Dr McNiel is presently affiliated with College of Veterinary Medicine, Michigan State University, East Lansing MI 48824. Presented in part at the 25th Annual American College of Veterinary Internal Medicine (ACVIM) Forum, Seattle, WA, 2007. The study was performed at The University of Minnesota Veterinary Medical Center, Saint Paul, Minnesota.
Corresponding author: Elizabeth McNiel, Department of Small Animal Clinical Sciences, Michigan State University, East Lansing, MI 48824; e-mail: firstname.lastname@example.org.
Background: The accumulation of frame-shift mutations in microsatellites (MS), termed microsatellite instability (MSI), is associated with certain tumors. MSI and its detection in urine samples has been used to aid in the detection of human bladder cancer.
Hypothesis: Evaluation of MSI in urine is a useful assay test for diagnosis of transitional cell carcinoma (TCC) in dogs and is more specific than the commercially available, veterinary bladder tumor analyte (V-BTA) test.
Methods: Prospective observational study. Urine samples collected from each animal were evaluated for MSI and using the V-BTA. For MSI detection, 22 MS sequences were polymerase chain reaction amplified from urine and blood, subjected to capillary electrophoresis, and the MS genotypes were compared. Aberration in ≥15% of MS was considered indicative of MSI.
Results: MSI was detected in 11 of 23 (48%) urine samples from dogs with TCC. MSI was also detected in 12 of 50 (24%) of the control animals, including 29, 16, and 24% of healthy, proteinuric, and lower urinary disease dogs, respectively. In this population, sensitivity and specificity of MSI analysis was 48 and 76%, respectively, compared with 83 and 64%, respectively, for the V-BTA test.
Conclusions: MS analysis as performed in this study is not useful in the diagnosis of TCC.
Transitional cell carcinoma (TCC) of the lower urinary tract accounts for 1–2% of all malignancies in dogs.1,2 Clinical signs and laboratory parameters in dogs with TCC are nonspecific and virtually indistinguishable from benign lower urinary tract diseases (LUTD), which often delays diagnosis. At the time of diagnosis, most TCC have extensive bladder wall invasion, and up to 40% show evidence of metastasis3 resulting in guarded to poor prognosis.
Neoplastic cells are identified by cytologic evaluation of the urine in only approximately 30% of dogs with TCC.3 The veterinary bladder tumor analyte (V-BTA) test is a commercially available, urine based latex agglutination assay for the detection of TCC. The sensitivity and specificity of the V-BTA test are reportedly 85–90 and 94%, respectively, when comparing normal and tumor bearing dogs.4,5 Glucosuria, proteinuria, pyuria, bacteriuria, and hematuria can all cause false positive test results.5,6 When dogs with non-TCC associated LUTD are included in the test population, the specificity of the V-BTA test decreases to 35–41%,4,5 which limits the clinical utility of this test. There is still a need for a specific, noninvasive diagnostic test for TCC in dogs with signs of LUTD.
Microsatellites (MS) are short tandem repeats, varying from 1 to 6 DNA bases, that are scattered throughout the genome, primarily in noncoding regions.7,8 Accumulation of mutations within MS is termed microsatellite instability (MSI) and can promote tumorigenesis. MSI can be detected in urine samples by comparing DNA from exfoliated urothelial cells to DNA from blood (lymphocytes) of the same individual. While MS abnormalities can develop in nonneoplastic cells, the identification of MS abnormalities is usually restricted to tumor tissue where there is clonal expansion of cells containing aberrant MS loci. Based on results of studies in humans,9–11 we hypothesized that the evaluation of MSI in urine samples in dogs would be a useful diagnostic tool for the noninvasive detection of TCC that would be more specific than the V-BTA test. The objectives of this study were (1) to determine the sensitivity and specificity of MSI as a diagnostic test for TCC by measuring MSI in dogs with TCC, in dogs with non-TCC associated urinary tract disease and in healthy dogs and (2) to compare the sensitivity and specificity of MSI with the V-BTA test in the same population of dogs.
Materials and Methods
Dogs enrolled in this prospective clinical study were evaluated by the University of Minnesota Veterinary Medical Center. All procedures were approved by the Institutional Animal Care and Use Committee and informed consent was obtained from each owner before enrollment.
Dogs were recruited if they met the inclusion criteria for any 1 of 4 groups defined below. Dogs were excluded from the study if they had been treated with antibiotics in the 3 days before presentation in order to accurately determine if a urinary tract infection was present, or were <5 years of age (in order to acquire age-appropriate controls for the TCC group).
Group 1: Healthy controls. This group included dogs with no clinical signs of LUTD (hematuria, stranguria, or pollakiuria), normal urinalysis results, and no growth on aerobic bacterial urine culture. Dogs of varying age (range 5.1–14.9 years; median = 8.9 years), sex (9 spayed females [FS], 11 neutered males [MN], and 1 intact male [MI]), and breed (3 Scottish Terriers, 3 West Highland White Terriers, 2 Greyhounds, 2 mixed breeds, and 1 each of Shetland Sheepdog, Border Collie, Boston Terrier, Cairn Terrier, Cocker Spaniel, Dalmatian, Fox Terrier, Rottweiler, Doberman Pinscher, Corgi, and Labrador Retriever) were included.
Group 2: Proteinuric dogs. Dogs were enrolled in this group based on the presence of an inactive urine sediment (negative for urine casts, crystals, red or white blood cells), no growth on aerobic bacterial urine culture and susceptibility, and 1+ or greater result for protein on the sulfosalicylic acid precipitation test. A urine protein to creatinine ratio was performed on all animals. Dogs of varying age (range 5.5–11.7 years; median = 8.7 years), sex (10 FS, 2 MN), and breed (1 each of West Highland White Terrier, Australian Shepherd, Collie, Coonhound, Giant Schnauzer, Jack Russell Terrier, Labrador Retriever, Australian Cattle Dog, Manchester Terrier, German Short Haired Pointer, Bearded Collie, and Cocker Spaniel mix) were included.
Group 3: LUTD not associated with TCC. Dogs were enrolled in this group if they presented for signs of LUTD and either had growth on aerobic bacterial urine culture and susceptibility or had the presence of inflammatory cells or bacteria in the urine. The lower urinary tract was evaluated by ultrasound or cystoscopy in all dogs in this group. Dogs of varying age (range 5–14.1 years; median = 9.3 years), sex (10 FS, 1 intact female, 7 MN, 1 MI), and breed (3 Labrador Retrievers, 2 English Springer Spaniels, 2 Miniature Schnauzers, and 1 each of Scottish Terrier, American Bulldog, Pug, Vizsla, Weimaraner, Rottweiler, Jack Russell Terrier, Cocker Spaniel, German Shepherd, mixed breed, and Golden Retriever) were included.
Group 4: TCC of the bladder, prostate, and urethra. Dogs were initially enrolled in the TCC group if they had a history of lower urinary tract signs and had evidence of a bladder or prostatic mass on ultrasonography or cystoscopy. This group included dogs with TCC of the bladder, prostate, and/or urethra diagnosed by fine needle aspiration cytology or histologic evaluation of biopsies obtained during cystoscopy, ultrasound, surgically, or via incisional biopsy after euthanasia. Sample size was estimated based on the hypothesis that determination of MSI would be more specific than the V-BTA test in detecting TCC among a population of dogs with both TCC and non-TCC associated LUTD. Ages ranged from 6.2 to 17.2 years (mean 11.11 ± 2.18 years; median = 11.25 years). Eleven female spayed and 12 male neutered dogs of varying breed (3 English Setters, 4 mixed breed, 2 Shetland Sheepdogs, 2 Golden Retrievers, and 1 each of Scottish Terrier, West Highland White Terrier, Beagle mix, Pomeranian, Cocker Spaniel, Miniature Schnauzer, Shih Tzu, Siberian Husky, Alaskan Malamute mix, American Water Spaniel, Border Collie mix, Newfoundland) were represented.
Three milliliters of blood from each dog was collected into a 7.5% EDTA containing tubea and refrigerated at 4°C. Fifteen to twenty milliliters of urine was collected by cystocentesis, transurethral catheterization, or midstream free catch. Immediately after collection, a urinalysis and an aerobic bacterial urine culture and susceptibility were performed on all samples through the University of Minnesota Veterinary Diagnostic Laboratory. Remaining urine was stored at 4°C. Urine cytospinb preparations of all samples were made and stored within 3 hours of collection. Determination of tumor type was achieved by histopathology or cytology by a board certified pathologist or clinical pathologist.
All abdominal ultrasounds were performed by a board certified veterinary radiologist. Fine needle aspiration of suspected tumors was performed only at the time of ultrasound and only if the extent of the mass suggested that surgical resection would not be possible and pet owners were aware of potential complications associated with the procedure.
The animals were assigned to different groups based on clinical signs, urinalysis results, cytology, or histology. The V-BTAc test was then performed on all urine samples within 24 hours of collection of urine by 1 investigator (A.J.S.) in accordance with manufacturer's instructions. Tests were evaluated at 30 seconds, 1 minute, and every minute thereafter for a total of 5 minutes. The test was compared with the manufacturer's color schematic and recorded as either positive or negative. The positive and negative controls included by the manufacturer in the test kit were evaluated with each V-BTA test. Positive results were retested with the same urine sample. To ensure accurate classification of controls, an abdominal ultrasound was recommended if not already performed and samples were positive on the 2nd V-BTA test, regardless of group assignment.
Blood and urine were stored at 4°C within 10 minutes of collection, and DNA was isolated within 48 hours. DNA from all dogs was isolated from peripheral blood lymphocytes and nucleated cells exfoliated in urine with a commercially available DNA extraction kitd in accordance with manufacturer's instructions. For isolation of DNA from urine, the samples were centrifuged at 2,000 ×g and the cell pellet was used for DNA isolation. Extracted DNA was stored at −80°C.
With the exception of DNA source and extraction, detection of MSI was performed as described previously 12 and was based on a panel of 22 MS regions chosen to include various repeat motifs including di-, mono-, and tetranucleotide repeats representing a wide range of genomic loci on 15 chromosomes (Table 1). MS sequences of matched blood and urine DNA samples were fluorescently labeled and amplified via polymerase chain reaction (PCR) by a 3-primer technique. PCR was performed in 15 μL reactions including 12.5 ng of DNA, 1.5 mM MgCl2, 1 μM of the forward primer, and 0.3 μM of the reverse primer with tail, 156 nM of a fluorescent dye-labeled primer complementary to the tail sequence, 300 μM of each dNTPe and 1.5 U Taq Polymerasef This reaction was heated to 95°C for 20 minutes, then submitted to 40 cycles consisting of 1 minute at 94°C, 1 minute at 56°C, 2 minutes at 72°C, and a final extension step of 10 minutes at 72°C. PCR products were evaluated by polyacrilamide capillary electrophoresis on the CEQ 8000 Genetic Analysis System.g Chromatograms were compared and scored for the presence of instability by the system software. Instability at a locus was defined as the addition or deletion of alleles in the tumor genotype when compared with the genotype in blood. Samples demonstrating aberrations in at least 15% of MS markers were considered MS unstable. A 2nd set of paired samples (urine and blood) was collected and MSI analysis repeated if either of the following occurred: (1) the analysis showed MSI in at least 15% of loci or (2) if PCR for multiple markers failed.
Table 1. Canine microsatellite panel used for evaluating MSI in bladder tumors.
Genomic location indicates the canine chromosome on which the marker resides; repeat type indicates the repetitive motif for the locus, Tetra or Di.
For sample size calculations, the following assumptions were used: (1) α= 0.05, (2) β= 0.2 (80% power), (3) the specificity of the V-BTA test in a population containing dogs with lower urinary tract signs would be roughly 40%, and (4) to be a clinically useful diagnostic test, we would need MSI evaluation to have a specificity of at least 85% in this population. Based on these assumptions, we estimated a need for a minimum of 16 animals with TCC and non-TCC LUTD. We also required collection of at least 10 animals that were defined as healthy and 10 animals with proteinuria to evaluate the performance of MSI evaluation and V-BTA testing in these conditions.
Fisher's exact test was used to test for associations between V-BTA test results and diagnosis of TCC, MSI results and diagnosis of TCC, and V-BTA and MSI test results. The Mann-Whitney test was used to compare the age distribution between control dogs and dogs with TCC, the age distribution between dogs with and without detectable MSI, and urine specific gravity between dogs with and without detectable MSI. A Fisher's exact test was used to compare the distribution of males and females between TCC and controls and to test for an association between sex and the detection of MSI. Statistical significance was set at P < .05. The sensitivity, specificity, and positive and negative predictive values were calculated for both the V-BTA and MSI evaluation using diagnosis of TCC by either cytological or histological means as the reference standard.
A total of 73 dogs were enrolled in the 4 groups.
Group 1 comprised healthy controls (n= 21). Urine was collected by cystocentesis in 20/21 (95%) and transurethral catheterization in 1/21 (5%) of dogs. Abdominal ultrasonography was performed in 9/21 (43%) dogs in this group. The urine specific gravity ranged from 1.003 to 1.060 (median 1.023).
Group 2 comprised proteinuric animals (n= 12). Urine protein to creatinine ratios ranged from 3.5 to 20.1 (median = 7.1). The urine specific gravity ranged from 1.006 to 1.036 (median 1.013). Urine was collected by cystocentesis in all dogs and 9/12 (75%) had abdominal ultrasonography performed.
Group 3 consisted of dogs with non-TCC associated LUTD (n= 17). Urine was collected by cystocentesis in 14/17 (82%) of animals. Urine was collected by urinary catheter in 3/17 (18%) of animals. An aerobic bacterial urine culture was performed in 16/17 (94%) of animals. Of these 16 animals, 6/16 (38%) had negative cultures and 10/16 (63%) had growth of different species (3 nonhaemolytic Escherichia coli, 2 Staphylococcus intermedius, 1 Staphylococcus sp., 1 Corynebacterium, 1 colony-like Mycoplasma sp., 1 mixed infection with a nonhaemolytic Escherichia coli and a Klebsiella sp., 1 mixed infection with Enterococcus sp. and Klebsiella sp.). The 1/17 (6%) that did not have a urine culture performed had been previously diagnosed with hemorrhagic cystitis from cyclophosphamide use and had a negative urine culture at that time. In addition to the 10 dogs with culture-confirmed bacterial urinary tract infection, 4 dogs had cystourolithiasis, 1 had prostatitis, and 1 had bacteriuria on urine sediment examination with a negative culture. The urine specific gravity ranged from 1.004 to 1.055 (median 1.026). Evaluation of the urinary tract was performed by ultrasonography in 14/17 (82%) and by cystoscopy in 1/17 (6%) of animals. No evidence of a bladder mass was present on either ultrasonography or cystoscopy.
Group 4 consisted of dogs with TCC of the bladder, prostate, and/or urethra (n= 23). Initially, 27 dogs were included based on clinical evidence of a bladder mass and signs of LUTD. The bladder mass or prostate gland were sampled in each animal to confirm the diagnosis of TCC either cytologically or histologically (9/27 by fine needle aspirate; 7/27 at necropsy; 3/27 via endoscopic pinch biopsies obtained at cystoscopy; 5/27 via biopsies obtained at the time of ultrasound [3/5 via trucut needle biopsies of the prostate gland and 2/5 biopsies via endoscopic cup forceps]; 2/27 biopsies obtained surgically; and 1/27 via incisional biopsy of a bladder mass after euthanasia). Determination of tumor type was achieved by histopathology in 18/27 (67%) and by cytology in 9/27 (33%). Animals with lower urinary tract neoplasia complicated by a urinary tract infection were also included. Urine was collected by cystocentesis in 21/27 (78%) of animals, by transurethral catheterization in 4/27 (15%) of animals, and by midstream free catch in 2/27 (7%) of animals. All animals had an aerobic bacterial urine culture performed. Of these animals, 22/27 (82%) had >0–5 WBC/high-power field (hpf), 21/27 (78%) had no growth, and 6/27 (22%) had growth of various bacterial species (2 colony morphology like Mycoplasma sp., 1 each of Serratia marcescens, Proteus mirabilis, Enterococcus sp., and 1 combination of S. intermedius and Streptococcus sp.). The urine specific gravity ranged from 1.006 to 1.055 (median 1.026). Evaluation of the bladder was performed by ultrasonography alone in 23/27 (85%) and by cystoscopy alone in the remaining 4/27 (15%).
Four dogs were subsequently excluded from statistical analysis based upon histopathology results, which were inconclusive in 2 dogs and demonstrated rhabdomyosarcoma in 1 dog and histiocytic sarcoma in the remaining dog.
Dogs with TCC were significantly older than control dogs (P= .0038). However, there was no difference between the TCC group and control dogs with respect to the distribution of males and females or with respect to urine specific gravity.
V-BTA Test and MSI Analysis
In the course of performing the V-BTA test, certain samples produced ambiguous test results when manufacturer recommendations were followed. These samples, from animals in all groups, tested positive initially (at 30 seconds) but then converted to a negative result within 3 minutes; mean = 70 seconds ± 51 seconds; median = 60 seconds. Ambiguous V-BTA test results were found in groups 1, 3, and 4 (group 1 n= 6 dogs, group 3 n= 1 dog, and group 4 n= 3 dogs). A total of 10/ 79 (13%) tested positive at 30 seconds and later read as negative on the same test strip. Because of this variation, the V-BTA test result at 5 minutes was used in statistical calculations.
There was variability in MSI percentage and V-BTA test results among all 4 groups (Fig 1). The percentage of MSI across all groups ranged from 0 to 80%. Nineteen of 23 (83%) dogs diagnosed with TCC tested positive on the V-BTA test whereas 4/23 (17%) of animals diagnosed with TCC tested negative. Eighteen of 50 animals in the 3 control groups (including 2/21 [10%] healthy, 7/12 [58%] proteinuric, and 9/17 [53%] with LUTD) tested positive with the V-BTA test, whereas the remaining dogs tested negative.
Unfortunately, PCR amplification of MS was not successful in all samples. There were 15 of 73 samples (21%) from which MS markers could not be successfully amplified. Most of these dogs were from either the healthy group (n= 6) or the proteinuric group (n= 5) with fewer cases among the LUTD (n= 1) and TCC groups (n= 3). Of the 20 dogs diagnosed with TCC that were evaluable, no MS aberrations were observed in 6 (30%), whereas aberrant MS were observed in 14 (70%) dogs. Eleven of 20 (55%) TCC dogs were classified as demonstrating MSI (aberrations in at least 15% of MS markers), while 9/20 (45%) dogs demonstrated relative MS stability (aberration <15% of MS markers). In the 3 control groups (N= 38 evaluable), at least 15% aberrant MS were identified in 12 of 38 (32%) dogs including 6/15 (40%) healthy dogs, 2/7 (29%) proteinuric dogs, and 4/16 (25%) dogs with LUTD. Figure 2 provides examples of the identification of unstable MS markers urine samples of apparently healthy animals.
Samples failing MS PCR were excluded from calculation of sensitivity and specificity of both the V-BTA test and MSI detection. For the diagnosis of TCC, the sensitivity of MSI detection was low and only marginally more specific than the V-BTA test (Table 2). Consistent with previous reports, the V-BTA test demonstrated reasonable sensitivity but low specificity (Table 2). Although the V-BTA test was significantly associated with the diagnosis of TCC classification, using MSI was not. Furthermore, there was no association between the 2 assays (P= .1309). None of the other variables examined including age, sex, or urine specific gravity were significantly associated with detection of MSI.
Table 2. Sensitivity and specificity for the V-BTA test and MSI analysis in the diagnosis of canine TCC.
Several human studies evaluating alterations in MS sequences or MSI in upper urinary tract and bladder cancer have shown that MS alterations are detectable in both the tumor and the urine sediment in a high percentage of patients.9,10,13–16 Using a panel of 22 MS, we detected MSI in 55% of urine samples from dogs with TCC, but also in 32% of control dogs. MS sequences failed to amplify from urine samples in 21% of study dogs and PCR failure was more common in samples with an inactive sediment. Eleven of 15 of the failed samples were from healthy dogs or proteinuric dogs. Poor DNA recovery from urine samples without significant cellular content is likely to be a major contributing factor. Twelve of 23 (52%) of dogs in the TCC group did not demonstrate a positive test result, and only 3 of these were attributed to PCR failure. In its current form, this assay would be expected to perform best in animals in which LUTD is suspected; however, measurement of MSI in urine did not prove to be sensitive for identification of TCC in this study.
Some animals in all groups had MSI regardless of the presence or absence of bladder tumors. There are a number of possible explanations for the identification of MSI in urine samples from animals that were not diagnosed with neoplasms. First, this could represent DNA contamination, although our results were repeatable on fresh samples from the same animals. Another possibility is that there is decreased mismatch repair in the canine bladder or in the bladders of some dogs. Although it is possible that the identification of MSI in a urine sample from a healthy dog is merely coincidental and has no pathologic significance, it is also possible that it is associated with an increased risk for the development of TCC.
The population of dogs with TCC in this study showed similarities with previously published reports with respect to presenting complaints (hematuria, pollakiuria, and strangiuria) and age distribution (range 6.2–17.2 years; median 11.25 years).3,17–20 Inflammatory urine sediments3 and urinary tract infections20 are commonly associated with TCC. In our study, 18/23 (78%) of TCC dogs had inflammatory urine sediments (>0–5 WBC/hpf) but only 5/23 (22%) had positive aerobic urine cultures. This suggests that infection might be less common than reported previously and inflammation could be due solely to the presence of intraluminal (or intravesicular) neoplasia. Animals were excluded from enrollment if they had received antibiotics in the 3 days before presentation. It is possible that some of the dogs that had a negative aerobic urine culture would have cultured positive if retested several days later. This could have contributed to the low rate of concurrent urinary tract infections with TCC found in this study.
The V-BTA test is currently the only nonmorphological urine based diagnostic test for TCC in dogs. Numerous reports document interference because of glucosuria, proteinuria, pyuria, bacteriuria, and hematuria.4–6 In addition to these limitations, we also noted a variation in V-BTA test results depending on when the test was read. Manufacturer instructions suggest reading the test as early as 30 seconds and no longer than 5 minutes. Discrepancies with results over time could be due the diffusion of the urine on the V-TBA test strip, as this can often take as long as 5 minutes. Thus, a test that appears positive before the urine has completed its transit on the test strip (at 30 seconds) can be negative at 1 minute or 5 minutes or any time in between. The effect of scoring the tests at 5 minutes was a decrease in sensitivity and an increase in specificity. We sought to maximize the specificity of the V-BTA test for comparison with MSI. Therefore, the 5-minute result was used in statistical calculations. Clinicians using the V-BTA test to screen for TCC should be aware of this additional apparent test limitation.
MSI has not been used previously to detect canine TCC. Even though MSI was more frequently detected in TCC dogs compared with control groups, the difference was not statistically significant. Furthermore, the assay was insensitive and, in contrast to results from human studies, poorly specific for neoplasia (Table 2). Although refinement of the loci selected might improve the diagnostic utility of detecting MS abnormalities, the particular technique used in this study has little value as a diagnostic test for TCC in dogs. Further investigation is required to determine whether MSI represents an early preneoplastic change and therefore might identify dogs at risk of TCC or whether MSI is irrelevant as an indicator of urothelial carcinogenesis in animals
a Monoject, Tyco Healthcare, Mansfield, MA
b Cytopro Wescor Inc, Logan, UT
c V-BTA, Alidex Inc, Redmond, VA
d Purgene, Gentra Systems Inc, Minneapolis, MN
e Fischer Scientific, Pittsburgh, PA
f HotStarTaq, Qiagen, Valencia, CA
g Beckman coulter Inc, Fullerton, CA
The authors thank Jim Mickelson of the University of Minnesota for his technical assistance.
Supported by a Companion Animal Grant from the College of Veterinary Medicine, University of Minnesota, Saint Paul, MN.