|Home | About | Journals | Submit | Contact Us | Français|
Serologic and molecular evidence of Anaplasma phagocytophilum has been demonstrated in white-tailed deer (WTD; Odocoileus virginianus), and deer are an important host for the tick vector Ixodes scapularis. In this study, we describe experimental infection of WTD with A. phagocytophilum. We inoculated four WTD with a human isolate of A. phagocytophilum propagated in tick cells. Two additional deer served as negative controls. All inoculated deer developed antibodies (titers, ≥64) to A. phagocytophilum, as determined by an indirect fluorescent antibody test, between 14 and 24 days postinfection [p.i.]), and two deer maintained reciprocal titers of ≥64 through the end of the 66-day study. Although morulae were not observed in granulocytes and A. phagocytophilum was not reisolated via tick cell culture of blood, 16S reverse transcriptase nested PCR (RT-nPCR) results indicated that A. phagocytophilum circulated in peripheral blood of three deer through at least 17 days p.i. and was present in two deer at 38 days p.i. Femoral bone marrow from one deer was RT-nPCR positive for A. phagocytophilum at 66 days p.i. There was no indication of clinical disease. These data confirm that WTD are susceptible to infection with a human isolate of A. phagocytophilum and verify that WTD produce detectable antibodies upon exposure to the organism. Because adults are the predominant life stage of I. scapularis found on deer and because adult I. scapularis ticks do not transmit A. phagocytophilum transovarially, it is unlikely that WTD are a significant source of A. phagocytophilum for immature ticks even though deer have a high probability of natural infection. However, the susceptibility and immunologic response of WTD to A. phagocytophilum render them suitable candidates as natural sentinels for this zoonotic tick-borne organism.
Human granulocytic anaplasmosis is an acute, febrile disease that may be accompanied by headache, myalgia, pancytopenia, and elevated serum aminotransferase levels (13). Although the disease is often mild, delayed treatment, misdiagnosis, and/or immunosuppression may result in a severe or fatal outcome (28, 31). Approximately 1,220 cases of human granulocytic anaplasmosis (HGA) have been diagnosed in the United States since 1994, when the disease was first described (13, 14). Sporadic HGA cases have been reported in Europe (10).
Previous to its recognition as a human pathogen, the organism was known in veterinary medicine as Ehrlichia equi (causing equine granulocytic ehrlichiosis) and Ehrlichia phagocytophila (causing tick-borne fever in sheep, goats, and cattle in several European countries). Based on recent phylogenetic analysis, the etiologic agent of HGA and E. equi were synonomized with E. phagocytophila. Furthermore, the analysis indicated that this organism should be reassigned to the genus Anaplasma, thus resulting in the currently accepted designation, Anaplasma phagocytophilum (20). The marked differences in host preference, clinical manifestations, and geographical distribution are attributed to the existence of multiple variant strains of A. phagocytophilum (20).
North American strains of A. phagocytophilum cause clinical disease in domestic animals, notably horses and dogs (20). Experimentally, both equine and human A. phagocytophilum strains are pathogenic to horses (37) and cross protective (6). Disease due to A. phagoctyophilum in domestic cattle has not been reported in North America. Furthermore, two steers were not susceptible to experimental infection with human and equine North American strains; although the steers seroconverted, blood was negative for A. phagocytophilum by 16S real-time PCR at all sampling dates (51). Anaplasma phagocytophilum was first isolated from human patients by using a human promyelocytic leukemia (HL60) cell line (24) and from horses and dogs by using Ixodes scapularis cell lines (45, 46).
Knowledge of the natural history of A. phagocytophilum remains incomplete. In the eastern and midwestern United States, the white-footed mouse (Peromyscus leucopus) and the black-legged tick (I. scapularis) are a competent reservoir host and principal vector, respectively (19, 54). Although field studies indicate that numerous species of wild rodents and other wild mammals may be naturally infected with A. phagocytophilum (22, 23, 32, 40, 48, 59), the relative importance of these hosts as sources of A. phagocytophilum for ticks has not been determined.
White-tailed deer (WTD; Odocoileus virginianus) are the principal hosts for adult I. scapularis (30) and therefore predictably are exposed to A. phagocytophilum. Sequence-confirmed 16S rRNA genes identical to those of A. phagocytophilum and/or A. phagocytophilum-reactive antibodies have been demonstrated in wild WTD from Connecticut, Georgia, Indiana, Maryland, Missouri, South Carolina, and Wisconsin (3, 8, 33, 38, 39, 60); however, clinical disease due to A. phagocytophilum has not been reported in WTD. Thus, WTD have been identified as both a potential sentinel species for A. phagocytophilum (8, 33, 39, 40, 60), as well as a potential reservoir host of A. phagocytophilum (3, 5, 6, 8, 43); however, these potential roles have not been analyzed because data on susceptibility to and course of infection with A. phagocytophilum in WTD have previously been unavailable. Therefore, we proposed to analyze these potential roles by experimentally infecting WTD with A. phagocytophilum.
Six WTD fawns (three females and three males, orphaned in the state of Georgia) were hand raised and housed in tick-free facilities at the College of Veterinary Medicine, The University of Georgia. Prior to initiation of the experiment, at approximately 7 months of age, the deer were screened for various hematotropic microorganisms known to infect wild WTD in the southeastern United States. The deer were determined to be seronegative by indirect fluorescence antibody (IFA) test for A. phagocytophilum and Ehrlichia chaffeensis (8, 17) and PCR negative for A. phagocytophilum, E. chaffeensis, Ehrlichia ewingii, and an undescribed Anaplasma sp. of WTD (33, 44). All deer were determined to be culture negative for Trypanosoma cervi, a protozoan known to interfere with tissue culture of deer blood (35, 36, 44). Two deer (WTD132 and WTD133) were diagnosed with Theileria cervi infection by observation of intraerythrocytic protozoans in whole-blood smears.
For all procedures, deer were anesthetized as described previously (16). Deer blood samples were collected via aseptic jugular venipuncture at 0, 6, 10, 13, 17, 24, 31, 38, 45, 54, and 66 days postinfection (days p.i.). Approximately 9 ml of whole blood was collected for culture, reverse transcriptase nested PCR (RT-nPCR) and clinical hematology in three separate EDTA Vacutainer tubes (Becton Dickinson, Rutherford, NJ) and for serology in a tube containing no additive. On all sampling dates, Giemsa-stained blood smears were prepared, and whole blood was submitted to the Clinical Pathology Laboratory, College of Veterinary Medicine, The University of Georgia, for analysis of the following parameters: hematocrit, erythrocyte count, hemoglobin levels, platelet count, total and differential leukocyte counts, and fibrinogen levels. Sera and duplicate samples of whole blood were stored at −20°C.
Deer were observed twice daily for visible signs of clinical illness, including decreased feed intake, depression, and reluctance to move. Complete physical examination (11) of each deer was performed at each blood collection date. The two control deer were removed from the study at 31 days p.i. All deer were euthanized via intravenous sodium pentobarbital overdose and subjected to complete necropsies.
Preparatory to and concurrently with the experimental deer trial, female C3H-HeN mice (Harlan, Indianapolis, IN), approximately 6 weeks old, were used to test animal infectivity of the A. phagocytophilum isolate propagated in tick cells (46). Mice were anesthetized for all procedures via subcutaneous injection of 10 mg of xylazine (Phoenix Scientific, Inc., St. Joseph, MO)/kg of body weight and 100 mg/kg ketamine HCl (Ft. Dodge Labs, Inc., Fort Dodge, IA). Samples of blood taken 6 days p.i. for RT-nPCR were collected retro-orbitally from all mice. After euthanasia of mice, blood samples were collected for culture by intracardiac puncture, and spleen samples were taken for RT-nPCR. All animals involved in this study were cared for and used in accordance with guidelines established by the Institutional Animal Care and Use Committee of The University of Georgia.
An I. scapularis tick cell line (ISE6) was used to propagate a human isolate (HGE-1) of A. phagocytophilum (46). Stock ISE6 cultures were maintained as described previously (47). The inoculum consisted of five 12.5-cm2 Falcon culture flasks (Becton Dickenson, Franklin Lakes, NJ) of A. phagocytophilum-infected tick cells, in which approximately 35% of the cells were infected. Monolayers were resuspended in existing media, and sterile water (Sigma, St. Louis, MO) was added to bring the volume of each flask to 8.0 ml. Then, the contents of all five flasks were combined for a total volume of 40 ml, and the mixture was divided into 2.0-ml aliquots for deer and 0.5-ml aliquots for mice. Negative-control injection material was prepared in a similar manner, using uninfected stock ISE6 cultures.
Each of four deer (WTD131, -132, -133, and -134) was injected with 2.0 ml of the A. phagocytophilum inoculum by each of four routes: intradermal, subcutaneous, intravenous, and intraperitoneal, for a total of 8.0 ml of inoculum per deer. Concurrently with the deer inoculation, each of nine mice was injected intraperitoneally with 0.5 ml of the A. phagocytophilum inoculum. Two negative control deer (WTD127 and WTD139) and three negative control mice were injected in a similar manner with uninfected ISE6 cells. Remaining fractions of the A. phagocytophilum inoculum and uninfected tick cells were used to prepare Giemsa-stained cytospins and were tested by RT-nPCR and DNA sequencing (see below).
The IFA test was performed as described previously (8). In brief, sera were screened at a dilution of 1:64 in 0.01 M phosphate-buffered saline on commercially prepared HGE substrate slides (Focus Technologies [formerly MRL Diagnostics], Cypress, CA). Fluorescein isothiocyanate-labeled rabbit anti-deer immunoglobulin G (Kirkegaard & Perry Laboratories, Gaithersburg, MD), diluted 1:50 in phosphate-buffered saline, was used as a conjugate. When distinct fluorescent staining of organisms was observed at a 1:64 dilution, serial twofold dilutions were performed. Serologic results are reported as reciprocals of the highest dilution at which specific fluorescence was observed.
Reisolation attempts of A. phagocytophilum from mouse and deer blood were performed as previously described (44, 45) except that resuspended ISE6 stock cell monolayers and washed WTD buffy coat cells were mixed, pelleted at 720 × g for 20 min, and then allowed to stand at room temperature for 30 min before the pellet was resuspended in “ehrlichia medium” (46) and divided into two flasks. Duplicate flasks were monitored and maintained separately using different sets of reagents as a precaution against contamination; antibiotics were not used at any time. Cultures were monitored by visual observation for cytopathic effect (CPE) (46), light microscopy of Giemsa-stained cell spreads, and RT-nPCR of cell culture supernatant. Monolayers were examined daily for development and progression of CPE. Periodically, samples were prepared from all cultures by centrifugation (720 × g for 20 min) of the entire volume of spent medium removed during feeding of the culture, followed by resuspension of pelleted cells in approximately 1.0 ml of spent medium. For cell spreads, 100 μl of each sample was placed in a Cytofuge filter concentrator (StatSpin; Iris Co., Norwood, MA) and centrifuged at 27 × g for 4 min using a Cytofuge 2 cytocentrifuge (StatSpin). Slides were air dried, fixed in 100% methanol, and stained in a 4% solution of Giemsa (Karyomax; GIBCO, Grand Island, NY) in Sorensen buffer, pH 6.5, for 30 min in a 37°C water bath. Stained cells were examined microscopically (×400 to ×1,000) for intracytoplasmic organisms. Our cell culture protocol specified a 60-day monitoring period for CPE, followed by RT-nPCR testing.
Total RNA was extracted from fresh deer and mouse blood samples with the RNA Blood Minikit (QIAGEN, Inc., Valencia, CA) and from cell culture aliquots and postmortem tissue stored at −70°C with the QiAmp Viral RNA Extraction kit and the RNEasy Minikit (QIAGEN, Inc.), respectively. All extractions were performed according to the manufacturer's instructions and in an RNase-free environment. RT-nPCR was performed on RNA extracted from deer blood, cell culture, and tissues in the following manner. Reverse transcription of 16S r-RNA to cDNA and subsequent primary amplification using primers ECC and ECB were carried out in a single-tube reaction, followed by secondary amplification as described previously (44), except that secondary primers GE9F and GA1UR were used to generate an internal 411-bp fragment (13, 34).
Two additional gene targets (p44 and groESL) for detection of A. phagocytophilum RNA were used. For detection of p44 RNA, RT-PCR was performed using primers MSP3F and MSP3R (65). Reaction mixtures were subjected to 45°C for 20 min, followed by 39 cycles of the following profile: 94°C for 60 s, 55°C for 45 s, and 72°C for 60 s, using a PTC-100TM Thermal Cycler. For detection of groESL RNA, RT-nPCR was performed using primers APF1 (5′TAGTGATGAAGGAGAGTGAC) and APR1 (5′CCAGGIGCCTTIACAGCWGCAAC) in a primary reaction and primers APF10 (5′TATGCTACGGTTGTTTGTTC) and APR11 (5′GGCGAAAGATATCCGCGA) in a secondary reaction to generate a 652-bp product (primers were generously provided by John Sumner, Centers for Disease Control and Prevention). Primary reaction mixtures were subjected to 43°C for 15 min, followed by 95°C for 5 min and 39 cycles of the following profile: 95°C for 30 s, 52°C for 30 s, and 72°C for 60 s; and a final step of 72°C for 5 min. Secondary reaction mixtures were subjected to 95°C for 5 min; 29 cycles of 95°C for 30 s, 52°C for 30 s, and 72°C for 60 s; and a final step of 72°C for 5 min.
All RT-PCRs were carried out with a PTC-100TM Thermal Cycler. All reagent concentrations were identical to those used for A. phagocytophilum 16S rRNA RTn-PCR. Amplification products were separated by electrophoresis on a 2% agarose gel, stained in ethidium bromide, and visualized with UV transillumination.
Quality control measures included negative controls (water) that were extracted and amplified in parallel with all specimens. To minimize the potential for DNA contamination, three separate, designated areas were used for extraction of RNA and preparation of primary and secondary PCRs. Additionally, two thermal cyclers were used, designated for either primary or secondary amplification.
RT-nPCR products from the original inoculum, all deer blood samples, selected mouse blood samples and cell culture aliquots, and one postmortem tissue sample were subjected to DNA sequencing. Gene fragments were purified by gel electrophoresis, and bands were extracted and purified with the QIAquick gel extraction kit (QIAGEN, Inc.). DNA was sequenced in forward and reverse directions at the Molecular Genetics Instrumentation Facility, The University of Georgia, with an ABI 3100 automated sequencer (Applied Biosystems, Perkin Elmer Corp., Foster City, CA). The sequences were assembled and edited using the Sequencher software package, version 4.1.4 (Gene Codes Corp., Ann Arbor, MI). A nucleotide-nucleotide BLAST (blastn) search was performed to determine the most similar sequences of the target genes published in GenBank (http://www.ncbi.nlm.nih.gov/).
Due to incidental molecular detection of Bartonella spp. in deer blood cultures during the course of the experiment, cell culture samples were subjected to RT-PCR for the gltA gene of Bartonella sp. using primers CS140f and BhCS1137n (9). To increase sensitivity, a heminested RT-PCR was developed for the detection of Bartonella spp. directly from peripheral blood, using primer set CS140f and BhCS1137n in the primary reaction and Bh731p and BhCS1137n (49) in the secondary reaction. Reagent concentrations and reaction conditions were identical to those used for 16S rRNA A. phagocytophilum RT-nPCR.
Tissues collected for RT-nPCR included spleen, prescapular and prefemoral lymph nodes, bone marrow from the sternum and femoral head, and lung. The aforementioned tissues as well as heart, liver, kidney, adrenal gland, brain, bladder, haired skin, reproductive organs, and gastrointestinal tract also were collected in 10% neutral buffered formalin for histopathologic examination.
The animal infectivity of the A. phagocytophilum inoculum was confirmed via RT-nPCR and DNA sequencing of a subset of products obtained from blood samples taken 6 days p.i. from seven of nine inoculated mice by using both the 16S and p44 gene targets (data not shown). Furthermore, rare morulae were observed in granulocytes of blood taken 6 days p.i. from RT-nPCR-positive mice. Although postmortem spleen samples for the previously mentioned seven mice were positive for A. phagocytophilum by 16S RT-nPCR, the organism was not reisolated in cell culture from terminal blood pooled from these mice at 11, 15, and 20 days p.i. Negative RT-nPCR results of blood and spleen samples were obtained for mice inoculated with uninfected tick cells.
All experimental deer developed reciprocal antibody titers of ≥64 to A. phagocytophilum between 14 and 24 days p.i. (Table (Table1).1). Two experimental deer (WTD133 and WTD134) seroconverted by 17 days p.i. and remained seropositive throughout the 66-day study. The peak reciprocal titer of 2,048 was detected in one of these deer (WTD134) on 38 days p.i. The other two experimental deer (WTD131 and WTD132) had peak reciprocal titers of 128 and were seropositive only through 38 days p.i. The geometric mean (57) of all titers of ≥64 was 115. Anaplasma phagocytophilum IFA results for the control deer were negative (<64) on all sample dates.
Blood from all four experimental deer was positive for A. phagocytophilum by 16S RT-nPCR on one to five occasions between 6 and 38 days p.i. (Table (Table1).1). All 16S rRNA gene amplicons from deer blood were sequenced and found to be identical to the sequence of the original inoculum and 99.5% similar to a published sequence of A. phagocytophilum (GenBank accession number U02521). Blood from WTD132 and WTD133 was positive for A. phagocytophilum by groESL RT-nPCR on 13 and 17 days p.i. Sequenced groESL products were 100% identical to published sequences of A. phagocytophilum (GenBank accession numbers AY219849 and U96728). Use of the p44 RT-PCR did not yield A. phagocytophilum amplicons from deer blood. Blood from the two negative control deer (WTD127 and -139) was negative for A. phagocytophilum by 16S and groESL RT-nPCR and by p44 RT-PCR at 6, 10, 13, 17, 24, and 31 days p.i.
Anaplasma phagocytophilum was not reisolated in any of 36 culture attempts from experimental deer (Table (Table1).1). All 10 and 45 days p.i. cultures failed when tick cells did not reform a monolayer in the flask after the deer blood culture procedure was completed. Of 29 remaining culture attempts of A. phagocytophilum-inoculated deer, 23 cultures exhibited CPE, and intracellular bacteria were observed in Giemsa-stained cytospins. Although RT-nPCR of these cell culture samples with primer pairs ECC-ECB and GE9F-GA1UR yielded products of the expected size, DNA sequencing revealed that they were not 16S rRNA gene fragments of A. phagocytophilum, but rather of Bartonella sp. (see below).
Control deer blood cultures and experimental deer blood cultures not exhibiting CPE from 6, 17, 24, 31, and 38 days p.i. were lost due to a single bacterial contamination event involving use of a commercially purchased component of the tick cell media at 45, 34, 27, 20, and 13 days in culture (DIC), respectively. For the same reason, the three mouse blood cultures from 11, 15 and 20 days p.i. were lost at 40, 36, and 31 DIC, respectively.
Clinical signs attributable to infection with A. phagocytophilum were not apparent in any deer throughout the 66-day study. Morulae were not observed in granulocytes on Giemsa-stained deer blood smears. Although results of complete blood counts were not always within normal limits for all fawns on all sampling dates, no consistent pattern of hematologic abnormalities attributable to infection with A. phagocytophilum was apparent.
At four sampling dates between 30 and 55 days p.i., WTD133 exhibited low platelet counts (48 × 103 to 168 × 103/μl) relative to its own values before and after that time period and relative to all platelet counts of the other experimental deer and the negative control deer (mean = 685 ×103/μl). During this time period, WTD133 showed no other signs consistent with thrombocytopenia.
Gross and histopathologic lesions attributable to infection with A. phagocytophilum were not apparent in any deer. RT-nPCR assays of postmortem samples of femoral and sternal bone marrow, prescapular and prefemoral lymph node, spleen, and lung of experimental and control deer were negative for A. phagocytophilum, with the exception of a sequence-confirmed 16S rRNA gene amplicon from femoral bone marrow of WTD133.
Sequencing of 16S rRNA gene products amplified with primer sets ECC-ECB and GE9F-GA1UR from a randomly chosen subset (7 of 24) of 16S RT-nPCR-positive cell culture samples revealed that the products were 98.6% similar to a sequence of Bartonella schoenbuchensis (GenBank accession number AJ278190.1), an intraerythrocytic bacteria first isolated from the blood of wild roe deer (Capreolus capreolus) in Germany (18). Subsequently, cell culture samples were subjected to RT-nPCR for the citrate synthase (gltA) gene of Bartonella spp., and 21 of 24 samples yielded products of the expected size. Furthermore, heminested gltA RT-PCR for Bartonella spp. developed in this study for screening whole blood yielded products of the expected size from preinoculation blood of two deer, WTD134 and WTD127. DNA sequencing and alignment indicated that these two gltA fragments had 98.6% and 94.6% sequence similarity, respectively, to a sequence of B. schoenbuchensis (GenBank accession number AJ564633) recently isolated from the midgut of a deer ked (Lipoptena cervi). One 16S rRNA gene sequence of the Bartonella sp. amplified from tick cell culture and the gltA DNA sequences of the Bartonella sp. amplified from the blood of WTD134 and WTD127 are available under GenBank accession numbers AY805111, AY805109, and AY805110, respectively.
We demonstrated that WTD can support infection with a human-infective strain of A. phagocytophilum and that infection of WTD with A. phagocytophilum is accurately reflected by seroconversion. Circulation of A. phagocytophilum in peripheral blood of experimental deer was transient, similar to mice and horses experimentally infected with various North American isolates of the organism (2, 27, 50, 51). Anaplasma phagocytophilum RNA was infrequently detectable in peripheral blood of experimental deer after development of antibodies, suggesting that the humoral immune system may have played a role in bacterial clearance. Although the function of humoral immunity in host elimination of rickettsiae is not well understood, Winslow et al. (61) demonstrated that antibodies can affect the course of active infection of E. chaffeensis in SCID mice. Administration of immune sera 10 and 17 days after infection resulted in partial clearance of E. chaffeensis from infected mice; however, the organism eventually recolonized the liver. This indicated that antibodies failed to mediate complete bacterial clearance. Subsequently, the authors hypothesized that E. chaffeensis may have persisted at low levels in the liver or emigrated from tissues that were inaccessible to the antibodies.
In the present study, A. phagocytophilum was detected by RT-nPCR in the peripheral blood of one deer (WTD131) at 38 days p.i., following a 3-week period of negative RT-nPCR results. The occurrence of recrudescent rickettsemia has been documented in a 9-month study of WTD experimentally infected with E. chaffeensis (16). Recrudescent rickettsemia may result from release of organisms sequestered in tissue, but this hypothesis requires further evaluation with regard to A. phagocytophilum in WTD. The detection of A. phagocytophilum RNA in femoral bone marrow from WTD133 on 66 days p.i. suggests a potential site of latent infection.
While A. phagocytophilum was not visualized in granulocytes of experimental deer, the difficulty of finding and definitively identifying morulae in blood smears of humans with confirmed A. phagocytophilum infections is well known (5, 58), particularly in blood samples taken from afebrile patients (1, 4). Furthermore, studies demonstrate that blood of experimentally infected laboratory mice, although morula negative, was still infectious to naïve mice and sometimes PCR positive and culture positive as well (26, 53, 54). Thus, our negative light microscopy results for the deer are consistent with those of several previous experimental infection studies of mice.
Demonstration of gene transcription by the use of RT-nPCR is suggestive of A. phagocytophilum survival and replication within the deer (21). Therefore, our 16S RT-nPCR data imply that viable A. phagocytophilum circulated until at least 17 days p.i. in three of the four deer. Although limited, this time period is of sufficient duration hypothetically to infect ticks. Nonetheless, we contend it is unlikely that WTD play an epidemiologically significant role as a source of A. phagocytophilum for ticks. This conclusion is based on the fact that WTD are parasitized primarily by the adult forms of I. scapularis (30) and therefore are most likely to be exposed to A. phagocytophilum at the end of the tick life cycle. Furthermore, A. phagocytophilum is not known to be maintained transovarially in the tick.
Anaplasma phagocytophilum infection initiated by a single needle inoculation may differ substantially from infection naturally acquired by the bite of one or more infected ticks over the course of a season. In addition to the likelihood that wild WTD experience multiple exposures to A. phagocytophilum, various immunologically active components of tick saliva may be important factors influencing the outcome of natural exposures of WTD to A. phagocytophilum (25).
We succeeded in reisolating A. phagocytophilum from a mouse inoculated with tick cell culture in the development phase of our research (unpublished data); however, we were unsuccessful in three attempts to reisolate A. phagoctyophilum from mice inoculated concurrently with deer. Relatively few studies have attempted culture of blood from laboratory mice; some have also reported discrepancies between cell culture and PCR results (26, 53).
With regard to our attempts to reisolate A. phagocytophilum from experimental deer, we encountered significant difficulties related to three separate issues. First, on two culture days (10 and 45 days p.i.), tick cells failed to reattach after the monolayer was disrupted and the cells were admixed with deer buffy coat cells. On these days, cellular debris created during resuspension may have resulted in cytotoxicity to the tick cells, as reported previously (64). Alternatively, the deer buffy coat cells may have killed the tick cells (U. G. Munderloh, unpublished data). Second, although the experimental design specified monitoring all cultures not developing CPE for 60 days, bacterial contamination of a commercially purchased component of the tick cell media resulted in the loss of many cultures before the end of 60 days. Because all “negative” cultures initiated previous to the contamination event on 51 days p.i. were destroyed, “negative” culture results for WTD127, -132, -133, and -139 are equivocal. Third, for at least 21 isolation attempts, the presence of Bartonella spp. confounded cell culture. For example, our best opportunities to reisolate A. phagocytophilum would have been days when deer were RT-nPCR positive, as was the case at 6 and 17 days p.i. (WTD131, -132, and -133) and on 38 days p.i. (WTD131 and -134); however, of these eight culture opportunities, Bartonella spp. were isolated in all but two. Bartonella spp. replicated rapidly in deer blood cultures; a CPE was apparent to the unaided eye as early as 7 DIC and was nearly 100% between 8 and 23 DIC. Perhaps the vigorous growth of Bartonella sp. in the tick cell medium resulted in conditions unsuitable for the survival of A. phagocytophilum. If this is the case, future attempts to culture A. phagocytophilum from the blood of deer coinfected with Bartonella spp. may be facilitated by the use of an antibiotic to which the former is resistant but to which the latter may be susceptible, such as erythromycin (7, 29).
In the present study, 16S RT-nPCR was the most sensitive assay for detection of A. phagocytophilum RNA in deer blood, followed by the groESL RT-nPCR assay. Because both of these assays are nested, we expected them to be more sensitive than the nonnested p44 RT-PCR. Massung and Slater (41) demonstrated that specificity and sensitivity vary markedly among the numerous published PCR assays and primer pairs for A. phagocytophilum. With regard to specificity, our cell culture results illustrate that even in a controlled experimental setting, PCR-based methods of detection should be confirmed by use of alternative gene targets or by DNA sequencing. We were aware of the existence of an undescribed Anaplasma sp. of WTD that is amplified in 16S PCR assays by using primers GE9F and GA1UR (34). Additionally, it has been reported that in blood samples containing a high concentration of Anaplasma (Ehrlichia) platys DNA, E. equi primers have induced false priming (52). However, we were not aware of a Bartonella sp. infecting WTD in the southeastern United States until we sequenced RT-nPCR products amplified with primers GE9F-GA1UR from tick cell culture of deer blood. Of interest is the fact that although the deer were apparently coinfected with A. phagocytophilum and Bartonella sp., we never detected Bartonella spp. directly from deer blood with the aforementioned primers. Therefore, we believe that the copy number of Bartonella spp. circulating in peripheral blood of the deer was very low. In fact, in retrospective testing of all deer blood samples collected during this study, we were unable to detect Bartonella spp. directly in blood by a single-step gltA RT-PCR assay using primers cited in a standard PCR assay for amplification of DNA from cervid isolates of Bartonella spp. from Europe and the western United States (12, 18). Only after development of a heminested gltA RT-PCR were we able to detect RNA of Bartonella spp. in preinoculation blood samples from two of the deer.
Our serologic and molecular findings lend support to the premise that WTD should be suitable sentinels for human risk of exposure to A. phagocytophilum (8, 33, 39, 40, 60). Because two experimental deer maintained detectable antibodies (titer, ≥64) for at least 49 days through the end of the 66-day study, we suggest that wild WTD repeatedly exposed to A. phagocytophilum may exhibit relatively long-lasting serologic response, a desirable trait in a potential sentinel species. Together with existing field data, our experimental findings, including use of a 1:64 dilution for serologic screening, lay the foundation for the development and validation of an A. phagocytophilum sentinel system using WTD. Although serologic cross-reactivity between E. chaffeensis and A. phagocytophilum in humans has been previously reported (15, 55), recent work suggests that this phenomenon is not a significant limitation to the use of WTD as sentinels for A. phagoctyophilum and E. chaffeensis when surveillance data sets are validated by confirmatory tests such as immunoblotting, PCR, and culture (60, 63).
Recently, Massung et al. (42) reported a genetic variant of A. phagocytophilum (AP-variant 1) from wild WTD in Wisconsin and Maryland and from I. scapularis in Rhode Island and Connecticut. Because this genovariant was not infectious for mice, the authors hypothesized that AP-variant 1 may be specific to WTD and may cycle independently of the human-infective strain of A. phagocytophilum (AP-ha) that is maintained in white-footed mice. If proven, this hypothesis would have implications for the use of WTD as A. phagocytophilum sentinels. Our serologic and RT-nPCR findings confirm that WTD are susceptible to a human-infective strain of A. phagocytophilum by needle inoculation, suggesting that both genetic variants, AP-ha and AP-variant 1, could be present in wild WTD. Future research related to A. phagoctyophilum infection among WTD should include identifying the genovariants present in deer on a broad geographic scale and determining the infection dynamics of simultaneous or sequential A. phagocytophilum genovariants in WTD, as recently reported for E. chaffeensis (56, 62).
This work was supported primarily by the National Institutes of Allergy and Infectious Diseases (5 R01 AI044235-02). Further support was provided by the Federal Aid to Wildlife Restoration Act (50 Stat. 917) and through sponsorship of the fish and wildlife agencies of Alabama, Arkansas, Florida, Georgia, Kansas, Kentucky, Louisiana, Maryland, Mississippi, Missouri, North Carolina, Oklahoma, Puerto Rico, South Carolina, Tennessee, Virginia, and West Virginia.
We thank Andrea Varela for assistance with acquiring and prescreening deer, Andrea Davidson for assistance during mouse inoculation, Chris King and Nat Seney for support regarding animal management, Jeff Tucker for animal handling expertise, and Molly Murphy for phylogenetic input during this project.