• Open Access

Molecular Prevalence of Bartonella,Babesia, and Hemotropic Mycoplasma sp. in Dogs with Splenic Disease


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: ed_breitschwerdt@ncsu.edu.



Among diseases that cause splenomegaly in dogs, lymphoid nodular hyperplasia (LNH), splenic hemangiosarcoma (HSA), and fibrohistiocytic nodules (FHN) are common diagnoses. The spleen plays an important role in the immunologic control or elimination of vector-transmitted, blood-borne pathogens, including Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp.


To compare the prevalence of Bartonella sp.,Babesia sp., and hemotropic Mycoplasma sp. DNA in spleens from dogs with LNH, HSA, and FHN.

Materials and Methods

Paraffin-embedded, surgically obtained biopsy tissues from LNH (N = 50), HSA (N = 50), and FHN (N = 37) were collected from the anatomic pathology archives. Spleens from specific pathogen-free (SPF) dogs (N = 8) were used as controls. Bartonella sp., Babesia sp., and Mycoplasma sp. DNA was amplified by PCR, followed by DNA sequencing.


Bartonella sp. DNA was more prevalent in FHN (29.7%) and HSA (26%) as compared to LNH (10%) (P = .019, .0373, respectively) or control spleens (0.0%). The prevalence of Babesia sp. and hemotropic Mycoplasma sp. DNA was significantly lower than Bartonella sp. DNA in HSA (P = .0005, .006, respectively) and FHN (P = .003, .0004, respectively). There was no statistically significant difference in DNA prevalence among the 3 genera in the LNH group.


The higher prevalence of Bartonella sp. in FHN and HSA warrants future investigations to determine if this bacterium plays a role in the development of these splenic diseases.


fibrohistiocytic nodules


splenic hemangiosarcoma


lymphoid nodular hyperplasia


mucosa-associated lymphoid tissue


polymerase chain reaction


vascular endothelial growth factor

The spleen, a highly vascular organ, plays an important role in the immunologic control and elimination of blood-borne pathogens.[1] Splenic macrophages phagocytose damaged and senescent erythrocytes to facilitate elimination of extracellular and intracellular blood-borne bacteria, protozoa, and viruses through both innate and adaptive immune responses.[1] Splenic marginal zone macrophages express various receptors that mediate antimicrobial immunity, including pattern recognition receptors (Toll-like receptors),[2] C-type lectin receptors (SIGNR1), and type-I scavenger receptors (MARCO).[3] SIGNR1 recognizes polysaccharide antigens,[4] whereas MARCO recognizes bacteria, including Escherichia coli and Staphylococcus aureus.[5]

Researchers have generated substantial evidence to support a role for infectious agents, including bacteria (including mycoplasmas), viruses, and protozoa as a cause or cofactor in the development of cancer in animals and human patients. Currently, the World Health Organization indicates that infectious organisms are responsible for nearly 20% of human cancers.[6] As a bacterial example, Helicobacter pylori causes some cases of mucosa-associated lymphoid tissue (MALT) lymphoma, which is an antibiotic reversible “neoplastic” condition.[7, 8]

Bartonella sp. establishes persistent infection in erythrocytes, endothelial cells, and professional macrophages, leading to NF-κB activation, promotion of a pro-inflammatory phenotype, and recruitment of inflammatory cells, including neutrophils and macrophages.[9, 10] Unique among bacteria, Bartonella sp. can induce in vitro and in vivo proliferation of endothelial cells, leading to vasoproliferative lesions in both immunocompromised human beings and dogs.[11, 12] Although unproven, splenic tissues may provide a permissive environment for the growth and perpetuation of Bartonella sp., resulting in chronic, low-grade inflammation.

Babesia sp. and hemotropic Mycoplasma sp. are vector-transmitted blood-borne pathogens of dogs, in which the spleen influences the disease pathogenesis.[13, 14] Clinical presentations of canine babesiosis vary from asymptomatic to multiorgan failure and death.[15] The spleen is of central importance for the innate and adaptive immune response that controls Babesia sp. infection.[16] Infection with hemotropic Mycoplasma sp. has been reported in several mammalian species, including dogs and human patients.[14, 17] The most commonly reported hemotropic Mycoplasma sp. in dogs include Mycoplasma haemocanis and Mycoplasma haematoparvum, both of which have been associated with immune-mediated hemolytic anemia.[18] Splenectomy or other forms of immunosuppression also play an important role in the infectious pathogenesis of hemotropic Mycoplasma sp.[13]

In this study, we compared the prevalence of 3 vector-transmitted, intravascular, blood-borne pathogens (Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp.) in dogs with histologically confirmed splenic disease. We chose to study an angiogenic neoplastic condition (HSA), a nonangiogenic neoplastic condition (FHN), and a non-neoplastic splenic inflammatory condition (LNH). These lesions span a pathologic spectrum involving endothelial proliferation, lymphoid and fibrohistiocytic proliferation, and chronic lymphoid stimulation. Histologically normal splenic tissues from specific pathogen-free (SPF) dogs were used as tissue and laboratory processing controls.

Materials and Methods

Sample Collection

Using the NCSU-CVM Pathology Data Base, paraffin-embedded surgical biopsy samples from 3 groups of dogs were retrieved from pathology archive storage facilities. Group I included splenic FHN (n = 37), group II HSA (n = 50), and group III LNH (n = 50). Group IV consisted of histologically unremarkable spleens (n = 8) from SPF dogs. For cases with more than 1 block available, the block containing the largest component of well-preserved mass (approximately 50% or more of the entire section) was chosen. Fresh tissues were prospectively collected from group IV SPF dogs at the time of euthanasia and processed into paraffin blocks during the course of this study. Although not processed at the same time, control samples were processed and paraffin embedded in the histology laboratory with the same equipment and techniques as the samples from the 3 study groups. All paraffin blocks were coded and processed for PCR in a blinded manner. Archival tissues used in this study were collected between 2004 and 2010. Splenic tissue samples were independently reviewed by a pathologist to confirm the histopathologic diagnosis and the FHN grade. After DNA extraction, all paraffin-embedded splenic tissues were tested by PCR for the presence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA.

DNA Extraction

For each specimen, approximately 25 mg of tissue was manually excised from paraffin-embedded spleen, with a sterile disposable no. 10 scalpel blade. Tissues were processed in small batches and the work surface was thoroughly cleaned between each tissue block to avoid Bartonella sp., Babesia sp. or hemotropic Mycoplasma sp. DNA to carry over between the samples.[19] DNA was extracted with QIAamp DNA blood mini kit1 following manufacturer's instructions. A blank paraffin block was used as a reagent control with each set of DNA extractions. DNA concentrations and purity were determined with a spectrophotometer.2 Extracted DNA was stored at −20°C.

Polymerase Chain Reaction

Samples were tested for the presence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA,[20-22] utilizing the primers listed in Table 1. Bartonella genus primers targeted the 16S-23S rRNA intergenic transcribed spacer region (ITS 438s-1100as and ITS 325s-1100as) or pap31 (1s-688as) gene. Bartonella koehlerae ITS species primers were also used (Bk1s-1125as). A 180-bp region of the Babesia 18S rRNA gene was amplified by real-time PCR and primers B.caniss and Bcommonas. The Babesia sp. was determined using primers Bab455s and Bab772as followed by DNA sequencing. In the Mycoplasma sp., PCR was performed using the primers 322s and 938as targeting a 600-bp region of the 16S rRNA gene. The Mycoplasma sp. PCR primers were developed after alignment of twelve 16SrRNA gene sequences from hemotropic Mycoplasma species cited in GenBank. The Mycoplasma species used for the development of the hemotropic mycoplasma PCR primers included Mycoplasma coccoides mouse isolate (AY171918), M. haematoparvum (GQ129113), Mycoplasma haemobos cow isolate (EF616468), M. haemocanis (AY529641), Mycoplasma haemofelis cat isolate H83 (EF198144), M. haemofelis strain USP-24 (EU442639), Mycoplasma haemominutum (AM691834), Mycoplasma haemovis (EU828581), Mycoplasma hominis (NC_013511), Mycoplasma ovis (from human) (GU230144), Mycoplasma sp. from deer (AB558899), Mycoplasma turicensis cat isolate (EU789559), and Mycoplasma wenyonii isolate cattle (EU367964). Optimization of PCR conditions, including primer concentrations, PCR profile, and the type of enzyme employed in the assay, were performed using a 560-bp fragment of the 16SrRNA region of Candidatus Mycoplasma hematoparvum (obtained from naturally infected dog blood) cloned into the pGEM-T Easy vector. Under the conditions described in this article, the PCR assay detected 10 DNA copies/μL 100% of the time and 5 DNA copies/μL 78% of the time. The novel primers and PCR protocol described in this study were further validated by amplifying and sequencing M. haemocanis and Candidatus M. hematoparvum from infected dog blood samples and M. haemofelis from infected cat blood samples, obtained from diagnostic accessions to the NCSU-CVM Vector Borne Diseases Diagnostic Laboratory. Reaction conditions used for each PCR assay are given in Table 2. Reaction mixtures contained 12.5 μL of the Takara premix Ex Taq (perfect real time),3 7 μL of molecular grade water, 0.25 μL each of forward and reverse primers (30 μM), and 5 μL of template DNA. Amplified products were analyzed on a 2% agarose gel stained with ethidium bromide. DNA extracted from the blood of a healthy dog was used as a negative control. Positive controls included 0.001 pg/μL (equivalent to 2.5 genome copies per μL) of B. henselae DNA, M. hemocanis DNA from a naturally infected dog, and 0.01 pg/μL of B. gibsoni DNA. All PCR positive amplicons were sequenced directly or by cloning into pGEM-T Easy vector4 to establish the species, strain or genotype. Three independent clones were sequenced from each PCR positive sample.

Table 1. Bartonella,Babesia, and hemotropic Bycoplasma spp. PCR primers used in this study.
PrimerDNA Sequence (5′–3′)Target Gene
Pap31 1 sGAC TTC TGT TAT CGC TTT GAT TTBartonella sp. pap31
Pap31 688asCACCACCAGCAAMATAAGGCATBartonella sp. pap31
Myco 322sGCCCATATTCCTACGGGAAGCAGCAGTHemotropic Mycoplasma sp. 16S
Myco938asCTCCACCACTTGTTCAGGTCCCCGTCHemotropic Mycoplasma sp. 16S
Table 2. PCR conditions used in this study for the amplification of Bartonella, Babesia, and Mycoplasma spp. target genes. All the reactions were performed with an Eppendorf Mastercycler epgradient (Eppendorf, Westbury, NY), except for Babesia genus real-time PCR which was performed in a Bio-Rad CFX96 real time system (Bio-Rad Laboratories, Hercules, CA).
Target GeneReaction ConditionsNumber of Cycles
Bartonella sp. ITS94°C X 30 s, 68°C X 10 s, 72°C X 15 s, 72°C X 30 s55
B.koehlerae ITS94°C X 15 s, 66°C X 15 s, 72°C X 18 s, 72°C X 30 s55
Bartonella sp. Pap 3194°C X 15 s, 62°C X 15 s, 72°C X 18 s, 72°C X 1 min55
Hemotropic Mycoplasma 16S94°C X 15 s, 68°C X 10 s, 72°C X 15 s, 72°C X 30 s55
Babesia genus 18S95°C X 45 s, 58°C X 45 s, 72°C X 45 s, 72°C X 5 min49
Babesia sp. 445-77295°C X 45 s, 58°C X 45 s, 72°C X 45 s, 72°C X 5 min50

Statistical Analysis

Statistical analysis was accomplished by SAS 9.2 software. The prevalence of Bartonella sp., Babesia sp., and Mycoplasma sp. DNA was compared between and among the study groups by chi-squared test of independence and the Fisher's exact test. The level of significance was set as P < .05.


Study Animals

Group I included FHN biopsy tissues from 22 male (60%) and 13 female (35%) dogs. Nineteen breeds were represented, including Labrador Retrievers (16%), Terriers (11%), and American Cocker Spaniels (8%). Ages of the animals in this group ranged from 6 to 15 years, with 35% of the animals being >12 years of age. Sex, breed, and age were unknown for 2, 3, and 4 dogs, respectively.

Group II included HSA biopsy tissues from 32 male (64%) and 17 female (34%) dogs. Sex was unknown for 1 dog. Twenty-one different breeds were represented, including Labrador Retrievers (20%), Golden Retrievers (18%), and Boxers (8%). Ages ranged from 5 to 15 years, with 42% of the dogs ranging between 9 and 12 years of age at the time of diagnosis.

Group III included LNH biopsy tissues from 26 male (52%) and 20 female (40%) dogs. Nineteen breeds were represented in this group, including Labrador Retrievers (24%), Terriers (12%), and Poodles (8%). Ages ranged from 7 to 14 years with 42% of the dogs ranging between 9 and 12 years of age. Sex, breed, and age were unknown for 4, 4, and 5 dogs respectively.

Group IV included histologically unremarkable splenic tissues from 5 male and 3 female SPF dogs that were euthanized at the conclusion of unrelated studies. Seven dogs were Beagles, 1 was a Bloodhound, and ages ranged from 2 to 7 years.

PCR Analysis and Sequencing

The prevalence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA in each study group, as determined by PCR and DNA sequencing, is summarized in Figure 1. All extraction, PCR negative controls (including the blank paraffin block reagent control) tested negative throughout this study. PCR positive controls were positive on each PCR gel.

Figure 1.

Prevalence of Bartonella, Babesia, and Mycoplasma sp. in splenic biopsies from dogs with fibrohistiocytic nodules (FHN), hemangiosarcoma (HSA), lymphoid nodular hyperplasia (LNH), and histologically unremarkable spleens from specific pathogen-free dogs (SPF). Results are presented as percent positive. N = the number of dogs in each group. *P < .05.

Group 1, FHN

The prevalence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA was 29.7%, 2.7%, and 0.0%, respectively. Biopsies from FHN dogs were re-reviewed by a veterinary pathologist (KEL) and graded using previously published criteria.[23] There were 13 grade I, 14 grade II, and 10 grade III FHN. Prevalence of Bartonella sp. DNA was 23%, 42.9%, and 20% in grades I, II, and III FHN, respectively. Dogs were infected with B. henselae (SA2 strain) (n = 4), B. vinsonii subsp. berkhoffii genotype I (n = 2), and B. koehlerae (n = 2). One dog was coinfected with B. vinsonii subsp. berkhoffii genotype III, and B. henselae SA2 and another with B. vinsonii subsp. berkhoffii genotype III, and B. koehlerae. For 1 amplicon, the Bartonella sp. could not be established lacking adequate clean sequence. One B. henselae SA2 ITS-positive sample also was B. henselae SA2-positive based on the Pap31 gene sequence. The Bartonella sp. Pap31 gene was not amplified from any other FHN tissues. Babesia gibsoni was sequenced from the spleen of 1 dog, whereas no hemotropic Mycoplasma sp. DNA was amplified from group I splenic tissues.

Group II, HSA

The prevalence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA in HSA dogs was 26%, 2%, and 6% respectively. Dogs were infected with B. henselae (SA2 strain) (n = 7), B. koehlerae (n = 1), and B. vinsonii subsp. berkhoffii genotype III (n = 1). One dog was coinfected with 2 B. henselae ITS strains (SA2 and Houston I), a second with B. henselae SA2 and B. koehlerae, and a third with B. henselae SA2 and B. vinsonii subsp. berkhoffii genotype I. The DNA sequence from 1 amplicon had low similarity when compared with available GenBank Bartonella sequences (closest similarity was 36% to B. alsatica). Two B. henselae ITS SA2 strains were PCR positive for B. henselae SA2 pap31 gene. By 2 independent DNA extractions, 1 HSA tissue was real-time PCR positive for a Babesia sp., but attempts to determine the species by amplifying a larger fragment of the 18S rRNA gene were not successful. Mycoplasma ovis DNA was amplified and sequenced from 3 dogs.

Group III, LNH

The prevalence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. was 10%, 2%, and 0%, respectively in splenic tissues from dogs with LNH. Dogs were infected with B. koehlerae (n = 3) and B. henselae SA2 (n = 2). One LNH tissue was real-time PCR positive for Babesia sp., but multiple attempts to determine the species by amplifying a bigger fragment of the 18S rRNA gene were not successful.

Group IV Histologically Unremarkable Spleens

All of the splenic tissues from SPF dogs were PCR negative for Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. Blood cultured from SPF dogs with Bartonella alphaproteobacteria growth medium followed by Bartonella sp. PCR was negative.[24]Bartonella antibodies were not detected by indirect fluorescent antibody testing.[24]

Statistical Analysis

Because SPF dogs were PCR negative for all 3 targeted genera, group IV was not included in the statistical analysis. When the prevalence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA was compared by a chi-squared test of independence, there was a statistical difference among groups I, II, and III. Subsequently, each group was compared with each of the other 2 remaining groups by a Fisher's exact test. The prevalence of Bartonella sp. DNA was higher in group I (FHN) compared with group III (LNH) (P = .019) and higher in group II (HSA) as compared with group III (LNH) (P = .037), but there was no difference in the prevalence of Bartonella sp. DNA between the FHN and the HSA groups (P = .700). Prevalence of Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. was also compared among the 3 groups. Bartonella sp. infection was statistically higher than infection with Babesia sp. and hemotropic Mycoplasma sp. in the FHN group (P = .003 and .0004 respectively), whereas there was no significant difference in Babesia sp. and hemotropic Mycoplasma sp. (P = 1) prevalence. Bartonella sp. DNA also was statistically more prevalent in the spleen of HSA dogs when compared with infection with Babesia sp. and Mycoplasma sp. (P = .0005 and .006, respectively). There was no difference between the prevalence of Babesia sp. and hemotropic Mycoplasma sp. in the HSA group (P = .31). There were no statistical differences in prevalence among Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. in the LNH group (P > .05). There were no significant differences in prevalence among each of the 3 FHN grades (P > .05).


In this study, there was a higher prevalence of Bartonella sp. DNA in paraffin-embedded splenic biopsy tissues from dogs with HSA and FHN as compared with splenic tissues from dogs with LNH. In the context of angiogenesis, 3 Bartonella spp. have been associated with vasoproliferative tumor-like lesions in human patients, including B. bacilliformis (verruga peruana), B. quintana (bacillary angiomatosis), and B. henselae (bacillary angiomatosis and bacillary peliosis hepatis).[11, 25] In dogs, B. henselae was amplified from peliosis hepatis and B. vinsonii subsp. berkhoffii was isolated from an immunosuppressed dog with bacillary angiomatosis.[12, 26] Also, because B. vinsonii subsp. berkhoffii genotype II was isolated from a dog with a hemangiopericytoma, and from a human patient with epithelioid hemangioendothelioma, viable intravascular bacteria have been documented in dogs and people with vasoproliferative tumors.[27] Splenic rupture and granulomatous splenitis also have been associated with Bartonella sp. infection in people.[28, 29] Based on these preliminary findings, the detection of Bartonella sp. DNA in tissues from a subset of dogs with HSA was anticipated. Unexpectedly, however, there was no difference in the prevalence of Bartonella sp. DNA between HSA and FHN groups. Based on in vitro studies using B. henselae as a prototypical representative of the genus, viable intracellular infection has been documented in several macrophage-like cells, including monocytes, dendritic cells, and microglial cells.[30-32] Therefore, based upon previous studies, Bartonella sp. may localize within splenic histiocytes and endothelial cells, but DNA also could be amplified from erythrocytes circulating through the spleen at the time of sample collection. Cellular localization of Bartonella sp. within the spleen is an important next step, but efforts in our laboratory to localize the bacteria by Warthin–Starry silver staining, B. henselae immunohistochemistry, and in situ hybridization on paraffin-embedded splenic tissues were only marginally successful (Varanat, M; data not shown).

The cause of splenic FHN in dogs is unknown. Fibrohistiocytic cells are considered to be components of reticular meshwork in the splenic red pulp. These cells are intermixed with plasma cells, macrophages, and lymphocytes.[23] Once erythrocytes are infected with a Bartonella sp., the bacteria remain within the infected cell for the entire life span of the cell.[33] Senile erythrocytes are removed from the bloodstream by the filtering action of the reticular meshwork in the red pulp and then are phagocytized by the splenic macrophages.[1] Therefore, Bartonella sp. DNA amplified from FHN, HSA, and LNH groups could be related to phagocytized, dead bacteria in the splenic tissue. Alternatively, because Bartonella sp. can infect macrophages and prevent apoptosis,[30] it is plausible that chronic Bartonella infection of splenic histiocytes leads to proliferation and formation of FHN. Splenic nodules appear to form a continuum from lymphoid nodular hyperplasia (LNH) to malignant splenic stromal neoplasms (malignant fibrous histiocytoma).[23] In the context of histiocytic infection, Bartonella sp. DNA was lower in Grades I (23%) and III (20%) FHN as compared with Grade II (42.9%), but these differences were not statistically significant. As the overall prevalence of Bartonella sp. DNA in the FHN group (29.7%) was slightly higher than in HSA group (26%), prospective studies are required to determine if Bartonella sp. are a cause or cofactor in the progression of FHN in dogs.

Although the etiology and pathogenesis of canine HSA is incompletely understood, multifactorial, hypoxia, inflammation, and uncontrolled angiogenesis are considered to be important pathogenic factors.[34] Canine HSA can develop spontaneously in the absence of any known mutations and is not a heritable condition, although breed predilections have been reported.[35] Mutations of PTEN (phosphatase and tensin homolog) have been found in some canine HSAs[36] and alterations in the p16-cyclin D-Rb pathway also have been demonstrated.[37] In human beings, 25% of all adult cancers develop as a result of chronic inflammation with or without documentation of infection.[38] Canine HSA is characterized by the presence of inflammatory cells intermixed with the tumor cells. Macrophages and their cytokines promote tumor survival and facilitate metastasis.[38]Bartonella infection activates the NF-κB pathway leading to the recruitment of inflammatory cells including neutrophils, lymphocytes, and macrophages to the site of infection.[10] Hypoxia activates hypoxia inducible factor Iα (HIF-Iα), which is a transcription factor for many downstream genes that promote angiogenesis, tumor growth, and metastasis. Bartonella infection has been shown to activate the HIF-Iα resulting in increased expression of the vascular endothelial growth factor (VEGF).[39] Therefore, Bartonella spp. increase VEGF concentrations in conjunction with a direct mitogenic effect on endothelial cells, while also inhibiting endothelial cell apoptosis and thereby potentially contributing to the development of vasoproliferative tumors.

Infection by bacteriophages is another unique feature of several Bartonella spp. Bacteriophages are found in B. henselae, B. bacilliformis, and B. vinsonii subsp. berkhoffii.[40-42] Phage-associated chromosomal rearrangement has been postulated to play a crucial role in host specialization of B. quintana,[43] which appears to be a genomic derivative of the zoonotic pathogen B. henselae. In fact, the majority of genes found in B. henselae that are not found in B. quintana are located in 4 genetic clusters flanked by several phage integrases. This finding supports a potential role for bacteriophages as vehicles for B. henselae DNA rearrangements, which could prove important in promoting oncogenesis. However, based on PCR amplification of the Pap31 bacteriophage-associated gene, only 3 B. henselae SA2 strains had evidence supporting the presence of phages. Additional studies are required to investigate whether bacteriophages, inflammation, hypoxia or deregulated angiogenesis associated with chronic Bartonella infection contribute to HSA oncogenesis and metastasis.

We determined if Babesia sp. and Mycoplasma sp. DNA, representing other intravascular, vector-transmitted erythrocyte-associated pathogens that might also sequester within splenic histiocytes, were associated with FHN, HSA or LNH. The prevalence of Babesia sp. and Mycoplasma sp. DNA in the FHN and HSA groups was significantly lower than the prevalence of Bartonella sp. DNA, whereas there was no difference between Babesia sp. and Mycoplasma sp. prevalence among FHN, HSA or LNH groups. In the context of vector-borne intravascular pathogenic bacteria and protozoa, the spleen plays an important immunomodulatory role in controlling the infection, frequently referred to as infection immunity or premunition. Babesiosis occurs commonly in splenectomized individuals,[44] and in many cases, patients with occult babesiosis become symptomatic after splenectomy.[45] Splenic lesions also are a common finding in patients with systemic cat scratch disease caused by B. henselae infection. Splenic rupture was reported in a 65-year-old man due to B. henselae-induced splenitis[28] and persistent infection with B. bacilliformis was diagnosed in a chronically thrombocytopenic patient after splenectomy.[46] By real-time PCR targeting, an 180-bp segment of the 18S rRNA gene, amplification of Babesia sp. DNA was obtained from only 1 dog in each splenic pathology group. A 6-year-old Chihuahua with FHN was infected with B. gibsoni, whereas a 14-year-old Labrador Retriever with LNH and a 13-year-old Golden Retriever with HSA both were infected with an undetermined Babesia sp. Attempts to define the Babesia sp. by amplifying a larger gene segment were unsuccessful, potentially due to a low number of target organisms in splenic tissue samples or because of DNA degradation caused by formalin fixation.[47]

Mycoplasma ovis was amplified and sequenced from the spleen of 3 HSA dogs, whereas no Mycoplasma sp. DNA was amplified from the control, FHN or LNH groups. By the conventional Mycoplasma sp. PCR assay described in this study, Candidatus M. haematoparvum and M. haemocanis were amplified and sequenced from the blood of only 7 of 546 (1.3%) healthy and sick dogs located predominantly in the southeastern United States (Breitschwerdt et al unpublished data). We obtained no M. ovis-like sequences from those blood samples. Recently, coinfection with M. ovis and B. henselae SA2 was reported in a veterinarian with a progressive neurologic illness.[17] One of the 3 M. ovis infected dogs in this study was also coinfected with a B. henselae SA2 strain. Both B. henselae and potentially hemotropic Mycoplasma sp. are thought to be transmitted by the cat flea, Ctenocephalides felis.[48] In numerous animal species, Mycoplasma infections typically are characterized by a long period of latency. In vitro, mycoplasmas are able to transform normal cells to cancer cells.[49]M. genetalium and M. hyorhinis are capable of inducing malignant transformation of human epithelial cells and are thought to play a role in the development of various cancers, including prostate cancer and gastric carcinoma.[50, 51] We are not aware of a previous report of M. ovis infection in dogs or whether M. ovis might cause oncogenic transformation of mammalian cells. Amplification of M. ovis DNA from these tissues may represent DNA carryover during the tissue processing, but this seems unlikely because sheep tissues are rarely processed in our histopathology laboratory and each of these tissues was processed at different time points in the histopathology laboratory. The mode of infection and significance of M. ovis DNA in dogs with splenic HSA is unknown, but warrants further investigation.

There are several limitations of this retrospective study. Obtaining optimal age and sex-matched control tissues was not possible because splenectomy is performed only when there is overt evidence of splenic disease and splenic biopsies are infrequently obtained in the clinical setting. Presumably, the prevalence of Bartonella sp., Babesia sp., and Mycoplasma sp. DNA reported in this study represents an underestimate of the actual prevalence of these intravascular infections because of the small quantity of tissue placed in paraffin blocks from markedly enlarged spleens and the limited quantity of extracted DNA (host and pathogen) that can be incorporated into each PCR reaction. Varying periods of formalin fixation also could induce denaturation of DNA targets resulting in a false negative PCR result. Another factor that could have resulted in negative PCR results is a possible antibiotic treatment before obtaining the biopsy, which could have resulted in decreased bacterial load and a negative PCR result. Previously, our laboratory reported Bartonella sp. DNA carryover during the collection and processing of paraffin-embedded tissues.[19] In that study, Bartonella sp. DNA was amplified from the necropsy room, tissue processor, and from several microtomes. To minimize potential DNA carryover in this study, we only included surgical biopsy specimens. Individual sterile scalpel blades and a clean work area, not microtomes, were used to obtain tissue scrolls for DNA extraction. A blank paraffin block was used as a reagent control with each set of DNA extractions. Also, at multiple time points, using the same collection and extraction protocols used for splenic samples, we processed splenic tissues from a small number of SPF dogs with histologically unremarkable spleens. Bartonella sp., Babesia sp., and hemotropic Mycoplasma sp. DNA were not amplified from negative control tissues, decreasing concern for laboratory-related DNA cross contamination.

In conclusion, we found a significantly higher prevalence of Bartonella sp. DNA in splenic tissues from dogs with FHN and HSA when compared with LNH. Also, dogs with FHN and HSA were more likely to have Bartonella sp. DNA in splenic tissues, as compared with Babesia or Mycoplasma sp. DNA. Prospective studies are needed to isolate viable Bartonella organisms by enrichment blood culture in dogs with FHN and HSA, and to determine the cellular localization of the bacteria within splenic tissues. Also, performing Bartonella sp. PCR from FHN and HSA tissues obtained from different geographic locations may further establish a temporal and spatial association between Bartonella infection and these splenic diseases in dogs.


This study was funded by the American Kennel Club-Canine Health Foundation (AKC-CHF) Acorn Grant No.01231A. This work was conducted in the Intracellular Pathogens Laboratory, College of Veterinary Medicine, North Carolina State University, Raleigh, NC.

We thank Bayer Animal Health for stipend support to Mrudula Varanat, who is a PhD graduate student in the Intracellular Pathogens Laboratory, College of Veterinary Medicine, North Carolina State University. We thank Dr Maria Correa for the assistance with the statistical analysis and the Histology Laboratory for the assistance with acquisition of the paraffin blocks.

Conflict of Interest: This research was supported by the State of North Carolina, the American Kennel Club and in part through PhD graduate student stipend support provided by Bayer Animal Health to Dr Mrudula Varanat. None of the funding sources had any role in study design; data collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the manuscript for publication consideration. In conjunction with Dr Sushama Sontakke and North Carolina State University, Dr Breitschwerdt holds US Patent No. 7,115,385; Media and Methods for cultivation of microorganisms, which was issued October 3, 2006. He is the chief scientific officer for Galaxy Diagnostics, a newly formed company that will provide advanced diagnostic testing for the detection of Bartonella species infection in animals.


  1. 1

    Qiagen, Valencia, CA

  2. 2

    Nanodrop, Wilmington, DE

  3. 3

    Takara Bio USA Inc, Madison, WI

  4. 4

    Promega, Madison, WI