**3. Discussion and conclusion**

#### **3.1 Discusion**

In our study, cancers accounted for 89.8% of tumors of epithelial origin. This is consistent with the results of some authors, although the results of studies on the incidence and types of tumors of the mammary gland in canine females given by various authors differ.

Fig. 26. Average number of cells showing Hsp90 expression depending on the tumor grade

Fig. 27. Average number of cells expressing Hsp90, depending on the degree of

between means was statistically significant (P≤0.05).

**3. Discussion and conclusion** 

**3.1 Discusion** 

cyclooxygenase – 2 expression. Letters (a, b) above the columns show that the difference

In our study, cancers accounted for 89.8% of tumors of epithelial origin. This is consistent with the results of some authors, although the results of studies on the incidence and types of tumors of the mammary gland in canine females given by various authors differ. Löhr (1997) found that 50% of the mammary gland tumors in female canines are malignant tumors (Löhr et al, 1997). Hellmén (1993) determined the incidence of malignant tumors to be 68% (Hellmén et al 1993). There are also studies showing, that about a half of tumors are benign (Bostock et al, 1992; Gilbertson et al, 1983). Moulton (1990) believes that benign tumors represent about 80% of cases and the majority of tumors studied are "benign mixed tumors" (65%) (Moulton, 1990). They are tumors in which, apart from epithelial and mesenchymal tissue, there is also cartilage and bone tissue. Their histopathological examination reveals no sign of malignancy. Nerurkar (1990) says that benign tumors represent about 27% of the mammary gland tumors in female dogs (Nerurkar et al, 1990). In our study, benign tumors of epithelial origin accounted for only 10.1% of all cases. The diversity of these results relates to the lack of uniform diagnostic criteria and lack of uniform classification of tumors of the mammary gland in dogs. In human medicine, age is a very important prognostic factor. Detection of breast cancer in women at a very young age and at the age over 60 is associated with worse prognosis (Host & Lund, 1986). In Philibert's (2003) studies there were no significant differences in survival of young and older dogs (Philibert et al, 2003). Hellmén (1993) presented completely different results (Hellmén et al, 1993). She found that age may be an important prognostic factor and showed that older bitches had a shorter survival time after surgery. Results obtained by Benjamin corresponded to that (1999) (Benjamin et al, 1999). Own results support the concept of age as an important prognostic factor, as there is very little information regarding survival or recurrence of malignancy in female dogs after surgery. Current study shows that the presence or absence of cellular inflammatory infiltrates may be a prognostic factor. Data on the incidence of cellular infiltration in mammary gland tumors in female dogs is also scarce. Gilbertson (1983) found that cell infiltration plays a role in the process of development of precancerous and invasive carcinomas (35%) (Gilbertson et al, 1983). In our study, the percentage of tumors with cellular infiltration was higher (88%). The highest intensity of cellular infiltration was observed in tumors with a high histological grade of malignancy. Skrzypczak (2004) (Skrzypczak, 2004) presented similar results in his study . Authors believe that presence of cellular infiltration plays a positive role in inhibiting tumor growth. Some studies have shown that presence of cellular infiltration is associated with good (Rilke et al, 1991) or poor (Parl & Dupont, 1982) prognosis and others, that it carries no prognostic value. In our study, no relationship was found between cellular infiltration and other tumor markers, consistent with the studies by Rodo (2007) (Rodo, 2007; Roses et al, 1982). It is worth noting the presence of necrosis in tumor foci, which may obscure the accuracy of the results. Necrosis in the tumor is the result of disparities between high proliferative activity and tumor vascularization. It may trigger cellular inflammatory reactions. We also analyzed the distribution of cellular infiltration. Our own results differ from the results of other authors, because infiltration was observed in the stroma, not scattered (Lee et al, 1996), and was more pronounced on the periphery of the tumor. The role of cellular infiltration within the tumor remains unclear and controversial. Estrogen receptors are recognized markers in the diagnosis of breast cancer in women. It is estimated that about 70-80% of breast cancers in women exhibit the expression of estrogen receptors. These tumors are characterized by slower growth, higher diversity, better prognosis with a suitable treatment regimen and correlate with the length of survival after

Prospective Study of Tumor Markers as Prognostic Factors

in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines 221

confirmed that the expression of Ki-67 is high in tumors with a higher histological grade of malignancy (Nieto et al, 2000). Similar results were obtained by Peña (1998), Giziński (2003) and Szczubiał (2002) (Peña et al, 1998; Giziński et al, 2003; Szczubiał & Łopuszański, 2002). They studied the expression of nuclear antigen Ki-67 in mammary gland tumors in dogs and came to the conclusion that higher expression of Ki-67 is an important prognostic factor and is associated with a higher risk of metastasis, a shorter period to recurrence and a shorter overall survival period. The rate of tumor growth is influenced by factors related to inhibition of cell cycle and promotion of apoptosis of cancer cells., Under physiological conditions it is a function of, among others, p53 gene and its protein product. Mutation of the p53 gene plays an important role in the uncontrolled proliferation and resistance to apoptosis in cancer cells , leading to functional changes in proteins. In dogs, expression of the p53 gene was found in both benign (Muto et al, 2000) and malignant (Mayr et al, 1994; Veldhoen et al, 1999) tumors of mammary gland. Muto (2000) in his study showed no association between p53 expression and histological type of cancer, but claimed that this protein may play an important role in carcinogenesis and may be a negative prognostic factor (Muto et al, 2000). Chung-Ho (2004) came to a similar conclusion, arguing that the presence ofp53 in tumor cells is associated with their malignancy and bad prognosis in mammary gland tumors (Chun-Ho et al, 2004). In our study, p53 expression was observed in benign as well as malignant tumors. Among all tumors, more than 50% exhibited a positive reaction for 53 protein. It contradicts the data obtained in his study by Gamblin (1997), who carried out an analysis of 16 adenocarcinomas and found a positive reaction for p53 protein in only 12.5% of cases (Gamblin et al, 1994). In our study, the highest expression of p53 was observed in complex carcinomas and in 1st and 2nd grade malignant tumors. These results are consistent with the results obtained by Rungsipipata (1999), who found the highest levels of p53 expression in simple and complex carcinomas and also showed expression of this protein in adenomas (16%) (Rungspipata, 1999). Rodo (2007) obtained results on the expression of p53 , stating that there is a positive correlation between proliferative activity and the number of cells expressing p53 protein (Rodo, 2007). In our study, we found no statistically significant differences between the expression of p53 and nuclear antigen Ki-67. Obtained data in part explain why the activation of oncogenes does not always lead to uncontrolled proliferation when normal signal transduction leads to stabilization of p53 and activation of programmed cell death. There is a need for further research in this area, because the data in the literature is ambiguous and does not allow for a definite conclusion as to the importance of p53 protein and its role in mammary gland neoplasms. Varying results undermine the claim that a positive reaction for p53 protein would predict worse prognosis in cancer and that p53 mutation may be responsible for increased proliferation in tumors with advanced malignancy. Research on COX-2 expression in canine mammary cancers is scarce (Heller et al, 2005; Nowak et al, 2005; Doré et al, 2003) despite it being an attractive and motivating topic. Overexpression of COX-2 is known to occur in 56-100% of canine mammary carcinomas (Doré et al, 2003; Millanta et al, 2005; Queiroga et al, 2007; Heller et al, 2005; Mohammed et al, 2004), but there is a marked variation in the percentage of COX-2-positive tumor cells and the intensity of its expression. Doré et. al (2003) did not find any COX-2 expression in four samples of normal mammary tissue (Doré et al, 2003), whilst Mohammed et. al (2004) (Mohammed et

surgical removal (Bacus et al, 1989; Barzanti et al, 2000). Study of the expression of estrogen receptors in mammary gland neoplasms in female canines was, as usual, inconclusive. Martin et. al (1984) studied 228 tumors of the mammary gland in bitches and showed the expression of estrogen receptors in only 2.1% of tumors (Martin et al, 1984). Pena et. al (1998), on the other hand, diagnosed 21 cancers with an ongoing inflammatory process and found no expression of estrogen receptors (Pena et al, 1998). According to Sartin et. al (1992) the greatest chance of long-term survival after surgery , is associated with tumors expressing ER alone or together with PgR. Indeed, in the absence of ER and PgR, researchers observed the shortest period of survival (Sartin et al, 1997). Millanta et. al (2005) study on 47 mammary gland neoplasms in bitches reported that ER and PgR expression did not correlate with survival or histological parameters of tumors (Millanta et al, 2005). Similarly, Sobczak-Filipiak and Malicka (1997) have shown no correlation between expression of ER and mitotic index (Sobczak-Filipiak & Malicka, 1997). In our study, the expression of estrogen receptors was demonstrated in 40% of the tested tumors. The highest expression of estrogen receptors was found in simple carcinomas and tumors with the highest histological grade of malignancy. There was a significant correlation between mitotic index and expression of estrogen receptors. Nowak (2007) found expression of estrogen receptors in only 6% of cancers (Nowak et al, 2007), but Mulas (2005) obtained different results- he found that estrogen receptors were expressed in benign tumors (Mulas et al, 2005). Similarly, McEwen et. al (1982) have demonstrated the expression of estrogen receptors in about 50% of cases, with a significantly higher levels of expression present in benign tumors (McEwen et al, 1982). Results obtained by Nieto et. al (2000) are different - they show a correlation between the expression of nuclear antigen Ki-67 and estrogen receptors (Nieto et al, 2000). Authors of these studies observed the highest levels of ER expression in simple and complex carcinomas, as we did in our study. Otherwise, Rutteman (1998), Sartin (1992), and Geraldes (2000), claim that high expression of ER is present in adenomas compared to adenocarcinomas, which exhibit a lower level of expression of this marker (Sartin et al, 1992; Rutteman et al, 2001; Geraldes et al, 2000). Opinions about the value of estrogen receptors as a prognostic factor are divided. Some authors consider ER expression a positive factor, but there are also voices postulating it is a negative prognostic marker. Interestingly, it seems that in our study, low expression of ER negatively correlated with high expression of nuclear antigen Ki-67. It is in agreement with the research done by Peña (1998), who led the study on the mammary gland neoplasms in canines (Peña et al, 1998). From our research we conclude, that the expression of estrogen receptors may be important in assessment of malignancy, but does not show significant correlation with other markers. An important marker of malignancy is the proliferative activity. Proper evaluation of proliferative activity of tumor cells is crucial for the evaluation of its biological activity and is used in determining the treatment of cancer. High mitotic index correlated with tumor size and presence of lymph node metastases (Niwińska, 1995; Mirecka et al, 1993). In our study, proliferative activity depended on both, the type of tumor and the degree of histological malignancy. The highest values of mitotic index were recorded in simple and solid carcinomas and in tumors with the highest histological grade of malignancy. Similar results were reported for the expression of Ki-67. The highest expression of Ki-67 was seen in solid, simple carcinomas and in 3rd grade tumors. Similar results were obtained by Nieto (2000), who

surgical removal (Bacus et al, 1989; Barzanti et al, 2000). Study of the expression of estrogen receptors in mammary gland neoplasms in female canines was, as usual, inconclusive. Martin et. al (1984) studied 228 tumors of the mammary gland in bitches and showed the expression of estrogen receptors in only 2.1% of tumors (Martin et al, 1984). Pena et. al (1998), on the other hand, diagnosed 21 cancers with an ongoing inflammatory process and found no expression of estrogen receptors (Pena et al, 1998). According to Sartin et. al (1992) the greatest chance of long-term survival after surgery , is associated with tumors expressing ER alone or together with PgR. Indeed, in the absence of ER and PgR, researchers observed the shortest period of survival (Sartin et al, 1997). Millanta et. al (2005) study on 47 mammary gland neoplasms in bitches reported that ER and PgR expression did not correlate with survival or histological parameters of tumors (Millanta et al, 2005). Similarly, Sobczak-Filipiak and Malicka (1997) have shown no correlation between expression of ER and mitotic index (Sobczak-Filipiak & Malicka, 1997). In our study, the expression of estrogen receptors was demonstrated in 40% of the tested tumors. The highest expression of estrogen receptors was found in simple carcinomas and tumors with the highest histological grade of malignancy. There was a significant correlation between mitotic index and expression of estrogen receptors. Nowak (2007) found expression of estrogen receptors in only 6% of cancers (Nowak et al, 2007), but Mulas (2005) obtained different results- he found that estrogen receptors were expressed in benign tumors (Mulas et al, 2005). Similarly, McEwen et. al (1982) have demonstrated the expression of estrogen receptors in about 50% of cases, with a significantly higher levels of expression present in benign tumors (McEwen et al, 1982). Results obtained by Nieto et. al (2000) are different - they show a correlation between the expression of nuclear antigen Ki-67 and estrogen receptors (Nieto et al, 2000). Authors of these studies observed the highest levels of ER expression in simple and complex carcinomas, as we did in our study. Otherwise, Rutteman (1998), Sartin (1992), and Geraldes (2000), claim that high expression of ER is present in adenomas compared to adenocarcinomas, which exhibit a lower level of expression of this marker (Sartin et al, 1992; Rutteman et al, 2001; Geraldes et al, 2000). Opinions about the value of estrogen receptors as a prognostic factor are divided. Some authors consider ER expression a positive factor, but there are also voices postulating it is a negative prognostic marker. Interestingly, it seems that in our study, low expression of ER negatively correlated with high expression of nuclear antigen Ki-67. It is in agreement with the research done by Peña (1998), who led the study on the mammary gland neoplasms in canines (Peña et al, 1998). From our research we conclude, that the expression of estrogen receptors may be important in assessment of malignancy, but does not show significant correlation with other markers. An important marker of malignancy is the proliferative activity. Proper evaluation of proliferative activity of tumor cells is crucial for the evaluation of its biological activity and is used in determining the treatment of cancer. High mitotic index correlated with tumor size and presence of lymph node metastases (Niwińska, 1995; Mirecka et al, 1993). In our study, proliferative activity depended on both, the type of tumor and the degree of histological malignancy. The highest values of mitotic index were recorded in simple and solid carcinomas and in tumors with the highest histological grade of malignancy. Similar results were reported for the expression of Ki-67. The highest expression of Ki-67 was seen in solid, simple carcinomas and in 3rd grade tumors. Similar results were obtained by Nieto (2000), who confirmed that the expression of Ki-67 is high in tumors with a higher histological grade of malignancy (Nieto et al, 2000). Similar results were obtained by Peña (1998), Giziński (2003) and Szczubiał (2002) (Peña et al, 1998; Giziński et al, 2003; Szczubiał & Łopuszański, 2002). They studied the expression of nuclear antigen Ki-67 in mammary gland tumors in dogs and came to the conclusion that higher expression of Ki-67 is an important prognostic factor and is associated with a higher risk of metastasis, a shorter period to recurrence and a shorter overall survival period. The rate of tumor growth is influenced by factors related to inhibition of cell cycle and promotion of apoptosis of cancer cells., Under physiological conditions it is a function of, among others, p53 gene and its protein product. Mutation of the p53 gene plays an important role in the uncontrolled proliferation and resistance to apoptosis in cancer cells , leading to functional changes in proteins. In dogs, expression of the p53 gene was found in both benign (Muto et al, 2000) and malignant (Mayr et al, 1994; Veldhoen et al, 1999) tumors of mammary gland. Muto (2000) in his study showed no association between p53 expression and histological type of cancer, but claimed that this protein may play an important role in carcinogenesis and may be a negative prognostic factor (Muto et al, 2000). Chung-Ho (2004) came to a similar conclusion, arguing that the presence ofp53 in tumor cells is associated with their malignancy and bad prognosis in mammary gland tumors (Chun-Ho et al, 2004). In our study, p53 expression was observed in benign as well as malignant tumors. Among all tumors, more than 50% exhibited a positive reaction for 53 protein. It contradicts the data obtained in his study by Gamblin (1997), who carried out an analysis of 16 adenocarcinomas and found a positive reaction for p53 protein in only 12.5% of cases (Gamblin et al, 1994). In our study, the highest expression of p53 was observed in complex carcinomas and in 1st and 2nd grade malignant tumors. These results are consistent with the results obtained by Rungsipipata (1999), who found the highest levels of p53 expression in simple and complex carcinomas and also showed expression of this protein in adenomas (16%) (Rungspipata, 1999). Rodo (2007) obtained results on the expression of p53 , stating that there is a positive correlation between proliferative activity and the number of cells expressing p53 protein (Rodo, 2007). In our study, we found no statistically significant differences between the expression of p53 and nuclear antigen Ki-67. Obtained data in part explain why the activation of oncogenes does not always lead to uncontrolled proliferation when normal signal transduction leads to stabilization of p53 and activation of programmed cell death. There is a need for further research in this area, because the data in the literature is ambiguous and does not allow for a definite conclusion as to the importance of p53 protein and its role in mammary gland neoplasms. Varying results undermine the claim that a positive reaction for p53 protein would predict worse prognosis in cancer and that p53 mutation may be responsible for increased proliferation in tumors with advanced malignancy. Research on COX-2 expression in canine mammary cancers is scarce (Heller et al, 2005; Nowak et al, 2005; Doré et al, 2003) despite it being an attractive and motivating topic. Overexpression of COX-2 is known to occur in 56-100% of canine mammary carcinomas (Doré et al, 2003; Millanta et al, 2005; Queiroga et al, 2007; Heller et al, 2005; Mohammed et al, 2004), but there is a marked variation in the percentage of COX-2-positive tumor cells and the intensity of its expression. Doré et. al (2003) did not find any COX-2 expression in four samples of normal mammary tissue (Doré et al, 2003), whilst Mohammed et. al (2004) (Mohammed et

Prospective Study of Tumor Markers as Prognostic Factors

in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines 223

significance between the two groups (Petterino et al, 2006). Furthermore, he found that simple and complex carcinomas were the most numerous groups among the examined cancers and that P-gp expression was absent from the tissues of healthy mammary glands. We achieved similar results in our research: we found the P-gp expression in 76% of the examined canine mammary cancers. The most numerous groups exhibiting a positive Pgp immunohistochemical reaction were composed of complex carcinomas (90.9%) and simple carcinomas (73%). High expression was found in cancers with the highest histological grade of malignancy. Due to fact that there are hardly any works investigating into the P-gp expression in canine mammary cancers, we attempted not only to confirm the presence of P-gp expression and location, but also to prove the relationship between P-gp and other neoplastic markers. Statistical analysis confirmed a positive correlation between P-gp and COX-2 expression: it demonstrated a statistical significance in the case of examined characteristics (P = 0.021). It is worth mentioning that the expression of two markers was significant in carcinomas featuring high histological grades of malignancy. Furthermore, P-gp expression was studied in canine lymphomas; additionally, attempts were made to conduct studies on carcinomas treated by means of chemotherapy; the results were compared with the results obtained in the control group of untreated carcinomas. Higher expression was observed in the treated cancer cells. On the other hand, increased expression in untreated cancers appears to be a negative prognostic factor, given that survival rate shall decrease in such cases (Lee et al, 1996). Due to the fact that this line of research has not brought about many works, there are several questions and ambiguities concerning the role of P-gp in the process of neoplastic genesis and the importance of this factor from the clinical point of view. Additionally, the question arises whether, or not P-gp can be classified into the significant prognostic marker category. We are not aware of defensive mechanisms of the cells equipped with membrane transporters against apoptosis generated by the compounds, that are not pump substrates. There are suggestions that the phenomenon may be connected with evacuation (pumping out) of a certain important mediator of apoptosis or with the impact of P-gp on the intracellular pH (Johnson et al, 1998). Identification of a mechanism of cell resistance to apoptosis is a key instrument in selection of suitable therapy. According to studies in humans, heat shock proteins may be important predictors of breast cancer (Park & Dupont, 1982). There is not much published data in the literature on heat shock protein expression in mammary gland tumors of female dogs. The role of these proteins in carcinogenesis has not been clearly defined either. Studies have only shown that expression of heat shock proteins takes place in canine breast cancers , but the linkage between these proteins and other tumor markers was not confirmed. Seymour (1990) studied endometrial cancer in women and found that heat shock proteins were useful markers in the diagnostics of these tumors (Saymour et al, 1990). Expression of Hsp27 was also investigated in breast cancer in women (Ciocca et al, 1993). A relationship between the expression of this protein and the degree of differentiation of the tumor cells was found. Similar studies were conducted by Storm (1996), who stated that tumors with Hsp27 expression showed a higher histological grade than tumors negative for the expression of this protein (Storm et al, 1993). Kumaraguruparan (2006) studied the expression of Hsp70 and Hsp90 in breast cancer in women and found a correlation between the expression of both proteins and proliferative activity (Kumaraguruparan et

al, 2004) and Queiroga et. al (2007) (Queiroga et al, 2007) reported expression of COX-2 in one of seven and two of four samples of normal mammary glands, respectively. Our own research demonstrated that as much as 91.7% of all carcinomas in the study displayed COX-2 expression. Doré et.al (2003) obtained similar results: he found a positive COX-2 reaction in 67% of complex cancers and in 47% of simple carcinomas (Doré et al, 2003). Our research demonstrated the highest COX-2 expression in simple cancers and the lowest in solid cancers. Similar data were presented by Heller et.al (2005), who found COX-2 expression in adenocarcinomas, whereas he failed to detect it in solid carcinomas (Heller et al, 2005). Ristimäki et. al (2002) demonstrated the relationship between COX-2 expression and certain clinical and pathological features of the tumor (Ristimäki et al, 2002). Her research into mammary gland tumours proved that COX-2 expression was positively correlated with tumour size as well as the strength of Ki-67 nuclear antigen staining and p53 protein expression. Results of our study are consistent with the results of works by Ristimäki et. al (2002) : increased COX-2 expression in carcinomas is related to higher mitotic index, i.e. with the proliferating activity (Ristimäki et al, 2002). Expression of Ki–67 nuclear antigen was demonstrated in carcinomas exhibiting increased COX-2 expression, yet research failed to prove the correlation between these markers; On the other hand, it proved the existence of statistical correlation between COX-2 expression and expression of p53 and Hsp70 proteins. Average number of cells exhibiting p53 protein expression was higher in carcinomas with higher levels of COX-2 expression. High expression level of Hsp70 protein and COX-2 was shown in carcinomas with the 3rd histological grade of malignancy, as described by low apoptotic index. Similar results were obtained by Lanza-Jacoby et. al (2004), who conducted her studies on experimental animals (Lanza-Jacoby et al, 2006), and by Liu and Rose (1996) who studied the COX-2 expression in cell cultures (Liu & Rose, 1996). Furthermore, Ristimäki et al (2002) and Dempke et al (2001) came to a conclusion that COX-2 promoted the mammary gland neoplasm growth, invasive capacity, and probability of metastasis (Ristimäki et al, 2002; Dempke et al, 2001). COX-2 expression may be relevant to a number of physiological processes within this tissue including proliferative activity, inhibition of apoptosis, increased angiogenesis and activation of matrix metalloproteinases (Dempke et al, 2001). Mammary physiology is a complex process regulated by hormones, estrogens, progesterone, growth hormone, prolactin, and epidermal growth factor (Howlin et al, 2006). As COX-2 expression is induced by different stimuli including cytokines, oncogenes, hormones and growth factors (Thomas et al, 2008), the same factors that control mammary growth and differentiation may act as trigger stimuli for COX-2 expression. The results of our work are consistent with the results obtained by other researchers. P-gp expression was discovered in epithelium and neoplasms of mesenchymal origin in dogs (Ginn, 1996). Author of these studies argues that continuation of this line of research will deliver additional prognostic data. There is a small number of works attempting to evaluate the P-gp expression in canine mammary cancers. A considerable progress in the investigation into P-gp expression in canine mammary cancers was made by Petterino et. al (2006), who attempted to define the P-gp expression in mammary gland carcinomas in female canines (Petterino et al, 2006). His study covered cases of both malignant and benign neoplasms; Petterino et al (2006) found the expression of the examined marker in two test groups and confirmed the statistical

al, 2004) and Queiroga et. al (2007) (Queiroga et al, 2007) reported expression of COX-2 in one of seven and two of four samples of normal mammary glands, respectively. Our own research demonstrated that as much as 91.7% of all carcinomas in the study displayed COX-2 expression. Doré et.al (2003) obtained similar results: he found a positive COX-2 reaction in 67% of complex cancers and in 47% of simple carcinomas (Doré et al, 2003). Our research demonstrated the highest COX-2 expression in simple cancers and the lowest in solid cancers. Similar data were presented by Heller et.al (2005), who found COX-2 expression in adenocarcinomas, whereas he failed to detect it in solid carcinomas (Heller et al, 2005). Ristimäki et. al (2002) demonstrated the relationship between COX-2 expression and certain clinical and pathological features of the tumor (Ristimäki et al, 2002). Her research into mammary gland tumours proved that COX-2 expression was positively correlated with tumour size as well as the strength of Ki-67 nuclear antigen staining and p53 protein expression. Results of our study are consistent with the results of works by Ristimäki et. al (2002) : increased COX-2 expression in carcinomas is related to higher mitotic index, i.e. with the proliferating activity (Ristimäki et al, 2002). Expression of Ki–67 nuclear antigen was demonstrated in carcinomas exhibiting increased COX-2 expression, yet research failed to prove the correlation between these markers; On the other hand, it proved the existence of statistical correlation between COX-2 expression and expression of p53 and Hsp70 proteins. Average number of cells exhibiting p53 protein expression was higher in carcinomas with higher levels of COX-2 expression. High expression level of Hsp70 protein and COX-2 was shown in carcinomas with the 3rd histological grade of malignancy, as described by low apoptotic index. Similar results were obtained by Lanza-Jacoby et. al (2004), who conducted her studies on experimental animals (Lanza-Jacoby et al, 2006), and by Liu and Rose (1996) who studied the COX-2 expression in cell cultures (Liu & Rose, 1996). Furthermore, Ristimäki et al (2002) and Dempke et al (2001) came to a conclusion that COX-2 promoted the mammary gland neoplasm growth, invasive capacity, and probability of metastasis (Ristimäki et al, 2002; Dempke et al, 2001). COX-2 expression may be relevant to a number of physiological processes within this tissue including proliferative activity, inhibition of apoptosis, increased angiogenesis and activation of matrix metalloproteinases (Dempke et al, 2001). Mammary physiology is a complex process regulated by hormones, estrogens, progesterone, growth hormone, prolactin, and epidermal growth factor (Howlin et al, 2006). As COX-2 expression is induced by different stimuli including cytokines, oncogenes, hormones and growth factors (Thomas et al, 2008), the same factors that control mammary growth and differentiation may act as trigger stimuli for COX-2 expression. The results of our work are consistent with the results obtained by other researchers. P-gp expression was discovered in epithelium and neoplasms of mesenchymal origin in dogs (Ginn, 1996). Author of these studies argues that continuation of this line of research will deliver additional prognostic data. There is a small number of works attempting to evaluate the P-gp expression in canine mammary cancers. A considerable progress in the investigation into P-gp expression in canine mammary cancers was made by Petterino et. al (2006), who attempted to define the P-gp expression in mammary gland carcinomas in female canines (Petterino et al, 2006). His study covered cases of both malignant and benign neoplasms; Petterino et al (2006) found the expression of the examined marker in two test groups and confirmed the statistical significance between the two groups (Petterino et al, 2006). Furthermore, he found that simple and complex carcinomas were the most numerous groups among the examined cancers and that P-gp expression was absent from the tissues of healthy mammary glands. We achieved similar results in our research: we found the P-gp expression in 76% of the examined canine mammary cancers. The most numerous groups exhibiting a positive Pgp immunohistochemical reaction were composed of complex carcinomas (90.9%) and simple carcinomas (73%). High expression was found in cancers with the highest histological grade of malignancy. Due to fact that there are hardly any works investigating into the P-gp expression in canine mammary cancers, we attempted not only to confirm the presence of P-gp expression and location, but also to prove the relationship between P-gp and other neoplastic markers. Statistical analysis confirmed a positive correlation between P-gp and COX-2 expression: it demonstrated a statistical significance in the case of examined characteristics (P = 0.021). It is worth mentioning that the expression of two markers was significant in carcinomas featuring high histological grades of malignancy. Furthermore, P-gp expression was studied in canine lymphomas; additionally, attempts were made to conduct studies on carcinomas treated by means of chemotherapy; the results were compared with the results obtained in the control group of untreated carcinomas. Higher expression was observed in the treated cancer cells. On the other hand, increased expression in untreated cancers appears to be a negative prognostic factor, given that survival rate shall decrease in such cases (Lee et al, 1996). Due to the fact that this line of research has not brought about many works, there are several questions and ambiguities concerning the role of P-gp in the process of neoplastic genesis and the importance of this factor from the clinical point of view. Additionally, the question arises whether, or not P-gp can be classified into the significant prognostic marker category. We are not aware of defensive mechanisms of the cells equipped with membrane transporters against apoptosis generated by the compounds, that are not pump substrates. There are suggestions that the phenomenon may be connected with evacuation (pumping out) of a certain important mediator of apoptosis or with the impact of P-gp on the intracellular pH (Johnson et al, 1998). Identification of a mechanism of cell resistance to apoptosis is a key instrument in selection of suitable therapy. According to studies in humans, heat shock proteins may be important predictors of breast cancer (Park & Dupont, 1982). There is not much published data in the literature on heat shock protein expression in mammary gland tumors of female dogs. The role of these proteins in carcinogenesis has not been clearly defined either. Studies have only shown that expression of heat shock proteins takes place in canine breast cancers , but the linkage between these proteins and other tumor markers was not confirmed. Seymour (1990) studied endometrial cancer in women and found that heat shock proteins were useful markers in the diagnostics of these tumors (Saymour et al, 1990). Expression of Hsp27 was also investigated in breast cancer in women (Ciocca et al, 1993). A relationship between the expression of this protein and the degree of differentiation of the tumor cells was found. Similar studies were conducted by Storm (1996), who stated that tumors with Hsp27 expression showed a higher histological grade than tumors negative for the expression of this protein (Storm et al, 1993). Kumaraguruparan (2006) studied the expression of Hsp70 and Hsp90 in breast cancer in women and found a correlation between the expression of both proteins and proliferative activity (Kumaraguruparan et

Prospective Study of Tumor Markers as Prognostic Factors

marker in the diagnostics of breast cancers in female dogs.

*Veterinary Pathology*, Vol. 36, pp. 423-436

*Veterinary Pathology*, Vol. 29, pp. 381-385

**4. Acknowledgments** 

N30800632/0667

**5. References** 

75-89

in the Histopathological Differential Diagnosis of Mammary Gland Neoplasms in Female Canines 225

Literature concerning the expression of p53 protein in mammary tumors in canines is scarce. It deals only with the expression of p53, but leaves no answers in respect to the expression of other factors that are considered prognostic markers, such as proliferative activity, histological grade of malignancy and expression of estrogen receptors. We conclude that COX-2 is an important prognostic factor and may be applied as a marker of canine mammary neoplasm malignancy, given the fact that higher expression of COX-2 was found in adenocarcinomas and in cancers featuring the highest histological malignancy grades in comparison to simple adenomas. Obtained results suggest that cyclooxygenase-2 may be a prognostic factor, but it requires clinical confirmation. The P-gp expression is also positively correlated with the degree of histological malignancy. This suggests a prudent approach to decisions concerning chemotherapy. Expression of Hsp90 can be considered as a marker of the degree of differentiation and of histological grade of mammary gland tumors in dogs, because the highest level of expression was found in solid carcinomas and in cancers exhibiting the highest histological grades of malignancy. Hsp70 expression was confirmed, but no correlation with other factors was found. This may suggest that Hsp70 is not a useful

Study was carried out at the Institute of Pathology, Department of Clinical Sciences, Faculty of Veterinary Medicine, Warsaw School of Life Sciences, 159C Nowoursynowska Street, 02-766 Warsaw. Research was conducted as a part of a doctoral dissertation, partially funded by a grant from the Ministry of Science and Information Technology, No

Bacus, S.S.; Goldschmidt, R.; Chin, D.; Moran, G.; Weinberg, D.; Bacus, J.W. (1989).

Benjamin, S.A.; Lee, A.C.; Saunders, W.J. (1999). Classification and behavior of canine

Bostock, D.E.; Moriarty, J.; Crocker, J.(1992). Correlation between histologic diagnosis mean

Chung-Ho, L.; Wan-Hee, K.; Ji-Hey, L.; Min-Soo, K.; Dae-Yong, K.; Oh-Kyeong, K. (2004).

tumors. *Journal of Veterinary Science*, Vol.5, No.1, pp. 63-69

markers. *The American Journal of Pathology* , Vol.135, No.5, pp. 783-792 Barzanti, F.; Dal Susino, M.; Volpi, A.; Amadori, D.; Riccobon, A.; Scarpi, E.; Medri, L.;

Biological grading of breast cancer using antibodies to proliferating cells and other

Bernardi, L.; Naldi, S.; Aldi, M.; Gaudio, M.; Zoli, W. (2000). Comparison between different cell kinetic variables in human breast cancer. *Cell Proliferation*, Vol.33,:pp.

mammary epithelial neoplasms based on life-span observations in beagles.

nucleolar organizer region count and prognosis in canine mammary tumors.

Mutation and overexpression of p53 as a prognostic factor in canine mammary

al, 2006). He suggested that heat shock proteins are important prognostic factors. Presence of Hsp70 was also reported in normal gastrointestinal tissues, but an increased expression of this protein was observed in gastrointestinal tumors (Isomoto et al, 2003). Isomoto (2003) believed that Hsp70 plays an important role in the degree of cellular differentiation in tumors, which would indicate that Hsp70 is an important prognostic factor. Similar studies are conducted in veterinary medicine, but to a lesser extent (Isomoto et al, 2003). A similar study was conducted by Rommanucci (2005) in mammary gland tumors in female dogs (Romanucci et al, 2006). She found Hsp70 and Hsp90 expression in simple and complex, as well as solid-type adenocarcinomas. Expression of these proteins was observed in the cytoplasm, as well as in the nuclei of tumor cells. Rommanucci (2006), in her study of breast cancers in female dogs, has not attempted to verify the possible correlations between heat shock proteins and other tumor markers (Rommanucci et al, 2006). She focused on establishing the degree of protein expression and their localization in tumor cells. Basing on the results of her study, the author suggested that these proteins may play an important role in the process of carcinogenesis. In our study, Hsp70 protein expression was found in 86.4% of tumors, while the expression of Hsp90 was observed in 66.2% of tumors. The aim of our study was to demonstrate the relationship between the expression of heat shock proteins and other prognostic factors. When comparing the expression of Hsp70 to the expression of Hsp90 in different types of tumors, we found a high statistical significance between the expression of both proteins. Analysis of the relationship between the expression of nuclear antigen Ki-67 and the expression of heat shock proteins showed that the highest expression of Ki-67 as well as Hsp90 was present in solid tumors. The highest expression of Hsp90 protein and Ki-67 was also found in cancers of the highest histological grades. Based on these data it can be concluded that Hsp90 is an important factor, which could be considered a marker of malignancy and which may be useful in the diagnostics of cancers. Analysis of the results showed high expression of Hsp70 in cancers with grade 1 and 3 of COX-2 expression; the lowest expression of Hsp70 was found in cancers with grade 2 of COX-2 expression. Statistical significance of P = 0.009 was found between the expression of Hsp70 and expression of COX-2. An increased expression of COX-2 was observed in tumors with a low mean number of cells showing positive reaction to Hsp90 protein. Correlation between the expression of Hsp70 protein and p53 protein in tumors of epithelial origin was also confirmed. Based on the analysis of expression of both proteins and taking into account the mean apoptotic index, it was found that in cancers with the third (the highest) histological grade, expression of p53 protein is significantly lower than the expression of Hsp70, and the value of apoptotic index in these cancers was the lowest.

#### **3.2 Conclusions**

Occurrence of mammary tumors in female dogs and their potential use as a model in comparative pathology makes it an interesting research material. It is known that the degree of malignancy is positively correlated with proliferative activity. The relationship between expression of estrogen receptors and the degree of malignancy is not entirely clear, although most authors believe that the expression of these receptors is higher in benign tumors. Literature concerning the expression of p53 protein in mammary tumors in canines is scarce. It deals only with the expression of p53, but leaves no answers in respect to the expression of other factors that are considered prognostic markers, such as proliferative activity, histological grade of malignancy and expression of estrogen receptors. We conclude that COX-2 is an important prognostic factor and may be applied as a marker of canine mammary neoplasm malignancy, given the fact that higher expression of COX-2 was found in adenocarcinomas and in cancers featuring the highest histological malignancy grades in comparison to simple adenomas. Obtained results suggest that cyclooxygenase-2 may be a prognostic factor, but it requires clinical confirmation. The P-gp expression is also positively correlated with the degree of histological malignancy. This suggests a prudent approach to decisions concerning chemotherapy. Expression of Hsp90 can be considered as a marker of the degree of differentiation and of histological grade of mammary gland tumors in dogs, because the highest level of expression was found in solid carcinomas and in cancers exhibiting the highest histological grades of malignancy. Hsp70 expression was confirmed, but no correlation with other factors was found. This may suggest that Hsp70 is not a useful marker in the diagnostics of breast cancers in female dogs.
