for E. canis were seronegative (100%) (Table 3). Three dogs that were retested for A. phagocytophilum 7 to 12 months after the initial presentation were seropositive, and two dogs had similar or higher titers than those that were previously measured (Table 3). 16S rrna gene sequence analysis. PCRs for seven of eight dogs that were tested with the primer pairs HGE1 and 706R were positive (Fig. 2), with 771-bp amplicons, whereas PCRs for five of seven dogs that were tested with primers HE521 and HE790 yielded 300-bp amplicons. Both the sense and antisense strands of each amplicon were sequenced. Sequences were manually edited to remove or confirm ambiguities. A consensus sequence was obtained for each amplicon by comparing both the sense and antisense sequences. The consensus sequence from each dog sample had a high degree (99.71% or higher) of overall identity with the reported sequence of the A. phagocytophilum prototype (8). Given that several of the individual samples yielded mixed sequences in the initial analysis, we reamplified and sequenced the 16S rrna gene from each sample to confirm the identities of genetic variants. We subsequently obtained five- to sevenfold coverage of the 16S rrna gene sequences (nucleotide positions 50 to 750) of A. phagocytophilum from each of the seven samples. Genetic variants of A. phagocytophilum from Washington State. Based on the redundancy of individual sequencing data, we discovered five genetic variants of A. phagocytophilum and one prototype strain sequence (GenBank accession no. U02521). Nucleotide differences were observed at positions 54 (A to C), 84 (G to A or C), 86 (T to C), and 120 (A to G), as shown in Table 4. These genetic differences at positions 54, 86, and 120 have not been described before. There were three variants with single nucleotide changes and two variants with two nucleotide changes. WA variant 4, the most common variant, was identified in six of the seven dogs, followed by variant 2, which was present in two dogs. Coinfection with more than one variant was also observed, including one case in which a dog was coinfected with the prototype strain and variant 3. Treatment. Six dogs were treated with doxycycline (5 to 10 mg/kg of body weight [2.3 to 4.5 mg/lb], administered orally every 12 h [q12h] or q24h) for 10 to 21 days, and two dogs were treated with tetracycline (22 mg/kg [10 mg/lb], administered orally q8h) for 14 or 21 days. Additional therapies included intravenous or subcutaneous fluids for two dogs, pyrantel pamoate for one dog, and levothyroxine for one dog. Transient therapies that were initiated prior to rickettsial disease identification but discontinued once the diagnosis was available included cephalexin for one dog and both carprofen and methylsulfonylmethane for another dog. A favorable and rapid response was reported for all eight dogs (100%). Six of the eight dogs appeared to have a complete resolution of the historical complaints and physical exam abnormalities within FIG. 1. A. phagocytophilum morulae in a neutrophil. The figure shows a peripheral blood smear from a dog stained with Wright stain. Magnification, 700.

VOL. 43, 2005 A. PHAGOCYTOPHILUM VARIANTS IN DOG INFECTIONS 799 TABLE 3. Serologic, PCR, and sequencing findings for eight dogs from western Washington State with A. phagocytophilum infection Dog no. Anti- A. phagocytophilum titer a Anti- A. phagocytophilum recovery titer a E. canis serology a 16S rrna PCR result a % Sequence homology between nucleotide positions 50 and 750 a 1 NA 1:40,960 (12 mos) b NA NA NA 2 1:640 1:640 (10 mos) b Negative Positive 99.7 3 1:256 NA Negative Positive 99.7 4 1:2,560 NA NA Positive 99.7 5 Negative, 1:16 NA Negative Positive 99.7 6 1:128 1:320 (7 mos) b Negative Positive 99.7 7 NA NA NA Positive 99.7 c 8 1:10,240 NA Negative Positive 99.7 c a NA, not available. b Postinfection. c Only nucleotide positions 50 to 700 were sequenced for cases 7 and 8. 24 to 48 h after the initiation of doxycycline or tetracycline. Two dogs (dogs 5 and 7) had progressive improvement, with a complete resolution within 5 to 6 days. There was no delay in the diagnosis or initiation of treatment for these two dogs compared to that for the other dogs. All eight dogs remained clinically healthy 7 to 19 months after the initial diagnosis. DISCUSSION Geographic, clinical, clinicopathologic, and molecular evidence from the eight described dogs confirmed the occurrence of canine infections with A. phagocytophilum, designated canine granulocytic anaplasmosis (CGA), in western Washington State. In the United States, occurrences of CGA have been previously reported for Minnesota, Wisconsin, Rhode Island, and California (18, 21, 28, 33). This is the first clinical report describing molecular evidence of A. phagocytophilum infection in dogs from western Washington State. We have observed serologic and cytologic evidence of granulocytic anaplasmosis in dogs and horses in this area in previous years (data not shown). Madigan and Gribble (22) previously reported the presence of EGA in Washington State in 1987, but CGA has never been reported in this state. We suspect that infection with A. phagocytophilum has been historically underdiagnosed in dogs from western Washington State. There was no recorded history of tick exposure for any of the dogs, but we suspect that A. phagocytophilum was transmitted to dogs by I. pacificus because this tick species has been linked to the transmission of EGA in horses in California and British Columbia. Attempts to detect A. phagocytophilum in other species of ixodid ticks in California have been unsuccessful (2, 19). The transstadial transmission of A. phagocytophilum in I. pacificus and experimental transmissions of A. phagocytophilum from infected to susceptible horses through bites of I. pacificus ticks have been demonstrated (30). The seasonality and geographic distribution of A. phagocytophilum infections are probably determined by the feeding habits of each particular tick vector. I. scapularis is known to be the vector of A. phagocytophilum in the upper Midwest and Northeast United States, but it is not prevalent in the Western United States. All cases from western Washington State were observed between April and July, with one case observed in October. A. phagocytophilum infections in Minnesota and Wisconsin have been observed mostly in October and November, with few cases between April and August (18). The difference in seasonality between the two areas may be due to the differences in vectors and weather. People and domestic animals such as dogs are incidental hosts and are unlikely to play an important role in the natural maintenance cycle of this pathogen. A. phagocytophilum is maintained in wildlife hosts. Natural reservoirs for A. phagocytophilum in the United States include white-tailed deer, white-footed mice, Eastern chipmunks, dusky-footed wood rats, and Southern red-backed voles (4, 31, 34). In the Western United States, dusky-footed wood rats, mule deer, and black-tailed deer serve as reservoirs (25, 26). In California, mountain lions and coyotes are also infected, but the role of these animals as reservoirs is undefined (14, 27). A. phagocytophilum has not been reported thus far in regions where E. ewingii or its vector tick Amblyomma americanum is endemic. TABLE 4. Variation within nucleotide region 54 to 120 in 16S rrna gene sequences of A. phagocytophilum identified in dogs from western Washington State A. phagocytophilum variant Nucleotide at indicated position b 54 84 86 120 FIG. 2. Agarose gel showing HGE1-706R amplicons. Canine sample numbers are listed above each lane., negative control;, A. phagocytophilum positive control. A. phagocytophilum ha a A G T A WA variant 1 A C T A WA variant 2 A A C A WA variant 3 A G T G WA variant 4 A A T A WA variant 5 C A T A a The number designations for the 16S rrna gene sequence are based on the reported sequence under GenBank accession no. U02521. b Changes from the reference sequence are indicated in bold.

800 POITOUT ET AL. J. CLIN. MICROBIOL. In this study, no breed, sex, or size was overrepresented. The ages of dogs ranged from 4 to 13 years, with a mean of 8 years. Age may be a susceptibility factor for A. phagocytophilum infection. Other authors have reported a similar age distribution, with all infected dogs being older than 1 year, whereas younger dogs had other tick-borne infections (11). In dogs, disease incubation is 1 to 2 weeks after tick transmission. Reported clinical findings are almost exclusively found during the bacteremic phase. Seven of the eight dogs had demonstrated moderate to severe overt clinical signs for only 5 days or less. More than 87% of our cases presented with nonspecific clinical signs, including fever, lethargy, and/or anorexia. Five dogs presented with slight to acute lameness. One dog had a painful joint, whereas the others had no evidence of any orthopedic or neurological abnormalities. Three dogs presented with occasional coughing which appeared to resolve after treatment for A. phagocytophilum infection. One dog was reported to have labored breathing. All eight dogs were thrombocytopenic and lymphopenic, and three dogs had nonregenerative anemia. Seven of the eight dogs had increased alkaline phosphatase values. These findings are very similar to those described for dogs in Minnesota, Wisconsin, and Sweden (11, 17, 18). None of the dogs had hypoproteinemia, hypoalbuminemia, hyperamylasemia, or regenerative left shift, as previously reported (11, 18). All but two of the dogs recovered rapidly, within 24 to 48 h after the initiation of treatment with doxycycline or tetracycline. The other two dogs had progressive improvement, with a complete resolution within 5 to 6 days. No relapse was observed, and all dogs remained healthy 7 to 19 months after the initial diagnosis. The eight naturally infected cases reported here were identified by the finding of A. phagocytophilum inclusions in circulating neutrophils. The granulocytic inclusions observed in neutrophils were morula-like structures. Inclusions of A. phagocytophilum have also been reported to occur in eosinophils. Bacteria of the family Anaplasmataceae are gram negative, nonmotile, coccoid-to-ellipsoid organisms. All are obligate intracellular parasites. They inhabit membrane-lined vacuoles, forming morulae in hematopoietic cells of mammalian hosts. Both polyclonal indirect immunofluorescence and immunoblotting serologic assays have been developed for the detection of these organisms. For the use of only immunofluorescence assays, a fourfold increase in the antibody titer to A. phagocytophilum between the acute and convalescent phases is necessary to confidently diagnose an active infection because up to 40% of acutely ill, morula-positive dogs can be seronegative at the time of presentation and because dogs may have detectable levels of antibodies from a previous exposure (18). Five of the six dogs for which serological data were available were seropositive for A. phagocytophilum. Experimental studies with horses and dogs have documented that seroconversion to A. phagocytophilum occurs 2 to 5 days after the first appearance of morulae in the peripheral blood (11). The dog that tested negative for A. phagocytophilum by serology had appeared sick for 2 days, and an adequate acute-phase titer may not have been reached by that point. In one study (12), antibody titers to A. phagocytophilum reverted to nondetectable levels by 7 to 8 months after acute experimental canine infections. For our cases, however, the antibody titers remained elevated 7 and 10 months after initial infections. Natural chronic A. phagocytophilum disease has not been documented yet. All six dogs tested for E. canis were seronegative. Although some degree of cross-reactivity does exist between A. phagocytophilum and E. canis, serological reactivity is uncommon between A. phagocytophilum and E. ewingii or Ehrlichia chaffeenssis (12, 18). For dogs, granulocytic inclusions caused by A. phagocytophilum must be differentiated from granulocytic inclusions caused by E. ewingii. E. ewingii infections are observed in the Southern United States and are responsible for an acute polyarthritic disorder characterized by fever, anorexia, lameness, mild thrombocytopenia without bleeding diatheses, and mild nonregenerative anemia (16). Currently, dogs with E. ewingii infections are identified by the presence of antibodies that are reactive to E. canis antigens and by E. ewingii-specific PCR, and they do not generally produce antibodies that are reactive with A. phagocytophilum antigens. A thorough understanding of the significance of genetic variants of A. phagocytophilum is still far from complete. The genetic variants described for samples from Rhode Island and Connecticut by Massung et al. (23) were speculated to interfere with the transmission and maintenance of strains that cause the disease in humans. AP variant 1 was unable to establish infections in mice under laboratory conditions (24). Massung et al. speculated that a lower incidence of human cases would occur in areas where variants predominate by competitively infecting certain reservoir or vector populations. Any variants that have a different physiology may have a different pathogenicity as well. Some variants may be more, less, or as pathogenic as the prototype strain that causes disease in humans. WA variants 1 to 5 were identified from dogs that presented clinical syndromes consistent with granulocytic anaplasmosis. These variants have yet to be identified from tick vectors or other hosts, including humans. Further research will be required to assess how the presence of these variants affects the prevalence rate of human anaplasmosis cases in Washington State. WA variants in Washington were different from A. phagocytophilum variants reported for sheep in Norway (32), two people in northern California (13), and llama-associated ticks (3). WA variant 4 is identical to the E. equi CA human variant reported by Massung et al. (23). The infection of a dog by a variant that has been identified in both horses and people suggests the possibility of additional variants having multiple hosts as well. Genetic variants of A. phagocytophilum are expected to reveal more about their pathogenicity as clinical cases or their presence in tick vectors is reported. A. phagocytophilum genogroup rickettsiae infecting I. pacificus ticks from California thus appear to be far less prevalent than those in I. scapularis from the Eastern and Midwest United States (2). From a public health standpoint, the risk of exposure to granulocytic anaplasmosis in California and perhaps in other western states that are devoid of I. scapularis may be lower than that in eastern and midwestern states (2). Dogs and horses may contribute to the enzootic cycle and to human infections. However, direct infections from dogs to people have not been identified, and tick transmission is believed to be the major means of human infection (25). Human granulocytic anaplasmosis has been reported in northern California (7) but,

VOL. 43, 2005 A. PHAGOCYTOPHILUM VARIANTS IN DOG INFECTIONS 801 to our knowledge, has not yet been reported in western Washington State. Infections of dogs with A. phagocytophilum can induce moderate to severe illness. A. phagocytophilum infection should be considered part of a differential diagnosis for dogs presenting with lethargy, anorexia, and fever, particularly in the context of lameness, lymphopenia, thrombocytopenia, and increased alkaline phosphatase. The identification of canine granulocytic anaplasmosis in western Washington State should support increased surveillance for A. phagocytophilum in people and horses from the same region. ACKNOWLEDGMENTS S.K.S. acknowledges Amy Sekorski for her contributions to this project. We thank Barbara Greig for coordinating our PCR and serologic assays and for her advice. We are also thankful to Kurt Reed and Faye Sturtevant for their support. We are grateful to the referring veterinarians, especially Lisa Woods, who supplied us with the medical records and blood samples of infected patients. REFERENCES 1. Asanovich, K. M., J. S. Bakken, J. E. Madigan, M. Aguero-Rosenfeld, G. P. Wormser, and J. S. Dumler. 1997. Antigenic diversity of granulocytic Ehrlichia isolates from humans in Wisconsin and New York and a horse in California. J. Infect. Dis. 176:1029 1034. 2. Barlough, J. E., J. E. Madigan, V. L. Kramer, J. R. Clover, L. T. Hui, J. P. Webb, and L. K. Vredevoe. 1997. Ehrlichia phagocytophila genogroup Rickettsiae in ixodid ticks from California collected in 1995 and 1996. J. Clin. Microbiol. 35:2018 2021. 3. Barlough, J. E., J. E. Madigan, D. R. Turoff, J. R. Clover, S. M. Shelly, and J. S. Dumler. 1997. An Ehrlichia strain from a llama (Lama glama) and llama-associated ticks (Ixodes pacificus). J. Clin. Microbiol. 35:1005 1007. 4. Belongia, E. A., K. D. Reed, P. D. Mitchell, C. P. Kolbert, D.H. Persing, J. S. Gill, and J. J. Kazmierczack. 1997. Prevalence of granulocytic Ehrlichia infection among white-tailed deer in Wisconsin. J. Clin. Microbiol. 35:1465 1468. 5. Berrington, A., R. Moats, and S. A. Lester. 1996. A case of Ehrlichia equi in an adult horse in British Columbia. Can. Vet. J. 37:174 175. 6. Bjoersdorff, A., L. Svendenius, J. H. Owens, and R. F. Massung. 1999. Feline granulocytic ehrlichiosis a report of a new clinical entity and characterization of the infectious agent. J. Small Anim. Pract. 40:20 24. 7. Chae, J. S., J. E. Foley, J. S. Dumler, and J. E. Madigan. 2000. Comparison of the nucleotide sequences of 16S rrna, 444 Ep-ank and groesl heat shock operon genes in naturally occurring Ehrlichia equi and human granulocytic ehrlichiosis agent isolates from Northern California. J. Clin. Microbiol. 38:1364 1369. 8. Chen, S. M., J. S. Dumler, J. S. Bakken, and D. H. Walker. 1994. Identification of a granulocytotropic Ehrlichia species as the etiological agent of human disease. J. Clin. Microbiol. 32:589 595. 9. Dumler, J. S., K. M. Asanovich, and J. S. Bakken. 2003. Analysis of genetic identity of North American Anaplasma phagocytophilum strains by pulsedfield gel electrophoresis. J. Clin. Microbiol. 41:3392 3394. 10. Dumler, J. S., A. F. Barbet, C. P. Bekker, G. A. Dasch, G. H. Palmer, S. C. Ray, Y. Rikihisa, and F. R. Rurangirwa. 2001. Reorganization of genera in the families Rickettsiaceae and Anaplasmataceae in the order Rickettsiales: unification of some species of Ehrlichia with Anaplasma, Cowdria with Ehrlichia and Ehrlichia with Neorickettsia, descriptions of six new species combinations and designation of Ehrlichia equi and HGE agent as subjective synonyms of Ehrlichia phagocytophila. Int. J. Syst. Evol. Microbiol. 51:2145 2165. 11. Egenvall, A., A. Bjoersdoff, I. Lilliehook, E. O. Engvall, E. Karstam, K. Artusson, A. Hedhammar, and A. Gunnarson. 1998. Early manifestations of granulocytic ehrlichiosis in dogs inoculated experimentally with a Swedish Ehrlichia species isolate. Vet. Rec. 143:412 417. 12. Egenvall, A. E., A. A. Hedhammar, and A. I. Bjoersdorff. 1997. Clinical features and serology of 14 dogs affected by granulocytic ehrlichiosis in Sweden. Vet. Rec. 140:222 226. 13. Foley, J. E., L. Crawford-Miksza, J. Stephen Dumler, C. Glaser, J.-S. Chae, E. Yeh, D. Schurr, R. Hood, W. Hunter, and J. E. Madigan. 1999. Human granulocytic ehrlichiosis in Northern California: two case descriptions with genetic analysis of the Ehrlichiae. Clin. Infect. Dis. 29:388 392. 14. Foley, J. E., P. Foley, M. Jecker, P. K. Swift, and J. E. Madigan. 1999. Granulocytic ehrlichiosis and tick infestation in mountain lions in California. J. Wildl. Dis. 35:703 709. 15. Foley, J. E., P. Foley, and J. E. Madigan. 2001. Spatial distribution of seropositivity to the causative agent of granulocytic ehrlichiosis in dogs in California. Am. J. Vet. Res. 62:1599 1605. 16. Goldman, E. E., E. B. Breitschwerdt, C. B. Grindem, B. C. Hegarty, J. J. Walls, and J. S. Dumler. 1998. Granulocytic ehrlichiosis in dogs from North Carolina and Virginia. J. Vet. Intern. Med. 12:61 70. 17. Greig, B. 2000. Granulocytic ehrlichioses, p. 298 300. In R. W. Kirk and J. D. Bonagura (ed.), Kirk s current veterinary therapy XIII. The W. B. Saunders Co., Philadelphia, Pa. 18. Greig, B., K. M. Asanovich, P. I. Armstrong, and J. S. Dumler. 1996. Geographic, clinical, serologic, and molecular evidence of granulocytic ehrlichiosis, a likely zoonotic disease, in Minnesota and Wisconsin dogs. J. Clin. Microbiol. 34:44 48. 19. Holden, K., J. T. Boothby, S. Anand, and R. F. Massung. 2003. Detection of Borrelia burgdorferi, Ehrlichia chaffeensis, Anaplasma phagocytophilum in ticks (Acari: Ixodae) from a coastal region of California. J. Med. Entomol. 40:534 539. 20. Lotric-Furlan, S., T. Avsic-Zupanc, M. Petrovec, W. L. Nicholson, J. W. Sumner, J. E. Childs, and F. Strle. 2001. Clinical and serological follow-up of patients with human granulocytic ehrlichiosis in Slovenia. Clin. Diagn. Lab. Immunol. 8:899 903. 21. Madewell, B. R., and D. H. Gribble. 1982. Infection in two dogs with an agent resembling Ehrlichia equi. J. Am. Vet. Med. Assoc. 180:512 514. 22. Madigan, J. E., and D. H. Gribble. 1987. Equine ehrlichiosis in northern California: 49 cases (1968 1981). J. Am. Vet. Med. Assoc. 190:445 448. 23. Massung, R. F., M. J. Mauel, J. H. Owens, N. Allan, J. W. Courtney, K. C. Stafford III, and T. N. Mather. 2003. Genetic variants of Ehrlichia phagocytophila, Rhode Island and Connecticut. Emerg. Infect. Dis. 8:467 472. 24. Massung, R. F., R. A. Priestley, N. J. Miller, T. N. Mather, and M. L. Levin. 2003. Inability of a variant strain of Anaplasma phagocytophilum to infect mice. J. Infect. Dis. 188:1757 1763. 25. McQuiston, J. H., C. L. McCall, and W. L. Nicholson. 2003. Ehrlichiosis and related infections. J. Am. Vet. Med. Assoc. 223:1750 1756. 26. Nicholson, W. L., M. B. Castro, V. L. Kramer, J. W. Sumner, and J. E. Childs. 1999. Dusky-footed wood rats (Neotoma fuscipes) as reservoirs of granulocytic Ehrlichiae (Rickettsiales: Ehrlichieae) in Northern California. J. Clin. Microbiol. 37:3323 3327. 27. Pusterla, N., C. C. Chang, B. B. Chomel, J. S. Chae, J. E. Foley, E. DeRock, V. L. Kramer, H. Lutz, and J. E. Madigan. 2000. Serologic and molecular evidence of Ehrlichia sp. in coyotes in California. J. Wildl. Dis. 36:494 499. 28. Pusterla, N., J. B. Pusterla, P. Deplazes, C. Wolfensberger, W. Muller, A. Horauf, C. Reusch, and H. Lutz. 1998. Seroprevalence of Ehrlichia canis and of canine granulocytic Ehrlichia infection in dogs in Switzerland. J. Clin. Microbiol. 36:3460 3462. 29. Remy, V., Y. Hansmann, S. De Martino, D. Christmann, and P. Brouqui. 2003. Human anaplasmosis presenting as atypical pneumonitis in France. Clin. Infect. Dis. 37:846 848. 30. Richter, P. J., R. B. Kimsey, J. E. Madigan, J. E. Barlough, J. S. Dumler, and D. L. Brooks. 1996. Ixodes pacificus (Acari: Ixodidae) as a vector of Ehrlichia equi (Rickettsiales: Ehrlichieae). J. Med. Entomol. 33:1 5. 31. Stafford, K. C., R. F. Massung, and L. A. Magnarelli. 1999. Infection with agents of human granulocytic ehrlichiosis, Lyme disease, and babesiosis in wild white-footed mice (Peromyscus leucopis) in Connecticut. J. Clin. Microbiol. 37:2887 2892. 32. Stuen, S., I. Van De Pol, K. Bergstrom, and L. M. Schouls. 2002. Identification of Anaplasma phagocytophila (formerly Ehrlichia phagocytophila) variants in blood from Sheep in Norway. J. Clin. Microbiol. 40:3192 3197. 33. Walker, D. H., and J. S. Dumler. 1996. Emergence of the ehrlichioses as human health problems. Emerg. Infect. Dis. 37:2887 2892. 34. Walls, J. J., B. Greig, and D. F. Neitzel. 1997. Natural infection of small mammal species in Minnesota with the agent of human granulocytic ehrlichiosis. J. Clin. Microbiol. 35:853 855. 35. Weber, R., N. Pusterla, M. Loy, C. M. Leutenegger, G. Schar, D. Baumann, C. Wolfensberger, and H. Lutz. 2000. Serologic and clinical evidence for endemic occurrences of human granulocytic ehrlichiosis in north-eastern Switzerland. Schweiz. Med. Wochenschr. 130:1462 1470.