Infectious Disease Pathology

Human granulocytic anaplasmosis (Anaplasma phagocytophilum)

J. Stephen Dumler
The Johns Hopkins University School of Medicine
Baltimore, MD


Clinical Summary
An 80 year-old white man from northwestern Wisconsin developed a fever and was admitted to the hospital 4 days later with additional complaints of confusion and cough. He had been in good health until this incident. After treatment with a third-generation cephalosporin without clinical benefit, he was transferred to a tertiary care facility where he additionally complained of muscle pain and hyperventilating. Aside from his recent antibiotic, he took no medicines and had no allergies, had oral surgery one month before, had a history of a tick bite, and had not recently traveled. On physical examination, his temperature was 39.7°C, pulse 100 beats/min, blood pressure was 140/80 mmHg, and respiratory rate was 34/min. He was tachypneic and coughing thick mucus, but the physical examination revealed only significant proximal leg muscle pain, delirium, and heme-positive stools, without other respiratory, cardiovascular, or neurological signs. Laboratory studies revealed hyponatremia (127 meq/L), WBC 14.3/µL, platelets 15,000/µL, hemoglobin 15.4 g/dL, blood urea nitrogen 56 mg/dL, and creatinine 4.2 mg/dL; arterial blood gasses revealed pH 7.48, pCO2 25 mmHg, bicarbonate 18 mmol/L, pO2 53 mmHg with 89.9% saturation. LDH was 1,855 IU/L, AST 491 IU/L, and total bilirubin 3 mg/dL. Chest X-ray and EKG examinations were unremarkable. The patient was treated for apparent sepsis with ceftriaxone, gentamicin, and metronidazole. Sputum gram stain, culture and cytology were unrevealing; blood cultures were sterile. Urine output dropped and did not respond to either fluid or diuretic challenges, and he was transferred to the ICU. Profound thrombocytopenia continued and PT and aPTT were prolonged to 14.4 and 43 seconds. A peripheral blood smear evaluation revealed possible intracellular cocci within leukocytes, supporting a diagnosis of sepsis. One day after admission, hypoxia continued and he had several episodes of hematemesis; fibrin split products were elevated, but his fibrinogen level was 438 mg/dL, the WBC count dropped to 10.6/µL, and PT and aPTT climbed to 16.5 and 57 sec. He was treated with platelet transfusions and fresh frozen plasma, and antimicrobial coverage was expanded by adding clindamycin. BUN, creatinine, bilirubin, and AST levels continued to climb, and hemoglobin dropped to 6.9 g/dL; he developed multiorgan failure and ARDS.

Slide Provided:
Peripheral blood smear from day of admission (Wright stain).


Slide 1
Click to view with ImageScope
Click to view with a Web-Based Viewer



Figure 1
Anaplasma phagocytophilum "morulae" or clusters within a band neutrophil of the patient with human granulocytic anaplasmosis. Note the basophilic staining and stippled appearance of the cluster that conforms to the size and shape of an intracytoplasmic vacuole. Wright stain, original magnification 1,000X
Figure 2
Gastroesophageal junction ulcer with hemorrhage in the patient with human granulocytic anaplasmosis. H&E, original magnification 25X
Figure 3
Gastroesophageal ulcer demonstrating budding yeast with pseudohyphae, most consistent with Candida albicans esophagitis. Giemsa stain, original magnification 1,000X
Figure 4
Lung demonstrating intra-alveolar hemorrhage and mild interstitial lymphohistiocytic infiltrate and alveolar wall edema. H&E, original magnification 400X

Figure 5
Liver demonstrating the mild lobular and periportal infiltrates composed predominantly of lymphocytes and histiocytes and Kupffer cell hyperplasia. H&E, original magnification 40X
Figure 6
Liver demonstrating focal hepatocyte apoptosis among a mild lymphohistiocytic lobular infiltrate and Kupffer cell hyperplasia. H&E, original magnification 200X
Figure 7
Spleen from patient with human granulocytic anaplasmosis that demonstrates lymphoid cell depletion with a relative increase in mononuclear phagocyte populations, many of which appear activate or are actively hemophagocytic. H&E, original magnification 400X
Figure 8
Spleen from the patient with human granulocytic anaplasmosis demonstrating intracytoplasmic inclusions or morulae within neutrophils. Giemsa, original magnification 1000X

Figure 9
Splenic neutrophil with intracytoplasmic morula. Note the membrane-bound inclusion that contains several pleomorphic bacteria that range in shape from coccoid to coccobacillary. The preparation was from formalin-fixed tissue previously embedded in paraffin and retrieved for transmission electron microscopy. Original magnification 15,000X
Figure 10
Spleen demonstrating the presence of A. phagocytophilum (red color) within neutrophils in sinusoids. Immunohistochemistry using rabbit-anti-A. phagocytophilum.
Figure 11
Immunohistochemical demonstration of A. phagocytophilum within neutrophils infiltrating the esophageal ulcer (left panel) and lung (right panel). Immunohistochemistry using rabbit-anti-A. phagocytophilum.
Figure 12
A. phagocytophilum cultivated from the blood of a patient with human granulocytic anaplasmosis using the HL-60 promyelocytic cell line. The left panel demonstrates the typical morulae that appear within the cytoplasm of the infected HL-60 cell (LeukoStat stain; original magnification 1,000X). The right panel demonstrates the ultrastructural appearance of the morulae and bacterial contents in transmission electron microscopy (original magnification; 12,000X).


Diagnosis:
Human granulocytic anaplasmosis (Anaplasma phagocytophilum)

Description
The patient was strongly suspected from the outset to have sepsis secondary to Staphylococcus aureus bacteremia, despite the fact that all blood cultures had no growth. At the time, the presence of cocci within the neutrophils was an unfamiliar finding now more easily recognized as ehrlichiosis. In this case, the first recognized human infection by the obligate intracellular bacteria now classified as Anaplasma phagocytophilum, there was ample evidence of a bacterial infectious agent. It was observed readily in 41% of circulating neutrophils and band neutrophils at the time of active infection (Image 1). Extracellular bacteria were not observed. At post-mortem examination, the key findings showed that the patient exsanguinated from an esophageal ulcer (Image 2) that proved to be caused by a yeast (Image 3), probably Candida albicans. In addition, there was evidence of hemorrhage in the lung (Image 4), and a systemic inflammatory process that involved the liver (lobular hepatitis) (Images 5 and 6), the spleen (lymphoid depletion and hemophagocytosis) (Image 7), the lung (mild to moderate interstitial pneumonitis) (Image 4), and the adrenals (lymphohistiocytic infiltrates). Careful observation of Giemsa-stained tissues revealed the presence of small cocci-like clusters within neutrophils (Image 8), especially prominent in the spleen. Serologic tests for Ehrlichia chaffeensis were negative, as was a PCR test for E. chaffeensis DNA in the patient's blood and tissues. Splenic tissue was examined by transmission electron microscopy to determine the nature of the intraneutrophilic inclusions, and this verified the morphological appearance of coccoid to coccobacillary bacteria sequestered within a vacuole (Image 9), as would be anticipated for ehrlichiosis. To determine the identity of the bacterium, broad range PCR using primers able to amplify all eubacterial 16S rRNA genes (rrs) was used on DNA from the patient's blood, revealing a 1,500 bp fragment. When sequenced, this revealed >99.9% identity to 2 veterinary pathogens, Ehrlichia phagocytophila and Ehrlichia equi, both known as obligate intracellular bacteria that infect neutrophils in ruminants or horses, respectively. Since, these organisms and the bacterial cause of this infection in humans have been reclassified into a single species, Anaplasma phagocytophilum. Based on these data, A. phagocytophilum-specific rrs primers were prepared and used in a PCR on blood DNA from the patient, successfully amplifying a 919 bp fragment, that when sequenced, was identical to that previously amplified using broad range PCR primers. To confirm this identification, antibodies to A. phagocytophilum were used in immunohistochemistry to demonstrate its presence in tissues for the patient (Images 10 and 11). [1, 2] Since this initial description, the causative agent was isolated in culture (Image 12) and used to create serological tests for routine diagnosis, and more than 5,181 cases of HGA have been reported to the CDC since 1994.

Etiology.
Intracellular inclusions of bacteria have been observed in a variety of conditions in humans, including Neisseria infections Chlamydia infections, Rickettsia diseases, Q fever, Donovanosis, Bartonellosis, and ehrlichiosis, among others. The latter category, ehrlichiosis, is a surrogate name used to simplify the variety of human infections caused by members of the Anaplasmataceae family, largely limited to Ehrlichia chaffeensis, Ehrlichia ewingii, and Anaplasma phagocytophilum. [3] It is also this group of organism which is most likely to present as an inclusion within circulating human leukocytes, monocytes for E. chaffeensis and neutrophils for E. ewingii and A. phagocytophilum. Despite the shared neutrophil host niche between E. ewingii and A. phagocytophilum, E. ewingii is genetically and antigenically more similar to E. chaffeensis. Regardless of which organism infects a human patient, the disease manifestations can be quite similar and highly nonspecific – typically fever, headache, myalgia, and malaise – mostly reminiscent of influenza or a viral syndrome.

Epidemiology and Clinical Presentation.
Since its first description in 1994, over 5,000 cases of HGA have been reported to the CDC, linearly increasing over the entire period. [3] Most infections in the U.S. occur in the Northeast and Upper Midwest where the nymphal stage of the major tick vector, Ixodes scapularis (deer or black-legged tick) is active during late Spring to Summer months. HGA has also been recognized in northern latitudes of Europe and eastern parts of Asia, including China and Korea. [4, 5] The tick vectors in these regions are Ixodes ricinus and I. persulcatus – along with I. scapularis, these ticks also transmit Borrelia burgdorferi that cause Lyme disease and that simultaneously infects up to 10% of patients with HGA as well. [6, 7] As A. phagocytophilum is not transmitted through the tick ovary, it is maintained by passage from tick to tick through infected vertebrate hosts that develop bacteremia that can persist. Reservoir hosts likely include white footed mice (Peromyscus leucopus) and possibly white tailed deer (Odocoileus virginianus) in the eastern U.S. [8, 9] In the U.S. cross-section seroprevalence studies in endemic regions reveal rates from 1 to 15%, yet active surveillance in Wisconsin and Connecticut demonstrate a much less frequent occurrence of disease – 0.05 to 0.06%. [3, 10] These data suggest that infection exceeds disease by 20 to 250 fold. Among individuals with disease manifestations, most present to a physician after 4 to 8 days of fever after an average incubation period of 10 days. Over 75% of patients recognize tick exposure or tick bite in the weeks prior to infection. Among those who present with an undifferentiated febrile illness, after tick exposure/bites in endemic regions during spring or summer, the laboratory features can further facilitate diagnosis since the majority have thrombocytopenia, leukopenia, or both, and evidence of mild to moderate liver injury is present in over 90% of individuals as reflected by increased levels of serum aminotransferase activities. Clinical disease is severe enough to require hospitalization in 50%, intensive care unit admission in 75, and death in 0.5%. A variety of complications have been reported, and most are inflammatory or post-inflammatory manifestations. Severe complications can also include a toxic- or septic-shock syndrome, adult respiratory distress syndrome, coagulopathy and hemorrhage, neuropathy, and opportunistic infections. Fatalities are almost uniformly associated with opportunistic infections, such as Candida esophagitis, Herpes simplex virus esophagitis, cryptococcosis, and invasive pulmonary aspergillosis. These opportunistic infections can occur among hosts without pre-existing immune compromise before HGA.

Pathology and Pathogenesis.
The clinical hallmark of many rickettsial infections is increased vascular permeability, recognized clinically as "vasculitis". From a histopathologic view, the classical Rickettsia species, such as R. rickettsii (Rocky Mountain spotted fever) and R. prowazekii (louse-borne typhus), infect endothelial cells and induce true lymphohistiocytic vasculitis with endothelial cell injury and erythrocyte extravasation that explains petechial skin lesions and organ ischemia. Although from a clinical standpoint, HGA sometimes also presents as "vasculitis", careful examinations have not identified significant histopathologic vasculitis since infection occurs in leukocytes and neutrophils rather than in endothelial cells. Neutrophil infection by A. phagocytophilum occurs only after initial events successfully bring the bacteria into contact with sialylated/fucosylated CD162 on the neutrophil surface, after which endocytosis occurs. [11] The engulfed bacteria continue to be found within an endosome that lacks any maturation markers, and thereby does not fuse to lysosomes. [12] However, the neutrophil is altered in very important ways, including downregulation of antimicrobial response, upregulation of proinflammatory response, and inhibition of apoptosis. [13] Eventually, the endocytic cluster of bacteria are released from or rupture the cell and spread hematogenously to other neutrophils. In spite of the lack of histopathologic vasculitis, increased vascular permeability has been documented as a feature of A. phagocytophilum neutrophil infection in vitro, and the expression of vasoactive cytokines and chemokines from infected neutrophils supports this observation. [14]

Few histopathologic examinations post-mortem or as diagnostic biopsies of humans with HGA have been conducted. [1, 15, 16] Among the preselected severe group where this has been done, common findings include increased mononuclear phagocyte organ (spleen, lymph node, bone marrow, liver) infiltration with macrophages and lymphocytes, including occasional leukoerythrophagocytic cells. In addition to leukopenia and thrombocytopenia, anemia is also a common finding, yet examinations reveal either normocellular or hypercellular bone marrows with normal trilinear maturation. On this basis of this finding, it is presumed that the pancytopenia occurs mostly owing to peripheral consumption, destruction, or sequestration of marrow elements, as demonstrated with hemophagocytosis in some tissues. A mild to moderate degree of lobular infiltration of liver by lymphocytes and histiocytes is frequent, and occasional apoptotic hepatocytes can be found, some in association with lymphohistiocytic aggregates. Where available, mild to moderate interstitial lymphohistiocytic are observed in lung tissues, with damage to the alveolar pneumocytes.

In most cases, only rare A. phagocytophilum-infected neutrophils can be found in tissues, illustrating histopathologic injury out of proportion to infection load. In fact, given a normal leukocyte count of 7.8 x 109 cells/L and the mean leukocyte count of 3.7 x 109/L in A. phagocytophilum infected patients, on average 4.1 x 109 leukocytes/L are lost with infection; however, the mean infected leukocyte count among patients is 0.27 x109/L, indicating that 15-fold more leukocytes are lost during infection than can be accounted for by direct bacteria-mediate lysis. [6, 10] Moreover, platelets and erythrocytes are not targets of infection, yet are reduced to similar or greater degrees, arguing for an immunological basis of many of the disease manifestations in HGA. Evidence to the contrary is limited to 1) increased severity with delayed diagnosis and treatment, and 2) rapid resolution of infection and disease with tetracycline antibiotic therapy, except when very late in the disease course. In fact, in both humans and animal models of A. phagocytophilum infection, levels of macrophage activating and modulating cytokines, interferon-g (IFN- g ) and interleukin-10 (IL-10) are significantly higher than in age-sex-gender matched febrile subjects or mock-infected controls. [17, 18] In accordance with these findings, the levels of both interleukin-12 (IL-12) that regulates IFN- g production, serum ferritin and triglycerides (products of activated macrophages), and a high IL-10:IFN-γ ratio a significantly associated with increased disease severity, suggesting that the activation of macrophages to produce proinflammatory mediators and effectors may be a major contributor to disease manifestations. [17] In support of the immunopathogenesis of HGA is the observation that pathogen load in increased 7-fold in IFN-γ knockout mice, yet the absence of IFN-γ in these animals leads to inhibition of hepatic inflammatory injury. [19] In contract, infection of IL-10 knockout mice that lose the ability to downregulate IFN-γ mediated macrophage activation does not alter pathogen load but significantly intensifies hepatic inflammation to the extent of marked hepatocyte apoptosis and focal liver necrosis. Overall, these findings indicate that HGA is likely a disease that is initiated by the obligate intracellular infection of neutrophils by A. phagocytophilum, but that this process triggers an immunopathological response that not can eradicate the pathogen but also is the predominant mediator of organ injury and disease.

Differential Diagnosis and Laboratory Confirmation.
Owing to the usual non-specific presentation as an undifferentiated febrile illness, HGA can be very difficult to diagnosis on clinical grounds alone. [7] The differential diagnosis for those presenting with febrile illness alone includes viral syndromes, upper respiratory tract infections, urinary tract infections, sepsis, toxic shock syndrome, meningococcemia, leptospirosis, viral hepatitis, enterovirus infections, influenza, murine typhus, Q fever, typhoid fever, bacterial endocarditis, malaria, and dengue. If recent tick exposure has been elicited, the differential also includes human monocytic ehrlichiosis (E. chaffeensis), Ehrlichia ewingii ehrlichiosis, Rocky Mountain spotted fever (R. rickettsii), relapsing fever, tularemia, Lyme borreliosis, and Colorado tick fever. Additional considerations include Kawasaki disease in children, bacterial pneumonias, collagen vascular diseases, even leukemia. The differential diagnosis when intracellular structures are observed within circulating leukocytes includes E. chaffeensis (monocytes), E. ewingii (neutrophils), toxic granulation, Döhle bodies, storage disorders, Auer rods (myelogenous leukemia), and very rarely true bacteria such as staphylococci or Neisseria spp or even yeast or parasites such as Toxoplasma gondii.

Once suspected based on clinical and historical features, HGA can be confirmed by a variety of laboratory approaches including blood smear examination for intracytoplasmic inclusions (morulae), amplification of A. phagocytophilum DNA from peripheral blood, and serological tests that usually require acute and convalescent serum antibody titers. Blood culture is also a useful method, but it is available in very few laboratories. The overall sensitivity of each approach at the time of illness varies, with the most sensitive approaches during the active phase of infection (<7 days after onset) including blood smear examination for morulae at 61%, PCR 69-76%, and serology 24%. [6, 7, 20] After the first 10 days, the likelihood of a positive blood smear or PCR rapidly diminish, whereas serological testing continues to improve in sensitivity, such that 99% are seropositive by 30 days, most displaying either a 4-fold increase in antibody titer or a clear seroconversion. Once suspected, even prior to final diagnostic laboratory confirmation, treatment with doxycycline is warranted in both adults and children alike. With pregnancy and with tetracycline allergy, limited investigation supports the use of rifampin as an alternative.

References
  1. Bakken JS, Dumler JS, Chen SM, Eckman MR, Van Etta LL, Walker DH. Human granulocytic ehrlichiosis in the upper Midwest United States. A new species emerging? JAMA 1994;272(3):212-8.

  2. Chen SM, Dumler JS, Bakken JS, Walker DH. Identification of a granulocytotropic Ehrlichia species as the etiologic agent of human disease. J Clin Microbiol 1994;32(3):589-95.

  3. Dumler JS, Madigan JE, Pusterla N, Bakken JS. Ehrlichioses in humans: epidemiology, clinical presentation, diagnosis, and treatment. Clin Infect Dis 2007;45 Suppl 1:S45-51.

  4. Zhang L, Liu Y, Ni D, et al. Nosocomial transmission of human granulocytic anaplasmosis in China. JAMA 2008;300(19):2263-70.

  5. Heo EJ, Park JH, Koo JR, et al. Serologic and molecular detection of Ehrlichia chaffeensis and Anaplasma phagocytophila (human granulocytic ehrlichiosis agent) in Korean patients. J Clin Microbiol 2002;40(8):3082-5.

  6. Dumler JS, Choi KS, Garcia-Garcia JC, et al. Human granulocytic anaplasmosis and Anaplasma phagocytophilum. Emerg Infect Dis 2005;11(12):1828-34.

  7. Bakken JS, Dumler S. Human granulocytic anaplasmosis. Infectious disease clinics of North America 2008;22(3):433-48, viii.

  8. Levin ML, Nicholson WL, Massung RF, Sumner JW, Fish D. Comparison of the reservoir competence of medium-sized mammals and Peromyscus leucopus for Anaplasma phagocytophilum in Connecticut. Vector Borne Zoonotic Dis 2002;2(3):125-36.

  9. Michalski M, Rosenfield C, Erickson M, et al. Anaplasma phagocytophilum in central and western Wisconsin: a molecular survey. Parasitol Res 2006;99(6):694-9.

  10. Dumler JS. Anaplasma and ehrlichia infection. Ann N Y Acad Sci 2005;1063:361-73.

  11. Carlyon JA, Fikrig E. Invasion and survival strategies of Anaplasma phagocytophilum. Cell Microbiol 2003;5(11):743-54.

  12. Carlyon JA, Fikrig E. Mechanisms of evasion of neutrophil killing by Anaplasma phagocytophilum. Curr Opin Hematol 2006;13(1):28-33.

  13. Borjesson DL, Kobayashi SD, Whitney AR, Voyich JM, Argue CM, Deleo FR. Insights into pathogen immune evasion mechanisms: Anaplasma phagocytophilum fails to induce an apoptosis differentiation program in human neutrophils. J Immunol 2005;174(10):6364-72.

  14. Choi KS, Park JT, Dumler JS. Anaplasma phagocytophilum delay of neutrophil apoptosis through the p38 mitogen-activated protein kinase signal pathway. Infect Immun 2005;73(12):8209-18.

  15. Lepidi H, Bunnell JE, Martin ME, Madigan JE, Stuen S, Dumler JS. Comparative pathology, and immunohistology associated with clinical illness after Ehrlichia phagocytophila-group infections. Am J Trop Med Hyg 2000;62(1):29-37.

  16. Walker DH, Dumler JS. Human monocytic and granulocytic ehrlichioses. Discovery and diagnosis of emerging tick-borne infections and the critical role of the pathologist. Arch Pathol Lab Med 1997;121(8):785-91.

  17. Dumler JS, Barat NC, Barat CE, Bakken JS. Human granulocytic anaplasmosis and macrophage activation. Clin Infect Dis 2007;45(2):199-204.

  18. Martin ME, Bunnell JE, Dumler JS. Pathology, immunohistology, and cytokine responses in early phases of human granulocytic ehrlichiosis in a murine model. J Infect Dis 2000;181(1):374-8.

  19. Martin ME, Caspersen K, Dumler JS. Immunopathology and ehrlichial propagation are regulated by interferon-gamma and interleukin-10 in a murine model of human granulocytic ehrlichiosis. Am J Pathol 2001;158(5):1881-8.

  20. Bakken JS, Haller I, Riddell D, Walls JJ, Dumler JS. The serological response of patients infected with the agent of human granulocytic ehrlichiosis. Clin Infect Dis 2002;34(1):22-7.