CMR FigSearch
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow An erratum has been published
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Paddock, C. D.
Right arrow Articles by Childs, J. E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Paddock, C. D.
Right arrow Articles by Childs, J. E.

 Previous Article  |  Next Article 

Clinical Microbiology Reviews, January 2003, p. 37-64, Vol. 16, No. 1
0893-8512/03/$08.00+0     DOI: 10.1128/CMR.16.1.37-64.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Ehrlichia chaffeensis: a Prototypical Emerging Pathogen

Christopher D. Paddock* and James E. Childs

Viral and Rickettsial Zoonoses Branch, Division of Viral and Rickettsial Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia

SUMMARY
INTRODUCTION
MICROBIOLOGY
    Taxonomy and Phylogenetic Placement
    Morphology
    Isolates of E. chaffeensis
    Genetic, Antigenic, and Phenotypic Characteristics
PATHOGENESIS
    Factors Relating to Disease Severity
    Pathology
    Immunology
    Entry and Survival of E. chaffeensis in the Cell
    Animal Models of Disease
CLINICAL FEATURES
    Characteristics of Disease
        General clinical features.
        Hematologic and biochemical abnormalities.
        Severe or unusual manifestations.
        Dual infections.
    ''Asymptomatic'' Infection
    Differential Diagnoses
    Comparison with Other Ehrlichioses
    Treatment and Prevention
LABORATORY DIAGNOSIS
    Serologic Testing
        Indirect immunofluorescence assay.
        Western blotting.
        Other assays.
    Visualization of Morulae and Staining Methods
    PCR Amplification
    Isolation
EPIDEMIOLOGY AND ECOLOGY
    Geographic Distribution
        United States.
        Other locations.
    Surveillance for HME
        Passive surveillance.
        Active surveillance.
    Mechanisms of Transmission and Seasonality of Infection
    Patient Demographics and Risk Factors for HME
    Tick Vectors
        A. americanum.
        Other tick species.
    Vertebrate Reservoirs
        White-tailed deer.
        Goats.
        Domestic dogs.
        Coyotes.
        Other species.
    Factors Influencing the Emergence of HME
        Recent clinical recognition or new disease?
        Evidence for recent emergence.
        Growth and geographic expansion of reservoir and vector populations.
        Improved diagnostics and surveillance.
        Growth of susceptible human populations.
CONCLUSION
REFERENCES

   SUMMARY
Top
 Next
 References
 
Ehrlichia chaffeensis is an obligately intracellular, tick-transmitted bacterium that is maintained in nature in a cycle involving at least one and perhaps several vertebrate reservoir hosts. The moderate to severe disease caused by E. chaffeensis in humans, first identified in 1986 and reported for more than 1,000 patients through 2000, represents a prototypical "emerging infection." Knowledge of the biology and natural history of E. chaffeensis, and of the epidemiology, clinical features, and laboratory diagnosis of the zoonotic disease it causes (commonly referred to as human monocytic ehrlichiosis [HME]) has expanded considerably in the period since its discovery. In this review, we summarize briefly the current understanding of the microbiology, pathogenesis, and clinical manifestations associated with this pathogen but focus primarily on discussing various ecological factors responsible for the recent recognition of this important and potentially life-threatening tick-borne disease. Perhaps the most pivotal element in the emergence of HME has been the staggering increases in white-tailed deer populations in the eastern United States during the 20th century. This animal serves as a keystone host for all life stages of the principal tick vector (Amblyomma americanum) and is perhaps the most important vertebrate reservoir host for E. chaffeensis. The contributions of other components, including expansion of susceptible human populations, growth and broadening geographical distributions of other potential reservoir species and A. americanum, and improvements in confirmatory diagnostic methods, are also explored.


   INTRODUCTION
Top
 Previous
 Next
 References
 
In April 1986, a medical intern scanning the peripheral blood smear of a severely ill man with an unexplained illness observed peculiar intracytoplasmic inclusions in several of the patient's monocytes. The patient described multiple tick bites sustained approximately 2 weeks earlier during a visit to a rural area in northern Arkansas, and a presumptive diagnosis of Rocky Mountain spotted fever had been made (104, 174). Clinicians and scientists subsequently identified these inclusions as clusters of bacteria belonging to the genus Ehrlichia, previously known in the United States solely as veterinary pathogens (174). Within the next 5 years, the organism was isolated in cell culture, characterized by molecular techniques, and formally named Ehrlichia chaffeensis (9, 73). During this interval, surveillance efforts identified several hundred cases of moderate to severe, and occasionally fatal, ehrlichiosis in patients with unexplained illnesses following tick exposures (97, 106, 107, 125, 233, 263). These findings indicated that ehrlichiosis was a widespread and significant public health problem of increasing but undefined magnitude.

During the 1990s, two additional Ehrlichia spp., Anaplasma (formerly Ehrlichia) phagocytophila (the agent of human granulocytic ehrlichiosis [HGE]) and E. ewingii (a cause of granulocytic ehrlichiosis in dogs), were identified as human pathogens, and these reports greatly expanded the geographic region and the size of the human population at risk for acquiring one of these potentially lethal infections (19, 42). While most of the cases of ehrlichiosis caused by E. chaffeensis were being identified in the southeastern and south central United States, within a relatively few years of the initial recognition of HGE the number of cases of human ehrlichiosis identified in the northeastern and north cental states surpassed other regional totals (188).

Although the term "emerging infection" has become almost hackneyed, the Ehrlichia spp. that cause human disease in the United States epitomize the intended application of this designation (156). Not only are these pathogens new to science, but their maintenance in nature requires the complex interactions of tick vectors and vertebrate hosts that are sensitive to environmental influences that can drive epidemics (6). Changes in host susceptibility within a population can be a critical factor in disease emergence (193). Ehrlichiosis caused by E. chaffeensis has increasingly been identified in population segments immunosuppressed through aging, infectious causes, malignancy, or medical therapy (206). Reports of severe and fatal ehrlichioses in these population segments will increase as an unavoidable consequence of environmental forces that increase the risk of exposure to these pathogens, coupled with dramatic changes in human demography and the geographic distribution of AIDS cases (63, 154). This entire process has been fueled by technical developments and the application of sensitive and versatile diagnostic methods, particularly PCR, and a renewed interest in tick-borne and other zoonotic diseases (156, 212, 272).

Several reviews have been written on the microbiology and molecular biology of ehrlichiae and the clinical characteristics of the human ehrlichioses (86, 87, 93, 111, 186, 205, 227, 228). This review of E. chaffeensis summarizes much of this material but focuses primarily on the ecological and epidemiological factors that have contributed to its recognition as an agent of human disease. Although disease caused by E. chaffeensis has been termed human monocytic ehrlichiosis or human monocytotropic ehrlichiosis (i.e., HME), designations of ehrlichioses based on cell tropism may become less useful monikers as additional ehrlichial pathogens are recognized. However, this nomenclature is firmly established in the literature, and to avoid confusion in this review, the acronym HME is used to designate disease caused by E. chaffeensis.


   MICROBIOLOGY
Top
 Previous
 Next
 References
 

Taxonomy and Phylogenetic Placement

E. chaffeensis is an obligately intracellular bacterium in the family Anaplasmataceae and is a member of the {alpha} subdivision of the Proteobacteria. Until 2001, the genus Ehrlichia was composed of a heterogeneous collection of several recognized species (e.g., E. canis, E. phagocytophila, E. sennetsu, E. equi, E. risticii, E. chaffeensis, E. ewingii, and E. muris) and various other taxa that do not have current standing in bacterial nomenclature. This assemblage of species demonstrates considerable molecular diversity based on phylogenetic analyses of 16S rRNA genes, surface protein genes, and groESL heat shock protein operon sequences. On the basis of these differences, Ehrlichia spp. were until recently segregated into three informal "genogroups" (86). A contemporary taxonomic revision reassigned several of these species to other genera (E. sennetsu and E. risticii to Neorickettsia and E. phagocytophila and E. equi to Anaplasma) and emended the genus Ehrlichia to include Cowdria ruminantium, a closely related tick-borne pathogen that causes severe disease ("heartwater") in ruminants in Africa and the Caribbean. In this classification, all members of the tribe Ehrlichieae were reassigned to the family Anaplasmataceae (88). The bacteria that cause human "ehrlichioses" are now represented by three genera rather than the single genus Ehrlichia; they include Neorickettsia sennetsu (the agent of sennetsu fever) Anaplasma phagocytophila, E. ewingii, and E. chaffeensis.

The emended genus Ehrlichia includes E. canis, E. chaffeensis, E. ewingii, E. muris, and E. ruminantium. These ehrlichiae share various genetic, morphologic, clinical, and ecological features: all are at least 97.7% similar in 16S rRNA gene sequences, all reside and multiply in cytoplasmic vacuoles of host cells (the principal cell types include mononuclear and polymorphonuclear leukocytes and endothelial cells, depending on the particular species); all cause disease in animals, humans, or both; and all are transmitted by hard-tick vectors (88).

Morphology

Light microscopic and ultrastructural descriptions of E. chaffeensis have been based on observations of the pathogen in human leukocytes and tissues and in various cell lines of mammalian origin. In these habitats, these small, nonmotile bacteria reside and grow in cytoplasmic vacuoles derived from an early endosome, forming loose to condensed aggregates of bacteria termed morulae. By light microscopy, these morulae appear as mulberry-like, bosselated intracytoplasmic inclusions that stain dark blue to purple with Romanovsky-type stains (Fig. 1) (227).



View larger version (127K):
[in this window]
[in a new window]
  		FIG. 1. Peripheral blood smear from a patient with HME, demonstrating variably sized basophilic inclusions (morulae) within the cytoplasm of a monocyte (lower cell). Each morula consists of a cluster of E. chaffeensis contained with a vacuole. Modified Wright's stain. Magnification, x1,000.

 
By electron microscopy, two distinct morphologic cell types are identified: coccoid and coccobacillary forms with ribosomes and nucleoid DNA fibrils uniformly dispersed throughout the cytoplasm (reticulate cells) (Fig. 2), and predominantly coccoid bacteria with centrally condensed nucleoid DNA and ribosomes (dense-cored cells). Reticulate cells measure 0.4 to 0.6 µm by 0.7 to 1.9 µm, and dense-cored cells measure 0.4 to 0.6 µm in diameter. Both cell types replicate by binary fission, and both demonstrate a gram-negative-type cell wall, characterized by a smooth-contoured cytoplasmic membrane and a generally ruffled outer membrane, separated by a periplasmic space. Members of the genus Ehrlichia do not appear to contain significant amounts of peptidoglycan (227). Both cell types have been demonstrated in clinical samples (209), although the microbiological significance of these distinct morphological forms is unknown. Morulae range from 1.0 to 6.0 µm in width and contain 1 to >40 organisms of uniform or mixed cell types (218, 228). The intramorular space may contain a fine, striated fibrillar matrix and intramorular tubules 25 nm in diameter and as long as 1.5 µm, which originate from the outer membrane of reticulate cells. In cell culture and infected human cells, host cell mitochondria are frequently apposed to the margins of morulae (209, 218).



View larger version (166K):
[in this window]
[in a new window]
  		FIG. 2. Electron photomicrograph of intracytoplasmic vacuoles containing reticulate forms of E. chaffeensis (r) in a DH82 cell (continuous canine histiocytoma cell line) (267). Reticulate cells demonstrate prominent ruffled outer cell membranes (arrowheads) and divide by binary fision (arrow). Lead citrate-uranyl acetate stain. Magnification, x18,000; bar, 1 µm. Reproduced with permission from V. Popov, University of Texas Medical Branch at Galveston.

 
Isolates of E. chaffeensis

At least 21 isolates of E. chaffeensis have been obtained from patients with HME, infected in Arkansas (73, 90), Oklahoma (59), Florida and Georgia (209, 259), Tennessee (206, 255), and Maryland (262). Isolates of E. chaffeensis from sources other than human tissues are few and include five from white-tailed deer (169) and one from a domestic goat (85), each obtained in Georgia.

Described isolates have been obtained in primary culture by using a continuous canine histiocytoma cell line (DH82 cells) and less frequently, human embryonic lung fibroblasts (HEL 299 cells) (59, 73, 85, 90, 169, 206, 209, 255).

In vitro, E. chaffeensis has been adapted to grow in various other cell lines, including human microvascular endothelial cells (HMEC-1 cells), African green monkey kidney cells (Vero cells), human cervical epithelioid carcinoma cells (HeLa cells), human monocytic leukemia cells (THP-1 cells), HEL 299 cells, mouse embryo cells, buffalo green monkey cells, and murine fibroblasts (25, 38, 58, 128, 187).

Genetic, Antigenic, and Phenotypic Characteristics

The genome size of E. chaffeensis is approximately 1250 kb (239). Among the nucleotide sequences that have been characterized are the 16S rRNA gene (9), various genes coding for immunoreactive proteins including the variable-length PCR target (VLPT) (259) and the 120-kDa (289), 106-kDa, and 37-kDa (290) protein genes, the groESL heat shock operon (260), a quinolate synthetase A gene (292), and a locus that contains 22 homologous but not identical genes (the p28 multigene family) (201, 286).

Two antigen-expressing genes that contain repetitive elements have been identified. The 120-kDa protein gene contains a series of 240-bp serine-rich tandem repeat units; the number of repeat units varies among isolates. To date, three variants of the gene (represented by two, three, or four repeats) have been identified in DNA extracts of E. chaffeensis obtained from patients with HME and from infected ticks (255, 259, 287, 289). The 120-kDa gene encodes a heavily glycosylated, immunodominant surface protein that is preferentially expressed on dense-cored forms of E. chaffeensis and as a component of the intramorular fibrillary matrix (183, 219). This gene demonstrates interstrain variation, and p120 proteins expressed by different isolates of E. chaffeensis vary in molecular weight; however, immune sera from patients with HME react with p120 antigens from various strains regardless of variations in the number of repeat units (290). The VLPT gene demonstrates even greater interstrain diversity (209, 259). This gene is also characterized by a series of direct tandem repeats, whose number may vary among isolates. DNAs of VLPT genes amplified from cultured isolates of E. chaffeensis or from ticks or patient blood samples infected with this pathogen have shown two to six repeats. Qualitative differences in the nucleotide sequences of the imperfect 90-bp repeats results in at least seven different types of repeat units. Additional genetic diversity is produced by differences in the linear order of the individual repeats and by various deletions and substitutions along the length of the gene. Based on a relatively small number of DNAs evaluated, VLPT patterns of E. chaffeensis in the southeastern United States are most frequently represented by three or four repeats, and the six-repeat variant appears to be the rarest version of this gene (206, 255, 257, 259). The biological function of this gene has not yet been elucidated; however, VLPT sequences code for immunoreactive proteins with apparent molecular masses of 30 to 60 kDa (259). Collectively, the occurrence of genetic heterogeneity among several of the recognized genes of E. chaffeensis suggests that considerable molecular diversity exists within this bacterium: evaluation of 18 patient isolates by using genetic composites created by polymorphisms in the VLPT gene and the 120-kDa protein gene reveal eight distinct genotypes (255, 259). No distinct biological, clinical, or epidemiological correlates have been associated with a particular genotype, although future studies may be more revealing.

Isolate-dependent sequence polymorphisms have also been described for a locus of E. chaffeensis genes that encodes major outer membrane proteins (OMP), described as the omp cluster or the p28 multigene family (286). Detailed analysis of this locus in the Arkansas isolate of E. chaffeensis reveals 22 complete, paralogous genes from 813 to 900 bp distributed along a 27-kb segment of the genome (201). The p28 genes code for mature proteins with predicted molecular sizes of approximately 26 to 32 kDa; none of the proteins are identical, and the amino acid sequence identity varies from approximately 20 to 80% (286). Sequences of individual p28 genes also vary among different isolates of E. chaffeensis (171, 286, 291). At least 16 p28 alleles are actively transcribed, and it is likely that the antigenic diversity of E. chaffeensis results from differential expression within this gene family (171). Homologous immunodominant proteins encoded by multigene families have been identified in closely related bacteria, including E. canis, E. ruminantium, A. phagocytophila, and A. marginale (183, 201, 226).

Several major immunoreactive proteins of the Arkansas isolate of E. chaffeensis have been identified by using human antisera in immunoblot analyses. These include polypeptides with relative molecular masses of approximately 120, 66, 58, 55, 44, 29, 28, and 22- kDa (57, 59). Genetic correlates have been established for several of these antigens, including the 120-kDa protein, the GroEL protein (58 kDa), and the p28 proteins.

Variations in reactivity among different isolates of E. chaffeensis have been demonstrated by using monoclonal antibody (MAb) analyses. MAb 1A9 reacts with epitopes of various p28 proteins with different molecular sizes; however, it does not react with all isolates of E. chaffeensis (59, 285), reflecting heterogeneity in the antigenic composition among isolates created by the diversity of p28 proteins (286, 291). Isolate-specific reactivity is also demonstrated by MAb 6A1, which reacts with a surface-exposed, 30-kDa antigen of the Arkansas isolate but does not react with the 91HE17 isolate (56). Variation in sizes of apparently homologous proteins have also been detected by using MAbs (59) and immunoblot analyses demonstrating isolate-specific expression of the 120-kDa protein and VLPT repetitive-element gene products (56, 259). Biological correlates for these variably sized proteins to pathogen virulence or clinical disease in humans are incompletely characterized, although MAbs directed against specific epitopes of p28 OMPs can mediate the clearance of E. chaffeensis in a SCID mouse model (160).

Other than descriptions of the antigenic composition and immunolocalization of these proteins, relatively few phenotypic characteristics of E. chaffeensis have been identified. Experiments with a low-passage, culture-adapted isolate show that this strain can survive for at least 11 days in anticoagulated human whole blood and for as long as 21 days in cell culture media at 4 to 6°C (187).


   PATHOGENESIS
Top
 Previous
 Next
 References
 

Factors Relating to Disease Severity

The pathogenesis and determinants of disease severity for HME are incompletely understood. Soon after the recognition of this disease, it became apparent that a wide range of clinical outcomes were possible in persons infected with E. chaffeensis (263). In a study of 149 patients diagnosed during 1985 to 1990, logistic regression was used to demonstrate that age (>=60 years) operated as an independent risk factor for severe or fatal illness (105). However, many cases of severe or fatal disease have been described in apparently healthy children (32, 103, 109, 246) and young adults (155, 179, 181). In this context, disease severity may ultimately depend on a complex interaction of several components relating to the host, the pathogen, and perhaps therapeutic interventions.

Severe or fatal HME has been described in persons with compromised immunity from various causes including human immunodeficiency virus (HIV) disease (23, 179, 206, 208), immunosuppresive therapies (12, 14, 177, 180, 241, 242, 262), monoclonal gammopathy (79), asplenia (98, 103), sickle ß-thalassemia (246), and Down's syndrome (96).

For some patients, the severity of clinical manifestations appears directly correlated with the level of bactermia, particularly among severely immunocompromised patients infected with HIV. In these individuals, morulae are often detected in peripheral blood leukocytes in relatively large numbers and the organism is detected in cell culture relatively rapidly (206). However, this correlate does not apply to all patients, since peripheral blood smears and bone marrow aspirates of some critically ill patients fail to reveal morulae (32, 91, 209, 262). There are no data to specifically associate distinct molecular or antigenic features among strains of E. chaffeensis with variations in disease severity or particular disease manifestations; however, it is possible that intrinsic markers for these outcomes will emerge as additional isolates are obtained and a broader repertoire of genetic identifiers are evaluated (90, 209, 255).

Among published descriptions of patients with particularly severe or fatal HME are reports of individuals who received long-term sulfa drug therapy for ulcerative colitis (194, 211) or as prophylaxis for opportunistic infections (14, 206, 242, 262) and reports of patients for whom trimethoprim-sulfamethoxazole was administered for several days or weeks before ehrlichiosis was correctly diagnosed (1, 27, 28, 84, 94, 103, 119, 236, 237, 243). An association between the use of sulfa-containing antibiotics and exacerbation of disease severity has been described for other rickettsial infections, and the frequency of similar reports of this association among patients with HME warrants further investigation (215).

Pathology

In vertebrate hosts, E. chaffeensis infects predominantly mononuclear phagocytic cells. The most frequently infected blood cells are monocytes; however, infections in other cell types have been described, including lymphocytes, atypical lymphocytes, promyelocytes, metamyelocytes, and band and segmented neutrophils (1, 91, 174, 209). Although E. chaffeensis appears capable of inhabiting other phagocytic cells (e.g., granulocytes), it is likely that mononuclear phagocytes maintain the productive infection (91). Infected cells typically contain only 1 or 2 morulae, although as many as 15 have been observed in leukocytes of immunosuppressed patients (23, 179, 208).

There are relatively few histopathologic data describing lesions in tissues and organs of persons with HME. The most extensively sampled and described tissue has been bone marrow. However, no consistent histopathologic patterns have emerged from these examinations, possibly because the biopsy specimens have been obtained during different stages in the course of the illnesses. The most frequently reported finding is a normocellular or hypercellular marrow with myeloid hyperplasia, megakaryocytosis, or both (91, 121, 253). Bone marrow biopsy specimens may reveal aggregates of foamy histiocytes or small noncaseating granulomas (91, 99, 121) or may show hemophagocytosis (1, 84, 91, 180) or may be unremarkable or normal (91, 99, 127, 138). Morulae have been detected in fewer than half of the described bone marrow biopsy specimens but are frequently visualized in marrow of patients infected with HIV (23, 208, 209). Hypocellular bone marrow is seldom observed in patients with acute disease, and diminished peripheral blood cell counts are characteristically far out of proportion to the absolute numbers infected leukocytes, implying that cytopenias associated with HME result from peripheral events that may include sequestration, consumption, or destruction of infected and noninfected cells (91, 125).

Pathologic findings in other tissues have been described most frequently in patients with fatal disease (79, 89, 91, 180, 208, 209). Because persons who die of HME often represent specialized patient cohorts (e.g., immunocompromised patients), quantitative and qualitative features of the histopathologic findings in these patients may not be directly comparable to features of disease in the general patient population. Findings in the lungs of these patients may include intra-alveolar hemorrhage, diffuse alveolar damage, and interstitial pneumonitis and edema (89, 109, 180, 208, 209). Perivascular, predominantly lymphohistiocytic infiltrates without evidence of endothelial damage or thrombosis can occur in many organs, including the meninges (180, 209, 275). Other findings may include hemophagocytosis and microvesicular steatosis in the liver (89, 208, 209), focal necroses in spleen, liver, and lymph nodes (89, 208), and diffuse hemorrhages involving soft tissues, kidneys, urinary bladder, diaphragm, and meninges (180, 209).

Localization of ehrlichiae and ehrlichial antigens by immunohistochemical and in situ hybridization techniques reveal systemic, multiorgan involvement in patients with fatal HME. The greatest distribution of bacteria occurs in tissues containing abundant mononuclear phagocytic cells, including splenic cords and periarteriolar sheaths, lymph nodes, and bone marrow (23, 79, 180, 208). Morulae are less frequently observed in macrophages in the pulmonary microvasculature and in the liver within Kupffer cells (79, 89, 208) (Fig. 3C and D). E. chaffeensis is detected occasionally and in lower abundance in mononuclear cell aggregates or perivascular infiltrates in the brain, heart, pancreas, adrenals, kidneys, gastrointestinal tract, omentum, ovaries, and connective tissue (23, 79, 180).




View larger version (245K):
[in this window]
[in a new window]
  		FIG. 3. Immunohistochemical localization of E. chaffeensis in mononuclear cells of the spleen (A) and bone marrow (B), in pulmonary macrophages (C), and in hepatic Kupffer cells (D) in patients with fatal HME; tissues represented in panels A through C were obtained from patients coinfected with HIV. Bacterial burdens in severely immunocompromised individuals are generally far greater than those observed in immunologically intact patients. Ehrlichiae and ehrlichial antigens appear as red inclusions within the cytoplasms of infected cells. Immunoalkaline phosphatase stain with naphthol phosphate-fast red substrate and hematoxylin counterstain. Magnifications, x630. Reproduced with permission from S. Zaki, CDC.

 
Immunology

As with many aspects of the pathogenesis of HME, there is only a nascent understanding of the immune mechanisms that follow infection with E. chaffeensis in a human host. The relative contributions of humoral and cell-mediated immunity have not been definitively established, although both appear to play important roles in host defenses against this pathogen. Because ehrlichiae are intracellular pathogens, it is intuitive that cellular immunity is an important component of successful clearance of E. chaffeensis. This paradigm is suggested directly by descriptions of particularly severe disease in HIV-infected patients (206) and indirectly by observations of the profound lymphoproliferative responses described for patients recovering from HME (45, 99, 105).

Various inbred mouse strains have been used to dissect the impact of cellular and humoral processes following infection. Wild-type mouse strains infected with E. chaffeensis clear the bacteria within 16 days, while mice with defective macrophage and T-cell functions maintain infections that may persist for one to several months (112). Mice lacking functional toll-like receptor 4 (tlr4) alleles, whose gene product is responsible for macrophage stimulation following exposure to lipopolysaccharide of gram-negative bacteria, produce significantly decreased levels of nitric oxide and interleukin-6 (IL-6) and develop infections with E. chaffeensis that persist for at least 2 weeks beyond the duration of infection observed in wild-type mice. However, macrophage activation alone does not appear to be sufficient for successful clearance of this pathogen. The role of major histocompatibility complex class II (MHC-II) genes appear to be even more profound, and mice lacking functional MHC-II genes are unable to clear E. chaffeensis following infection. These findings suggest that CD4+ T lymphocytes are essential for complete clearance of this intracellular pathogen (112). These observations are supported by the results obtained with other murine models using immunodeficient animals. In contrast to tlr4 and MHC-II mutants which do not become ill or die following infection with E. chaffeensis, SCID mice deficient in T and B lymphocytes develop persistent, overwhelming infections and become moribund within 24 days postinfection (280). However, animals with functional B cells but deficient for {alpha}/ß T cells or both {alpha}/ß and {gamma}/{delta} T cells remain persistently infected but do not become ill. Similarly, immune serum from immunocompetent mice or MAbs recognizing an immunodominant outer membrane protein (p28) of E. chaffeensis, administered passively to SCID mice prior to or during active infection, results in protection from disease but does not effect complete bacterial clearance (160, 281). Collectively, observations in murine systems suggest that antibodies contribute to the elimination of this pathogen during active infection and may ameliorate disease and that intact cellular immunity, particularly involving processes coordinated by CD4+ T cells, appears to be the crucial determinant of complete recovery following infection with this agent.

Paradoxically, the relative paucity of bacteria detected in the blood and tissues of most patients infected with E. chaffeensis, even those with severe illnesses, suggests that clinical manifestations of HME may also be mediated by host immune responses, and possibly amplified by specific cytokine production (228). In vitro studies have shown that human monocytes infected with E. chaffeensis produce only two proinflammatory interleukins, IL-1ß and IL-8, and an immunosuppresive cytokine, IL-10 (157). However, when infected cells are exposed to hyperimmune serum containing anti-E. chaffeensis IgG antibodies, additional proinflammatory cytokines, including tumor necrosis factor alpha and IL-6, are generated by the cells (158). Binding of the E. chaffeensis-antibody complex to human monocytes via the Fc{gamma} receptor is required for expression of TNF-{alpha} and IL-6 mRNAs and enhances the expression of IL-1ß mRNA. The presence of immune complexes also activates nuclear factor kappa-B, further stimulating secretion of these cytokines. In concert, these processes generate levels of major proinflammatory cytokines as high as the levels observed in cells stimulated with Escherichia coli lipopolysaccharide. These findings suggest that the generation of antibodies to E. chaffeensis may trigger pathophysiologic responses detrimental to the host through a mechanism similar to endotoxic shock (158). In this context, cytokine production and modulation by anti-E. chaffeensis antibodies may play critical roles in processes involving both elimination of the pathogen and generation of systemic disease (228).

Patients with HME typically develop a lymphocytosis during recovery that is disproportionately represented by CD3+ CD4- CD8- T cells expressing a T-cell receptor composed of {gamma} and {delta} chains. Expansion of lymphocytes with this relatively unusual phenotype has been associated with immune responses to various intracellular pathogens, including Mycobacterium, Listeria, and Leishmania spp. However, patients with HME display the most profound {gamma}/{delta} T-cell lymphocytosis reported, with levels as high as 97%. Because this response is temporally associated with resolution of infection, it is uncertain if this peculiar lymphocytosis is directly involved in host defense against ehrlichiae or represents an epiphenomenon of the infection (45). Resolution of the {gamma}/{delta} T-cell lymphoproliferation involves apoptotic cell death of the lymphocytes, which may represent an important mechanism for modulating the T-cell immune response during recovery from the infection (46).

Extensive genetic variability exhibited by the p28 multigene locus of E. chaffeensis and in expressed surface antigen proteins has been proposed as a mechanism of immune evasion (225, 226, 286). Although quantitative transcriptional analyses remain to be performed, it is possible that E. chaffeensis may differentially and sequentially express the p28 multigene family to rapidly alter the composition of one or more of its immunodomminant surface proteins and thereby escape immune surveillance. Only 12 (38%) of 32 convalescent-phase serum samples from patients with HME demonstrated reactivity with a recombinant p28 cloned from the Arkansas isolate (p28-19) of E. chaffeensis, supporting the concept that differential expression of p28 genes results in proteins with substantially different antigenic properties (291). Inoculation with recombinant p28 protects mice from E. chaffeensis infection (202), raising hopes that this antigen will have uses as vaccines for ehrlichial pathogens. MAbs directed against epitopes within the amino terminus of a hypervariable region of the OMP-1g protect SCID mice from otherwise fatal E. chaffeensis infection, and humans with HME produce antibodies reactive with the same OMP-1g hypervariable region (160).

There are few data available that evaluate long-term immunity to E. chaffeensis in persons infected with this pathogen. A single case of sequential infections with two genetically distinct strains of E. chaffeensis has been described. The patient was a liver transplant recipient receiving immunosuppressive therapy, who developed illnesses characteristic of HME during each episode (161). However, the susceptibility of previously infected, immune-intact individuals to reinfection with different strains or even the identical strain remains undetermined. Similarly, a single case of persistent infection with E. chaffeensis has been documented in a 68-year-old debilitated patient in whom ehrlichiae were detected in intrasinusoidal histiocytes of the liver at the time of death, 68 days following the onset of illness (92). The prevalence or clinical significance of persistent infection with this pathogen in human hosts is unknown.

Asymptomatic infection with E. chaffeensis has not been conclusively demonstrated; however, isolation of an ehrlichia closely related or identical to E. canis from the blood of an asymptomatic human from Venezuela suggests that infections with some Ehrlichia spp. may remain clinically silent (214).

Entry and Survival of E. chaffeensis in the Cell

Because E. chaffeensis lacks pili or a capsule, it may bind to its host cell via its outer membrane (229). In vitro studies showing attachment and invasion of HeLa cells by E. coli containing a plasmid expressing the 120-kDa OMP of E. chaffeensis suggest that the p120 is an adhesin that might also enhance the internalization of ehrlichiae (219). Internalized ehrlichiae are invested by the host cell membrane, forming endosomes that maintain distinct cytoplasmic compartments that do not fuse with lysosomes (229).

Survival of ehrlichiae within the cell may be influenced by complex molecular and biochemical pathways involving iron acquisition. Iron is essential for cytochromes and other iron-containing enzymes of E. chaffeensis (229). The iron chelator deferoxamine completely inhibits the growth of E. chaffeensis, indicating that this bacterium is sensitive to cytoplasmic iron depletion (25). Early endosomes containing E. chaffeensis selectively and progressively accumulate transferrin and transferrin receptors (26, 195). Because these endosomes are slightly acidic, ehrlichiae may acquire iron directly from transferrin-iron complexes present in the endosome. Infection of cells by E. chaffeensis further modulates iron uptake by activating a cytoplasmic protein (iron-responsive protein 1), which increases host cell transferrin receptor mRNA levels (24). In vitro studies using recombinant gamma interferon show that this cytokine activates the intracellular killing of E. chaffeensis in human monocytes early in the course of infection by markedly diminishing the number of host cell transferrin receptors, thereby reducing the intracellular labile iron pool (25); however, E. chaffeensis appears to rapidly block the ehrlichiacidal activity of gamma interferon by increasing protein kinase A activity in host cells within 30 min following infection (159). E. chaffeensis also expresses a 37-kDa protein homologous to iron binding proteins of gram-negative bacteria; however, the exact role of this protein remains to be determined (290).

Animal Models of Disease

Investigations of the pathogenesis and immunology of HME have been hampered by the lack of a convenient, reproducible, and generalizable animal model of disease. White-tailed deer, which serve as important natural reservoirs of E. chaffeensis, maintain persistent bacteremias capable of infecting lone star ticks but do not demonstrate clinical manifestations of disease (70, 76, 100).

E. chaffeensis causes naturally occurring disease among dogs that is indistinguishable clinically from diseases caused by E. canis and E. ewingii (37). Experimental infection of dogs also suggests that these animals may have E. chaffeensis circulating in blood for over 3 weeks (77) and may develop characteristic antibody responses (230). However, needle-inoculated animals appear to develop only mild febrile responses without hematologic abnormalities (77). Other factors limit the utility and versatility of canines as experimental hosts, including the absence of inbred syngeneic dogs (particularly animals with genetically defined immune defects) and commercially available canine-specific markers for immune effectors (252).

Several strains of inbred immunocompetent mice (Mus musculus) have been inoculated with E. chaffeensis. These animals appear to rapidly clear the infection and seldom develop illness consistent with HME (165, 264, 280); however, neutrophil infiltrates, hepatocyte apoptosis, and granuloma formation have been observed in the livers of some infected mice (112). Although relatively restricted in their applicability as models of pathogenesis of HME in immune-intact patients, various murine strains with defined immunological deficiencies have proved useful for exploring cellular and antibody-mediated host defenses to E. chaffeensis (112, 280). Small-scale studies using other rodents including white-footed mice, hamsters, and red-backed voles have been unsuccessful in reproducing disease (264).

Animal models of HME using closely related Ehrlichia spp. as surrogates for E. chaffeensis include infection of BALB/c mice with E. muris (144, 145) and infection of C57BL/6 or BALB/c mice with an as yet unnamed Ehrlichia sp. (Ixodes ovatus ehrlichia [IOE]) that is >98% similar to E. chaffeensis by 16S rRNA sequence analysis (203, 249, 252). Mice infected with E. muris develop a transient, mild illness and almost always recover from infection, decreasing the utility of this model as an instrument to study severe HME. However, animals infected with appropriate inocula of IOE consistently die within 9 days and demonstrate histopathologic lesions that resemble lesions identified in human patients with fatal HME, including interstitial pneumonitis, myeloid hyperplasia of bone marrow, and hepatic apoptosis and erythrophagocytosis. In this context, the IOE-mouse model represents a promising system for investigating the immunity and pathogenesis of HME (203, 252).


   CLINICAL FEATURES
Top
 Previous
 Next
 References
 

Characteristics of Disease

The early disease manifestations of HME are relatively constant and, with few exceptions, are shared by a vast array of infectious and noninfectious processes. As the disease progresses, involvement of multiple organ systems may complicate the clinical course and result in various life-threatening scenarios.

General clinical features. Within 1 to 2 weeks (median, 9 days) following exposure to an infecting tick, patients experience a prodrome characterized by malaise, low-back pain, or gastrointestinal symptoms or may develop sudden onset of fever (often >39°C). Patients with HME are most likely to seek medical attention within 3 to 4 days after the onset of symptoms, and the presenting clinical features frequently include fever (>95%), headache (60 to 75%), myalgias (40 to 60%), nausea (40 to 50%), arthralgias (30 to 35%), and malaise (30 to 80%) (99, 105). During the course of the illness, other manifestations of multisystem disease develop in approximately 10 to 40% of patients, including cough, pharyngitis, lymphadenopathy, diarrhea, vomiting, abdominal pain, and changes in mental status (99, 105, 204, 254). Less frequently reported manifestations include conjunctivitis (32, 250), dysuria (109, 179), and peripheral edema (98).

Large case series of HME in the general population report rashes in approximately 30 to 40% of patients, although a rash is reported more frequently among adult persons infected with HIV (206) and may occur in as many as two-thirds of pediatric patients (95, 139). In comparison, rash is a component of approximately 90% of cases of Rocky Mountain spotted fever (248). Rash patterns associated with HME are variable in character, distribution, and temporal occurrence. This pleomorphism includes petechiae (50, 103, 177, 206, 215, 250), macules (103, 126), maculopapules (32, 200, 241, 243), and diffuse erythema (28, 103, 173, 206). Rash generally occurs later in the course of disease (median of 5 days after onset) (105), may be fleeting or transient (27, 103), and may involve the extremities, trunk, face or, rarely, the palms and soles (96, 124).

Hematologic and biochemical abnormalities. Multilineage cytopenias are a hallmark laboratory feature of HME early in the course of the illness and may provide early presumptive clues to the diagnosis (105, 246). Mild to moderate leukopenia is observed in approximately 60 to 70% of patients during the first week of illness, with the largest decreases occurring in the total lymphocyte count (99, 105, 139, 255). A relative and absolute lymphocytosis (approximately 45 to 85% of the total leukocyte count) is seen in most patients during recovery and is characterized predominantly by the expansion of activated T cells expressing the {gamma}/{delta} T-cell receptor (45, 99). Thrombocytopenia is the most frequently identified cytopenia, being seen in 70 to 90% of patients during their illness (105, 206). Although some patients may develop very low platelet levels (e.g., <20,000/µl), platelet nadirs are generally between 60,000 and 120,000/µl (105). The majority of patients present with a normal hematocrit; however, anemia eventually develops in approximately half of HME patients, occurring within 2 weeks following the onset of illness (99, 105, 139, 254, 255).

Mildly or moderately elevated hepatic transaminase levels are noted in approximately 80 to 90% of patients at some point during their illness (99, 105, 200, 254). Alkaline phosphatase and bilirubin levels are less likely to be elevated; however, these markers can be elevated in 25 to 60% of patients (99, 200, 255). Mild to moderate hyponatremia has been reported in as many as 50% of adult patients and 70% of pediatric patients (99, 139). Serum sodium levels of <130 mEq/liter are frequently observed in persons with severe disease (16, 32, 177, 181, 206).

Various other biochemical abnormalities may occur, reflecting progression of the illness to multisystem involvement. These include prolonged activated partial thromboplastin and prothrombin times, increased levels of fibrin degradation products, elevations in the levels of serum creatinine, lactate dehydrogenase, creatine phosphokinase, and amylase, and electrolyte abnormalities including hypocalcemia, hypomagnesemia, and hypophosphatemia (94, 99, 138, 177, 206). The pathophysiologic processes responsible for electrolyte abnormalities are not well understood. In some patients, diminished concentrations of albumin and protein in serum are also noted (1, 126, 138), which may affect the measurement of some divalent cations.

Severe or unusual manifestations. HME generally manifests as a moderate to severe disease, and approximately 60 to 70% of patients in contemporary case series have been hospitalized (48, 105, 254, 255). In some patients, untreated disease may progress to death as early as the second week of illness (79, 109, 179, 209) or may cause a febrile illness lasting 2 to 3 weeks (106). Multisystem involvement often develops in patients with severe disease and may include acute renal failure, metabolic acidosis, respiratory failure, profound hypotension, disseminated intravascular coagulopathy, hepatic failure, adrenal insufficiency, and myocardial dysfunction (103, 138, 174, 177, 179, 194, 206, 246, 254, 261, 279). The factors responsible for disease severity and involvement of specific organ systems are incompletely understood.

Approximately 20% of persons infected with E. chaffeensis develop signs and symptoms of central nervous system disease (99, 105). Neurologic findings may suggest a meningitis syndrome (meningismus, photophobia, severe headache, lethargy, confusion, or cranial nerve palsies), or an encephalitis or encephalopathy syndrome (delirium, obtundation, coma, seizures, hyperreflexia, clonus, broad-based gait, or ataxia) (72, 222). Cognitive impairment is the most predictive indicator of abnormalities in the cerebrospinal fluid (CSF), which are generally characterized by a mild to moderate lymphocytic pleocytosis and a moderately elevated protein level (222). CSF white blood cell counts in adult patients with meningitis are generally lower than 250 cells/mm3, although counts in children may be higher, occasionally exceeding 500 cells/mm3 (32, 103, 246, 253). Morulae are rarely visualized in CSF mononuclear cells and, if found, are typically in severely ill patients (32, 94, 222). Long-term sequelae of central nervous system infections are not well documented; however, persistence of various symptoms, including headache and photophobia (32), facial or ocular palsies (50, 99), tremors (16), diminished memory (138) and confusion (222), for one to several weeks has been reported. Impairment of cognitive performance has been described for some pediatric patients following HME (139).

Cough or other respiratory symptoms are described in 20 to 25% of all patients with HME (105, 204); however, pulmonary manifestations, including interstitial pneumonitis (16, 59, 138), pleural effusions (109, 173, 243), pulmonary edema (109, 277), and acute respiratory distress syndrome (155, 211, 213, 215, 246, 271), are frequent components of severe disease.

Patients may develop profound thrombocytopenia and coagulopathies and occasionally develop hemorrhagic manifestations including epistaxsis (208), pulmonary hemorrhage (89, 109, 180, 208), gastrointestinal bleeding (1, 89, 174, 263, 275), subdural hematomas (180, 206), hematuria (263), and conjunctival hemorrhage (27, 103, 173, 261).

The estimated case-fatality ratio for HME is approximately 3% (188). Fatal disease has been described most frequently in males (approximately 70%), older patients (median age, 51 years; range, 6 to 80 years), and patients debilitated by underlying disease or immunodeficiencies including HIV infection, malignancy, asplenia, chronic ethanol abuse, and corticosteroid therapy. Half of all deaths occur during the second week of illness (range, 7 to 68 days), and death is generally attributed to multisystem organ failure, catastrophic hemorrhage, or secondary bacterial or fungal infections (23, 79, 89, 92, 109, 179, 180, 206, 208, 209, 222).

Secondary infections, including those caused by cytomegalovirus, Candida, and Aspergillus spp., have occurred in some severely ill patients (92, 104) suggesting that infection with E. chaffeensis may induce suppression of the host immune system (275). The occurrence of pathogen-mediated immune dysfunction has also been proposed for animals and patients infected with A. phagocytophila (87, 227).

Dual infections. Lone star ticks harbor or vector several other pathogenic or potentially pathogenic bacteria including the spirochete "Borrelia lonestari" (22, 44, 140), Francisella tularensis (132), various spotted fever group rickettsiae (118), and E. ewingii (282). However, there are relatively few well-documented laboratory confirmed cases of concurrent infection with E. chaffeensis and another tick-borne agent. PCR-confirmed HME occurring synchronously with a spotted fever rickettsiosis has been described (247), and several prospective epidemiologic studies have demonstrated simultaneous seroconversions to E. chaffeensis and spotted fever group rickettsiae among military personnel exposed to A. americanum-infested habitats (185, 284). Descriptions of patients with simultaneous HME and Lyme disease (3, 27, 220) require cautious interpretation (27), particularly because the lone star tick is not a competent vector of Borrelia burgdorferi (217). However, in states where populations of A. americanum may be sympatric with I. scapularis, antibodies reactive with E. chaffeensis have been detected in patients with well-documented erythema chronicum migrans (176), suggesting that patients with Lyme disease may be exposed simultaneously or sequentially to other tick species carrying E. chaffeensis or other antigenically related ehrlichiae. A fatal case of HME and babesiosis in an 85-year-old man has been reported from New Jersey (141).

"Asymptomatic" Infection

Military training exercises involving troops exposed to tick-infested habitats where E. chaffeensis is highly endemic have permitted prospective investigations of seroconversions among individuals to E. chaffeensis or related antigens following known tick exposure (185, 216, 284). In one study, more than two-thirds of persons demonstrating seroconversion reported no clinical illness (284) while a subsequent investigation conducted at the same location found that 80% of individuals developing antibodies reactive with E. chaffeensis reported an illness compatible with ehrlichiosis (185). Although these studies suggest that asymptomatic infections with E. chaffeensis occur, a definitive interpretation of the data is precluded by the potential for antigenically related ehrlichiae to elicit antibodies that cross-react in serologic tests.

The possibility for asymptomatic or subclinical HME is also suggested by the relative paucity of described cases of disease in children, as well as relatively high seroprevalences of antibodies reactive with E. chaffeensis among children residing in several regions of the southeastern and south-central United States where this agent is endemic. In one study of children 1 to 17 years of age evaluated at major medical centers in Arkansas, Kentucky, Missouri, North Carolina, Oklahoma, and Tennessee, age-adjusted prevalence rates of antibody reactive with E. chaffeensis (or antigenically related ehrlichiae) at titers of >=1:80 ranged from 2 to 22%. At most of these locations, the age-adjusted seroprevalence exceeded 10% (178). Among the first 250 cases of ehrlichiosis described, fewer than 10% were in individuals aged 2 to 13 years (95). Children are known to be exposed to tick-borne pathogens at levels similar to or greater than those for adults, as documented by the very high incidences of Rocky Mountain spotted fever and Lyme disease among children aged 5 to 9 years (55, 67). There is no reason to believe that children are less commonly exposed than adults to A. americanum, suggesting that infection with E. chaffeensis in the pediatric population in general results in less severe illness relative to HME in adults.

Differential Diagnoses

E. chaffeensis ehrlichiosis is a multisystem disease with protean manifestations, but because it lacks a pathognomonic clinical feature, the differential diagnosis of HME is often broad. Initial symptoms may be generalized and relatively vague, and diagnoses frequently include "viral syndrome" in the context of gastroenteritis, upper respiratory infection, pneumonia, or meningoencephalitis. Localized findings may lead to suspicion of pharyngitis, urinary tract infection, epididymitis, or prostatitis (99, 206). Abdominal pain may mimic cholecystitis, and cholecystectomies have been performed in some patients with HME before the correct diagnosis was made (50, 209). Marked hypotension or laboratory abnormalities associated with HME may be interpreted as indicators of sepsis, thrombotic thrombocytopenic purpura, or hematologic neoplasia (103, 138, 180).

A history of recent tick bite or exposure can be elicited from most patients; however, this feature is absent in approximately 10 to 30% of cases (99, 106, 204, 254). The clinical presentation of HME may be similar to that of other tick-associated illnesses, especially other ehrlichioses, caused by A. phagocytophila or E. ewingii, and Rocky Mountain spotted fever. Because most patients with HME who are treated with doxycycline show obvious clinical improvement within 42 to 72 h, failure to improve within this interval generally supports an alternative diagnosis (273).

Comparison with Other Ehrlichioses

The usual symptom complex of fever, headache, myalgia, and malaise, coupled with thrombocytopenia, leukopenia, and elevated hepatic transaminase levels, are features shared by HME, HGE, and E. ewingii ehrlichiosis. However, some differences in the frequencies of disease manifestations exist between HME and the other forms of human ehrlichiosis. Rash, central nervous system involvement, and gastrointestinal disturbances are reported more often for patients with HME than for patients with HGE. Comparisons of general measures of disease severity, including hospitalization rates (2, 20, 30) and case-fatality ratios (188), suggest that severe or life-threatening disease occurs more frequently among patients with HME than among persons with HGE (Table 1). Descriptions of E. ewingii ehrlichiosis exist for only eight patients, precluding broad comparisons of severity. However, no known deaths have been attributed to E. ewingii, and analysis of a small group of HIV-infected patients coinfected with E. chaffeensis or E. ewingii suggests that these patients develop fewer disease manifestations and complications than do HIV-infected patients with HME (206).


View this table:
[in this window]
[in a new window]
  		TABLE 1. Selected clinical characteristics and outcomes of ehrlichioses for patients infected with E. chaffeensis, A. phagocytophila, or E. ewingii, extracted from case series and state surveillance activities

 
Treatment and Prevention

In vitro susceptibility testing has shown that E. chaffeensis is resistant to representatives of most classes of antibiotics including aminoglycosides (gentamicin), fluoroquinolones (ciprofloxacin), penicillins (penicillin), macrolides and ketolides (erythromycin and telithromycin), and sulfa-containing drugs (co-trimoxazole) (18, 41, 234). Clinical experience supports the results of these tests and indicates that other classes of antibiotics, including cephalosporins, are equally ineffective. Rifampin exerts rapidly bactericidal effects in vitro (41); however, there are no clinical data that evaluate the use of this antibiotic in patients with HME. Relatively little is known of the biochemical mechanisms responsible for the resistance of E. chaffeensis to various antimicrobials; however, a molecular basis for resistance of E. chaffeensis and closely related Ehrlichia spp. to fluoroquinolones appears to be associated with natural mutations in the quinolone resistance-determining region of gyrA, the gene encoding the A subunit of DNA gyrase (182). Specifically, the presence of alanine residues at two positions of the dimer interface in the DNA binding area of the A subunit of this enzyme confers resistence to the activity of fluoroquinolone antibiotics (182).

E. chaffeensis is susceptible to tetracyclines and their derivatives, broad-spectrum antimicrobials which act by inhibiting protein synthesis of various bacterial species by reversibly binding to the 30S ribosomal subunit to prevent the addition of new amino acids during the formation of peptide chains. Many other human pathogens, including rickettsiae, chlamydiae, borreliae, mycoplasmas, Actinomyces spp., Vibrio spp., Bartonella spp., and some Mycobacterium spp. and protozoa, are also susceptible to tetracyclines. These drugs, particularly doxycyline, represent the treatment of choice for all persons with HME. Most patients become afebrile within 1 to 3 days following treatment with a tetracycline (99, 105, 106); however, fever may persist in some severely ill persons even after several days of therapy (105, 107, 221, 274, 277). The optimal duration of therapy has not been established definitively; however, a treatment course of 7 to 10 days, or at least 3 days after the abatement of fever, is widely accepted (18). For most patients, leukocyte and platelet counts and serum sodium levels correct to normal values within 3 to 7 days and hepatic transaminase levels normalize within 1 to 4 weeks (99, 105, 263). Although susceptibility data are limited to evaluations of a single isolate (41), there is no clinical evidence to suggest that tetracycline-resistant strains of E. chaffeensis exist.

In vitro data have shown that E. chaffeensis is resistant to chloramphenicol (41), and several anecdotal reports describe treatment failures with this antibiotic (103, 174, 250). Paradoxically, there are also reports of apparent treatment successes with chloramphenicol, particularly in children (28, 96, 124, 221). However, because the efficacy of chloramphenicol remains incompletely defined, this drug should not be considered primary therapy for HME, even in young children (8).

Reducing contact with infected ticks lowers the risk of acquiring HME. Because it is unreasonable to assume that a person can eliminate all activities that may result in these contacts, prevention techniques primarily involve personal protection. Wearing light-colored clothing that facilitates the detection of crawling or attached ticks and the use of repellents containing DEET (n,n-diethyl-m-toluamide) can minimize the risk of tick bites. However, the best protective measure consists of a thorough body examination for ticks after returning from potentially tick-infested areas. It is not known how long A. americanum must remain attached before it can transmit E. chaffeensis to a host; however, because other tick species generally require several hours of attachment before bacteria are transmitted (131, 143), frequent inspections for and prompt removal of attached ticks by using forceps or tweezers is an important method to minimize the risk of HME.


   LABORATORY DIAGNOSIS
Top
 Previous
 Next
 References
 
A diagnosis of HME can be confirmed by several laboratory methods. In order of their routine application, these are serologic tests to measure specific antibody titers, detection of morulae in peripheral blood or in CSF leukocytes, detection of ehrlichial DNA by PCR of blood or CSF, direct detection of ehrlichiae in tissue samples by immunohistochemistry (discussed above under Pathology), and isolation of bacteria.

Serologic Testing

The most widely available laboratory diagnostic tests detect and measure antibody reactive with E. chaffeensis (273). Although these assays remain the most frequently utilized confirmatory methods, there are several caveats to their use. Currently available serologic tests may return negative results for the majority of patients during the first week of illness. Additionally, the discovery of other pathogenic species of related bacteria that share cross-reacting antigens (e.g., A. phagocytophila and E. ewingii) requires careful interpretation and correlation of diagnostic test results with clinical and epidemiologic findings to avoid incorrect designation of the specific agent.

Indirect immunofluorescence assay. Most patients with HME have been diagnosed by the indirect immunofluorescence assay (IFA). The original IFA format for detecting antibodies reactive with E. chaffeensis used a surrogate antigen, E. canis, as substrate (78). Currently, the standard IFA for HME uses the Arkansas strain of E. chaffeensis (9) cultivated in DH82 cells or Vero cells (75) as substrate. Paired sera collected during a 3- to 6-week interval represent the preferred specimens for serologic evaluation of HME. Both immunoglobulin M (IgM) and IgG antibodies can be measured using the IFA (61); however, the IgG IFA test is negative in as many as 80% of patients during the first week of illness and the IgM titers may also be uninformative at this time (61). It is important to obtain a convalescent-phase serum specimen since most (>80%) patients have developed diagnostic IFA titers by 6 weeks postinfection (61, 255). Unfortunately for the purposes of diagnosis, individuals with HME initially present for care a median of 4 days after disease onset (105), and often this initial visit is the only time at which a serum sample is obtained. The impact of these observations on surveillance and underreporting of HME has been discussed (60, 61). Few data are available to describe the kinetics of IFA-detectable antibody for E. chaffeensis infections (78), and none have been published using E. chaffeensis antigen as a testing substrate.

The diagnosis of ehrlichiosis in a person with a clinically compatible illness can be confirmed by seroconversion or a fourfold or greater change in antibody titer (sometimes limited to a rise in antibody titer [273]) between acute- and convalescent-phase samples (51). Recommendations for diagnosing HME in a patient with compatible illness promulgated by the Task Force on Consensus Approach for Ehrlichiosis include a single reciprocal titer of >=256 as sufficient to confirm disease and a titer of 64 as indicating probable HME (273); however, national surveillance efforts have considered cases with a single IFA titer of >=64 as only probable HME, regardless of the magnitude of the end-point titer (188).

Antibodies cross-reactive with a number of ehrlichial antigens (57) and different Ehrlichia species are well documented in humans (48, 64). Western immunoblotting analyses using purified E. chaffeensis and A. phagocytophila proteins suggest that patient antibodies dually reactive with these agents are recognizing homologous heat shock proteins, not major OMPs (268). Because cross-reactivity among ehrlichial species is seen in 10 to 30% of patient sera, sera should be tested against both E. chaffeensis and A. phagocytophila antigens when ascribing specific etiology (64). In general, a fourfold or higher end-point IFA titer is useful in discriminating between etiologic agents when PCR has been used to confirm the diagnosis. Many end-point IFA titers to E. chaffeensis and A. phagocytophila antigens are within a twofold range, precluding the use of IFA serologic testing from ascribing specific etiology (64). Cross reactivity among antibodies to a number of ehrlichial species has been an important feature in defining new human pathogens. Just as E. canis provided a useful surrogate antigen for diagnosing HME until E. chaffeensis was cultured, E. chaffeensis antigens have been used to diagnose some cases of ehrlichiosis caused by E. ewingii (42).

Negative serologic results for the acute-phase sample do not necessarily exclude the diagnosis. Similarly, the lack of seroconversion does not rule out HME. Some small fraction of patients do not develop measurable antibody following infection with E. chaffeensis. In some instances, this failure to seroconvert can be attributed to immune impairment (208, 209) or to early death due to rapidly progressive disease, as seen with some cases of Rocky Mountain spotted fever (207), but in other instances the reasons are unclear (255). Early treatment with a tetracycline-class antibiotic occassionally reduces or abrogates the antibody response to R. rickettsii (245), and it appears that a similar phenomenon occurs with E. chaffeensis (254).

Western blotting. The use of Western blotting has permitted the identification of antigenic variability among isolates of E. chaffeensis and identified variability in the reactivity of patient sera to a number of E. chaffeensis antigens (40, 56). The majority of HME patients with detectable IFA reactivity to whole E. chaffeensis preparations have antibody reactive with the 120-kDa protein (56), and a recombinant 120-kDa protein has potential application as a serodiagnostic antigen (288). A recombinant major OMP of E. chaffeensis (rP30) has been used as antigen in immunoblot analysis to determine specific reactivity to E. chaffeensis among serum samples dually reactive with this agent and with A. phagocytophila (268). Western blotting of E. chaffeensis antigen has also been useful in diagnosing human infections with E. ewingii, which was first recognized as a human pathogen in 1999. Because E. ewingii has not yet been cultured, homologous IFA antigens are unavailable. However, human serum samples from patients infected with E. ewingii fail to react with the 28-kDa antigen of E. chaffeensis (42).

Other assays. Enzyme linked-immunosorbent assays using whole cell antigen or recombinant protein antigens hold promise for the future diagnosis of HGE (134) but are still in the developmental stage for diagnosis of HME. An E. chaffeensis homolog of the major antigenic protein 2 of E. ruminantium has been cloned and sequenced (35). This 21-kDa protein from E. chaffeensis has been expressed in Escherichia coli and successfully adapted to an enzyme-linked immunosorbent assay format to detect antibodies from patients with HME (5). Preliminary findings with 20 human serum samples indicated a diagnostic sensitivity of 95%, using the IFA as the "gold standard," and a diagnostic specificity of 100%.

Visualization of Morulae and Staining Methods

Morulae have been identified in smears of peripheral blood, buffy coat preparations, and bone marrow aspirates by using various eosin-azure (Romanovsky)-type stains, including Wright's, Diff-Quik, Giemsa's, and Leishman's. Although this technique offers the most rapid method of diagnosis, it is considered relatively insensitive and is seldom confirmatory in clinical practice. In this context, morula-positive smears