In veterinary medicine, a variety of assays and techniques have been used in an attempt to identify the prognostic markers for neoplastic diseases, including clinical parameters, histologic features, immunohistochemistry, quantitative PCR, evaluation for genomic mutations and specific serum protein levels or enzyme activity analyses .
To investigate the potential functional relevance of CA 15–3 as a serum tumour marker in veterinary oncology, we performed a multidisciplinary study on canine neoplastic mammary tissues and cell lines.
The main finding in Western blot analysis regarded the positive interaction between the human antibodies and the canine MUC1 protein. The protein sequence of epitope DF3 in humans is TRPAPGS , whereas the canine DF3 epitope sequence is SRPAPSS (protein sequences have been obtained from NCBI with accession numbers [canine NP_001181906.1 and human P15941.3] and then compared using the MultAlin on-line platform ). We used the BLOSUM62 amino acid substitution matrix [29, 30] to score alignments between evolutionarily divergent protein sequences. It revealed a high similarity between these two epitopes in humans and in dogs (with only one point of difference). Thus, theoretically, the similarity of the protein sequence of the DF3 epitope is very high in these two species. Nevertheless, the results of WB showed that anti 115d8 and anti DF3 antibodies display cross-reactivity with canine samples. The ICC analysis of the MUC1 expression (using both antibodies) in canine mammary cancer cell lines showed staining patterns similar to those observed in humans.
In WB, a double protein band with a high molecular weight was observed in MCF-7 cells [31–33]. In the canine tumour cell lines CMT-W1, CMT-W2, CMT-W1M and CMT-W2M, we found a single protein band of MUC1 with a molecular weight similar to the higher human form. However, when we extended the time of the chemical exposure, some films revealed an apparent reactive point with a molecular weight of approximately 250 kDa. These data are still under investigation because it is difficult to determine whether it is an actual protein band or an artefact of the analysis, considering the low sensitivity of the test.
Our immunocytochemistry results differ from those obtained using Western blot. The immunocytochemical examination revealed MUC1 expression in all of the examined cell lines. However, only the CMT-W1, CMT-W1M, CMT-W2 and CMT-W2M cell lines expressed MUC1 at high levels, similar to MCF7 (no statistically significant differences were found). The remaining cell lines exhibited a weaker expression of MUC1, which differed significantly from the MCF7 cell line. According to the subject data [34, 35], a higher expression of MUC1 is observed in more invasive metastatic cancer. Our findings support this hypothesis, as CMT-W1, CMT-W1M, CMT-W2 and CMT-W2M cell lines with invasive phenotypes [17, 18, 20] (Figure 2) showed a higher MUC1 expression (detected in WB and ICC). Thus, a lack of MUC1 expression in non-invasive P114, CMT-U27 and CMT-U309 cell lines in Western blot analysis may be due to low levels of this antigen expression and/or the low sensitivity of the Western blot analysis. Although both methods detect a specific protein expression, they significantly differ from each other on what may explain the differences in the obtained results. Similar variations between the results obtained using Western blot and immunocytochemistry have previously been reported , showing that immunocytochemistry is a more sensitive method than the Western blot in cases of weak antigen expression.
Statistical analysis allowed us to demonstrate that CA 15–3 serum levels correlate positively with the degree of tumour proliferation and differentiation, as a significantly higher concentration of CA 15–3 was found in grades II-III than in grade I carcinomas (p = 0.0019). Tumour size, skin ulceration, necrosis, inflammation and the histological type of mammary cancer were not significantly related to the serum levels of CA 15–3.
The existence of a relationship between CA 15–3 serum levels and tumour grading appears to be logical because we assumed that tumour markers reflect the number and the activity of neoplastic cells as well as their products that access the circulation where the marker is measured.
Therefore, this parameter could be taken into consideration as potentially enabling a better evaluation of a tumour’s malignancy, particularly for establishing the preoperative stage of the neoplasm and providing useful guidelines when determining the surgical approach.
Another finding of our study, as expected, is the expression of MUC1 in normal canine mammary tissues and in mammary gland benign lesions. The reaction was primarily restricted to the apical membrane of epithelial cells. However, in benign tumours, the pattern of reaction was extended to the cytoplasm of the luminal epithelial cells. These results are in absolute parallel to those observed in human mammary tissues. .
Among the carcinomas, we observed MUC1 activity in 68% of the analysed cases that displayed a strong immunostaining in a mixed pattern (involving the entire cell surface and the cytoplasm), which comprises the whole cell population in several carcinoma samples. This finding is in accordance with the observation of Lacunza , which demonstrated a correlation between MUC1 protein over-expression and gene amplifications in benign and malignant breast tumours. Many studies have revealed that this event could be linked to the presence of membrane-associated MUC1 molecular portions that are shed from the surface of tumour cells to the cytosol by modulation of its glycosylation state [39, 40].
In the present study, there was no significant correlation between the over-expression of MUC1 and CA 15–3, the circulating form of the DF3 antigen. However, a significant correlation has been found between the histopathological grade of the tumour malignancy and MUC1 expression and between the histopathological grade of the tumour malignancy and the circulating CA 15–3 level. We suppose that a lack of a correlation between MUC1 over-expression in tissues and the serum level of CA 15–3 may be related to the different antibodies used in the IHC and CA 15–3 tests.