Hemangiosarcoma (HSA), a locally invasive and highly metastatic endothelial cell neoplasm, accounts for two-thirds of all cardiac and splenic neoplasms in dogs. Bartonella spp. infection has been reported in association with neoplastic and non-neoplastic vasoproliferative lesions in animals and humans.

The objective of this study was to determine the prevalence of Bartonella spp. in conjunction with two other hemotropic pathogens, Babesia spp. and hemotropic Mycoplasma spp., in tissues and blood samples from 110 dogs with histopathologically diagnosed HSA from throughout the United States.

This was a retrospective, observational study using clinical specimens from 110 dogs with HSA banked by the biospecimen repository of the Canine Comparative Oncology and Genomics Consortium. Samples provided for this study from each dog included:fresh frozen HSA tumor tissue (available from n = 100 of the 110 dogs), fresh frozen non-tumor tissue (n = 104), and whole blood and serum samples (n = 108 and 107 respectively). . Blood and tissues were tested by qPCR for Bartonella, hemotropic Mycoplasma, and Babesia spp. DNA; serum was tested for Bartonella spp. antibodies.

Bartonella spp. DNA was amplified and sequenced from 73% of dogs with HSA (80/110). In contrast, hemotropic Mycoplasma spp. DNA was amplified from a significantly smaller proportion (5%, p<0.0001) and Babesia spp. DNA was not amplified from any dog. Of the 100 HSA tumor samples submitted, 34% were Bartonella PCR positive (32% of splenic tumors, 57% of cardiac tumors, and 17% of other tumor locations). Of 104 non-tumor tissues, 63% were Bartonella PCR positive (56% of spleen samples, 93% of cardiac samples, and 63% of skin/subcutaneous samples). Of dogs with Bartonella positive HSA tumor, 76% were also positive in non-tumor tissue. Bartonella spp. DNA was not PCR amplified from whole blood.

This study documented a high prevalence of Bartonella spp. DNA in dogs with HSA from geographically diverse regions of the United States. While 73% of all tissue samples from these dogs were PCR positive for Bartonella DNA, none of the blood samples were, indicating that whole blood samples do not reflect tissue presence of this pathogen. Future studies are needed to further investigate the role of Bartonella spp. in the development of HSA.

Study design and sample sources

This was a retrospective, observational, descriptive study of 110 dogs with HSA. Specimens used for this study were previously collected and banked by the Canine Comparative Oncology and Genomics Consortium (CCOGC) based on previously published standard operating procedures.[34]. Briefly, the CCOGC collected samples from eight participating veterinary university teaching hospitals starting in 2006 with the goal of creating a repository of clinical samples from dogs with common naturally occurring cancers. Prior to sample submission to CCOGC, dogs were given a definitive diagnosis of neoplasia based on histopathology performed by a board-certified veterinary pathologist at the diagnostic laboratory of each participating university. Tissue samples for the CCOGC biospecimen repository were obtained from the primary tumor via surgical biopsy or post-mortem collection (HSA tumor tissue). As an internal control, tissue samples from each dog were also obtained from adjacent grossly normal tissue in the same organ, or if no grossly normal tissue in the affected organ was apparent, skin biopsies were obtained (non-tumor tissue). Tissue submission (HSA tumor and non-tumor) from each dog required provision of both formalin-fixed (subsequently stored in ethanol) and non-fixed specimens snap-frozen in liquid nitrogen (subsequently stored at -80° C). Dogs also had serum and whole blood collected at the time of tissue sampling for submission to the CCOGC.

For this study, samples from dogs diagnosed with HSA were provided by the CCOGC repository to the investigators. Four sample types were provided for pathogen testing: fresh frozen HSA tumor tissue, fresh frozen non-tumor tissue, whole blood, and serum. Two sample types were provided for histopathological confirmation of tissue type and tumor presence included: formalin-fixed HSA tumor tissue and formalin-fixed non-tumor tissue. Any dog that had a diagnosis of HSA (as determined by the CCOGC SOP), and had an adequate amount of fresh frozen tissue stored in the biorepository at the time of the investigators’ request for specimens, was included. This study, therefore, included 110 dogs with samples collected between May 2008 and November 2011.

Because of previous sample requests or lack of submission of all requested samples to the CCOGC, there were samples missing when provided to the investigators. Of the 110 dogs, 91 had a complete set of the 4 sample types for pathogen testing (fresh frozen HSA tumor tissue, fresh frozen non-tumor tissue, whole blood, serum) provided to the investigators. No dog was missing more than one sample type, and all dogs had at least one fresh-frozen tissue sample (HSA tumor, non-tumor tissue, or both) available for testing. For this reason, dogs with missing samples for pathogen testing were not excluded from the study. Similarly, of the 110 dogs only 37 had a complete set of the 2 sample types for histopathological confirmation (formalin-fixed HSA tumor tissue, formalin-fixed non-tumor tissue) provided to the investigators. Of the 110 dogs, 37 had formalin-fixed HSA tumor tissue and 93 had formalin-fixed non-tumor tissue provided. There were no formalin-fixed samples of HSA tumor or non-tumor tissue from cardiac tumors available for independent histopathologic review. Because a large proportion of dogs were missing formalin-fixed samples for independent histopathological confirmation, the subgroups of tissues with independent histopathologic confirmation of tissue type and tumor presence was analyzed separately.

The CCOGC provided demographic information for each dog, including age (years), breed, weight (kg) and sex and neuter status. The date and geographic location (university teaching hospital) of sample collection was also provided. The anatomic location of tumor and non-tumor tissue samples for each dog was also provided.

Independent confirmation of demographic and histologic data

When possible, information provided by the CCOGC was independently confirmed by the investigators for this study. Histopathology reports providing the diagnosis of hemangiosarcoma wwere provided for 102 dogs; these were reviewed by one author (EL) to confirm hemangiosarcoma diagnosis and tumor location.

For dogs with formalin-fixed biopsy samples provided (see above), biopsies were independently evaluated by a board-certified veterinary pathologist (KL) to confirm the tissue and tumor origin of the biopsy. Formalin fixed tissues were submitted to the NCSU College of Veterinary Medicine histology laboratory (Raleigh NC) and tissues were embedded in paraffin blocks (FFPE blocks). Slides containing 5 um sections were prepared from the FFPE blocks and stained with hematoxylin and eosin (H&E). Samples were categorized by organ type and HSA tumor or non-tumor tissue based on H&E staining. If the tissue of origin was not able to be determined from the biopsy (5 tumor samples and 2 non-tumor tissues), or if no formalin-fixed sample was provided (68 tumor samples and 14 non-tumor tissues), the tissue of origin was categorized by the information provided by the CCOGC and the original diagnostic histopathology reports. Non-tumor tissues from skin or various subcutaneous tissues (including hair, skin, adipose, skeletal muscle, or mammary gland, or any combination of these tissues) were categorized as skin/SQ for analysis.

Pathogen detection methods

DNA was extracted from EDTA anti-coagulated blood and fresh frozen tissue samples using a Qiagen DNeasy® Blood and Tissue kit (Qiagen, Valencia, CA) following the manufacturer’s protocols. For each tissue sample (HSA tumor and non-tumor), a 25 mg piece of tissue was excised from the entire sample using a new, prepackaged sterile scalpel. DNA yield and quality was assessed by spectrophotometry (Nanodrop, Wilmington, DE).

Each DNA sample was screened for the presence of Bartonella spp. DNA using conventional and qPCR, and Babesia spp. and hemotropic Mycoplasma spp. using qPCR. Bartonella qPCR was performed using primers targeting the 16S-23S intragenic transcribed spacer (ITS) region of Bartonella species as described previously[35], in conjunction with a BsppITS438 FAM-labeled hydrolysis probe (TaqMan, Applied Biosystems, Foster City, CA, USA). Bartonella qPCR was performed at two dilutions for each sample (using 1 uL and 5 uL of template DNA respectively). PCR screening for Babesia spp. and hemotropic Mycoplasma spp. were carried out as described previously.[22] Briefly, oligonucleotides Myco16S-322s (5’ GCCCATATTCCTACGGGAAGCAGCAGT 3’) and Myco16S-938as (5’ CTCCACCACTTGTTCAGGTCCCCGTC 3’) were used as forward and reverse primers respectively for hemotropic Mycoplasma spp. DNA amplification. Oligonucleotides Piro18S-144s (5’ GATAACCGTGSTAATTSTAGGGCTAATACATG 3’) and Piroplasma18S-722as (5’ GAATGCCCCCAACCGTTCCTATTAAC 3’) were used as forward and reverse primers respectively for Babesia spp. DNA amplification.

Amplification was performed in a 25-µl final volume reaction containing 12.5 µl of MyTaq Premix (Bioline), 0.2 µl of 100 µM of each forward primer, reverse primer (IDT® DNA Technology), 7.1 µl of molecular–grade water, and 5 µl of DNA from each sample tested.  PCR negative controls were prepared using 5 µl of DNA from blood of a healthy dog. Positive controls for PCR were prepared by using 5 µl of DNA from previously characterized positive dog (clinical cases).  Conventional PCR was performed in an Eppendorf Mastercycler EPgradient® under the following conditions: a single hot-start cycle at 95°C for 2 minutes followed by 55 cycles of denaturing at 94°C for 15 seconds, annealing at 68°C for 15 seconds, and extension at 72°C for 18 seconds.  Amplification was completed by an additional cycle at 72°C for 1 minute, and products were analyzed by 2% agarose gel electrophoresis with detection using ethidium bromide under ultraviolet light.

Validation of positive results was performed by Sanger sequencing of amplicons followed by chromatogram evaluation and sequence alignment using Contig-Express and Align X software (Vector NTI Suite 10.1, Invitrogen Corp, CA, USA). For bacterial species identification, DNA sequences were analyzed for nucleotide sequence homology at NCBI nucleotide database using BLAST version 2.0. A sample was considered Bartonella spp. PCR positive if one or more PCR tests (qPCR or conventional PCR) were positive (tests run in parallel). Stringent processing methods were used to avoid DNA carryover during tissue processing.[36] Specifically, tissue samples were processed independently using manual DNA extraction. For all batches of DNA extractions, between 2 and 4 blanks samples (water) were used as negative controls.  All negative controls for DNA extractions rendered negative results on all PCR assays. DNA carryover after PCR amplification was avoided by processing each sample in three separate laboratory rooms (one for sample sorting and DNA extraction, a second for PCR processing, and a third for PCR analysis post amplification), and strict use of personal protective equipment for sample handling by laboratory personnel.

For IFA testing, Bartonella antibodies were determined using 3 cell culture grown Bartonella spp. (Bartonella henselae, Bartonella vinsonii subsp. berkhoffii, and Bartonella koehlerae) as antigens and following standard immunofluorescent antibody assay (IFA) techniques.[37,38] Briefly, bacterial colony isolates were passed from agar plate grown cultures into permissive cell lines.  For each antigen, heavily infected cell cultures were spotted onto 30-well Teflon-coated slides (Cel-Line/Thermo Scientific), air-dried, acetone-fixed, and stored frozen. Fluorescein conjugated goat anti-dog IgG (KPL, SeraCare, Milford MA) was used to detect bacteria within cells using a fluorescent microscope (Carl Zeiss Microscopy, LLC, Thornwood NY). Serum samples were diluted in phosphate-buffered saline (PBS) solution containing normal goat serum, Tween-20, and powdered nonfat dry milk to block nonspecific antigen binding sites. Sera were first screened at dilutions of 1:16 to 1:64. All sera that are reactive at 1:64 were further tested with two-fold dilutions to 1:8192.

Dog_ID: A unique anonymized identifier for each dog

Collection_Site: The veterinary hospital at which the dog was enrolled

CensusRegion: Classification of the Collection_Site into a census region as follows. Pacific, Pac (WA, OR, CA); Mountain (ID, NV, MT, WY, UT, CO, AZ, NM); West North Central, WNC (ND, SD, NE, KS, MN, IA, MO); West South Central, WSC (OK, AR, TX, LA); East North Central, ENC (WI, IL, IN, OH, MI); East South Central, ESC (KY, TN, MS, AL); South Atlantic (MD, DE, WV, VA, NC, SC, GA, FL, DC); Middle Atlantic, MA (NY, PA, NJ); New England, NE (ME, NH, VT, MA, CT)

Collection_Date: Date dog was enrolled

Collection_Season: Classification of the Collection_Date into a season as follows. summer (Jun, July, Aug), autumn (Sept, Oct, Nov), winter (Dec, Jan, Feb), spring (Mar, Apr, May)

Breed: Dog breed reported by veterinary hospital

Age_yr: Dog age in years, rounded to integer

Sex: Dog sex and neuter status as follows. Male intact, MI; Male castrated, MC; Female intact, FI; Female spayed, FS

Weight_kg: Dog weight in kilograms

Tumor_Organ: The anatomic location of the HSA tumor submitted to CCOGC. "EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review.

NonTumor_Organ: The anatomic location of the non-tumor tissue submitted to CCOGC. "EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review. "MISSING" indicates tissue was not available from CCOGC.

Bart_IFA: Bartonella IFA result. A dog was considered positive if seroreactivity was 1:64 or greater to one or more Bartonella spp. antigen. 0 = negative, 1 = positive. "MISSING" indicates serum was not available from CCOGC.

Any_Sample_Bart_PCR: Bartonella PCR result in any sample. A dog was considered positive if PCR was positive on any one or more sample (HSA tumor, non-tumor tissue, blood). 0 = negative, 1 = positive

Blood_Bart_PCR: Bartonella PCR result on blood sample. 0 = negative, 1 = positive. "MISSING" indicates blood was not available from CCOGC.

NonTumor_Bart_PCR: Bartonella PCR result on non-tumor tissue. 0 = negative, 1 = positive. EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review. "MISSING" indicates tissue was not available from CCOGC.

Tumor_Bart_PCR: Bartonella PCR result on HSA tumor tissue. 0 = negative, 1 = positive. "EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review.

Blood_Myco_PCR: Hemotropic Mycoplasma spp. PCR result on blood sample. 0 = negative, 1 = positive. "MISSING" indicates blood was not available from CCOGC.

NonTumor_Myco_PCR: Hemotropic Mycoplasma spp. PCR result on non-tumor tissue. 0 = negative, 1 = positive. EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review. "MISSING" indicates tissue was not available from CCOGC.

Tumor_Myco_PCR: Hemotropic Mycoplasma spp. PCR result on HSA tumor tissue. 0 = negative, 1 = positive. "EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review.

Blood_Bab_PCR: Babesia spp. PCR result on blood sample. 0 = negative, 1 = positive. "MISSING" indicates blood was not available from CCOGC.

NonTumor_Bab_PCR: Babesia spp. PCR result on non-tumor tissue. 0 = negative, 1 = positive. EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review. "MISSING" indicates tissue was not available from CCOGC.

Tumor_Bab_PCR: Babesia spp. PCR result on HSA tumor tissue. 0 = negative, 1 = positive. "EXCLUDE" indicates the tumor tissue submitted was not HSA tumor based on independent pathology review.

Tumor_HE_Organ: Anatomic origin of HSA tumor tissue, based on independent evaluation of formalin-fixed tissues by a single investigator (board-certified veterinary pathologist (KL)). "MISSING" indicates formalin-fixed tissue was not available from CCOGC.

Tumor_HE_Path: Pathology in HSA tumor tissue, based on independent evaluation of formalin-fixed tissues by a single investigator (board-certified veterinary pathologist (KL)). Pathology was confirmed as HSA (HSA, HSA equivocal); if pathology was not able to be confirmed (Undetermined) the sample was excluded from analysis. "MISSING" indicates formalin-fixed tissue was not available from CCOGC.

NonTumor_HE_Organ: Anatomic origin of non-tumor tissue, based on independent evaluation of formalin-fixed tissues by a single investigator (board-certified veterinary pathologist (KL)). "MISSING" indicates formalin-fixed tissue was not available from CCOGC.

NonTumor_HE_Path: Pathology in non-tumor tissue, based on independent evaluation of formalin-fixed tissues by a single investigator (board-certified veterinary pathologist (KL)). Lack of pathology was confirmed (Normal); if tumor pathology was seen (HSA) the sample was included in the HSA tumor tissue group. If there was pathology that was not HSA (Undetermined), the sample was included in the non-tumor tissue group. "MISSING" indicates formalin-fixed tissue was not available from CCOGC.