**Prospective Study of Tumor Markers as Prognostic Factors in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines**

Anna M. Badowska-Kozakiewicz *Department of Biophysics and Human Physiology, Medical University of Warsaw, Poland* 

#### **1. Introduction**

198 A Bird's-Eye View of Veterinary Medicine

Tomsa, K. , Major, A. & Galus , T. (2011). Treatment of atresia ani type I by balloon

Tsioli, V., Papazoglou, L., Anagnostou, T., Kouti, V. & Papadopoulou, P. (2009). Use of a

Van den Broek, A., Else, R. & Hunter, M. (1988). Atresia ani and urethrorectal fistula in a kitten. *Journal of Small Animal Practice*, Vol. 29, No. 2, (February 1988), pp. 91-94 Vianna, M. & Tobias, K. (2005). Atresia ani in the dog: a retrospective study. *Journal of the* 

Webb, G., McCord, K. & Twedt, D. (2007). Rectal stricture in 19 dogs: 1997-2005. *Journal of* 

No. 6, (June 2011), pp. 277 -280

317-322

2007), pp. 332-336

*Surgery*, Vol. 11, No. 12, (Dec 2009), pp. 1011-1014

dilatation in 5 kittens and one puppy. *Schweizer Archiv fur Tierheilkunde*, Vol. 153,

temporary incontinent end-on colostomy in a cat for the management of rectocutaneous fistulas associated with atresia ani. *Journal of Feline Medicine and* 

*American Animal Hospital Association*, Vol. 41, No. 5. (September-October 2005), pp.

*the American Animal Hospital Association*, Vol. 43, No. 6, (November-December

Cancer constitutes a major problem in animal pathology and is a subject of intensive research. The aim of this study was to gain comprehensive knowledge of cancer biology. Some animal tumors are also considered good research models in comparative oncology. Literature and own studies show that dogs often suffer from skin cancer and mammary gland tumors (Moulton, 1990). The incidence of mammary gland tumors in female canines is second only to skin cancer (Misdorp & Meuten, 2002; Ramalho et al, 2006). They most commonly appear between 6 and 10 years of age (Misdorp & Meuten, 2002; Hampe & Misdorp, 1974; Kubiak, 200). There are also documented cases of such tumors as early as at the age of two years and, very rarely, in males (Moulton, 1990). Histopathological examination in about 40-50% of cases shows malignant tumors, most of which are of epithelial origin (Rutteman, 2001). These cancers originate from epithelial follicles or ducts and take forms of papillary adenocarcinomas, simple or complex tubular and solid cancers (Misdorp & Meuten, 2002). There are also tumors derived from myoepithelial cells and mesenchymal tissue. In veterinary medicine there is a constant search for prognostic factors and predictors that allow for proper evaluation of the disease state, survival, susceptibility to treatment and risk of relapse. Recognized factors are: age of the animal, sentinel lymph node status, histological type of tumor and a histological grade of differentiation of malignancy. Beside the fundamental method in cancer diagnosis, which is the histopathological examination, additional information useful in this regard is provided by research on cancer biology, including, among others, the study of tumor markers. In histological examination of a tumor, markers of proliferative activity such as mitotic index, thymidine labeling index, percentage of cells in the S phase of cell cycle, expression of nuclear antigen Ki-67 and proliferating cells nuclear antigen (PCNA) are also taken into account. The following factors are also examined: expression of estrogen receptors, growth factor receptors such as: epidermal growth factor receptor (EGFR), insulin growth factor receptor (IGFR), growth hormone (GH),markers of high risk of metastasis: plasminogen activator and cathepsin D, oncogenes and tumor suppressor genes: c -erbB-2, c-myc, p53,

Prospective Study of Tumor Markers as Prognostic Factors

**2.2 Methods** 

statistically significant at P ≤ 0.05.

**2.3 Results** 

in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines 201

Then, they were stained using proper methods. In sections stained with the routine H&E method, the following determinations were carried out: type of neoplasm (Misdorp et al, 1999), tumor grade including tubule formation, intensity of division and the degree of neoplastic cell differentiation (Misdorp & Meuten, 2002), mitotic index defined as a mean number of mitoses in neoplastic cells counted in 10 fields of vision at the lens magnification of 400x (surface field 0.17mm2). Paraffin sections on slides covered with 2% saline solution in acetone, at the temperature of 420C were used for immunohistochemical methods. In immunohistochemical reactions, the following antibodies, properly diluted in 1% BSA (Sigma), were used: mouse monoclonal antibody against human nuclear antigen Ki-67 (Dako) diluted 1:75 (Nieto et al, 2000), mouse monoclonal antibodies against p53 (Dako) human protein, diluted at a ratio 1:25 (Gamblin et al, 1994; Rodo, 2007), mouse monoclonal antibodies against alpha (Dako) human estrogen receptor, diluted at a ratio 1:35 (Mulas et al, 2005), mouse monoclonal antibodies against COX-2 (Dako) human receptor, diluted at a ratio 1:100 (Queiroga et al, 2007; Doré et al, 2003; Soslow et al, 2000), mouse monoclonal antibodies against Hsp70 (Novocastra) human heat shock proteins, diluted at a ratio 1:40 (Romanucci et al, 2006), mouse monoclonal antibodies against Hsp90 (Novocastra) human heat shock proteins, diluted at a ratio 1:40 (Romanucci et al, 2006 ) and mouse monoclonal antibodies against P-glycoprotein (Sigma), diluted at a ratio 1:100 (Petterino et al, 2006). Sections were deparafinized in xylene and rehydrated in increasing concentrations of alcohol. Then, they were placed in a buffer with a pH of 6 (Dako). In order to uncover the antigen epitope, sections were warmed up in a microwave oven (1x5 min at 600 W, 2x5 min at 300 W). Then, they were cooled for 20 min. After two washings in distilled water (5 min), slides were washed in TRIS-buffered saline with pH of 8 (Sigma) and incubated with the primary antibody for 1h at room temperature. Then, the preparations were washed with TRIS buffer for 10 min. The En Vision + TM System (Dako) was used for visualization. After 30 min of incubation with reagent, the slides were washed with TRIS buffer and a solution of diaminobenzidine (DAB, Dako), prepared according to the procedure supplied by the producer, was put on slides. The degree of slide staining was controlled. They were washed in tap water and stained with Ehrlich hematoxylin for 5 min, contrasted in 1% acid alcohol and washed again in tap water. Then, slides were dehydrated in graded alcohol concentrations, passed through xylene and fixed in DPX mounting medium (Gurr®). Computer image analysis and Lucia v. 4.21 software were used for interpretation of the results of Ki-67, ER, p53, COX-2, Hsp70, Hsp90 expression; using those facilities we could count the number of neoplastic cells featuring stained cytoplasm per 1,000 neoplastic cells. Results were analyzed using the SPSS 12.0 program. To determine whether differences for a few independent traits were significant, Kruskal-Wallis test was used. This test is an equivalent to the test of variance for traits without normal distribution. Two-sided correlations were performed using Spearman correlation test. The differences were deemed

Investigated material contained 14 adenomas (Fig.1), 66 complex carcinomas (adenocarcinomas), 47 simple carcinomas (adenocarcinomas) (Fig. 2, Fig. 3) and 6 solid carcinomas. The number of cancers with a defined grade amounted to, respectively, 1st grade – 48, 2nd grade – 39 and 3rd grade – 32. Mammary gland neoplasms were excised from

female dogs belonging to 18 breeds in the ages between 3 and 16 years.

BRCA-1, BRCA-2, markers of tumor angiogenesis, and heat shock proteins (Koda et al, 2004; Niwińska, 1995; Olszewski, 1994). Diagnosis of neoplasms of the mammary gland in canines does not usually pose a problem, but in some cases there may be difficulties. Targeted diagnostic immunohistochemistry using a panel of antibodies makes it easier to determine the correct diagnosis. Difficulties encountered during routine diagnosis of neoplasms of the mammary gland in female dogs relate to assessing the origin of the tumor and its biology. Immunophenotyping can be helpful in the diagnostic process and tumor markers (expression of nuclear antigen Ki-67 and expression of estrogen receptors and the p53 protein) carry useful information, but determination of malignancy is still based on morphological criteria, such as mitotic index, which together with the histopathological criteria are the most important factors differentiating benign tumors from those of malignant nature;. In veterinary oncology there is a constant search for new tumor markers that could aid in differential diagnosis of mammary gland neoplasms in female canines. The aim of research on the mammary gland neoplasms in female dogs is to extend the panel of tumor markers by addition of new markers such as COX-2, P-gp, Hsp90 and Hsp70 and to incorporate them into routine diagnostics. Occurrence of mammary tumors in females as well as potential to use canine models in comparative pathology of these tumors makes it an interesting research material. It is known that the degree of malignancy is positively correlated with proliferative activity. Relationship between the expression of estrogen receptors and the degree of malignancy is not entirely clear, although most authors believe that expression of these receptors is greater in benign tumors. COX-2 is expressed in many canine tumors such as canine kidney, bladder, prostate, and mammary gland cancers, but there is no data on the relationship between expression of COX-2 and other markers, even though such relation would seem logical because the expression of COX-2 may be related to prognosis by indirect immunosuppressive effect and inhibition of apoptosis. Literature on the expression of heat shock protein or glycoprotein - P in breast cancer is very scarce. There is evidence of Hsp27, Hsp70 and Hsp90, as well as glycoprotein – P expression, which may influence susceptibility to therapy. One would expect a correlation between proliferative activity, degree of malignancy and expression of estrogen receptors, but no data is available on such relationship. The aim of the study was to analyze the role of classical and new tumor markers in the histopathological differential diagnosis of mammary gland neoplasms in female dogs. The objective of the present study was to investigate the expression of cyclooxygenase - 2 and heat shock proteins in canine mammary gland tumors and their correlations with histological type of tumor, degree of its histological malignancy, proliferative activity, estrogen receptor expression, infiltration by cells, as well as P-gp and p53 protein expression.

#### **2. Materials, methods and results**

#### **2.1 Materials**

Material for the investigation comprised 133 tumor samples of the mammary gland collected from female canines during surgical procedures performed at Warsaw Veterinary Clinics and Small Animal Clinic of the Department of Clinical Sciences, Faculty of Vetrinary Medicine, Warsaw University of Life Sciences – SGGW. Tumor samples were fixed in 8% formalin buffered with phosphates. After 24h fixing, material was dehydrated in a series of increasing concentrations of alcohol and embedded in paraffin. Paraffin blocks were cut into serial sections of 4µm in thickness.

#### **2.2 Methods**

200 A Bird's-Eye View of Veterinary Medicine

BRCA-1, BRCA-2, markers of tumor angiogenesis, and heat shock proteins (Koda et al, 2004; Niwińska, 1995; Olszewski, 1994). Diagnosis of neoplasms of the mammary gland in canines does not usually pose a problem, but in some cases there may be difficulties. Targeted diagnostic immunohistochemistry using a panel of antibodies makes it easier to determine the correct diagnosis. Difficulties encountered during routine diagnosis of neoplasms of the mammary gland in female dogs relate to assessing the origin of the tumor and its biology. Immunophenotyping can be helpful in the diagnostic process and tumor markers (expression of nuclear antigen Ki-67 and expression of estrogen receptors and the p53 protein) carry useful information, but determination of malignancy is still based on morphological criteria, such as mitotic index, which together with the histopathological criteria are the most important factors differentiating benign tumors from those of malignant nature;. In veterinary oncology there is a constant search for new tumor markers that could aid in differential diagnosis of mammary gland neoplasms in female canines. The aim of research on the mammary gland neoplasms in female dogs is to extend the panel of tumor markers by addition of new markers such as COX-2, P-gp, Hsp90 and Hsp70 and to incorporate them into routine diagnostics. Occurrence of mammary tumors in females as well as potential to use canine models in comparative pathology of these tumors makes it an interesting research material. It is known that the degree of malignancy is positively correlated with proliferative activity. Relationship between the expression of estrogen receptors and the degree of malignancy is not entirely clear, although most authors believe that expression of these receptors is greater in benign tumors. COX-2 is expressed in many canine tumors such as canine kidney, bladder, prostate, and mammary gland cancers, but there is no data on the relationship between expression of COX-2 and other markers, even though such relation would seem logical because the expression of COX-2 may be related to prognosis by indirect immunosuppressive effect and inhibition of apoptosis. Literature on the expression of heat shock protein or glycoprotein - P in breast cancer is very scarce. There is evidence of Hsp27, Hsp70 and Hsp90, as well as glycoprotein – P expression, which may influence susceptibility to therapy. One would expect a correlation between proliferative activity, degree of malignancy and expression of estrogen receptors, but no data is available on such relationship. The aim of the study was to analyze the role of classical and new tumor markers in the histopathological differential diagnosis of mammary gland neoplasms in female dogs. The objective of the present study was to investigate the expression of cyclooxygenase - 2 and heat shock proteins in canine mammary gland tumors and their correlations with histological type of tumor, degree of its histological malignancy, proliferative activity, estrogen receptor expression, infiltration by cells, as well as P-gp and p53

Material for the investigation comprised 133 tumor samples of the mammary gland collected from female canines during surgical procedures performed at Warsaw Veterinary Clinics and Small Animal Clinic of the Department of Clinical Sciences, Faculty of Vetrinary Medicine, Warsaw University of Life Sciences – SGGW. Tumor samples were fixed in 8% formalin buffered with phosphates. After 24h fixing, material was dehydrated in a series of increasing concentrations of alcohol and embedded in paraffin. Paraffin blocks were cut into

protein expression.

**2.1 Materials** 

**2. Materials, methods and results** 

serial sections of 4µm in thickness.

Then, they were stained using proper methods. In sections stained with the routine H&E method, the following determinations were carried out: type of neoplasm (Misdorp et al, 1999), tumor grade including tubule formation, intensity of division and the degree of neoplastic cell differentiation (Misdorp & Meuten, 2002), mitotic index defined as a mean number of mitoses in neoplastic cells counted in 10 fields of vision at the lens magnification of 400x (surface field 0.17mm2). Paraffin sections on slides covered with 2% saline solution in acetone, at the temperature of 420C were used for immunohistochemical methods. In immunohistochemical reactions, the following antibodies, properly diluted in 1% BSA (Sigma), were used: mouse monoclonal antibody against human nuclear antigen Ki-67 (Dako) diluted 1:75 (Nieto et al, 2000), mouse monoclonal antibodies against p53 (Dako) human protein, diluted at a ratio 1:25 (Gamblin et al, 1994; Rodo, 2007), mouse monoclonal antibodies against alpha (Dako) human estrogen receptor, diluted at a ratio 1:35 (Mulas et al, 2005), mouse monoclonal antibodies against COX-2 (Dako) human receptor, diluted at a ratio 1:100 (Queiroga et al, 2007; Doré et al, 2003; Soslow et al, 2000), mouse monoclonal antibodies against Hsp70 (Novocastra) human heat shock proteins, diluted at a ratio 1:40 (Romanucci et al, 2006), mouse monoclonal antibodies against Hsp90 (Novocastra) human heat shock proteins, diluted at a ratio 1:40 (Romanucci et al, 2006 ) and mouse monoclonal antibodies against P-glycoprotein (Sigma), diluted at a ratio 1:100 (Petterino et al, 2006). Sections were deparafinized in xylene and rehydrated in increasing concentrations of alcohol. Then, they were placed in a buffer with a pH of 6 (Dako). In order to uncover the antigen epitope, sections were warmed up in a microwave oven (1x5 min at 600 W, 2x5 min at 300 W). Then, they were cooled for 20 min. After two washings in distilled water (5 min), slides were washed in TRIS-buffered saline with pH of 8 (Sigma) and incubated with the primary antibody for 1h at room temperature. Then, the preparations were washed with TRIS buffer for 10 min. The En Vision + TM System (Dako) was used for visualization. After 30 min of incubation with reagent, the slides were washed with TRIS buffer and a solution of diaminobenzidine (DAB, Dako), prepared according to the procedure supplied by the producer, was put on slides. The degree of slide staining was controlled. They were washed in tap water and stained with Ehrlich hematoxylin for 5 min, contrasted in 1% acid alcohol and washed again in tap water. Then, slides were dehydrated in graded alcohol concentrations, passed through xylene and fixed in DPX mounting medium (Gurr®). Computer image analysis and Lucia v. 4.21 software were used for interpretation of the results of Ki-67, ER, p53, COX-2, Hsp70, Hsp90 expression; using those facilities we could count the number of neoplastic cells featuring stained cytoplasm per 1,000 neoplastic cells. Results were analyzed using the SPSS 12.0 program. To determine whether differences for a few independent traits were significant, Kruskal-Wallis test was used. This test is an equivalent to the test of variance for traits without normal distribution. Two-sided correlations were performed using Spearman correlation test. The differences were deemed statistically significant at P ≤ 0.05.

#### **2.3 Results**

Investigated material contained 14 adenomas (Fig.1), 66 complex carcinomas (adenocarcinomas), 47 simple carcinomas (adenocarcinomas) (Fig. 2, Fig. 3) and 6 solid carcinomas. The number of cancers with a defined grade amounted to, respectively, 1st grade – 48, 2nd grade – 39 and 3rd grade – 32. Mammary gland neoplasms were excised from female dogs belonging to 18 breeds in the ages between 3 and 16 years.

Prospective Study of Tumor Markers as Prognostic Factors

pathologic mitotic figures

**histological type of tumor of epithelial origin** 

occurred most often (55.0%) (Table 2).

in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines 203

Fig. 3. Adenocarcinoma simplex, H&E (40x). The following mature tissues are visible:

Dogs were divided into three age groups: <8 years, 8 -12 years and >12 years. In the group of bitches below the age of 8, majority (61.1%) consisted of tumors with the lowest histological grade of malignancy (1st). In the oldest group, 1st and 2nd grade tumors in the 1st and 2nd accounted for 77.8%. In the entire pool of studied tumors in all age groups, the largest share consisted of tumors with the lowest degree of histological malignancy (40.4%) (Table 1). Assessment of the contribution of individual types of tumors at different ages showed that in bitches younger than 8 years the most common findings were adenomas (21.7%) and complex carcinomas (56.5%) and in those over 12 years simple carcinomas

Age of bitches Tumor grade Total

<8 lat (n=18) 11 (61,1%) 3 (16,7%) 4 (22,2%) 18 8-12 lat (n=83) 30 (36,0%) 29 (35,0%) 24 (29,0%) 83 >12 (n=18) 7 (38,9%) 7 (38,9%) 4 (22,2%) 18 total (n=119) 48 (40,4%) 39 (32,8%) 32 (26,8%) 119 Table 1. Incidence of malignancies of various grades in bitches in different age groups

Io IIo IIIo

**2.3.1 Relationship between age of bitches and the grade of malignancy and** 

Fig. 1. Adenoma, H&E method x 20

Fig. 2. Adenocarcinoma simplex, H&E method x 20

Fig. 1. Adenoma, H&E method x 20

Fig. 2. Adenocarcinoma simplex, H&E method x 20

Fig. 3. Adenocarcinoma simplex, H&E (40x). The following mature tissues are visible: pathologic mitotic figures

#### **2.3.1 Relationship between age of bitches and the grade of malignancy and histological type of tumor of epithelial origin**

Dogs were divided into three age groups: <8 years, 8 -12 years and >12 years. In the group of bitches below the age of 8, majority (61.1%) consisted of tumors with the lowest histological grade of malignancy (1st). In the oldest group, 1st and 2nd grade tumors in the 1st and 2nd accounted for 77.8%. In the entire pool of studied tumors in all age groups, the largest share consisted of tumors with the lowest degree of histological malignancy (40.4%) (Table 1). Assessment of the contribution of individual types of tumors at different ages showed that in bitches younger than 8 years the most common findings were adenomas (21.7%) and complex carcinomas (56.5%) and in those over 12 years simple carcinomas occurred most often (55.0%) (Table 2).


Table 1. Incidence of malignancies of various grades in bitches in different age groups

Prospective Study of Tumor Markers as Prognostic Factors

in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines 205

Fig. 5. Adenocarcinoma simplex, expression of Ki-67, immunohistochemical method. (40x)

In most cases, infiltrates were located predominantly in the stroma at the periphery of the tumor, and, in rare cases, occurred in the glandular alveoli. Inflammatory infiltrates consisted of lymphocytes, neutrophils and macrophages. The intensity of cellular infiltration was assessed on a four-level scale:: -/+; +; ++; +++. Statistical analysis indicates no correlation between the age of the dog and the intensity of cellular infiltration in the tumor. It is worth noting that, in bitches between the ages of 8 to 12 years, most tumors exhibited the intensity of infiltration at the first (+) level , which constituted 50% of tumors in this age group. A significant relationship was found between the intensity of cellular infiltration and the type of tumor. In 50% of solid cancers, there was no cellular infiltration. Only 33% of these tumors exhibited the presence of cellular infiltration at the first level (+), and 16.7% on the second level (++). Among the complex cancers, the largest group consisted of those, in which cellular infiltration was found at the first (+) and the second (++) level. The highest average intensity of cellular infiltration was found in complex carcinomas, followed by simple carcinomas, adenomas and solid carcinomas. Specific differences were seen between adenomas and complex carcinomas (P=0.007) and between solid carcinomas and complex carcinomas (P=0.005). While examining the relationship between intensity of cellular infiltration and histological grade of malignancy, no significant differences were revealed. But it may be noted that among tumors of all histological grades of malignancy, the largest group consisted of cancers with the first level of cellular infiltration. While analyzing the average intensity of cellular infiltration in tumors with various histological grades of malignancy no significant differences were found between groups with the exception of the

**2.3.3 Results of cellular infiltration in tumors of epithelial origin** 

tumors in the 3rd grade, where a significant difference was observed (P=0.023).


Table 2. Occurrence of individual types of epithelial neoplasms in different age groups in bitches

#### **2.3.2 Results of proliferative activity**

The value of mitotic index differed significantly between types of tumors. The lowest proliferative activity was observed in adenomas, the highest in simple and solid carcinomas. The highest proliferative activity was found in tumors in the 3rd grade of malignancy, the lowest in the tumors in the 1st grade (Fig.4). Expression of Ki-67 protein was observed in the nuclei of neoplastic cells that have undergone division (Fig.5). Statistical analysis shows significant differences between particular types of tumors. The lowest number of cells exhibiting Ki-67 protein expression was observed in adenomas, the highest in the solid and simple carcinomas and in tumors with the highest histological grade of malignancy (3rd).

Fig. 4. Average number of cells showing Ki-67 expression depending on the tumor grade. Letters (a, b, c) above the columns show that the differences between mean values were statistically significant (P≤0.05).

<8 lat 5 (21,7%) 1 (4,3%) 4 (17,4%) 13 (56,5%) 23 8-12 lat 7 (7,8%) 4 (4,4%) 32 (35,6%) 47 (52,2%) 90 >12 lat 2 (10,0%) 1 (5,0%) 11 (55,0%) 6 (30,0%) 20 Total 14 (10,5%) 6 (4,5%) 47 (35,3%) 66 (49,6%) 133 Table 2. Occurrence of individual types of epithelial neoplasms in different age groups in

The value of mitotic index differed significantly between types of tumors. The lowest proliferative activity was observed in adenomas, the highest in simple and solid carcinomas. The highest proliferative activity was found in tumors in the 3rd grade of malignancy, the lowest in the tumors in the 1st grade (Fig.4). Expression of Ki-67 protein was observed in the nuclei of neoplastic cells that have undergone division (Fig.5). Statistical analysis shows significant differences between particular types of tumors. The lowest number of cells exhibiting Ki-67 protein expression was observed in adenomas, the highest in the solid and simple carcinomas and in tumors with the highest histological grade of malignancy (3rd).

Fig. 4. Average number of cells showing Ki-67 expression depending on the tumor grade. Letters (a, b, c) above the columns show that the differences between mean values were

Adenoma Carcinoma

**2.3.2 Results of proliferative activity** 

statistically significant (P≤0.05).

solidum

Types of tumors

Adenocarcinoma simplex

Total

Adenocarcinoma complex

Age of bitches

bitches

Fig. 5. Adenocarcinoma simplex, expression of Ki-67, immunohistochemical method. (40x)
