• Open Access

Canine Granulocytic Anaplasmosis: A Review

Authors


Corresponding author: J.E. Sykes Department of Medicine and Epidemiology, School of Veterinary Medicine, University of California, Davis, Davis, CA, 95616 e-mail: jesykes@ucdavis.edu

Abstract

Anaplasma phagocytophilum is an emerging pathogen of humans, horses, and dogs worldwide that is transmitted by Ixodid ticks and maintained in a variety of small wild mammal species. Recent studies suggest that multiple strains of A. phagocytophilum may be circulating in wild and domestic animal populations, and these strains may have differential host tropisms and pathogenicity. The organism infects and survives within neutrophils by disabling key neutrophil functions, including neutrophil motility, phagocytosis, the oxidative burst mechanism, and neutrophil-endothelial cell interactions, as well as interfering with neutrophil apoptosis. Coinfections with other tick-borne pathogens may occur, especially Borrelia burgdorferi. A. phagocytophilum causes an acute febrile illness in dogs with lethargy and inappetence. Less frequent signs include lameness, coughing, polydipsia, intermittent vomiting, and hemorrhages. Diagnosis is based on finding morulae within granulocytes in the peripheral blood, the combination of acute and convalescent serology using immunofluorescent antibody techniques, and detection of the DNA of A. phagocytophilum using specific polymerase chain reaction assays. Whether persistent infection or reinfection with A. phagocytophilum occurs after natural infection requires additional study, with most reports suggesting that anaplasmosis is a self-limiting disease in dogs that responds well to a 2-week course of doxycycline therapy.

Abbreviations:
DNA

deoxyribonucleic acid

GMP

guanosine monophosphate

HGE

human granulocytic ehrlichiosis

IFA

immunofluorescent antibody

IFN

interferon

IL

interleukin

NADPH

nicotinamide adenine dinucleotide phosphate

PCR

polymerase chain reaction

PSGL-1

P-selectin glycoprotein ligand-1

RNA

ribonucleic acid

Anaplasma phagocytophilum (formerly Ehrlichia equi, Ehrlichia phagocytophila, and in humans, the human granulocytic ehrlichiosis [HGE] agent) is an obligate, intracytoplasmic coccus that belongs to the family Anaplasmataceae (Fig 1).1 The outer cell wall structure of the bacterium ultrastructurally resembles that of Gram-negative bacteria. It infects granulocytes, predominantly neutrophils but also eosinophils, where it exists and reproduces in membrane-bound vesicles, forming microcolonies called morulae (Latin for mulberry) (Fig 2A). A. phagocytophilum is transmitted by ticks, which include Ixodes pacificus in the western United States, Ixodes scapularis in the upper midwestern and northeastern United States, Ixodes ricinus in Europe, and Ixodes persulcatus and Dermacentor silvarum in Asia and Russia.2–4 Other Ixodes spp. ticks also have been implicated in transmission, including Ixodes trianguliceps, Ixodes hexagonus, and Ixodes ventalloi in Europe.5–7 The organism was first described as a veterinary pathogen, after its identification in the cytoplasm of leukocytes of sheep in Scotland.8,9 The disease in domestic ruminants, which occurs in Europe, is also known as tick-borne fever, and has been reported in sheep, cattle, goats, and deer. Reports of equine granulocytic anaplasmosis occurred as early as 1968 in California.10 Dogs were first identified with A. phagocytophilum infection in California in 1982.11 The 1st reports of granulocytic infections in humans in the United States came from the upper midwest in 1993,12 and since then, human granulocytic anaplasmosis has been increasingly recognized in the United States, Europe, and Asia.

Figure 1.

 Phylogenetic tree, based on 16S rRNA gene sequences, showing relationships between α-proteobacteria belonging to the order Rickettsiales. The dataset was resampled 1,000 times and bootstrap percentages are given at the nodes of the tree. Bar, 10 substitutions per 100 nucleotides. Species and family names (in parentheses) are given on the right.

Figure 2.

 (A) Canine neutrophil containing an Anaplasma phagocytophilum morula. Wright's stain, × 100 oil magnification. (B) HL-60 cell infected with A. phagocytophilum that has been transformed genetically to express mCherry, a red fluorescent protein (Image generously provided by MJ Herron, University of Minnesota).

Cases of canine granulocytic anaplasmosis in North America have been reported from California, Washington, Illinois, Minnesota, Wisconsin, Missouri, and British Columbia.11,13–20 In Europe, infected dogs have been reported from Austria, Italy, Sweden, Switzerland, Germany, Poland, and the United Kingdom.21–28 Evidence of possible exposure of dogs to A. phagocytophilum has been detected in at least 39 US states, with the highest seroprevalence being reported in dogs from the upper midwestern, northeastern, and western states.19,29–33 This documented canine A. phagocytophilum exposure parallels the geographic distribution of human cases, which most commonly occur in the Northeast, upper Midwest, and northern California. In addition to horses, cattle, sheep, goats, and humans, serological cross-reactivity with other Anaplasma spp. such as Anaplasma platys can occur, potentially overestimating the true distribution of A. phagocytophilum infection. A. phagocytophilum infection has been reported, albeit less commonly, in camelids and domestic cats.34–37 The renaming of E. equi, E. phagocytophilum, and the HGE agent as A. phagocytophilum based on sequence results of the 16S rRNA genes has been controversial because these organisms differ from previously named Anaplasma species, such as Anaplasma marginale, in their host cell tropism.38 In addition, the former Ehrlichia species differed in their virulence and in their ability to cause disease in different host species, with cross-transmission experiments showing failure of European E. phagocytophila to cause disease in horses, and failure of the HGE agent isolated from humans in the United States to cause disease in cattle.39 These are now considered to represent phenotypic variations among A. phagocytophilum strains in different geographical locations. Other than llamas, domestic ruminants in the United States have not been reported to be infected with A. phagocytophilum. Strains infecting domestic ruminants in Europe and white-tailed deer in the United States (Ap-V1) appear to be distinct from those infecting horses, humans, and dogs on the basis of sequence results of major surface protein genes.40,41

Ecology and Epidemiology

A. phagocytophilum is transmitted primarily by ticks in the I. persulcatus-complex. Experimental transmission after I. pacificus and I. scapularis bites has been demonstrated.3,4,42 In a recent study, tick exposure was reported for 14 of 18 dogs with granulocytic anaplasmosis in Europe.26 Depending on geographical location, a variety of small wild mammals, including mice, woodrats, chipmunks, voles, and shrew, as well as deer and possibly birds act as reservoir hosts for A. phagocytophilum. For the midwestern and eastern United States, white-footed mice (Peromyscus leucopus) and eastern chipmunks likely act as reservoirs,4,43 whereas in the western states, dusky-footed woodrats, gray squirrels, and chipmunks have been implicated.44–47 Evidence is accumulating that some strains circulating in the rodent population may not infect other species, such as horses.48 Larger wild animal species including mountain lions and coyotes also act as hosts, but function primarily as sentinels, as they do not develop chronic infection and are unlikely to infect juvenile ticks.47 Dogs and humans are accidental hosts. Bacteremia appears to be of short duration (<28 days), and as such dogs and humans are not important in transmission to other host species.49 A study in Slovenia showed that approximately 15.4% of the human sample population had antibodies for A. phagocytophilum, although there was no difference in antibody prevalence of those people with exposure to dogs and those without exposure.50

Rarely, transmission of A. phagocytophilum has been reported in the absence of a tick vector. Nosocomial transmission of the organism was recently reported in China.51 An outbreak investigation suggested that transmission occurred through direct, close contact with blood and respiratory secretions of a patient who died of human granulocytic anaplasmosis. In a report from the upper Midwest, 3 adults who had each butchered over 250 deer and gave no report of tick bites developed granulocytic anaplasmosis, possibly after exposure to deer blood.52 More recent studies, however, suggest that the A. phagocytophilum strain infecting deer in the United States is distinct from that infecting humans,40,41 so it was suggested that the individuals may have been unwittingly bitten by ticks.53 Rare case reports also describe possible transmission of A. phagocytophilum to humans transplacentally54 and after blood transfusion.55 Perinatal infection with A. phagocytophilum was described in a bitch from Wisconsin with a retained, mummified fetus. No evidence of infection was documented in the surviving puppies from the litter.56

The seroprevalence of A. phagocytophilum in dogs has been evaluated worldwide (Table 1), although neither molecular evidence nor culture positive cases of A. phagocytophilum have been documented in the southern hemisphere.19,20,29–33,58–74 The prevalence of seropositive test results depends on whether the dog population sampled was sick or healthy, whether clinical suspicion existed for the presence of a vector-borne infectious disease, and geographical variation in exposure to tick vectors and reservoir hosts. In addition, because antibodies to A. phagocytophilum cross-react with other Anaplasma species, such as A. platys, seropositivity may not necessarily reflect previous exposure to A. phagocytophilum. For example, a recent study examining 233 dogs from northern Arizona showed moderate agreement between positive ELISA serology and polymerase chain reaction (PCR) positivity to A. platys, whereas no dog tested PCR positive for A. phagocytophilum.70 Reported seroprevalences in dogs from regions of North America and Europe are as high as 55 and 50%, respectively.20,57 A recent large study that used a commercially available recombinant p44/MSP2 ELISA assaya to determine the prevalence of seroreactivity to A. phagocytophilum in over 400,000 dogs from the United States also revealed wide variation in seroprevalence, with some counties in the upper midwestern and northeastern United States having seroprevalence of >40%, although the overall seroprevalence in these regions was 6.7 and 5.5%, respectively.33 In northern California, between 0 and 50% of the dogs in 8 small towns were seropositive for A. phagocytophilum, showing wide variation in the extent of exposure even over relatively small geographic areas.75 Among 996 serum samples collected from Swiss dogs between March 1991 and March 1998, 7.5% were positive for A. phagocytophilum antibodies, with seroprevalence significantly higher in the southern part of the Swiss Alps.67 Although A. phagocytophilum has been detected in ticks from China and Russia, no prevalence studies in dogs have been reported from these countries to date.2,76,77 Although molecular evidence of A. phagocytophilum infection in dogs was detected in Venezuela and Thailand using PCR, it was subsequently determined that these dogs were in fact infected with A. platys, and false PCR priming had occurred as a result of high organism loads within the blood.78,79 Evidence of A. phagocytophilum exposure and infection in dogs recently was detected in Tunisia,74 and a bacterium closely related to A. phagocytophilum was detected in dogs in South Africa.80

Table 1.   Prevalence of antibodies reacting to Anaplasma phagocytophilum, and the DNA of A. phagocytophilum, in blood samples collected from dogs worldwide.
LocationNumber of
Dogs
PopulationMethodPrevalence
(%)
Reference
  1. IFA, immunofluorescent antibody; PCR, polymerase chain reaction.

Europe
Germany563Sick dogs, suspected to have anaplasmosisIFA serology50.1Kemperman et al55
Germany111Healthy and sick dogsIFA serology43.2Plier et al56
  PCR6.3 
Central Italy1,232Not statedIFA serology8.8Barutzki et al57
Italy460Dogs suspected to have tick-borne illnessPCR0Jensen et al58
Sicily, Italy342Not statedIFA serology32.8Ebani et al59
Sicily, Italy344Pet dogs, pound dogs, hunting dogsPCR0Solano-Gallego et al60
Northwest Poland192Dogs from a region with endemic Lyme borreliosisPCR1.0Skotarczak et al24
Portugal55Dogs suspected to have tick-borne illnessIFA serology55Torina et al61
  PCR0 
Spain466Sick and healthy dogsIFA serology11.5de la Fuente et al62
Northwest Spain479Dogs presenting to veterinary clinicsIFA serology5Santos et al63
Sweden611Dogs suspected to have sarcoptic mangeIFA serology17.7Solano-Gallego et al64
Switzerland996Sick and healthy dogsIFA serology7.5Amusategui et al65
United Kingdom120Dogs suspected to have tick-borne illnessPCR0.8Egenvall et al66
North America
South Ontario and Quebec, Canada53Dogs suspected to have tick-borne illnessIFA serology0Pusterla et al67
Northern Arizona (Hopi Reservation), USA233Pet and stray dogsELISA serology11.6 
  PCR0 
Connecticut and New York, USA106Sick, privately owned dogsIFA serology, Western immunoblotting9.4Manna et al27
Oklahoma, USA259Dogs suspected to have tick-borne illnessIFA serology33Kirtz et al28
Baxter, Minnesota, USA731Sick and healthy pet dogsIFA serology55.4Beall et al20
273 PCR9.5 
North Carolina and Virginia, USA1,845Sick dogs presenting to a referral hospitalIFA serology1.1Magnarelli et al29
Hoopa, Humboldt County, California, USA182Sick and healthy pet dogsIFA serology40Henn et al19
  PCR7.6 
California, USA1,082Clinically healthy dogsIFA serology8.7Rodgers et al30
Missouri, USA88Sick dogs and dogs suspected to have ehrlichiosisPCR1.1Liddell et al14
USA479,640Sick and healthy dogs presenting to veterinary clinicsELISA serology4.8Suksawat et al31
South America
Southeastern Brazil198Dogs suspected to have tick-borne illnessPCR0Shaw et al68
Asia
Israel195Apparently healthy pet dogs, stray and shelter dogsIFA serology9Gary et al69
Yamaguchi and Okinawa, Japan154Sick and healthy dogs presenting to a teaching hospitalPCR0Diniz et al70
Africa
Tunisia286Healthy and sick pet and kenneled dogsIFA serology25.2de Paiva Diniz et al71

Five genetic variants of A. phagocytophilum have been identified in dogs, with 1–2 nucleotide differences in the 16S rRNA gene sequences at nucleotide positions 54, 84, 86, and 120.17 Organisms detected in 2 dogs from Switzerland had 16S rRNA gene sequences that were identical, and these sequences were found to be identical to the 16S rRNA gene sequences found in human granulocytic Anaplasma species.23 DNA sequencing of Swedish canine isolates also showed 100% homology with human isolates.81 The degree to which genetic variation contributes to altered pathogenicity of different strains of A. phagocytophilum currently is poorly understood.

Risk factors for A. phagocytophilum infection in dogs include season of the year, coinfection with other tick-borne pathogens, and signalment. In the western United States, infection with A. phagocytophilum has been diagnosed most frequently in dogs between April and July, with some infections occurring in October.17 For the states of Minnesota and Wisconsin, a bimodal seasonal distribution of illness has been reported that reflects the activity of the I. scapularis tick, in late spring (May, June) and fall (October, November).13,20 In Berlin, 17 of 18 cases were reported between the months of April and September, with the remaining case occurring in November.26 The seasonal distribution of disease most likely reflects periods of peak nymphal and adult tick activities, as well as periods when humans and their dogs are engaged in increased outdoor activities. The proportion of seropositive dogs increased with the age of the dog population in the Swedish study,66 reflecting an increased likelihood of exposure over time. The median age of clinically affected dogs has been approximately 6–8 years (range, 6 months to 14 years).13,17,26,82 Golden Retrievers comprised almost half of affected dogs in 1 study, possibly reflecting the popularity of these dogs for outdoor activities,82 although other studies have reported disease in a variety of different breeds.13,17,26

Because of shared arthropod vectors, concurrent exposure to multiple vector ticks, or both, coinfections with A. phagocytophilum and other arthropod-borne pathogens may occur and complicate the clinical picture. Conversely, the presence of antibodies to other tick-borne pathogens may represent a risk factor for A. phagocytophilum infection. A kennel of Walker Hounds in southeastern North Carolina showed a high degree of coinfection with Ehrlichia spp., Bartonella spp., Rickettsia spp., and Babesia spp.83 Of the 27 dogs screened, 26 were Ehrlichia or Anaplasma spp. seropositive, 16 Babesia canis seropositive, 25 Bartonella vinsonii seropositive, 22 Rickettsia rickettsii seropositive, and 3 had antibodies that reacted to A. phagocytophilum. PCR results from the 27 dogs showed 15 infected with Ehrlichia canis, 9 with Ehrlichia chaffeensis, 8 with Ehrlichia ewingii, 3 with A. phagocytophilum, and 9 with A. platys. Because Borrelia burgdorferi is transmitted by the same Ixodid tick species and maintained in sylvatic cycles with the same rodent reservoirs, evidence of coinfection with B. burgdorferi and A. phagocytophilum frequently is detected, and the 2 organisms may enhance one another's pathogenicity.84,85 Across the United States, antibodies to both B. burgdorferi and A. phagocytophilum were most commonly found in dogs from the Midwest (2.0%), followed by the Northeast (1.4%).33 In a study of 731 naturally exposed dogs from 1 veterinary practice in Baxter, Minnesota, 29% of dogs had antibodies to A. phagocytophilum, 11% had antibodies to B. burgdorferi, and 25% of dogs had antibodies to both.20 Of 89 suspected ill dogs, antibodies to A. phagocytophilum were found in 25%, antibodies to B. burgdorferi in 9%, and antibodies to both pathogens in 43% of dogs. In a study from northern California, dogs that were seropositive for A. phagocytophilum were 18.2 times more likely to be seropositive for B. vinsonii subspecies berkhoffii than dogs that were Anaplasma seronegative.75

Pathogenesis

A. phagocytophilum is transmitted transtadially within the tick and the tick must attach for 36–48 hours for transmission to occur.86,87 The organism may alternate between 2 forms, small dense-cored cells, which bind to host cellular targets, and reticulate cells, which multiply intracellularly and then mature into dense-cored cells that are released upon cell lysis.88–90 Studies using human neutrophils have shown that the bacterium attaches to sialylated ligands on the surface of the neutrophil, such as P-selectin glycoprotein ligand-1 (PSGL-1) and the tetrasaccharide sialyl Lewis X, and enters host neutrophils through caveolae-mediated endocytosis (Fig 3).91–93 Caveolae are specialized lipid rafts enriched in proteins and lipids, which perform a number of signaling functions. Use of this port of entry allows the organism to bypass phagolysosomal pathways.93 Recent work suggests that A. phagocytophilum possesses a bacterial 2-component system utilizing the signal transducer cyclic di-GMP that is essential for infection of, and adaptation to, the host cell environment.90 In order to ensure its intracellular survival and replication, A. phagocytophilum actively dysregulates key neutrophil bactericidal functions. Of major importance, the bacterium is capable of inhibiting neutrophil superoxide production, by blocking the assembly of the multicomponent NADPH oxidase system on the inclusion membrane, and detoxifying superoxide anion.93,94 The bacterial protein AnkA apparently enters the nucleus and interacts with gene regulatory regions, downregulating expression of host defense genes including gp91phox, a component of NADPH oxidase.95A. phagocytophilum has been shown to decrease neutrophil motility and phagocytosis,96 and decreases endothelial adherence and transmigration of neutrophils, which normally occurs through selectin-mediated rolling, cellular activation, and binding via surface integrin molecules. In contrast, research evaluating neutrophil function in dogs experimentally infected with a Swedish isolate of A. phagocytophilum showed normal phagocytosis with enhancement of the respiratory burst.97 The decreased neutrophil adherence to endothelial cells appears to result from loss of PSGL-1 and L-selectin ligands by infected neutrophils, at a time when β2-integrin, immunoglobulin superfamily adhesion molecule, and intercellular adhesion molecule are upregulated.98 The latter may promote pathogen survival in peripheral blood.

Figure 3.

 (A) Typical neutrophil response to a bacterial pathogen. (1) Bacteria bind immunoglobulin or LPS-type receptors on the neutrophil surface. (2) Bacteria enter through phagocytosis. (3) Lysosomes fuse with phagosome. (4) Bacteria are killed by activation of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase system and generation of superoxide anions. (5) PMNs undergo apoptotic cell death and inflammation resolves. (6) Neutrophils interact with endothelial cells through rolling, attachment, and diapedesis. (B). Impaired neutrophil response to Anaplasma phagocytophilum. (1) A. phagocytophilum binds neutrophils using fucosylated ligands on the neutrophil surface. (2) The organism enters through receptor-mediated caveolae endocytosis. (3) Endosomes modified to prevent lysosome fusion. (4) A. phagocytophilum inhibits production of superoxide anion, as well as formation of NADPH complex. (5) A. phagocytophilum delays neutrophil apoptosis by delaying expression of apoptosis-associated genes. (6) A. phagocytophilum infection results in decreased neutrophil adherence to endothelial cells, which may help to maintain an intravascular presence of the organism so it is more accessible to feeding ticks.

Typically, neutrophils circulate for 10–12 hours before they enter tissues and undergo death through apoptosis, and as such make an unfavorable home for most pathogens. Remarkably, A. phagocytophilum is able to delay neutrophil apoptosis in vitro and in vivo, allowing it to survive and form morulae within a normally short-lived, terminally differentiated leukocyte.93,94,99–101.

Although the mechanism for cell-to-cell bacterial transfer is not known, the organism induces interleukin-8 (IL-8) release in infected humans, as well as increasing the surface expression of CXCR2, an IL-8 receptor that may lead to further recruitment of susceptible neutrophils for infection by the pathogen and subsequent uptake by uninfected ticks.102,103 In addition, A. phagocytophilum can infect human CD34+ bone marrow cells,104 endothelial cells,105 and cells of the megakaryocyte lineage.106 These cells may also serve as potential cell reservoirs for cell-to-cell transfer of infection to peripheral blood neutrophils. Recently A. phagocytophilum has been transformed in the laboratory such that it is capable of expressing fluorescent proteins, permitting direct visualization of the organism within cells with fluorescence microscopy (Fig 2B).107 This tool should facilitate further understanding of the tissue distribution, cellular binding, entry, and intracellular replication of A. phagocytophilum in vivo.

A. phagocytophilum infection is associated most commonly with the development of mild to moderate thrombocytopenia, although other cytopenias also may occur, including neutropenia, lymphopenia, and mild anemia. The mechanism of these hematologic abnormalities remains unclear. A. phagocytophilum can enter cells of the megakaryocyte lineage, which also express PSGL-1, although such infection does not appear to impair the ability of megakaryocytes to produce platelets.106 Antiplatelet antibodies have been detected in serum from humans108 and dogs26 with granulocytic anaplasmosis, and immune-mediated mechanisms may contribute to thrombocytopenia. However, thrombocytopenia occurs in acute disease, before antibodies are detected, and in mice with no B or T cells (severe combined immune deficiency), suggesting other mechanisms may be important.109

Impaired neutrophil function and leukopenia as a result of A. phagocytophilum infection may predispose to the development of secondary opportunistic infections. These have been best documented in sheep and cattle with tick-borne fever in Europe. A. phagocytophilum may predispose lambs to tick pyemia, which is characterized by lameness and paralysis caused by Staphylococcus aureus,110Pasteurella spp.,111,112 and Listeria monocytogenes.113 Opportunistic infections also have been documented and suggested as a potential cause of mortality in humans and dogs with granulocytic anaplasmosis,49,94,114 but are less well described than those in ruminants. Impaired neutrophil function in humans infected with A. phagocytophilum may impact the outcome of infection with B. burgdorferi.85

Because granulocytic anaplasmosis is largely a self-limiting infection in dogs, a paucity of information is available regarding the pathology of A. phagocytophilum infection. Tissue injury appears to result from the host inflammatory response, rather than the bacterial infection itself.115 Information regarding the pathology of human infection also is limited, although splenic lymphoid depletion, macrophage aggregates and apoptosis within the liver, paracortical lymphoid hyperplasia, and hemophagocytic cells within tissues of the reticuloendothelial system have been described.114 One report described the death of a horse after experimental infection with a Swedish isolate of A. phagocytophilum.116 At necropsy, widespread petechial, ecchymotic, and intrathoracic hemorrhages were documented. Microscopic pathology consisted of necrotizing vasculitis and hyaline thrombosis, with perivascular mononuclear cell infiltration. Within the liver, there were scattered aggregates of inflammatory cells within the hepatic sinusoids and evidence of hepatocyte apoptosis. The vascular changes in this case were suggestive of disseminated intravascular coagulation, which also has been reported in humans with granulocytic anaplasmosis.49 It was hypothesized that tissue damage might result from substances released by infected neutrophils, or through induction of monocyte tissue factor procoagulant activity.117 Macrophage activation during human granulocytic anaplasmosis is evidenced by high serum ferritin, IL-10, IL-12 p70, and interferon γ (IFN-γ) concentrations, as well as evidence of hemophagocytosis in infected individuals.115

The immune response to A. phagocytophilum infection is not fully characterized. IFN-γ production is thought to play an important role in the initial control of A. phagocytophilum infection,118,119 although it may also contribute to the inflammatory process associated with the disease.120 The production of IFN-γ may be triggered by IL-12 and IL-18 production.121,122 Natural killer T-lymphocytes may be the primary source of IFN-γ, with dendritic cells producing IL-12 and IL-18. In a mouse model of infection, ultimate bacterial clearance appeared to be dependent on CD4+ T cells, and did not depend on major Th1 cytokines such as IL-12 and IFN-γ.119 Although T-cell-mediated immunity generally is considered to be the most important arm of the immune system for clearance of intracellular pathogens, data generated from mouse models suggest a substantial role for humoral immunity in clearance of the ehrlichiae, including A. phagocytophilum.123,124 Antibodies may develop after exposure to bacteria during extracellular phases of infection. Antibodies may access intracellular components containing ehrlichiae or enhance phagocytic activity.123 Both B and T cells are needed to clear infection, because mice with severe combined immunodeficiency, which lack both B and T cells, remain persistently infected. In contrast, mice with isolated T-cell deficiencies can clear infection.119 In lambs, A. phagocytophilum is capable of evading the host immune response through differential expression of MSP2, an outer surface protein involved in immune recognition, as has been documented for A. marginale infections in cattle.125,126 Additional research is required to document whether persistence occurs through similar mechanisms in other host species.

Clinical Disease in Dogs

Most dogs naturally infected with A. phagocytophilum probably remain healthy, as indicated by widespread serological evidence of exposure in endemic areas in the absence of a history of clinical illness.20,32 To date, there are no case reports documenting fatalities in dogs. Four large case series have been reported describing the clinical features of natural infection in dogs examined in Germany (18),26 Scandinavia (14),82 Washington state (8),17 and the upper Midwest of the United States (17),13 together with several reports including 3 or fewer dogs from the United States, Italy, British Columbia, Switzerland, Austria, and the United Kingdom.11,15,16,22,23,25,27,28,56 Other reports describing illness in dogs associated with morulae within granulocytes have not distinguished between E. ewingii and A. phagocytophilum.127–131 Some of these were from areas within the United States where A. phagocytophilum infection is not endemic. One report from the southeastern United States described molecular detection of E. ewingii in 2 of 6 dogs with granulocytic morulae, and detection of an organism resembling E. canis in another, with PCR assays for A. phagocytophilum being negative in the 3 dogs tested.130

The most common clinical signs in dogs infected with A. phagocytophilum that develop illness are lethargy and fever, which occur after an incubation period of 1–2 weeks. Lethargy has been reported in almost all dogs affected.13,17,26,82 Three of the 4 larger case series reported fever in approximately 90% of affected dogs.13,17,82 The remaining case series reported fever in 66% of affected dogs.26 Reported rectal temperatures in febrile dogs have ranged in magnitude from 102.6 to 106.5°F (39.2–41.4°C). Inappetence or anorexia has been reported in 47–88% of dogs.13,17,26 Musculoskeletal signs, such as reluctance to move and lameness, also have been commonly reported. Lameness was reported in 1 of 17 dogs from the upper Midwest,13 5 of 8 dogs from Washington state,17 and 2 of 18 dogs from Europe.24 In some dogs, lameness may result from a neutrophilic polyarthritis.24,132

Infrequent coughing may be noted, which is typically soft and nonproductive (J.E. Sykes, unpublished data)17,56 as well as scleral injection (J.E. Sykes, unpublished data). The presence of a cough was associated with interstitial infiltrates on thoracic radiographs in 1 case report, with a focal area of alveolar infiltrates, and large numbers of neutrophils containing morulae were recovered from a tracheal wash.56

Less commonly, polydipsia, pale mucous membranes, gastrointestinal signs including vomiting and diarrhea, and hemorrhage, manifested as mucosal petechiae, melena, or epistaxis, have been reported.13,17,26 Protracted vomiting, hypersalivation, and fever was the presenting complaint in a dog that recently presented to our hospital that had morulae within granulocytes and PCR confirmation of infection. Splenomegaly and mild lymphadenopathy, detected during physical examination or after radiographic or ultrasonographic imaging, may also be present.13,26 In canine and murine models of infection, splenomegaly and lymphadenopathy are because of reactive lymphoid hyperplasia and, in the spleen, concurrent extramedullary hematopoiesis.133,134

Neurologic manifestations have been uncommonly described in people with human granulocytic anaplasmosis, and the organism has been detected in cerebrospinal fluid.135 Neurologic signs consisting of seizures and proprioceptive deficits were reported in 2 of 17 dogs from the upper midwest with granulocytic anaplasmosis, respectively, but the dog with seizures had a history of idiopathic epilepsy.13 An association between neurological signs and infection in dogs was not apparent in 1 retrospective study of 248 dogs with disorders of the nervous system.136 A granulocytic morula was detected using cytology in the CSF of a dog with meningitis from California. The reciprocal serum antibody titers in this dog were highest for E. canis (40,960), followed by R. rickettsii (5,120) and A. phagocytophilum (40).129 Further molecular studies were not carried out, so it was possible that this organism was E. ewingii.

Infection with A. phagocytophilum appears to be self-limiting in dogs, and dogs with chronic disease have not been described. The extent to which A. phagocytophilum can persist in tissues and contribute to chronic disease manifestations in human and canine patients is the subject of ongoing investigations. In 1 report, treatment of dogs that had been experimentally infected with A. phagocytophilum with prednisolone up to 6 months after infection was associated with development of positive PCR results for the organism, and in some dogs, reappearance of morulae on blood smears and thrombocytopenia.137 In another report, persistent infection was thought to occur in dogs despite treatment with doxycycline, because dogs developed positive PCR results after treatment with prednisolone 5 weeks after doxycycline therapy.b However, it is unknown whether viable organisms were present as determined using culture or animal inoculation studies.

The most consistent laboratory abnormality is thrombocytopenia, which occurs in approximately 90% of dogs.13,17,26 The platelet count in thrombocytopenic dogs has been reported to range from 5,000 to 164,000 platelets/μL.13,17,26,82 The bone marrow of infected dogs contains increased absolute numbers of megakaryocytes and immature megakaryocytes, suggesting that thrombocytopenia may occur due to platelet destruction.138 The majority of affected dogs have had lymphopenia, with lymphocytosis being reported rarely.13,17,26 Anemia was reported in 3/8 dogs from Washington and 2/15 dogs from the upper Midwest, and was typically mild and nonregenerative. In contrast, 11 of 18 dogs from Germany were anemic, with hematocrits that ranged from 19 to 39% (median, 32%).26 The anemia was found to be regenerative in 3 of 8 of these anemic dogs. Both neutrophilia and neutropenia have been reported, although the majority of affected dogs have had neutrophil counts in the lower half of the reference range.13,17,82 Monocytopenia has been reported in approximately one-third of infected dogs from the United States,13,17 whereas several dogs from Germany had monocytosis.26 Mild hypoalbuminemia, hyperglobulinemia, and a mild increase in hepatic enzyme activity (especially alkaline phosphatase) are most commonly observed on serum biochemistry bio-profile.13,17,22,82 High alkaline phosphatase activity was reported in 88% of affected dogs from North America, and 61% of dogs in the German study. Hyperbilirubinemia was present in 5 of 22 dogs with granulocytic anaplasmosis in Europe26 but has not been reported in dogs from North America. The differences in clinical presentation that have been reported from Europe and North America may be a reflection of strain differences in A. phagocytophilum, as has been reported in cattle and sheep.39,139

Diagnosis

The diagnostic criteria for confirmed human granulocytic anaplasmosis are clinical signs and laboratory findings suggestive of granulocytic anaplasmosis together with (1) detection of morulae within neutrophils combined with a single positive reciprocal antibody titer to A. phagocytophilum of ≥80; (2) a 4-fold increase or decrease in the antibody titer within 4 weeks; (3) a positive PCR test result using specific A. phagocytophilum primers; or (4) isolation of A. phagocytophilum from blood.49 These criteria also could be applied to dogs, although isolation is not used routinely for diagnosis.

Morulae may be seen within neutrophils during cytologic examination of Romanowsky-stained peripheral blood smears. Although the finding of morulae within neutrophils in peripheral blood from a dog in an endemic area is highly suggestive of infection with A. phagocytophilum, the morulae cannot be distinguished from those of Ehrlichia spp. such as E. ewingii, and serology or PCR are needed to confirm that the organism is A. phagocytophilum. Morulae were detected within neutrophils from 36, 56, 67, and 100% of dogs in each of the 4 larger case series reported.13,17,26,82 The percentage of neutrophils containing morulae during acute infection has ranged from 7 to 32%.13,17,82 In experimentally infected dogs, morulae appear as early as 4 days after inoculation, and persist for 4–8 days.133

Several conventional and real-time PCR assays have been developed for detection of A. phagocytophilum DNA in peripheral blood, buffy coat, bone marrow, or splenic tissue specimens. Some of these assays amplify DNA from other rickettsial species, and sequencing of the PCR product is required to determine whether A. phagocytophilum is the infecting species. Other assays only amplify the DNA of A. phagocytophilum. The targets of the majority of the assays used have been either the 16S rRNA gene or the outer surface protein gene msp2 (p44). Assays based on the msp2 gene are usually specific for A. phagocytophilum, whereas assays based on the 16S rRNA gene may detect other Anaplasma species, and even other bacteria such as Bartonella henselae.137 Other reported assays have targeted the msp4, groEL, rrs, epank1, or ankA genes of A. phagocytophilum.140–143 One real-time assay was designed to amplify DNA from the msp2 gene over a wide variety of strains isolated from varying locations.18A. phagocytophilum DNA was amplified in 1 study using PCR from 3% of healthy dogs and 37% of clinically ill dogs.20 After treatment, all dogs in 1 study became PCR negative,26 supporting an association between positive PCR results and clinical illness. In experimentally infected dogs, the results of PCR on whole blood were positive for 6–8 days before and 3 days after morulae appeared on blood smears.133,137

Diagnosis also can be accomplished using paired, acute, and convalescent serology. Most veterinary laboratories perform serologic testing using immunofluorescent antibody (IFA) techniques. IgG class antibodies are 1st detectable approximately 8 days after initial exposure, 2–5 days after the appearance of morulae. As a result, during acute illness, antibodies may be inapparent, so PCR may be more useful for diagnosis of acute infection when morulae are not present. Because positive titers may reflect previous exposure, demonstration of a 4-fold rise in titer is required. Antibody titers may persist for several months21,137 with 1 study showing seropositivity as long as 12 months after resolution of acute illness.17 In humans, persistent antibody titers lasting as long as 3 years have been described.49 A polyvalent ELISA using a recombinant p44 antigen has been developed that is suitable for testing serum from both dogs and horses.144 A point-of-care lateral flow ELISA device,a which also uses a recombinant Msp2/p44 protein, is also available for the detection of antibodies to A. phagocytophilum in dog serum. Positive test results using this assay do not imply that A. phagocytophilum is the cause of a dog's illness, and as with IFA testing, negative results may occur in dogs presenting with acute illness because of the lag in antibody production relative to the onset of clinical signs.

Serologic cross-reactivity among different Anaplasma species occurs. Using monoclonal antibodies directed at the highly conserved major surface protein 5 (Msp5), 100% serologic cross-reactivity was documented between A. phagocytophilum and A. marginale by an indirect ELISA assay, but no cross-reactivity was documented using a competitive ELISA.145 Dogs infected with A. platys also test positive using serologic assays for A. phagocytophilum, including the widely used recombinant Msp2 assay.33 Serological cross-reactivity between A. phagocytophilum and E. canis also has been reported, but appears to be uncommon and minor.17,56,146 Dogs presenting to the authors' teaching hospital that have antibodies to A. phagocytophilum generally lack antibodies that react to E. canis using IFA testing.

In a comparative study, blood samples taken from dogs suspected of being infected with A. phagocytophilum were tested by IFA, 16S rRNA gene-based PCR, and by blood smear examination.147 A total of 73% of animals that were positive using PCR and cytologic examination for inclusions also had positive antibody titers, whereas 34 of 36 samples (94%) that were positive using cytologic examination of blood smears were also PCR positive. Thus, use of multiple diagnostic modalities may be needed to confirm the diagnosis of granulocytic anaplasmosis in some cases.

A. phagocytophilum isolates from dogs and humans can be readily cultivated in human promyelocytic leukemia cell lines (HL-60) and tick embryo cell lines. Although culture may be the most sensitive diagnostic modality for detection of acute infection in human patients,148 this method is not routinely used in dogs for diagnostic purposes.

Treatment and Prevention

The treatment of choice for granulocytic anaplasmosis in dogs is doxycycline (5 mg/kg PO q12h for 2 weeks). Most dogs show clinical improvement within 24–48 hours of initial antibiotic treatment, with 1 study documenting that 2 of 8 infected dogs required up to 6 days of tetracycline treatment for full resolution of clinical signs.17 Infection may be prevented by keeping ticks from attaching to the host, and prompt removal of ticks. Ticks should be removed by gently grasping the tick close to the skin with tweezers, forceps, or a commercially available tick removal device and retraction using constant pressure. Combinations of imidacloprid and permethrinc and fipronil and (S)-methoprened have been shown to prevent the transmission of A. phagocytophilum to dogs from infected ticks for up to 25 days after initial treatment.149 Nevertheless, infections have been documented in dogs apparently receiving monthly tick preventatives.26

Reinfection has not been reported in dogs. In the human literature, reinfection has been documented in 1 patient.150 Horses have been shown to resist reinfection after recovery from initial infection with A. phagocytophilum.151 Natural infection may confer long-term protection against development of new disease.

Human Granulocytic Anaplasmosis and Public Health Significance

Human granulocytic anaplasmosis is a nonspecific febrile illness of varying severity that closely resembles the disease in dogs, and has been described as an “influenza-like illness after a tick bite.”152 Men are affected slightly more frequently than women,153 and up to 75% of affected humans report a history of a tick bite.49 As in dogs, coinfections with B. burgdorferi can occur. The overall seroprevalence in individuals bitten by ticks in the United States is 8.9–36%,49 with disease incidence being highest in humans from the upper midwestern and northeastern United States. In Europe, the disease has been most commonly reported in humans from Sweden and Slovenia, where it was first reported in 1997.154,155 Morulae may be observed less frequently in humans with granulocytic anaplasmosis in Europe and the disease may be somewhat milder than the North American counterpart.156

The most common clinical signs reported in human patients are myalgia, often severe headache, malaise, and shaking chills, but anorexia, nausea, arthralgias, and coughing also may occur.49,152,153 The disease typically is mild and self-limiting. In the absence of adequate antimicrobial therapy, active clinical illness beyond 2 months has not been reported.49 Occasionally, more severe disease may occur, with up to 17% of affected humans requiring admission to an intensive care unit in 1 study.152 Death occurs in ≤1% of clinically affected humans, usually as a result of complications such as a septic or toxic shock-like syndrome, respiratory insufficiency, opportunistic fungal or viral infections, rhabdomyolysis, acute renal failure, hemorrhage, and neurologic diseases.49 Severe illness tends to occur in humans of advanced age or in those with concurrent immunosuppressive illness or drug therapy. Laboratory testing of peripheral blood typically reveals normal or slightly decreased white blood cell and platelet counts, sometimes with a neutrophilic left shift, and mild to moderate increases in hepatic transaminase activities.49,152,153 In the United States, the disease is reportable to the Centers for Disease Control and Prevention. Dogs act as sentinels for human exposure, and have the potential to be a source of infection by bringing infected ticks in to contact with humans.

Footnotes

a4Dx SNAP test, IDEXX Corporation, Westbrook, ME

bAlleman A, Chandrashekar M, Beall M, et al. Experimental inoculation of dogs with a human isolate (NY18) of Anaplasma phagocytophilum and demonstration of persistent infection following doxycycline therapy. J Vet Intern Med 2006;20:763 (abstract)

cK9 Advantix, Bayer, Shawnee Mission, KS

dFrontline Plus, Merial, Duluth, GA

Acknowledgment

Ms Carrade was supported by a grant from IDEXX Laboratories during the preparation of this manuscript. The authors acknowledge Nathan Nieto and Michael Sullivan for their critical reviews on this manuscript, and Emir Hodzic for his assistance with creation of the phylogenetic tree.

Ancillary