SEARCH

SEARCH BY CITATION

Keywords:

  • Cancer genes;
  • Molecular biology;
  • Oncology;
  • PCR assays;
  • Renal neoplasia;
  • Urinary tract

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

Background: Similarities in human and canine renal cell carcinoma (RCC) epidemiology and biologic behavior suggest that molecular mechanisms of tumorigenesis may be similar in both species. Approximately 75% of RCC in people are of the clear cell subtype, up to 85% of which are associated with mutation of the von Hippel-Lindau (VHL) gene. The canine VHL coding deoxyribonucleic acid (DNA) shares 90% identity with the human VHL gene.

Objective: To determine whether or not RCC in dogs are associated with VHL mutations, and if so determine the prevalence, type, and location of these mutations.

Animals: Thirteen dogs with RCC, 2 dogs with primary renal sarcomas, and 10 dogs without neoplastic kidney disease.

Methods: DNA was extracted from paraffin-embedded RCC tissue; DNA extracts from paraffin-embedded and snap-frozen nonneoplastic canine kidneys and canine whole blood were used as negative controls. Polymerase chain reaction and sequencing of the 3 VHL exons was performed, and results compared with the accessioned canine sequence.

Results: All VHL exons were amplified from 9 of 13 canine RCC samples, both renal sarcomas, 8 of 10 nonneoplastic kidney samples, and canine whole blood; only exon 2 could be amplified from 2 RCC samples. Mutations were not identified in any exons. A maximal prevalence of 33.6% for VHL mutations in canine RCC was determined.

Conclusion and Clinical Importance: Although similarities between canine and human RCC merit further investigation of the dog as a model for some subtypes of renal tumors, the lower prevalence of VHL mutations suggests that oncogenesis in these 2 species differs.

Abbreviations:
BHD

Birt-Hogg-Dubé

DNA

deoxyribonucleic acid

HIF

hypoxia-inducible factor

mRNA

messenger ribonucleic acid

PCR

polymerase chain reaction

RCC

renal cell carcinoma

TGF

transforming growth factor

VHL

von Hippel-Lindau

Renal tumors account for 0.6–1.7% of canine neoplasms, of which renal cell carcinomas (RCC) account for 60–85%.1–3 Affected dogs usually are older (mean age at diagnosis, 8 years), with males slightly overrepresented (1.6–1.8 : 1 male : female ratio).3 Owners of affected dogs most commonly seek veterinary care for nonspecific signs including lethargy, anorexia, and weight loss.2,3 Less common findings include hematuria, abdominal pain, a palpable abdominal mass, and clinical signs associated with renal failure. Diagnosis typically occurs late in the disease, with effacement of a majority of the affected kidney (equivalent to stage IV using human primary tumor size/extent-lymph node spread-distant metastasis [TNM] staging guidelines).4,5 Histologic variants of canine RCC include tubular, papillary, tubulopapillary, and undifferentiated forms, with the tubular subtype predominating.2,3,6 Subclassification using immunohistochemical markers has not been performed, and the clear cell variant most commonly diagnosed in people is seen uncommonly in dogs.6

RCCs in people are similar in epidemiology, biologic behavior, and disease course to RCC in dogs. Approximately 2.0% of all malignant tumors in people are RCCs.7 Men are more commonly affected, with increased prevalence of disease in geriatric patients.4 Typical clinical signs and physical examination findings at diagnosis include fatigue, weight loss, flank pain, hematuria, a palpable abdominal mass, or some combination of these findings.4 Prognosis is closely associated with TNM stage, with T3 stage (tumor confined to within Gerota's fascia) and below having 5-year survival rates >50%, whereas only 20% of patients with T4 stage tumors (tumor extension beyond Gerota's fascia) survive beyond 5 years.4,8,9

RCCs in people are subdivided into several histologic subtypes based on microscopic appearance, immunohistochemical markers, and presumptive cell origin.4,10 The clear cell variant, thought to arise from proximal convoluted tubular cells, accounts for approximately 75% of cases, and is associated with biallelic inactivation of the von Hippel-Lindau (VHL) gene in 50–85% of tumors.11–15 The best-characterized function of the VHL protein in normal cells is inactivation or downregulation of hypoxia-inducible factor (HIF) via targeting of the HIF-α subunit for proteasomal degradation.4,16,17 Absence of VHL protein results in increased HIF activity and thus increased transcription of hypoxia-inducible genes, including vascular endothelial growth factor and transforming growth factor (TGF)-α.4,16 Numerous mutations have been described as well as promoter hypermethylation and larger-scale chromosomal loss or translocations.13,17,18 Although VHL mutations occasionally are found in other RCC histologic subtypes, their prevalence is very low.11,17

Based on the similarities in tumor epidemiology and biologic behavior, we hypothesized that some RCC in dogs may be associated with VHL mutations. If so, then risk factors identified in people and newer antiangiogenic/antigrowth factor treatment options should be investigated in dogs as well. Additionally, confirming VHL mutations in canine RCCs would establish the dog as a model for the study of spontaneous human clear cell RCC, or alternatively for one of the other histologic RCC subtypes if mutations in VHL are absent. The canine VHL messenger ribonucleic acid (mRNA) and predicted amino acid sequences (GenBank accession no. AY764285) share 90% identity with the human VHL sequence (GenBank accession no. NM_000551). The purpose of this study therefore was to develop a polymerase chain reaction (PCR) protocol for sequencing of the VHL gene in archived, paraffin-fixed tissue samples, determine whether or not RCC in dogs is associated with VHL gene mutations, and if so describe the frequency, type, and location of these mutations. As presented here, we were able to successfully amplify the complete VHL coding deoxyribonucleic acid (DNA) sequence in 9 of 13 canine RCC and partial VHL coding DNA sequence in an additional 2 RCC. We did not identify VHL mutations in any of the tissue samples examined, and determined that the prevalence of VHL mutations in canine RCC is likely lower than that reported in human RCC.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

Case Identification and Tissue Collection

The computerized patient and pathology databases at Purdue University's School of Veterinary Medicine (2000–2007) and North Carolina State University's College of Veterinary Medicine (1988–2005) were searched for dogs with a histologic diagnosis of “renal carcinoma” or “renal tumor.” Hematoxylin and eosin-stained sections from identified cases were reviewed by a single pathologist using published criteria for subclassification of renal tumors in dogs.6 No dogs with renal tumors were polycythemic at any time during evaluation. Ten 10-μm-thick tissue slices were cut from archived paraffin-embedded sections of renal tumors (selecting those sections that did not have any or only minimal amounts of nonneoplastic tissue present on histologic examination) as well as from any paraffin-embedded kidney tissue sections that did not have tumor cells identified by the pathologist. Snap-frozen normal kidney tissue from 6 healthy hound dogs, whole blood from 1 healthy hound dog, paraffin-embedded kidney tissue from any non-RCC renal tumors identified in the North Carolina State University database, and paraffin-embedded kidney tissue from 3 dogs without neoplastic kidney disease randomly selected at Purdue University were used as normal controls. Blood and kidney tissue collection from hound dogs was performed with approval from Purdue University's Animal Care and Use Committee.

DNA Extraction and PCR

DNA extraction from frozen and paraffin-embedded tissues and from whole blood was performed using a commercial kita and total extracted DNA quantified.b For paraffin-embedded tissues, the manufacturer's protocol was modified to include 2 xylene treatments followed by ethanol washes to improve paraffin removal before DNA extraction.

The canine VHL mRNA sequence (GenBank accession no. AY764285) was compared with the human VHL sequence (GenBank accession no. NM_000551) and canine whole genome shotgun sequence (GenBank accession no. NW_876271). The canine VHL coding DNA was determined to be 90% homologous to the human sequence, and divided into 3 exons on chromosome 20. Because of the differences in sequence between species, 3 canine VHL-specific primer sets were designedc to amplify each exon separately rather than using previously published human VHL primers (Table 1).d

Table 1.   Primers used for canine VHL DNA amplification.
ExonF/RTA(°C)Expected Amplicon Size (bp)Sequence (5′–3′)
  1. F, forward; R, reverse; VHL, von Hippel-Lindau.

1F45652CCG CCA TTT CTG CCA CAG TAG C
F45479CGT TGT CTA GGC TCC GGG AGG TAA TGC
R  CCA ACT GCA TCT GTC ACA AAC TCG
2F53343GAG AGG CGG GTA CTT CAG TG
R  GGG CAA CCT ACT TCT CCC TC
3F53645AGC CTC TTG CTC ATC CAT TG
R  ATC ACC AAC CAC AAC ACA GC

Amplification of the 3 VHL exons was accomplished using either AccuPrime Pfx DNA Polymerasee (exon 1) or AccuPrime Taq High Fidelity DNA Polymerasee (exons 2 and 3). Cycling conditions were: denaturation at 96 °C for 2 minutes (exon 1) or 94°C for 1 minute (exons 2 and 3) followed by 35 cycles of denaturation for 30 seconds, annealing at primer-specific temperature for 30 seconds, and extension at 68°C for 1 minute, with a 5-minute final extension at 68 °C.f PCR products were visualized on 1.5% agarose gels containing 0.1% ethidium bromide, purified, and sequenced.g,h

Statistical Analysis

The 97.5% 1-sided exact confidence intervals for prevalence of VHL mutations were determined assuming a binomial distribution.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

Database searches identified 22 dogs with primary renal tumors diagnosed during the search periods. Review of archived material determined that 4 dogs did not have paraffin-embedded tissue available for DNA extraction, 1 dog with a nonrenal cancer had been miscoded as having a renal tumor, 1 dog initially diagnosed with RCC had a nonneoplastic inflammatory renal disease, and 1 renal tumor was from a cat. The remaining 15 dogs included 13 with RCC and 2 with primary renal sarcomas. After review of hematoxylin and eosin-stained sections from 14 of these dogs, RCC dogs were subclassified as having tubulopapillary carcinoma (4 dogs), papillary carcinoma (1 dog), chromophobe carcinoma (5 dogs), clear cell carcinoma (1 dog), and an unclassifiable carcinoma (1 dog). Dogs with RCC included 2 mixed breeds and 9 pure breeds (3 Labrador Retrievers and 6 dogs of various other breeds). Tissue from 1 RCC was not subclassified because insufficient sample remained after collection of sample for DNA extraction. The 2 dogs with primary renal sarcomas (1 mixed breed, 1 Boxer) were subclassified as having myofibroblastic or anaplastic sarcoma. Two dogs with RCC and the dog with myofibroblastic sarcoma also had paired paraffin-embedded kidney tissue from cancer-free tissue sections.

Sufficient DNA for PCR was successfully extracted from 20 of the 21 paraffin-embedded tissue samples (13 RCC, 2 renal sarcomas, 3 paired normal kidney tissue sampled from dogs with renal tumors, and 2 kidney tissue samples from dogs without renal tumors). DNA could not be extracted from 1 of 3 kidney tissues from dogs without renal tumors. DNA also was successfully extracted from the 6 snap-frozen kidney tissue samples and the 1 whole blood sample from healthy hound dogs. Successful amplification of at least 1 of the 3 VHL exons was accomplished in 22 of these 24 samples, including 11 RCC, 2 primary renal sarcomas, 3 sections of normal kidney tissue from dogs with renal tumors, and 9 dogs without evidence of renal neoplasia (Table 2). None of the 3 VHL exons could be amplified from 2 RCC samples.

Table 2.   Tissue types from which VHL exons were successfully amplified.
TissueN# Specimens with Successful Exon Amplification
Exon 1Exon 2Exon 3
Renal cell carcinoma119119
 Carcinoma, unclassified2222
 Chromophobe carcinoma4343
 Papillary carcinoma1010
 Tubulopapillary carcinoma4444
Renal sarcoma2222
Nonneoplastic kidney tissue8888
 Normal, snap-frozen6666
 Nonneoplastic, formalin-fixed2222

Initial PCR attempts yielded VHL exons 2 and 3 amplicons of expected size for most DNA samples, but VHL exon 1 was successfully amplified in only approximately 50% of samples. Alignment of the forward primer for VHL exon 1 with the canine VHL mRNA sequence (GenBank accession no. AY764285) identified an intervening CpG island 3′ to the primer attachment site, and 5′ to the presumptive canine VHL methionine start codon, which was suspected to be inhibiting primer attachment or DNA extension during PCR. A second forward primer therefore was designed that would anneal immediately 3′ to the CpG island; the same reverse primer was used for both VHL exon 1 PCR reactions (Table 1).

VHL exons 1 and 3 were successfully amplified from 23 of the 25 DNA extracts, whereas exon 2 was amplified from all 25 samples. Amplicons all were of expected size, and of identical size in all samples (Fig 1). Sequencing results for all VHL DNA amplicons were 100% identical with the accessioned canine VHL nucleotide sequence. No mutations were found in the coding DNA sequence in any of the neoplastic or normal canine kidney tissues regardless of the method of preservation (formalin-fixed, paraffin-embedded versus snap-frozen), or in the whole blood DNA sample.

image

Figure 1.  Representative 1.5% agarose gel demonstrating polymerase chain reaction (PCR) amplicons for the 3 canine von Hippel-Lindau exons. Exon 1 amplicons in this gel were produced using the 652 bp predicted amplicon primer pair. Lane M: molecular weight marker; C1–C3, PCR amplicons from renal cell carcinoma samples; N1, amplicons from normal kidney sample. M, molecular weight; C, cancer; N, normal.

Download figure to PowerPoint

Based on our prevalence of zero mutations found in canine VHL exon 1 or 3 in 11 RCC samples, the 97.5% binomial 1-sided exact confidence interval for the prevalence of mutations in these exons was 0.0–33.6%. The 97.5% binomial 1-sided exact confidence interval for prevalence of mutations in canine VHL exon 2 was 0.0–28.5% based on zero mutations found in 9 samples.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

Although naturally occurring tumors in domestic dogs are receiving increased recognition as possible models for human cancer, further molecular characterization of canine tumors is required.19,20 This is particularly important before investigating novel anticancer therapeutics that target aberrant pathways in the analogous human diseases. For example, many newer therapies for RCC in people target vascular endothelial growth factor, platelet-derived growth factor, and HIF, all of which are dysregulated in VHL mutation-associated human clear cell RCC.21,22 Based on this need, we hypothesized that canine RCC may be associated with a similar prevalence of VHL mutations as described in human clear cell RCC. However, we did not find any mutations in the coding DNA sequences in our canine neoplastic or normal tissue samples, and were able to estimate a maximal prevalence for canine RCC-associated VHL mutations that is less than that reported in human clear cell RCC. Therefore, despite similarities in tumor epidemiology, clinical presentation, and disease progression, canine RCC do not appear to have VHL mutations present at the same prevalence as that found in people with RCC.

Despite our conclusion that the frequency of VHL mutations (and thus pathogenesis) in canine RCC likely differs from human clear cell RCC, absence of mutations in the canine VHL coding DNA does not exclude the possibility of decreased VHL protein expression in a higher percentage of tumors than predicted by our 97.5% 1-sided exact confidence interval. Although point mutations leading to truncated mutants are the most common VHL mutation found in human clear cell RCC, complete or partial chromosomal deletions or translocations are also possible.14,15 We suspect that our inability to amplify some or all exons in a subset of canine RCC was because of the degradation of DNA known to occur secondary to prolonged formalin fixation and paraffin-embedding, and resultant inability to amplify target sequences by PCR.23,24 However, we acknowledge that these tumors instead may have had these larger-scale DNA abnormalities with complete disruption of primer binding sites. Alternatively, abnormalities in the promoter region also may alter VHL protein expression despite lack of mutations in the coding DNA sequence. For example, hypermethylation of the human VHL promoter resulting in decreased gene transcription has been reported in 8.3–20.4% of clear cell RCC.13,15,25,26 Direct mutation or similar hypermethylation of the VHL promoter may be a cause of RCC development in dogs which our study was not designed to detect. Unfortunately, the canine VHL gene is preceded by a 5′ CpG island that interfered with exon 1 forward primer annealing or DNA polymerase extension. In a few cases, our reverse primer for exon 1 did yield partial sequences for the 5′ region that are homologous to the human VHL promoter, but the length of these amplified promoter regions sequences were insufficient to make any conclusions about possible mutations.

Unlike the somatic biallelic VHL mutations seen in spontaneous clear cell RCC, familial VHL disease in people is caused by a germline mutation in only 1 of 2 VHL alleles. Spontaneous acquired mutations in the remaining normal allele result in several tumor types, including clear cell RCC, pheochromocytomas, and central nervous system and retinal hemangioblastomas. An identical syndrome has not been described in dogs, but several related families of German Shepherd dogs and German Shepherd crossbreeds develop bilateral multifocal renal cystadenocarcinomas, nodular dermatofibrosis, and uterine leiomyomas, similar to Birt-Hogg-Dubé (BHD) syndrome in people.4,22,27 At least 1 pedigree of affected dogs has a mutation of the canine BHD gene homolog, which encodes folliculin, a protein of unknown function.28 Affected dogs are heterozygous, and abnormalities develop after acquired mutations in the normal BHD gene.29 Histopathologic changes are similar to those noted in vitro with TGF-β dysregulation.27,30 Therefore, although mutation in the BHD gene in people is associated with chromophobe and oncocytoma variants of RCC rather than clear cell RCC, determining whether or not VHL mutations also are present in these cystadenocarcinoma dogs may be merited based on the similarities to human VHL disease and the suggested role of TGF in tumorigenesis in both syndromes.

Although we only investigated the VHL gene in these canine renal tumors, a number of other genetic mutations have been associated with RCC in people. Mutations in MET occur in 13% of papillary RCC, and BHD mutations are found in 4% of patients with chromophobe or oncocytoma RCC.4 Interestingly, a germline MET mutation has been described in dogs, with a prevalence of approximately 70% in Rottweilers.31 A single report identified vimentin expression in 3 of 4 renal carcinomas, which is a marker for clear cell and papillary RCC subtypes in people.32,33 Therefore, it appears that mutations in canine papillary and tubulopapillary RCC also should be investigated.

None of the dogs from which renal tumors were included in this study were polycythemic at any point during evaluation. Transcription of the erythropoietin gene increases in response to HIF activation.4,16 HIF is targeted for proteasomal degradation by the VHL protein, and as expected VHL mutations have been associated with idiopathic erythrocytosis in people with and without RCCs.34,35 Whether or not an association between erythrocytosis and VHL mutations exists in dogs with primary or paraneoplastic polycythemia cannot be determined based on the results reported here.

The relatively small number of canine RCC available for this study resulted in wide confidence interval for the predicted prevalence of VHL mutations in these tumors. Nevertheless, we established here that the maximal prevalence is likely lower than that reported in most studies on human clear cell RCC.11–15 In fact, more recent studies investigating the prevalence of VHL mutations in human clear cell RCC have tended to produce higher prevalence rates, presumptively because of more sensitive detection methods and recognition of promoter hypermethylation as a cause of decreased VHL expression.13,14 Future studies that include larger numbers of the rarer canine clear cell subtype RCC should be conducted to more definitively determine if there is an association between VHL mutations and this histologic subtype in dogs.

Footnotes

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

aQiagen DNeasy Blood and Tissue Kit, Valencia, CA

bNano-Drop Spectrophotometer, Wilmington, DE

cPrimer 3, Biology WorkBench, http://workbench.sdsc.edu

dIntegrated DNA Technologies, Coralville, IA

eInvitrogen, Carlsbad, CA

fApollo ATC401 Thermal Cycler, GMI Inc, Ramsey, MN

gQiagen PCR Purification Kit, Valencia, CA

hBig Dye Terminator Chemistry, Purdue Genomics Core Facility, Purdue University, West Lafayette, IN

Acknowledgments

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References

The authors thank Beth Case and Sandra Horton for identification of cases and processing of tissue samples, and Dr George Moore for assistance with statistical analysis of data.

Grant support: This research was supported by an award to Dr Pressler by the International Renal Interest Society and by an equipment grant to Dr Pressler by the charitable giving office of Kindy French.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Footnotes
  7. Acknowledgments
  8. References