- Top of page
- Materials and methods
This study investigated the correlation between KIT gene expression determined by immunohistochemistry and real-time polymerase chain reaction (RT-PCR) and the rate of tumour recurrence and tumour-related deaths in dogs affected with mast cell tumour (MCT). Kaplan–Meier curves were constructed to compare tumour recurrence and tumour-related death between patients. The log-rank test was used to check for significant differences between curves. KIT-I, KIT-II and KIT-III staining patterns were observed in 9 (11.11%), 50 (61.73%) and 22 (27.16%) tumours, respectively. Tumour recurrence rates and tumour-related deaths were not associated with KIT staining patterns (P = 0278, P > 0.05), KIT (P = 0.289, P > 0.05) or KIT ligand (P = 0.106, P > 0.05) gene expression. Despite the lack of association between KIT staining pattern and patient survival time, the results suggest a correlation between aberrant KIT localization and increased proliferative activity of MCTs. RT-PCR seems to be a sensible method for quantitative detection of KIT gene expression in canine MCT, although expressions levels are not correlated with prognosis.
- Top of page
- Materials and methods
Oncogenes are genes that cause cancer. These genes are derived from proto-oncogenes that regulate cell growth and differentiation. Several proteins produced by oncogenes have been investigated, and studies show that growth factor receptors may be overexpressed or constantly activated in cancer. c-kit is a proto-oncogene that encodes KIT (CD117). This type III tyrosine kinase protein (KIT) acts as a receptor for stem cell factor (SCF), a cytokine responsible for growth factor activation in mast cell tumours (MCTs). KIT is a transmembrane tyrosine kinase receptor consisting of one intracellular tyrosine kinase and five extracellular immunoglobulin-like domains. The KIT receptor ligand SCF, or mast cell growth factor, is related to mast cell survival, proliferation, differentiation, chemotaxis, degranulation, suppression of apoptosis and adhesion to fibronectin.[2, 3]
Mutations in the juxtamembrane domain of c-kit have been identified and correlated with more aggressive MCTs of higher histological grade and poorer prognosis.[3-6] Kiupel et al. have observed that MCTs showing pure perimembrane KIT-positive staining were not associated with reduced survival times or tumour recurrence. However, diffuse cytoplasmic KIT-positive staining was directly associated with reduced survival times and high recurrence rates. These observations led to the proposal of a new classification method based on the localization of KIT immunohistochemical staining in MCTs. According to this method, tumours are classified as KIT-I, KIT-II and KIT-III, depending on the specific site of KIT expression. KIT-II and KIT-III staining patterns are significantly associated with local tumour regrowth and reduced survival time, respectively.
Turin et al. have evaluated c-kit mRNA expression using reverse transcription quantitative real-time polymerase chain reaction (RT-PCR) analysis of tumour tissue and blood of dogs affected with MCTs. The c-kit expression was not significantly related to tumour size, location, histological grade or biological behaviour, response to treatment or patient survival.
The evaluation of c-kit receptor ligand expression and mutation in cultures of different mast cell lineages has confirmed c-kit expression but not c-kit mutation.
Canine MCT is frequently diagnosed in clinical practice and extensively researched. However, prognostication is difficult because of great variability in biological behaviour, particularly of intermediate grade tumours.[9, 10] Although several prediction methods have been tested, histopathology remains the most reliable prognostic tool to date.
The purpose of this study was to investigate a potential relationship between the prognosis of canine MCT of different grades and aberrant KIT localization using immunohistochemistry, and between canine MCT prognosis and KIT protein and KIT ligand gene expression determined by RT-PCR. The results of this study may help clarify the role of KIT in canine MCT and improve prognostic accuracy.
Materials and methods
- Top of page
- Materials and methods
This research project was approved by the Bioethics Committee of the School of Veterinary Medicine and Animal Science, University of São Paulo – FMVZ/USP (Faculdade de Medicina Veterinária e Zootecnia da Universidade de São Paulo; protocol 1014/06).
Dogs of any gender, age or breed diagnosed with MCT by fine needle aspirate cytology performed at the FMVZ/USP Veterinary Hospital (HOVET) were selected for the study. The following inclusion criteria were adopted:
- dogs with a cytological diagnosis of MCT eligible for surgical resection;
- dogs whose owners agreed to follow-up for a minimum of 1 year following surgical resection;
- dogs with histopathological confirmation of MCT diagnosis following surgical resection.
Animals presenting with tumours that could not be successfully resected or that were lost to follow-up during the 1-year data collection period were excluded from the study.
Surgically resected tumours were submitted to histopathology for confirmation of previous cytologic diagnosis of MCT. Histological grading was based on cell differentiation according to the criteria proposed by Patnaik et al. Surgical margins were evaluated.
Immunohistochemistry and evaluation of KIT immunolabelling
Tumour samples were sectioned at 5 µm, mounted onto silanized glass slides, deparaffinized and rehydrated in xylene and alcohol, respectively. Antigen retrieval was performed in citrate buffer (pH 6.0) for 15 min using a microwave oven at maximum power. Sections were washed three times with gentle agitation in 1× phosphate-buffered saline (PBS) between procedures. Endogenous peroxidase activity was quenched by 30-min incubation in 5% hydrogen peroxide. Non-specific reactions were blocked with 5% skimmed milk/PBS solution for 15 min at 37 °C. Sections were then incubated overnight with anti-KIT primary antibody (1:100; Dako, Carpinteria, CA, USA) in a moist chamber at 8 °C prior to incubation with secondary antibody [Biotinolate Kit LSAB + System-HRP (Dako)] for 30 min, washing in PBS and incubation with tertiary antibody [Complex Streptoavidine-Biotine Kit LSAB + System-HRP (Dako)] for 30 min in a moist chamber at room temperature. Development was performed with diaminobenzene solution (DAB + Cromogen, Dako) (100 µL per slide). Slides were counterstained with haematoxylin for 2 min, washed in running water, dehydrated and mounted with coverslips in synthetic resin. Interstitial cells of Cajal were used as internal positive controls for KIT.
Immunohistochemical KIT expression was classified according to previously described staining patterns: KIT I – membrane-associated staining with little to no cytoplasmic staining of neoplastic mast cells; KIT II – intense focal or stippled cytoplasmic staining of neoplastic mast cells and KIT III – diffuse cytoplasmic staining of neoplastic mast cells, sometimes covering other cell structures. Immunostained cells were manually counted under light microscopy (40× magnification) using Image Pro Plus software (Image Pro Plus 4.5®, Media Cybernetics, Silver Spring, MD, USA). Ten fields were selected at random, and the prevailing KIT staining pattern was the pattern considered.
Evaluation of KIT and KIT ligand gene expression
MCT samples collected during surgery were frozen in liquid nitrogen and stored at −80 °C for subsequent RT-PCR analysis. Thirty grams of tumour tissue was macerated in autoclaved pestle and mortar washed in ribonuclease-free water. Trizol (500 µL) was added to the macerated tissue for RNA extraction, taking care to avoid RNA degradation and sample contamination. Following incubation for 5 min at room temperature, 100 µL of chloroform was added and the mixture was homogenized. The sample was incubated for 2 min at room temperature and then centrifuged at 152.9424 g relative centrifugal force (RCF) for 20 min at 4 °C. The supernatant was transferred to a new tube containing cooled isopropanol (500 µL), incubated on ice for 15 min and centrifuged at 1200 rpm for 20 min at 4 °C. The supernatant was removed and the RNA pellet was washed with 1 mL of 70% ethanol/diethylpyrocarbonate (DEPC)-treated water solution prior to centrifugation at 1200 rpm for 20 min at 4 °C. The supernatant was again removed and the RNA pellet was resuspended in DEPC-treated MilliQ water, vortexed for 10 min and incubated for 10 min at 56 °C. The RNA extracted was frozen and stored at −80 °C immediately after extraction.
RNA was diluted in autoclaved DEPC-treated MilliQ water (10:40 µL). Absorbance was determined using a BioPhotometer (Eppendorf, Hamburg, Germany) and the A260/280 nm ratio was determined for RNA quantification. Samples with RNA concentration between 1.7 and 2.0 were suggestive of low RNA contamination and considered eligible for reverse transcription. RNA viability was verified by 1.5% agarose gel electrophoresis using Tris-acetate–ethylenediamine tetraacetic acid (EDTA) buffer.
Viable RNA was reverse transcribed into complementary DNA (cDNA). For cDNA preparation, DNAse treatment was performed using a mixture of RNA (1000 ng), DNAse buffer (1 µL), DNAse (1 µL) and DEPC-treated MilliQ water (7 µL) incubated for 15 min at room temperature. EDTA (25 mM; 2 µL) was added prior to incubation for 10 min at 65 °C. Following cooling, 1 µL oligo-dT (short sequence of deoxy-thymine nucleotides; 0.5 mg mL−1) and 1 µL deoxyribonucleotide triphosphates (10 mM) were added and the mixture was incubated at 65 °C and left to cool. RNAse treatment was then started with a mixture of 1 µL 5× first-strand buffer, 2 µL dithiothreitol (0.1 M) and 1 µL RNAse OUT (ribonuclease inhibitor; 5000 UI), incubated for 2 min at 42 °C and then cooled. Superscript II (10 000 U; 1 µL) was added to the mixture prior to incubation for 50 min at 42 °C and for 15 min at 70 °C. The cDNA obtained was cooled and stored at −20 °C.
cDNA amplification was performed by RT-PCR using Platinum SYBR Green qPCR SuperMix-UDG kit and 25 µL of a solution containing MIX (12.5 µL), ROX (6-carboxy-X-rhodamine; 0.5 µL), 1:20 cDNA/MilliQ water solution (4 µL), MilliQ water (5.5 µL) and primer (2.5 µL) for each sample. RT-PCR was performed on ABI PRISM 7000, (Applied Biosystems, Foster City, CA, USA) a sequence detection system containing a CCD camera for fluorescence detection and corresponding data analysis software.
PCR products (25 µL) were placed in optical tubes (Applied Biosystems, Foster City, CA, USA) for determination of KIT and KIT ligand expression. Following complete mixing of reactants, the tubes were closed with MicroAmp Optical caps (Applied Biosystems). All reactions were performed in duplicate. The RT-PCR procedure consisted of 40 cycles: 2 min at 50 °C, 10 min at 95 °C, denaturing for 15 s at 90 °C and extension for 1 min at 60 °C. KIT ligand (QT 00895342—Quant Tect Primer Assay—QIAGEN Sciences, Germantown, MD, USA) and KIT receptor (QT 00897099—vkit Hardy-Zuckerman 4 Feline Sarcoma Viral Oncogene Homolog—QIAGEN Sciences) were employed as primers. Data analysis was performed using the 2DDCT method.
Following surgery, animals were evaluated monthly for the first 3 months and then every 3 months for at least 1 year.
Postoperative assessment was based on clinical history and physical examination. Abdominal ultrasonography and radiography were performed in cases where tumour regrowth and/or metastasis were suspected. Owners were instructed to bring their dogs for re-evaluation ahead of schedule if tumour regrowth was observed.
Maximum likelihood estimation (MLE) was used to verify the association between classificatory variables. Significant variables according to values of normality were compared using the Mann–Whitney or the Kruskal–Wallis test. Kaplan–Meier curves were constructed to compare relapse-free survival time and tumour-related death between patients, and the log-rank test was used to determine if there were significant differences between curves. Significant variables (log-rank test) were used for adjustment of the multivariate Cox regression model. The time frame adopted for statistical analysis was the time from surgical intervention. The level of significance was set at 5% (P < 0.05).
- Top of page
- Materials and methods
Eighty-one dogs, 40 females (49.38%) and 41 males (50.62%), with a positive cytologic diagnosis of MCT were selected among 156 patients. MCT was more common in Boxers (32 animals; 39.51% of patients), although other breeds were also represented. Patients were aged between 3 and 17 years (mean age: 8.43 years).
Histological grading and immunohistochemical staining were performed on the same samples for better comparison of results. Out of 81 animals, 51 (63%) had grade II, 15 (18.5%) had grade I and 15 (18.5%) had grade III MCTs. Forty-eight (59.25%) patients died during the follow-up period. Forty-three (53.08%) deaths were MCT-related – 13 (30.23%) due to grade III, 25 (58.13%) to grade II and 5 (11.62%) to grade I MCT. Tumour regrowth was observed in 24 of 51 (49.01%) animals with grade II, 11 of 15 (73.33%) animals with grade III and 6 of 15 (40%) animals with grade I MCT.
Margins were compromised in 27 of 81 (33.33%) animals. Margins were narrow in 8 (9.81%) and free in 46 (56.79%) animals. Compromised margins were not associated with higher recurrence rates or death (log-rank test; P = 0.783, P > 0.05) or when death (P = 0.264) and recurrence (P = 0.856) were treated as separate events.
Given the prospective nature of the study, case follow-up was easy and clinical re-evaluation was performed as frequently as possible. Whenever animals failed to return on schedule, owners were contacted by telephone and follow-up appointments were rescheduled if possible. Out of 81 patients, 66 (81.48%) could be followed up for more than 18 months.
Tumour staining patterns and respective histological grades are displayed in Table 1. A graph showing staining pattern distribution according to histological grade is presented in Fig. 1. Relapse-free survival time was not associated with staining pattern (log-rank test; P = 0.680, P > 0.05). KIT immunostaining patterns are shown in Fig. 2. The distribution of tumour-related deaths according to KIT immunostaining is displayed in Table 2.
Table 1. Distribution of KIT staining patterns according to histological grade (FMVZ/USP, São Paulo, 2010)
|KIT staining pattern||Histological grade|
|Grade I||Grade II||Grade III||Total|
Figure 1. Case distribution according to KIT staining pattern for each histological grade. Cases were classified as KIT I, KIT II or KIT III according to KIT immunostaining pattern (MLE; P = 0.208, P > 0.05) – São Paulo – 2010.
Download figure to PowerPoint
Figure 2. Photomicrographs of the different KIT immunostaining patterns (haematoxylin). (A) KIT I; (B) KIT II and (C) KIT III. Light microscopy; ×40 magnification.
Download figure to PowerPoint
Table 2. Distribution of MCT-related and unrelated deaths according to KIT immunostaining
|KIT staining pattern||Tumour-related deaths||Deaths unrelated to tumour||Animals alive at the end of the study||Total|
|KIT I||6 (7.41%)||0 (0.0%)||3 (3.7%)||9 (11.11%)|
|KIT II||26 (32.1%)||3 (3.7%)||21 (25.93%)||50 (61.73%)|
|KIT III||11 (13.58%)||2 (2.47%)||9 (11.11%)||22 (27.16%)|
|Total||43 (53.09%)||5 (6.17%)||33 (40.74%)||81 (100%)|
Medians, minimum and maximum values of KIT staining patterns were not associated with KIT or KIT ligand gene expression (Kruskal–Wallis test; P = 0.323 and P = 0.115, respectively). The frequency of distribution of KIT staining patterns did not differ significantly according to histopathological grade (MLE; P = 208, P > 0.05; Fig. 1).
KIT and KIT ligand gene expression
KIT and KIT ligand gene expression in one of the patients was used as a reference (Table 3; animal no. 53) for readings of KIT and KIT ligand gene expression in the remaining patients. Animals were named according to the archives of the FMVZ/USP Comparative Oncology Laboratories (LOC – VPT; Table 3). Thirty-seven randomly selected MCT samples were analysed; however, only 22 samples showed RNA variability following extraction. In all 22 animals, KIT and KIT ligand gene expression was greater than in the reference animal.
Table 3. Patient distribution according to registration within the Comparative Oncology Laboratories (LOC – VPT) and according to kit and kit ligand gene expression (FMVZ/USP, São Paulo, 2010)
|File no.||LOC registration no.||KIT gene expression||KIT ligand gene expression|
Data presented above were analysed as a function of the relapse-free survival time curve and were not associated with KIT gene expression (log-rank test; P = 0.289, P > 0.05). Although a positive association with KIT ligand gene expression was observed (log-rank test; P = 0.010, P > 0.05; Fig. 3), it was devoid of independent predictive value (Cox multivariate analysis).
Figure 3. Kaplan–Meier curve of relapse-free survival times as a function of kit gene expression, dichotomized in groups of patients with mean kit gene expression >18 and <18 (log-rank test; P = 0.010); censored observations to the right.
Download figure to PowerPoint
Tumour regrowth was observed in 13 of the 22 (59.1%) animals analysed above and 14 (63.63%) animals died. Death was tumour-related in 13 (59.1%) cases classified as KIT I (1 case; 7.69%), KIT II (9 cases; 69.23%) or KIT III (3 cases; 23.07%). KIT and KIT ligand gene expression means for each KIT classification are given in Table 4.
Table 4. Mean KIT and KIT ligand gene expression according to KIT immunostaining patterns (FMVZ/USP, São Paulo, 2010)
| ||KIT||Standard deviation||Coefficient of variation||KIT ligand||Standard deviation||Coefficient of variation|
- Top of page
- Materials and methods
This study evaluated the potential value of KIT expression detected using immunohistochemistry and PCR as a prognostic indicator for cases of canine MCT. Eighty-one dogs satisfied the inclusion criteria.
Selected animals were followed up for a minimum of 12 months and some for more than 40 months (mean follow-up time was 18.32 months). The minimum follow-up criterion adopted in this study was based on previous studies where a similar period was considered sufficient to evaluate the progression of MCTs in dogs.[12-15]
The c-kit oncogene is currently under investigation by several research groups.[2, 5, 8, 13, 16-21] Constitutive activation of c-kit oncogene receptor leads to uncontrolled cell proliferation and loss of apoptotic ability, suggesting a more aggressive behaviour of neoplastic cells. This activation mechanism is not completely understood and may be related to c-kit gene mutation. In a previous study, mutations were not identified in all tumours presenting with aberrant KIT localization studied, and a correlation between KIT localization and mutation was not always observed. Evidences that proliferation of mutating mast cells can be induced by constitutive activation without ligand participation have been provided, and the presence of such cells is associated with reduced survival times and increased recurrence rates. The c-kit mutations that were not associated with cell proliferation or neoplasia have been observed in mast cell cultures. KIT receptor may be activated by internal or external autocrine SCF/KIT receptor stimulation through circuit mechanisms involving HRMC cells. Activation of cell receptors by their own SCF secretion can be assumed when external autocrine stimulation occurs. A study involving the analysis of 60 skin tumours failed to identify c-kit exon 11 mutations.
Deeper cytoplasmic KIT signalling (KIT III) may be related to tumourigenesis. A possible explanation may be the increased metabolism in tumours presenting with cytoplasmic KIT immunostaining, which is correlated with increased cell proliferation following KIT receptor activation. However, increased cell proliferation per se is not enough to induce neoplastic transformation, although increased risk of spontaneous mutations because of cell proliferation may contribute to tumour malignancy.[3, 20]
Assuming a potential direct relationship between KIT gene expression and histological grade, a new classification has been proposed based on the staining pattern of tumour cells. Higher KIT expression and histological grade would reflect a higher degree of malignance and shorter survival time.[4, 16] However, associations between KIT staining pattern, survival time and MCT histological grade[13, 23] or between KIT localization and MCT histological grade have not been confirmed. Similarly, in this study, KIT localization was not associated with histological grade.
Conversely, an association between KIT localization, histological grade and tumour necrosis and ulceration has been reported. Necrosis would be related to increased cell proliferation devoid of angiogenesis, while ulceration would be indicative of more aggressive tumours, with greater histamine and serotonin release and consequent pruritus. KIT was also associated with an increased number of ki-67 marked nuclei, confirming increased cell proliferation. Recent work has shown that KIT and ki-67 staining, the presence of c-kit mutation and AgNOR staining are good prognostic indicators for MCT and can be used for MCT treatment evaluation. Similar findings were reported in a study by Thompson et al. but the conclusion was that none of these staining patterns alone can accurately predict MCT progression.
Aberrant KIT location may prove an accurate prognostic marker for canine MCT provided genetic (e.g. additional KIT mutations) or epigenetic causes of this phenomenon can be indentified in future molecular studies. Given diffuse cytoplasmic KIT staining suggests dysregulation, aberrant KIT localization may indicate the need for targeted therapy in affected dogs. However, despite high sensitivity, diffuse cytoplasmic KIT staining may not be a specific test. Cross-sectional studies (where the prevalence is not predetermined) are needed to verify this hypothesis and to establish more precise cutoff values of cell proliferation for comparison. Multivariable analyses of larger subset of tumours could help determine the best test combination for MCT diagnosis.
In this study, KIT staining patterns were not associated with tumour remission interval or death. Morphological staining patterns were not associated with histological grade, and KIT III staining patterns were observed in grade I MCTs. Our results differ from previously published data,[2, 19, 21, 24] possibly because of discrepancies in the interpretation of semiquantitative results. KIT staining patterns can be classified into I, II or III based on the predominant staining pattern observed in a minimum of 100 randomly selected mast cells, a subjective method that implies interobserver differences and possibly different interpretations. KIT expression may also differ in dogs living in São Paulo, such as those selected for this study. Given the low reproducibility of results in this study and the lack of supporting data to demonstrate a relationship between KIT staining pattern, histological grade and prognosis, KIT immunostaining cannot be recommended for accurate MCT prognostication. It can, however, be considered a good method for differentiation between MCTs and other types of round cell neoplasias.
Despite the lack of prognostic value, KIT expression has been considered a good method for quantitative and sensitive detection of c-kit gene expression in canine MCTs. Kit ligand gene expression has been reported in different MCT lineages and was not correlated with mutations. Our results support these findings. Gene expression was documented in neoplastic tissue, but could not be associated with tumour remission interval or death. We had suspected that KIT ligand expression would increase as expression of the previous gene decreased but, in fact, KIT ligand expression tended to follow KIT expression.
KIT ligand expression was associated with prognosis in this study. Affected animals had a 91% chance of recurrence when c-kit ligand expression was below 18 (mean value of ligand expression) against a 27% chance when expression was above 18. Although lower KIT ligand expression increased the chances of tumour recurrence, death was not a significant variable. The lack of similar data in literature precluded further discussion of results. An association between KIT receptor and KIT ligand gene expression, and histological grade could not be established using KIT immunostaining in this study. Gene expression using quantitative real-time PCR has been seldom investigated. The number of tumours evaluated by this method in this study was smaller, given some tumours could not be sampled immediately after surgery and had to be discarded. Also, non-viable RNAs were obtained following extraction in some cases.
Although MCTs may express large amounts of c-kit mRNA, KIT protein levels may be too low to be picked up by immunohistochemistry. Also, transcription pathways may not be intact, leading to early interruption of protein synthesis.
The study of potential associations between differences in KIT and KIT ligand gene expression and the occurrence of gene mutation requires the investigation of concurrent DNA mutations. Mutation analysis may also help support the indication of new treatments for MCTs based on tyrosine kinase inhibitors,[20, 25-30] which apparently reduce cell proliferative capacity.