Immunophenotypic characterization of canine malignant lymphoma: a retrospective study of cases diagnosed in Poland Lower Silesia, over the period 2011–2013
Abstract
Lymphoma is the most frequently diagnosed cancer of the canine haematopoietic system. In this study, the flow cytometry and polymerase chain reaction (PCR) analysis were used to characterize a series of canine lymphomas in detail. The aim of this study was to determine the incidence of B- and T-cell high-grade lymphomas and their immunophenotypic characterization in Lower Silesia, Poland. The results show that the frequency of each type of lymphoma is 71% for B-cell and 17% for T-cell lymphomas. In two cases the PCR techniques confirmed the presence of simultaneous double gene rearrangements of the BCR and TCR receptors.
Introduction
Nowadays cancer, beside infectious diseases, is the main cause of canine mortality,1 and lymphoma is one of the most common ones.2 Taking into account, the morphology, tumour genetics, disease progression and the response to treatment, the canine lymphomas show strong similarities to human high-grade non-Hodgkin's lymphomas (NHLs).3 The survival time for a dog with NHL is associated with the type of disease,4 therefore a proper diagnosis is of most importance.
The diagnosis of lymphoma in dogs, besides a full medical history, a clinical examination and additional tests (biochemical and morphological blood examination, ultrasound and diagnostic imaging) includes also a cytopathological or histopathological examination of the altered lymph nodes, blood or bone marrow. Owing to the specific nature of tumour growth (diffuse, dispersed changes), an assessment of a cytopathological smear obtained by fine-needle aspiration (FNA) biopsy is usually sufficient to give a diagnosis. Flow cytometry, polymerase chain reaction (PCR) and Western blotting are also very helpful techniques used for research and detailed lymphoma diagnosis.5
Multiparameter flow cytometry is a very useful tool, both in human and veterinary oncology. In small animal oncology, the most common application of this technique is associated with haematooncology, especially with immunophenotyping of different types of leukocytes.6 Flow cytometry is a very precise and accurate method, limited only by the diversity of antibodies applied. In addition to determining the phenotype of cancer cells, it also facilitates the diagnosis of metastases.
In some cases, the diagnosis of haematopoietic neoplasms in dogs can be a challenge (early stages of disease, low quality of the analysed material). As clonality is the hallmark of malignancy, molecular methods that determine the presence of clonal gene rearrangements of B- or T-cell receptors are very helpful. Such assays, based on the PCR technique, were developed at the Clinical Immunology Laboratory at Colorado State University.7, 8
The aim of our study was to determine, by flow cytometry, the incidence of high-grade B- and T-cell lymphomas in the population of examined dogs and to define their phenotypic characteristics.
Materials and methods
Collection of canine lymphoma cases and cytopathological examination
The research material was based on the cases of dogs with lymphoma diagnosed in the years 2011–2013 at the Clinic of the Diseases of Horses, Dogs and Cats, Faculty of Veterinary Medicine, Wroclaw University of Environmental and Life Sciences, Poland.
From among 109 collected samples, 86 lymph node FNA samples were included in this study. In three cases, the amount of material was insufficient to carry out the test and two samples were excluded because of low quality of the material. Also, the samples in which the diagnosis was doubtful or where further tests ruled out the malignancy (n = 18) were excluded from the study. The samples were collected from dogs of different breeds, age and sex, with various lymphoid malignancies. The material was taken from the dogs that have not been previously treated with any anti-cancer agents.
Cytological smears of all samples included in this study were air-dried, fixed and stained with haematoxylin and eosin (HE). All cytological slides were examined under Olympus BX 41 microscope by the same cytologist. The morphological criteria were based on the updated Kiel classification. The assessment was based on the lymphocyte morphology, with particular attention paid to the size of the cells and their nuclei (using erythrocytes present in the formulations as scale size), the quality and staining of chromatin, the presence and distribution of nucleoli, and the ratio of nuclei to cytoplasm (N/C ratio). The mitotic index was assessed at the highest magnification (×600), by counting atypical mitotic figures, and calculated as a mean number of cells from five fields of vision. This index was assessed according to the scale: low index, ≤1 mitosis in the fields of vision; average index, 2–4 mitoses and high index, ≥5 mitoses per high power fields.
Determination of the immunophenotype with flow cytometry
Sample preparation
Fine-needle aspirates (FNAs) of the canine lymphoma samples underwent several cycles of rinsing with phosphate-buffered saline (PBS; Institute of Immunology and Experimental Therapy, Polish Academy of Sciences, Wroclaw, Poland) and erythrocyte lysis (0.84% ammonium chloride, Sigma Aldrich, Steinheim, Germany) and then they were dispensed into 11 individual cytometry tubes (BD Falcon, Becton Dickinson, San Jose, USA). The antibodies used for immunophenotyping by means of the flow cytometry are shown in Table 1.
| Antigen | Antibody/clone | Producer |
|---|---|---|
| CD3 (FITC) | CA17.2A12 | AbD Serotec |
| CD4 (PE) | YKIX302.9 | AbD Serotec |
| CD8 (Alexa Fluor® 647) | YCATE55.9 | AbD Serotec |
| CD5 (FITC) | YKIX322.3 | AbD Serotec |
| CD14 (Alexa Fluor® 647) | TÜK4 | AbD Serotec |
| CD34 (PE) | 1H6 | AbD Serotec |
| CD45 (PE) | YKIX716.13 | AbD Serotec |
| CD45R (PE) | YKIX753.22.2 | AbD Serotec |
| MHC II (FITC) | YKIX334.2 | AbD Serotec |
| CD79A (PE) | HM57 | Dako |
| CD21 (Alexa Fluor® 647) | CA2.1D6 | AbD Serotec |
Staining of surface antigens
Briefly, the tubes containing 5 × 105 cells were incubated for 20 min at 4 °C, with a selected monoclonal anti-CD3, CD4, CD5, CD14, CD8, CD21, CD34, CD45, CD45R or MHC class II antibody directly conjugated with a fluorochrome (Table 1). After incubation, the cells were rinsed twice with FACS buffer (1% bovine serum albumin in PBS) and resuspended in PBS.
Staining of cytoplasmic antigens
Briefly, the tubes containing 5 × 105 cells were incubated with a permabilization and fixation buffer (Dako, Glostrup, Denmark) for 15 min, washed and stained with monoclonal anti-79α. After incubation, the cells were washed twice with FACS buffer (1% bovine serum albumin in PBS) and resuspended in PBS.
Isotype-matched controls were included for each labelling. When the acquisition could not be performed immediately, the samples were fixed with 2% paraformaldehyde and analysed within 2 days. The labelled cells were analysed using a flow cytometer (FACS Calibur, Becton Dickinson, San Jose, USA). Data acquisition and analysis were done using the Cell Quest 3.1f software with consistent instrument settings for forward and side scatter, compensation and threshold throughout the study. A minimum of 10 000 gated events per tube were acquired. For all samples the gate was established using a forward scatter (FSC) greater than that of the platelets, dead and red blood cells and included the predominant population of cells, with regard to the small lymphocytes typical for small cell-lymphoma and the large cells characteristic for immunoblastic lymphoma. For each antibody, the proportion of cells with fluorescence greater than the negative control sample was recorded and used in the analysis.
PCR for antigen-receptor gene rearrangement
PCR for antigen-receptor gene rearrangement (PARR) assay was performed using the method described by Burnett et al.,7 in the case where the diagnosis was questionable, or when the flow cytometry analysis did not give a clear answer regarding the type of cancer cells.
In short, DNA from the sampled lymph node was isolated using a commercially available kit (DNA Mini Kit, Qiagen, Hilden, Germany). The primer pairs used are given in Table 2. For amplification, HotStarTaq Master Mix Kit (Qiagen) was used. The reaction mixture was prepared in a volume of 25 µL containing 100–200 ng of DNA, 1 µL of each primer, 1x buffer and nuclease free water. The PCR reactions were performed in t3thermocycler (Biometra, Goettngen, Germany). The samples were initially heated to 95 °C for 15 min (specific conditions for the activation of HotStarTaq, Qiagen). The following cycle: 94 °C for 8 s, 60 °C for 10 s and 72 °C for 15 s, was repeated 35 times. No final extension time was used.7 The reaction products, in a volume of 6 µL PCR, were analysed on 10% native polyacrylamide gels (1.5 mm thick) with 13 Tris-borate-EDTA buffer. A reaction was considered positive if one or more dominants or discrete bands were present. A negative reaction occurred when no band or diffuse smears were detected.
| Tube no. | Antigen | Fluorochrome | Tube no. | Antibody/Isotype control |
|---|---|---|---|---|
| 0 | PBS | PBS | ||
| 1 |
CD3 CD45 CD21 |
FITC PE Alexa Fluor® 647 |
2 |
mIgG1 mIgG2b mIgG1 |
| 3 |
CD3 CD4 CD8 |
FITC PE Alexa Fluor® 647 |
4 |
mIgG1 rIgG2a rIgG1 |
| 5 |
CD34 CD14 |
PE Alexa Fluor® 647 |
6 |
mIgG1 mIgG2a |
| 7 |
MHC II CD45R CD5 |
FITC PE Alexa Fluor® 647 |
8 |
rIgG2a rIgG2b rIgG2a |
| 9 |
CD3 CD79a |
FITC PE |
10 |
mIgG1 mIgG1 |
Results
Cell morphology
Using updated Kiel classification we diagnosed 55 cases of centroblastic lymphoma (41 monomorphic and 14 polymorphic), 6 immunoblastic, 5 pleomorphic of mixed, small and large cells, and 10 unclassifiable lymphomas from small lymphocytes. The most frequent type of B-cell lymphoma was the centroblastic lymphoma and among T-cell lymphomas the most common were pleomorphic and immunoblastic types. Owing to the lack of detailed cytology results, the correlation between immunophenotype and specific morphological subtypes was not analysed.
Immunophenotyping characterization of canine lymphoma
Forward and side scatter analysis allowed us to easily identify a predominant population of neoplastic cells in all examined FNA samples. Residual population of non-neoplastic cells of all cases could serve as an internal positive control.
The frequency of different types of neoplasms was 71% for B-cell and 17% for T-cell lymphomas. All cases of B-cell lymphomas were positive for CD79α, CD45 and CD45R. Fifty-seven of these cases (93%) were positive for CD21 and MHC class II. Eight cases (13%) were positive for CD5. Among B-cell lymphomas, in 8 (13%) cases the expression of haematopoietic precursor antigen CD34 was found. Unusual phenotypes were identified in 26 (39%) cases of B-cell lymphoma. The most common findings, beyond the expression of CD34, included a diminished or absent expression MHC class II (6%) and positive expression of CD8 (8%). Detailed results are shown in Table 3.
| No. of cases | CD3 | CD4 | CD5 | CD8 | CD21 | CD34 | MHCII | CD45 | CD45R | CD79A |
|---|---|---|---|---|---|---|---|---|---|---|
| 41 | − | − | − | − | + | − | + | + | + | + |
| 6 | − | − | − | − | + | + | + | + | + | + |
| 1 | − | − | − | − | + | + | − | + | + | + |
| 1 | − | − | − | + | + | + | + | + | + | + |
| 2 | − | − | + | + | + | − | + | + | + | + |
| 1 | − | − | + | − | − | − | − | + | + | + |
| 3 | − | − | + | − | + | − | + | + | + | + |
| 1 | − | − | − | + | − | − | − | + | + | + |
| 1 | − | − | − | − | + | − | − | + | + | + |
| 1 | − | − | + | − | + | − | + | − | + | + |
| 1 | − | − | − | + | + | − | + | + | + | + |
| 1 | − | − | − | − | − | − | + | + | + | + |
| 1 | − | − | + | + | − | − | + | + | + | + |
- a + indicates positivity in more than 65% of the neoplastic cells; − indicates negativity in more than 65% of the neoplastic cells.
T-cell lymphomas were diagnosed in 15 (17%) dogs and all these cells were positive for CD45 and CD3. CD45R was expressed in 13 (87%) of the examined samples. Twelve samples (80%) were positive for CD4 and 9 (60%) for CD5. No lymphomas derived from cytotoxic T-cells (CD8+) were found. Aberrant phenotypes were reported more often in T-cell lymphomas than in B-cell lymphomas, accounting for 65%. The lack of CD5 expression, found in six of the examined cases (40%), constituted the predominant aberrancy. In two samples, a co-expression of T- and B-cell markers (13% of T-cell lymphomas expressed also CD79α) were found. Owing to the acute, severe clinical course of the disease and poor response to treatment typical for T-cell lymphomas, both cases were classified as T-cell lymphoma.
PARR analysis confirmed the presence of clonal gene rearrangements of B- and T-cell receptors in all double positive samples. In three (20%) cases, the cells were negative for both CD4 and CD8, and three (20%) T-cell lymphomas were positive for MHC class II. Detailed results are shown in Table 4.
| No. of cases | CD3 | CD4 | CD5 | CD8 | CD21 | CD34 | MHCII | CD45 | CD45R | CD79A |
|---|---|---|---|---|---|---|---|---|---|---|
| 7 | + | + | + | − | − | − | − | + | + | − |
| 1 | + | + | + | − | − | − | − | + | − | − |
| 1 | + | + | − | − | − | − | − | − | − | + |
| 1 | + | − | − | − | − | − | − | + | + | + |
| 2 | + | − | − | − | − | − | + | + | + | − |
| 2 | + | + | − | − | − | − | − | + | + | − |
| 1 | + | + | + | − | − | − | + | + | + | − |
- a + indicates positivity in more than 65% of the neoplastic cells; – indicates negativity in more than 65% of the neoplastic cells.
Among 86 lymphoma cases, 10 (12%) were classified as a mixed or a null cell type (cells without CD3 and CD79a expression). T- or B-cell type of the cancer cells was not determined based on the expression of other than CD3 or CD79a cellular markers. Classification was made when the predominant population of cells could not be determined based on the expression of lineage-specific markers, or when the expression level did not exceed 65%.
PCR for antigen-receptor gene rearrangement
The detection of cell clonality within a tumour is crucial for distinguishing a lymphoma from reactive lymphoid proliferations. In lymphomas, a clonal rearrangement occurs when the same Variable Diverse Joining (VDJ) locus is present in each cell, and this can be easily detected by PCR.
The samples with questionable cytopathological diagnosis and in which the flow cytometry did not enable a clear identification of the cancer cell type, were subjected to PCR for PARR. Our study confirmed that the clonal rearrangement of BCR or TCR receptor occurred in most of the diagnosed B- and T-cell lymphomas. PARR assay was performed for a total number of 45 samples. Ten cases were finally classified as a mixed or a null cell type and 3314 were excluded from the study. In the remaining samples (n = 19), the PARR assay results corresponded to the cytometric diagnosis. The clonal gene rearrangements of both B- and T-cell receptors were found in two cases.
Discussion
Flow cytometry has been one of the most important diagnostic methods for a long time now, both in human and veterinary haematooncology.9, 10 In human medicine, flow cytometry has been used not only to classify different lymphoma subtypes, but also to predict an individual patient prognosis and response to treatment.11 The use of a wide range of well-defined antibodies provides an opportunity to characterize the canine leukocyte antigens and lymphoid population, present in different parts of the body.12
The results of our study show that the prevalence of high-grade B- and T-cell lymphoma is 71 and 17%, respectively. The obtained results coincide with other studies carried out in Poland, stating that lymphomas, classified as derived from B-cells (CD79+ and CD21+) accounted for 73% of cases, and those derived from T lymphocytes (CD3+ and CD4+ or CD8+) for 27% of cases13 (results limited to canine multicentric lymphoma). A similar incidence of T-cell lymphoma was found in North America by Greenlee et al.14 – 19% and Carter et al.15 – 25%, although their results were not limited to high-grade lymphoma alone. Similarity of the results between different research groups may be due to the fact that the low-grade lymphoma was found very rarely – only in 5.3% of cases as reported by Carter et al.15 The research carried out in Austria showed that B-cell lymphomas occurred in 51.2% and T-cell lymphomas in 29.3% of cases,16 but the low-grade lymphoma accounted for 13.4% of cases, 11% of which were classified as T subtype. Other corresponding results of T-cell lymphomas incidence at the level of 27% were found by Ferrer et al.17 According to Fournel-Fleury et al.,18 the frequency of T-cell lymphomas within the high-grade lymphoma is 9%. Higher incidence of T-cell lymphoma (results limited to a high-grade lymphoma), amounting to 31%, was found by Teske et al.,19 and the values found in Italian20 and French studies21 were 31 and 28%, respectively. Analyzing the results of these four groups,18-21 we can conclude that among the B subtype high-grade lymphomas represented 82 ± 6.7% of all B-cell lymphomas, whereas the frequency of high-grade T-subtype was less common (60 ± 21%) and varied within wider range.
Studies conducted in humans and animals have shown that lymphoma prevalence varies between regions. However, the geographic differences should be interpreted with caution due to a relatively small number of patients studied and due to the use of different antibodies or antibody panels.
Collectively, these previous studies show that humans and dogs are more likely to develop B-cell lymphomas and B-cells are subjected to a stronger genomic rearrangement, due to the production of immunoglobulins. This phenomenon makes them more prone to neoplastic changes and more able to remain longer in the bone marrow.22 The Boxer breed is at higher risk of developing lymphoma than other breeds. Some independent studies have showed that Boxers are especially predisposed to T-cell lymphoma, with the T-cell phenotype in up to 85% of cases.23-25 The chromosomal abnormalities, including the loss of canine chromosome 11 that can be observed in Boxers, may be the reason for a higher incidence of T-type lymphoma in this breed.
The presence of aberrant phenotypes is a common feature, both in human26 and veterinary medicine.10 The prevalence of aberrant patterns in humans differs according to studies and the type of lymphoid neoplasia. In the Hastrup et al.27 study, unusual phenotypes were found in 82% of cases, whereas Sanchez et al.28 reported these changes in 98% of B-cell lymphomas. In the veterinary medicine, such features are also often found.
The abnormalities involve the absence or lack of expression of typical leukocyte markers, co-expression of T- and B-cell markers or changes in the quantitative (high or low) antigen expression. Unusual phenotypes in dogs were reported in 58% of B-cell lymphomas and in 75% of T-cell lymphomas,29 thus corresponding with the results of our study (39 and 65%, respectively), although our results are limited to the high-grade form of lymphoma. The most common aberrant findings in dogs include the co-expression of T- and B-cell markers (CD3 and CD79α or CD21), observed in two cases in our study, aberrant expression of CD34 (eight cases in our study), co-expression of CD4 and CD8 in T-cell lymphomas and lack of expression of CD45 or CD18.8, 16, 29
The study investigating the importance of CD18/CD45 expression in neoplastic cells was performed by Comazzi et al..30 This work shows that the ratio between the median fluorescence intensity for CD18 and CD45, for lymphoid and granulocytic populations, could be significantly decreased in the case of lymphoma. Possibly, this parameter may be of a prognostic value in the future canine lymphoma diagnostics.
The lack of MHC class II expression, linked by Rao et al.31 to poor prognosis, was found in 6% of B-cell lymphomas in our study. They showed that patients with low MHC class II expression were 2.9 times more likely to die than those with high antigen expression. At the same time, the presence of CD34 was not associated with clinical prognosis.
The PCR for PARRs is known as a useful and sensitive method for detecting the clonality within the lymphoid neoplasms. The results of our study show the co-expression of both B- and T-cell markers in two cases. The co-expression of CD3 and CD79 was found earlier, in both humans32 and dogs.8 Some authors explain this phenomenon by the lack of specificity of the pan-T-cell markers used, however, our results were confirmed by PARR assay, in which the clonal gene rearrangements of both B- and T-cell receptors were found. Such abnormality, detected by PARR, was also described both in human and veterinary medicines. One reason for such a result, apart from a true cross lineage rearrangement, may be the presence of two co-existing neoplastic populations, a phenomenon showed by Burnet et al..7 In human cross-lineage, the TCR gene rearrangements could occur in B-cell malignancies,33 and the frequency of this phenomenon is up to 90%. Gelain et al.29 have also found such rearrangements in T-cell lymphomas, with accompanying phenotype aberrancies like CD4 and CD8 or CD3 and CD79 double positive cells, and have concluded that Ig and TCR gene rearrangements could not be necessarily restricted to B or T cells.
Conclusion
By using additional diagnostic techniques such as flow cytometry or PCR it is possible to accurately determine the type of cancer cells, which may be important in the development of new therapeutic strategies for the treatment of lymphoma in dogs.
As the dog is a suitable model for the study of human NHL, such studies may also be valuable for human medicine.
Acknowledgements
This work was supported by Grant No. DEC-2011/01/N/NZ5/02833 from the National Science Centre, Poland. We would like to thank Wojciech Hildebrand, PhD, from the Clinic of the Diseases of Horses, Dogs and Cats, Faculty of Veterinary Medicine, Wroclaw University of Environmental and Life Sciences, Poland, for providing us with the material for this study. The authors wish to thank Anne Avery for her consultation in analysis of double positive (B/T) cells.
Conflict of interest
None of the authors has any financial or personal relationships that could inappropriately influence or bias the content of this article.
