Hop
Back to the Neurology Journal Club Page!
Acknowledgment
We thank Dr. J. Scott Briggs (Lawson, OK) for referral of this case.
Abnormalities noted at physical examination included depressed mentation, generalized weakness, hyperreflexia of all 4 limbs, mild conscious proprioceptive deficits of the rear limbs, and evidence of cervical pain. A CBC and biochemical profile revealed a severe nonregenerative anemia (PCV 13.9%, reticulocytes 0.2%), thrombocytopenia (61 x 103/µL), and hyperproteinemia (7.8 g/dL, reference range 6.0-7.5 g/dL) resulting from hyperglobulinemia (5.5 g/dL).
The dog was anesthetized for bone marrow aspiration, cerebrospinal fluid (CSF) collection, and spinal radiography. No radiographic abnormalities were observed on spinal radiographs. The bone marrow aspirate yielded numerous marrow flecks of normal cellularity. Normal numbers of megakaryocytes were present. The myeloid (M) series was prominently represented and displayed orderly maturation. Cells of the erythroid (E) series also displayed normal maturation, but appeared quantitatively decreased. The (M:E) ratio was estimated to be approximately 3:1. One to 8 plasma cells were present per 50x field in cellular areas of the smears and hemosiderin-laden macrophages were abundant. These findings were interpreted as mild erythroid hypoplasia with mild plasmacytosis and increased iron stores. The CSF was cloudy and had pleocytosis (total nucleated cell count 590/µL, reference range 0-8/µL), increased protein concentration (3,400 mg/dL, reference range <25 mg/dL), and a positive Pandy test. Cytologic examination revealed 88% nondegenerate neutrophils and 12% mononuclear cells. Many neutrophils contained basophilic to amphophilic, cytoplasmic inclusions, 1.5-5 µm in diameter, that were morphologically consistent with Ehrlichia spp. morulae (Fig 1). The CSF had a reciprocal canine distemper virus antibody titer of 32. A diagnosis of ehrlichiosis with neurologic involvement was made. Based on the cellular tropism of organisms observed, the infectious agent was assumed to be Ehrlichia ewingii, as this is the most common granulocytic ehrlichial agent to cause clinical disease in dogs in Oklahoma. Reevaluation of previously submitted blood smears failed to demonstrate any morulae. Because E. ewingii is sometimes associated with polyarthritis in dogs,1 synovial fluid was collected from several joints, but neither morulae nor any other cytologic abnormalities were observed in any of these samples.
The dog was treated with chloramphenicol (40 mg/kg, IV ql2h) for 24 hours after diagnosis and then switched to doxycycline (5 mg/kg PO ql2h for 21 days). Short-term dexamethasone (1 dose each of 0.1 mg/kg, 0.05 mg/kg, and 0.025 mg/kg IV ql2h) therapy was given to reduce central nervous system (CNS) inflammation. The weakness and ataxia improved over the next 4 days and the patient was discharged from the hospital on day 6. At reevaluation 2 months after discharge, the dog was clinically and hematologically normal. At last contact, 2.5 years after initial presentation, the dog was clinically normal.
Whole blood collected immediately after CSF evaluation was used to intravenously inoculate a female Treeing Walker Coonhound (dog 2100) in an attempt to isolate the causative organism as part of an ongoing study of canine ehrlichial disease. This dog (dog 2100) was monitored daily for signs of disease and the presence of morulae in peripheral blood smears. On day 13 postinfection (PI), the dog became febrile (40.5ÿC) and exhibited nasal and ocular discharge. Hematologic abnormalities included severe thrombocytopenia (<15,000/µL) and mild normocytic, normochromic anemia. Morulae were seen in a few peripheral blood leukocytes after the 16th day PI. Unlike those seen in the CSF of the index case, these morulae were located exclusively in mononuclear cells (primarily lymphocytes), typical of Ehrlichia canis rather than E. ewingii. Reexamination of slides from the original CSF sample revealed rare morulae in mononuclear cells as well as the neutrophils.
Ehrlichia chaffeensis is 1 etiologic agent of human ehrlichiosis and primarily invades lymphocytes and monocytes but can also be found in granulocytes.2 Because dogs have been shown to be susceptible to experimental and natural infection with E. chaffeensis,3,4 the index case possibly could have been infected with this organism. Personnel at the Centers for Disease Control and Prevention (CDC) were contacted and further studies were done.
Ethylenediaminetetraacetic acid (EDTA)-anticoagulated blood samples collected from the index dog (stored from the day of CSF examination) and dog 2100 (collected during acute parasitemia) were submitted to the CDC for further examination by nested polymerase chain reaction (PCR) for detection of Ehrlichia spp. 16S DNA. DNA was extracted and the outside amplification products were produced as described.5 Nested PCR was used to provide species-specific identification based on 16S DNA sequence. Each outside reaction product was used as template in a 2nd reaction identical to the 1st except for the use of species-specific primers to detect E. canis, E. equi, E. chaffeensis, and E. ewingii.4 Each positive control product was detected, and the E. canis-specific primers permitted amplification of a 391-base pair (bp) product from both the index case and dog 2100, indicating the presence of this organism in both animals. No other amplification products were observed with any of the other primer sets. Sequencing of the 391-bp PCR product produced with the E. canis-specific primers revealed only a single base pair difference (an A to G change at position 199 of GenBank accession M73221) from E. canis (Oklahoma strain).6 This isolate was designated "Ebony" to distinguish it from other isolates of E. canis.
E. canis is found almost exclusively in mononuclear cells (lymphocytes, monocytes) in blood samples from infected dogs as well as in CSF samples from dogs with ehrlichia-induced inflammatory CNS disease (Meinkoth and Cowell, personal observation).7 In an attempt to understand the predominantly neutrophilic location of morulae seen on the original CSF slides, a stored sample of cells from the CSF, collected in EDTA, from the index animal was submitted to the CDC. This sample was analyzed by repetitive element PCR (rep-PCR), a new technique that allows comparison of a larger fraction of the infecting organism genome than does 16S rDNA detection.8 This revealed a banding pattern identical to that of E. ewingii (Figs 2, 3).
To compare the effects of the Ebony isolate with those of a better characterized isolate of E. canis (Oklahoma isolate), 2 naive mixed breed, littermate pups (dogs 2430 and 2431) were inoculated intravenously with blood of either a carrier of the Ebony isolate or the Oklahoma isolate of E. canis, respectively. Both pups developed signs of clinical disease that was characterized by anorexia, pyrexia (39.4-40.4ÿC), and ocular and nasal discharges. Signs were more pronounced and of a longer duration (7 days versus 2 days) in the pup infected with the Ebony isolate than in the pup infected with the Oklahoma isolate. Both pups developed severe thrombocytopenia (10,000-20,000/µL) and moderate normocytic, normochromic anemia (PCV 22-29%). Morulae were observed in peripheral blood leukocytes of both dogs beginning on day 16 PI. Again, morulae were seen exclusively in mononuclear cells. Development of antibodies to E. canis, as measured by the FIAX system,6 occurred by day 21 for the pup infected with the Ebony isolate, whereas the pup infected with the Oklahoma isolate remained seronegative through 26 days of monitoring. Both pups were euthanized on day 26 PI. Gross and histologic lesions were similar in both pups. Gross lesions consisted of prominent enlargement of lymphoid structures throughout the body. Histologic examination revealed hyperplasia of lymph nodes and spleen and lymphoreticular cuffing of vasculature in the lung, kidney, portal areas of the liver, and various other loci. Both pups had severe nonsuppurative meningoencephalitis, uveitis, and retinitis.
The designation of species within the tribe Ehrlichieae has traditionally been based largely on host range, cell tropism, geographic distribution, and antibody response.9 These criteria are somewhat general and not always sufficient for unambiguous classification. More than 1 species of Ehrlichia may infect any given cell line for a particular host and some species may occasionally infect more than 1 cell type. Also, serologic cross-reactivity occurs among many species8 and some species, such as E. ewingii, have not been grown successfully in long-term cell culture, so no specific antibody tests exist.10 Recently, molecular genetic techniques have been used to define new Ehrlichia species as well as to clarify relationships among previously described species.9 The PCR technique can be used to identify the causative agent in a clinical specimen by amplification of the unique Ehrlichia sp. 16S rDNA.11,12 This is accomplished by using primers that are specific for variable regions within the 16S rRNA gene. The amplified DNA fragments can be sequenced and the species identified by sequence variability.
Dogs have been reported to host a variety of ehrlichial species, including some that have traditionally been considered primary pathogens of other species. E. canis,13 E. ewingii,1,14,15 E. platys,16 E. risticii,17 E. equi,18 E. chaffeensis,3,4 and the human granulocytic ehrlichiosis (HOE) agent19-21 have all been identified from natural or experimental infections in dogs. Recent studies indicate that the HGE agent is closely related to, if not a strain of, E. equi. These organisms show significant 16S rRNA homology,10 are cross-protective,22 and isolates from humans with HOE have been shown to cause clinical disease in horses.23
E. canis, an agranulocytic species, and E. ewingii, a granulocytic species, are the ehrlichial species that most commonly cause clinically apparent disease in dogs in Oklahoma. The dog of this report was first assumed to be infected with E. ewingii because all morulae initially seen in the CSF sample were within neutrophils. It was surprising, therefore, that attempts to experimentally infect naive dogs resulted in recovery of an organism expressed strictly in mononuclear cells. This apparent shift in tropism prompted consideration of infection with E. chaffeensis. However, this hypothesis was dismissed because no DNA was amplified from the blood of either the index dog (taken before treatment) or of dog 2100 (taken during parasitemia) by PCR using primers specific for E. chaffeensis. Furthermore, sequence analysis of segments of the 16S rRNA gene of the organism recovered from blood of the index dog showed it to be a strain of E. canis that differed at only 1 base pair from the reported sequence of the Oklahoma isolate of E. canis. This infection is unlikely to explain the presence of numerous morulae within neutrophils of the CSF sample, as E. canis is found predominantly within mononuclear cells. When the CSF sample was analyzed by a newly developed rep-PCR technique, a banding pattern identical to that of E. ewingii was obtained. Therefore, it seems likely that the dog was infected with both organisms. E. canis and E. ewingii are cross-reactive, but are not cross-protective,24 and combined infections can be produced experimentally.24
Several questions remain. The order of infection and duration of infection with each organism cannot be determined. The initial negative E. canis titer suggests that the dog was not chronically infected with this organism. The strain identified in this dog (Ebony) cross-reacted with standard E. canis serologic tests in subsequently infected dogs. It is also not clear why only 1 species was recovered when passaged via injection of whole blood, or why only E. canis DNA was identified in the peripheral blood. These 2 results taken together would imply a lack of, or an extremely low level of, E. ewingii in the systemic circulation. The identification of only E. ewingii DNA in the CSF is likely the result of the relative concentrations of organisms at this site. The vast majority of the morulae in the CSF were in neutrophils. The primers used in rep-PCR are not species specific (species and strains are differentiated based on the pattern of bands obtained) and the more abundant template DNA will be expressed.
Although not present in the majority of affected animals, necrologic disorders are a well-documented manifestation of E. canis infection.14,25 Ataxia, seizures, hyperesthesia, cranial nerve deficits, and central and peripheral vestibular dysfunction have all been reported. Such clinical manifestations are not surprising because a nonsuppurative meningoencephalitis is one of the more consistent pathologic findings associated with E. canis infection.14,26 Humans with E. chaffeensis infection have been reported to exhibit CNS disorders.27,28 Also, approximately one half of the reported cases in 1 study of people with HGE (an agent distinct from E. chaffeensis) have had neurologic signs.12 In contrast, clinical signs and histologic lesions referable to the necrologic system have not been prominent findings in most descriptions of E. ewingii infection. However, Maretzki et al29 recently described a case of granulocytic ehrlichiosis in a dog with meningitis. This indicates that at least 1 species of granulocytic ehrlichial infection can cause necrologic disease in dogs; however, the species of Ehrlichia was not positively identified in that case.
Few reports detail results of CSF analysis in dogs with E. canis-induced necrologic disease. Among these, a mild to moderate mononuclear (primarily lymphocytic) pleocytosis and a moderately increased protein concentration have been observed.7,25 We have noted mononuclear pleocytosis in 2 other clinical cases in which Ehrlichia morulae were present in CSF mononuclear cells (Meinkoth and Cowell, personal observations). Both of those dogs had strongly positive E. canis titers and were presumed to be infected with E. canis, but confirmatory studies were not done. Such cytologic findings reflect the nonsuppurative nature of the meningoencephalitis in afflicted animals. A similar lymphocytic pleocytosis has been reported in human patients with E. chaffeensis infection.27,28 In contrast, the CSF pleocytosis seen in the 2 cases involving granulocytic ehrlichial species (the present case and that of Maretzki et al29) was characterized by a higher percentage of neutrophils (88 and 48%, respectively). Additional studies will be necessary to determine whether this is a consistent finding that is dependent on the species of Ehrlichia, or whether it is related to some other factor such as timing of CSF evaluation with respect to the course of disease. The dog described in the present report had a substantially greater nucleated cell count in the CSF than has been reported previously with ehrlichial-induced CNS disease.
Demonstration of ehrlichial organisms within CSF specimens is rare, and we know of only 2 previous reports in the veterinary literature.7,29 The 1st report dealt with a dog that presented with seizures and nonregenerative anemia.7 CSF analysis revealed mild mononuclear pleocytosis, increased protein, and leukocytes that contained Ehrlichia morulae. The infected leukocytes were sparse and all were mononuclear cells, suggesting E. canis infection. In the 2nd instance,29 morulae were located primarily within neutrophils, suggesting an infection with E. ewingii, an organism of the HGE/E. equi group, or some other unidentified organism.
This is the 1st report of a naturally occurring, apparently concurrent, infection of a dog with 2 separate species of Ehrlichia and our findings underscore the usefulness of DNA sequence analysis in studies of ehrlichial disease. The report also supports the notion that granulocytic ehrlichiosis can produce CNS disease in the dog. Although we have evidence of a dual infection in this dog and no way to assess the relative importance of each pathogen to the disease process, the presence of large numbers of organisms in CSF neutrophils during the acute phase of illness (and their concurrent absence from blood) suggests that E. ewingii was an active component of the CNS disease.
1. Stockham SL, Schmidt DA, Tyler JW. Canine granulocytic ehrlichiosis in dogs from central Missouri: A possible cause of polyarthritis. Vet Med Rev 1985:6:3-5.
2. Maeda K, Markowitz N, Hawley RC, et al. Human infection with Ehrlichia canis, a leukocytic rickettsia. N Engl J Med 1987;316:853-856.
3. Dawson JE, Ewing SA. Susceptibility of dogs to infection with Ehrlichia chaffeensis, causative agent of human ehrlichiosis. Am J Vet Res 1992;53:1322-1327.
4. Dawson IE, Biggie KL, Warner CK, et al. Polymerase chain reaction evidence of Ehrlichia chaffeensis, an etiologic agent of human ehrlichiosis, in dogs from southeast Virginia. Am J Vet Res 1996,57:1175-1979.
5. Dawson JE, Warner CK, Baker B, et al. Ehrlichia-like 16S rDNA sequence from wild white-tailed deer (Odocoileus virginianus). J Parasitol 1996;82:52-58.
6. Mathew JS, Ewing SA, Barker RW, et al. Attempted transmission of Ehrlichia canis by Rhipicephalus sanguineus after passage in tissue cell culture. Am J Vet Res 1996;57:1594-1598.
7. Meinkoth JH, Hoover JP, Cowell RL, et al. Ehrlichiosis in a dog with seizures and nonregenerative anemia. J Am Vet Med Assoc 1989;195:1754-1755.
8. Dawson JE, Warner CK, Ewing SA, et al. Fingerprinting of Ehrlichia species by repetitive element PCR. Am J Trop Med Hyg 1997;57:109-114.
9. Anderson BE, Dawson JE, Jones DC, et al. Ehrlichia chaffeensis, a new species associated with human ehrlichiosis. J Clin Microbiol 1991;29:2838-2842.
10. Chen SM, Dumler JS, Bakken JS, et al. Identification of a granulocytotropic Ehrlichia species as the etiologic agent of human disease. J Clin Microbiol 1994;32:589-595.
11. Dunn BE, Monson TP, Dumler JS, et al. Identification of Ehrlichia chaffeensis morulae in cerebrospinal fluid mononuclear cells. J Clin Microbiol 1992;30:2207-2210.
12. Bakken JS, Dumler JS, Chen SM, et al. Human granulocytic ehrlichiosis in the upper midwest United States. A new species emerging? JAMA 1994;272:212-218.
13. Troy GC, Forrester SD. Canine ehrlichiosis. In: Greene CD, ed. Infectious Diseases of the Dog and Cat. Philadelphia, PA: WB Saunders; 1990:404-414.
14. Ewing SA, Roberson WR, Buckner RG, et al. A new strain of Ehrlichia canis. J Am Vet Med Assoc 1971;159:1771-1774.
15. Anderson BE, Greene CE, Jones DC, et al. Ehrlichia ewingii sp. nov., the enologic agent of canine granulocytic ehrlichiosis. Int J Syst Bacteriol 1992;42:299-302.
16. Harvey JW. Ehrlichia platys infection (infectious cyclic thrombocytopenia of dogs). In: Greene CD, ed. Infectious Diseases of the Dog and Cat. Philadelphia, PA: WB Saunders; 1990:415-418.
17. Kakoma I, Hansen RD, Anderson BE, et al. Cultural, molecular, and immunological characterization of the etiologic agent for atypical canine ehrlichiosis. J Clin Microbiol 1994;32:170-175.
18. Lewis GE, Huxsoll DL, Ristic M, et al. Experimentally induced infection of dogs, cats, and nonhuman primates with Ehrlichia equi, etiologic agent of equine ehrlichiosis. Am J Vet Res 1975;36:85-88.
19. Johansson KE, Pettersson B, Uhlen MU, et al. Identification of the causative agent of granulocytic ehrlichiosis in Swedish dogs and horses by direct solid phase sequencing of PCR products from the 16S rRNA gene. Res Vet Sci 1995;58:109-112.
20. Greig B, Asanovich KM, Armstrong PJ, et al. Geographic, clinical, serologic, and molecular evidence of granulocytic ehrlichiosis, a likely zoonotic disease, in Minnesota and Wisconsin dogs. J Clin Microbiol 1996;34:44-48.
21. Ewing SA, Dawson JE, Panciera RJ, et al. Dogs infected with a human granulocytotropic Ehrlichia spp. (Rickettsiales:Ehrlichieae). J Med Entomol 1997;34:710-718.
22. Barlough JE, Madigan JE, DeRock E, et al. Protection against Ehrlichia equi is conferred by prior infection with the human granulocytotropic ehrlichia (HGE agent). J Clin Microbiol 1995;33:3333-3334.
23. Madigan JE, Richter PJ, Kimsey RB, et al. Transmission and passage in horses of the agent of human granulocytic ehrlichiosis. J Infect Dis 1995;172:1141-1144.
24. Ewing SA, Fox JC, Johnson EM, et al. Canine ehrlichiosis: Differences between granulocytic and agranulocytic agents (abstr), Prog Abstr 69th Ann Meet Conf. Res. Work. Anim. Dis. 1988;40 (abstract).
25. Greene CE, Burgdorfer W, Cavagnolo R, et al. Rocky Mountain spotted fever in dogs and its differentiation from canine ehrlichiosis. J Am Vet Med Assoc 1985;186:465-72.
26. Hildebrandt PK, Huxsoll DL, Walker JS, et al. Pathology of canine ehrlichiosis (tropical canine pancytopenia). Am J Vet Res 1973;34:1309-1320.
27. Harkess JR, Stucky D, Ewing SA. Neurologic abnormalities in a patient with human ehrlichiosis. South Med J 1990;83:1341-1343.
28. Eng TR, Harkess JR, Fishbein DB, et al. Epidemiologic, clinical, and laboratory findings of human ehrlichiosis in the United States. 1988. JAMA 1990;264:2251-2258.
29. Maretzki CH, Fisher DJ, Greene CE. Granulocytic ehrlichiosis and meningitis in a dog. J Am Vet Med Assoc 1994:205:1554-1556.
Fig 1. Cytospin preparation of cerebrospinal fluid. An ehrlichia morula is present in a neutrophil near the top of the field (Wright's stain; bar = 10 µm).
Fig 2. Ehrlichia fingerprints generated by repetitive element polymerase chain reaction from 10 µL of extracted template DNA. Lane 1, 123-base pair molecular weight ladder; lane 2, E. sennetsu; lane 3. E. risticii; lane 4, E. canis Oklahoma strain; lane 5, E. chaffeensis; lane 6, E. ewingii; lane 7, E. platys; lane 8, human granulocytotropic Ehrlichia; lane 9, E. equi, lane 10, DNA molecular weight marker XI. Values on the right are in base pairs. (Reprinted from Dawson et al,8 with permission.)
Fig 3. Results of repetitive element polymerase chain reaction analysis of cerebrospinal fluid from the index dog. Lane 1, 123-base pair molecular weight ladder; lane 2, Ehrlichia ewingii; lane 3, isolate from the cerebrospinal fluid of the index dog; lane 4, DNA molecular weight marker XI. Values on the right are in base pairs. (Reprinted from Dawson et al,8 with permission.)