Bartonella spp. Infection in Healthy and Sick Horses and Foals from the Southeastern United States
Corresponding author: E.B. Breitschwerdt, DVM, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 1060 William Moore Dr, Raleigh, NC 27607; e-mail: email@example.com.
Bartonella species bacteremia has been identified in numerous animal species. These bacteria cause, or have been associated with, a spectrum of clinical manifestations in dogs and human patients. The frequency of exposure to or infection with Bartonella spp. among healthy and sick horses has not been reported.
To test healthy and sick horses and sick foals from the southeastern United States for serological, microbiological, and molecular evidence of Bartonella infection.
Forty-seven healthy horses, 15 sick foals, 22 horses with musculoskeletal manifestations, and 8 horses with colic were tested for Bartonella.
IFA serology and PCR before and after BAPGM (Bartonella alpha-Proteobacteria Growth Medium) enrichment blood culture.
Bartonella antibodies were not detected in foals or horses. Three Bartonella species, B. henselae, B. vinsonii subsp. berkhoffii (genotypes I and III), and a Bartonella species with closest homology to Candidatus Bartonella volans, were PCR-amplified and sequenced from blood or BAPGM enrichment blood culture samples from 1/47 healthy horses, 3/15 sick foals, 5/22 horses with musculoskeletal disease, and 0/8 horses with colic.
Conclusions and Clinical Importance
Horses in the southeastern United States are naturally infected with B. henselae, B. vinsonii subsp. berkhofii genotypes I and III, and a bacteria most similar to Candidatus Bartonella volans. Antibodies were not detectable by indirect fluorescent antibody assay (IFA) testing in bacteremic foals or horses, and prolonged enrichment culture for periods up to 21 days were necessary to document bacteremia in most horses. Further investigation into the pathogenic potential of Bartonella spp. infection in horses is warranted.
Bartonella alpha-proteobacteria growth medium
Bartonella spp. are fastidious, intracellular, gram-negative aerobic bacteria with a worldwide distribution. Transmission occurs via bites, scratches, needle sticks, blood-feeding arthropods, and, for some Bartonella spp., transplacental transmission.[2-4] At least 26 different species or subspecies of this highly adaptive intravascular bacterium have been isolated from domestic and wild animals. Members of the genus Bartonella are gaining recognition as pathogens of substantial clinical importance in several animal species, but limited data are available regarding Bartonella spp. infection in horses. Therefore, the purpose of this study was to test a convenience sample of healthy horses, sick foals, and horses presented with signs of lameness or colic for serological and molecular evidence of Bartonella infection.
Bartonella henselae bacteremia was first described in 2 horses in 2008 when the organism was isolated from a 7-year-old mare with presumptive vasculitis, and DNA was PCR-amplified and sequenced from the blood of an 11-year-old gelding with chronic arthropathy. Subsequently, Johnson et al reported B. henselae infection in an aborted equine fetus. In that study, molecular, histological, and immunohistochemical techniques were used to confirm B. henselae in the liver, lung, and kidney of the fetus. Subsequently, B. henselae infection was documented in a 2-year-old mare with hemolytic anemia from Germany. Recently, B. vinsonii subsp. berkhoffii DNA was amplified from hemangiopericytoma tissues from a horse, dogs, and a red wolf (Canis lupus rufus), and the organism was shown to induce secretion of vascular endothelial growth factor. In the context of vector transmission, the cat flea (Ctenocephalides felis) is the primary vector for B. henselae transmission among cats, whereas ticks, lice, and biting flies are potential vectors for transmission of Bartonella spp. to other hosts.[2, 10]
Materials and Methods
Healthy horses from North Carolina (n = 40) and Virginia (n = 7) and sick foals or horses examined at the North Carolina State University Veterinary Teaching Hospital (NCSU-VTH) were tested for serological, microbiological, or molecular evidence of Bartonella infection. Between July 2008 and June 2010, blood samples were collected from 47 healthy pastured adult horses (age range, 2–25 years; median age, 11 years) residing on 8 farms with no concurrent clinical signs of illness. In addition, blood or discarded blood culture samples from 15 sick foals (age range, 1–90 days; median age, 2 days) admitted to the NCSU-VTH with clinical signs indicative of sepsis, diarrhea, general failure to thrive or some combination of these and blood from 22 lame adult horses (age range, 1.5–30 years; median age, 15 years) donated and housed at the NCSU Teaching Animal Unit or admitted to the NCSU-VTH due to laminitis, arthritis, joint swelling or some combination of this and 8 horses presented with signs of colic (age range, 2–24 years; median age, 11 years) were tested. For the sick foals and horses, antibiotic treatment before referral to NCSU-VTH was not an exclusion criterion for this exploratory study. As would be expected in a referral setting, antibiotics had been administered to some horses for varying periods before referral (the class of antibiotic, dosage, and duration was not consistently docu-mented in the medical record), whereas others had not been treated with antibiotics or there was no notation in the medical record regarding prior antibiotic administration.
For detection of Bartonella antibodies, an IFA was performed using B. henselae (Houston 1 strain) and B. vinsonii subsp. berkhoffii genotype I organisms grown in DH82 cells as antigens as previously described. Serum was available from 40 of 47 healthy horses, 7 of 15 foals, 21 of 22 lame horses, and from all 8 colic horses. Sera (kindly provided by Dr Bruno Chomel, University of California, Davis) from 3 horses experimentally infected with B. henselae (SA2 strain, isolate designation NCSU 2008-EO-1) were used as IFA positive controls. In our assay, the IFA endpoint titers for these 3 horse sera were 1 : 512, 1 : 512, and 1 : 2048. The starting dilution was 1 : 16, with endpoint titers defined as the last dilution at which brightly stained organisms could be detected by fluorescence. A cut-off titer for seroreactivity was arbitrarily defined as 1:64.
Bartonella alpha-Proteobacteria Growth Medium (BAPGM) Enrichment Blood Culture
BAPGM enrichment blood culture was performed as previously described[13, 14] with slight modifications, using 2 mL of aseptically collected EDTA or ACD anti-coagulated blood inoculated into 10 mL BAPGM and incubated for 14–21 days at 35°C with 5% CO2. At 7, 14, and, for some cases, 21 days post-enrichment culture, 500 μL was withdrawn from the BAPGM blood culture flask and sub-inoculated onto 10% sheep blood agar plates and incubated for at least 4 weeks using identical culture conditions. At the same time points, DNA was extracted from 200 μL of culture followed by PCR amplification using Bartonella primers, as described below. For 8 of the 15 sick foals that had previously tested culture-negative in the NCSU-VTH Clinical Microbiology Laboratory using the BD BACTEC1 system, DNA was extracted and BAPGM enrichment cultures (2 mL from the BACTEC blood bottle into 10 mL of BAPGM) were established.
Bartonella PCR Testing
Using DNA extracted from blood, serum, and BAPGM subcultures, Bartonella spp. PCR was performed targeting the 16S–23S ribosomal RNA intergenic spacer (ITS) region. Primers and PCR conditions were used as previously described.[11, 14] PCR amplicons were sequenced directly or after cloning using pGEM-T Easy Vector System2 as previously described.
B. henselae IFA antibodies were not detected in 7 foals and 67 horses. Bartonella spp. DNA was PCR-amplified and sequenced from blood, BACTEC blood culture, or BAPGM enrichment blood cultures from 9 of 92 equine samples (9.8%). One of 47 (2.1%) healthy horses (Horse 1), 3 of 15 (20%) sick foals (Foals 2–4), 5 of 22 (22.7%) horses with musculoskeletal diseases (Horses 5–9), and 0/8 horses presented for colic were infected with a Bartonella sp. With the exception of Horses 1 and 6, and Foal 3, from which Bartonella DNA was amplified from blood, enrichment blood culture was necessary to achieve a positive Bartonella PCR in the remaining 6 infected foals and horses. For Horse 6, B. vinsonii subsp. berkhoffii genotype I DNA initially was sequenced from an enrichment blood culture, whereas B. henselae Houston-1 DNA was amplifed from an extracted blood sample obtained 15 months later. No isolates were obtained from any horse after subculture.
DNA sequences obtained directly from the blood of Horse 1 (9-year-old Arabian mare with no clinical signs of illness) and from the 21-day BAPGM blood enrichment culture of Foal 2 (3-month-old Arabian/Quarter Horse crossbreed colt with septic arthritis of the right tibiotarsal and left stifle joints caused by Salmonella sp.) were most similar to B. henselae San Antonio-2 strain (GenBank Accession# AF369529). B. henselae Houston-1 strain (GenBank Accession# BA016SRB) DNA was amplified directly from the blood of both Foal 3 (30-hour-old Quarter Horse colt with Staphylococcal septicemia and enterocolitis caused by Clostridium perfringens) and Horse 6 (11-year-old Quarter Horse mare with idiopathic polysynovitis of the left stifle, right carpal, and right hock joints). Horse 5 (14-year-old Quarter Horse mare with chronic laminitis) was infected with a B. henselae Fizz strain (GenBank Accession# AF369526) or a B. henselae Cal-1 strain (GenBank Accession# AF369527), as the readable sequence from the 14-day BAPGM enrichment blood culture did not differentiate between these 2 strains. B. vinsonii subsp. berkhoffii genotype III DNA (GenBank Accession# DQ059764) was amplified from the 21-day BAPGM enrichment blood culture of Foal 4 (12-hour-old Holsteiner colt with meconium impaction), whereas B. vinsonii subsp. berkhoffii genotype I DNA (GenBank Accession# AF167988) was amplified and sequenced from the 14-day BAPGM enrichment blood culture of Horse 6, the 7-day BAPGM enrichment blood culture of Horse 7 (22-year-old Quarter Horse mare with laminitis and arthritis), and the 14-day BAPGM enrichment blood culture of Horse 8 (30-year-old Arabian mare with arthritis). The 3 B. vinsonii subsp. berkhoffii genotype I sequences (457 bp) were 98.9–99.3% identical to the reference sequence.
A unique Bartonella sequence was amplified from the 7-day BAPGM enrichment blood culture of Horse 9 (15-year-old Quarter Horse crossbreed mare with laminitis). That sequence was most similar to Candidatus Bartonella volans, an isolate obtained from a southern flying squirrel, Glaucomys volans (GenBank Accession# EU294521), and Bartonella DNA amplified from a squirrel flea Orchopeas howardi (GenBank Accession# DQ336386). Sequences from Horse 9, the flying squirrel isolate, and the squirrel flea all contained a 17 base pair deletion at the 3′ end. In addition, the Horse 9 sequence was similar to the following GenBank sequences: (1) 98.6% identity (357/362 bp) with an amplicon from the blood of an 86-year-old man with encephalopathy, memory loss, and recent onset arthritis, (2) 98.9% identity (360/364 bp) with an amplicon from a post-enrichment blood culture from a dog with a 2-year history that included muzzle histiocytomas, hock hamartomas, shifting leg lameness, painful joints, and lymphadenopathy, and (3) 98.9% identity (360/364 bp) with an amplicon obtained in our laboratory from a sea otter with vegetative valvular endocarditis.3
In this study, DNA representing different B. henselae strains, 2 different B. vinsonii subsp. berkhoffii genotypes, and a sequence most similar to Candidatus Bartonella volans was amplified from blood or enrichment blood cultures from healthy and sick horses or sick foals. Because Bartonella spp. infection was documented previously in an aborted foal, a horse, dogs, and human patients, with lameness, joint pain or polyarthritis, sick foals and lame horses were sampled preferentially in this study. Three previous reports described B. henselae infection in horses.[5-7] In this study, B. henselae was amplified and sequenced from 1 healthy horse, 2 sick foals, and 2 horses with musculoskeletal manifestations. Because 1 B. henselae-infected foal was diagnosed with Salmonella septic arthritis and a second with Staphylococcal septicemia and Clostridium enterocolitis, one might speculate that pre-existing Bartonella bacteremia predisposes foals to opportunistic infections.
Based upon the results of this study and a previous report describing B. henselae infection in an aborted equine fetus, it appears that foals can be infected in utero, during the perinatal period, or shortly after birth. If foals are infected by transplacental transmission, one might expect to detect Bartonella maternal antibodies, unless the mare was anergic, which seems plausible because antibodies were not detected in any bacteremic horse in this study. Alternatively, if infected shortly after birth, perhaps by vector transmission (eg, biting flies), then antibodies might not have achieved a detectable concentration at the time of hospitalization. Convalescent antibody titers should be determined in future studies.
Three horses with musculoskeletal manifestations were infected with B. vinsonii subsp. berkhoffii genotype I, a genotype that was isolated originally from a dog with presumptive polyarthritis, epistaxis and endocarditis from North Carolina. The mode(s) of transmission for B. vinsonii subsp. berkhoffii to any animal species is (are) currently unknown. B. vinsonii subsp. berkhoffii genotype I has been isolated from coyotes and dogs and has been detected in human blood, dog saliva, and, more recently, from a feral pig in North Carolina.[20-25] Also, B. vinsonii subsp. berkhoffii genotype I DNA was amplified from hemangiopericytoma tissue from 14-year-old Arabian horse located in Alabama. The origin and travel history of that horse were not reported. For the first time to our knowledge, B. vinsonii subsp. berkhoffii genotype III DNA was amplified from a 21-day BAPGM enrichment blood culture from a foal with meconium impaction. B. vinsonii subsp. berkhoffii genotype III has been isolated from gray foxes in the United States, detected in a human patient with endocarditis in Europe, and isolated from a military working dog that originated from Germany and developed endocarditis after transport to the United States.[24, 26, 27] B. vinsonii subsp. berkhoffii genotype I DNA was PCR-amplified and sequenced from Horse 6 in November 2007, whereas B. henselae DNA was amplified and sequenced from this horse in March 2009, suggesting that coinfection or sequential infection with more than 1 Bartonella sp. can occur in horses, as has been reported for dogs and people.[20, 21] Given the wide range of hosts that this subspecies can infect and the potential for chronic bacteremia, further investigation is warranted to determine the mode(s) of transmission for dogs, horses, humans, and other animals.
Based upon the IFA assays used in this study, serology was not useful in establishing prior exposure to or infection with a Bartonella sp. in foals or adult horses. A poor correlation between PCR results and IFA seroreactivity has been documented previously in dogs and humans with Bartonella sp. bacteremia.[13, 28-30] False negative B. henselae IFA results in human patients can result from antigenic variation among B. henselae strains used for IFA testing. Alternatively, bacteremic horses may be anergic, resulting in false negative IFA results, as has been suggested in bacteremic dogs and human patients with Bartonella bacteremia.
Documentation of infection with a Bartonella species in horses is difficult to achieve by serology, PCR amplification, or by standard blood culture isolation of the bacteria. With the exception of 3 samples from which Bartonella DNA was amplified directly from blood, BAPGM enrichment blood culture for incubation periods of 7, 14, or 21 days was necessary to confirm Bartonella sp. bacteremia in horses. This finding in naturally infected horses is consistent with results obtained after experimental infection of horses with B. henselae. Similar diagnostic trends also have been documented in studies involving sick dogs and human patients.[13, 31] Although Bartonella DNA was PCR-amplified directly from blood samples from 2 horses, the 7, 14, and 21-day BAPGM enrichment culture PCR results were negative. Because neither of the horses was treated with antibiotics before sample collection, lack of growth in culture most likely is due to decreased viability of the bacteria in these particular samples and a dilutional effect when a small number of viable bacteria is inoculated into a large volume of BAPGM enrichment media, thereby making less template DNA available for amplification.
No isolates were obtained after subinoculation from the enrichment culture onto sheep blood agar plates. To date, only 1 subculture isolate has been obtained from a horse. Because this remains a common microbiological limitation when working with Bartonella species, future studies are necessary to identify a reliable means to obtain isolates from dogs, horses, and human patients. Because ours was a descriptive study based upon testing of convenience samples, no comparisons among groups can be made. Also, larger populations of foals and adult horses with well-characterized illnesses or lameness diagnoses should be used in future studies to clarify any potential pathogenic role for Bartonella spp. in the respective disease processes.
In conclusion, serology did not prove useful in identifying horses or foals that had been exposed to or were infected with a Bartonella sp. The results of this study do not provide evidence to support a causal role in the diseases reported in the sick foals or horses, but it supports the need for further investigation into the possibility that these organisms contribute to disease in foals with underdeveloped immune systems and in adult horses with musculoskeletal disease.
We thank Julie Bradley for serological testing, Barbara Hegarty for preparation of diagnostic antigens, and Koren Barnes Kartes and Dr Don Meuten for assistance with sample collection.
Conflict of Interest: This research was supported in part by Novartis Animal Health by providing stipend support for Natalie Cherry during her PhD program, and by additional grants from the North Carolina Horse Council.
Becton, Dickinson and Company
Promega, Madison, WI
Carrasco SE, Chomel BB, Mazet JA, et al. Investigation of Bartonella spp. in Northern Sea Otters (Enhydra lutris kenyoni) with Vegetative Valvular Endocarditis in Alaska. Poster #280, 5th International Meeting on Rickettsiae and Rickettsial Diseases, Marseille, 2008.