Differences of immunoreactivity for luminal and myoepithelial cell markers and oxytocin receptor in canine normal mammary ducts and ductal tumors

Abstract

Ductal tumors of canine mammary glands can be difficult to distinguish from simple tumors on routine histology. Because oxytocin receptor (OXTR) expression in myoepithelial cells has been implicated in human breast tumorigenesis, we immunohistochemically evaluated 8 ductal adenomas and 5 ductal carcinomas, including OXTR expression. Luminal cells were positive for cytokeratin 19 and pan-cytokeratin. Suprabasal myoepithelial cells were strongly positive for p63, weakly and focally positive for vimentin, and negative for smooth muscle actin and calponin, therefore indicating that p63 is the most suitable marker for these cells in ductal tumors. Unlike humans, OXTR was expressed in luminal cells of normal canine mammary ducts and in both luminal and suprabasal myoepithelial cells in ductal tumors, suggesting a role in ductal tumor development.

Ductal adenomas and carcinomas are relatively uncommon histological subtypes of canine mammary tumors, and may be difficult to distinguish from simple adenomas and carcinomas in hematoxylin and eosin (HE)-stained sections [7, 23]. While the malignancy of ductal carcinoma is low, that of simple carcinoma is extremely high and often associated with a poor prognosis [7, 16, 23]. Therefore, it is important to differentiate ductal tumors from simple tumors by histological examinations. Myoepithelial cells in canine mammary tumors have been classified as suprabasal and interstitial myoepithelial cells, with the former being difficult to differentiate from luminal cells based on cellular morphology [23]. Therefore, immunohistochemistry is required to confirm the proliferation of suprabasal myoepithelial cells [23]. Complex adenoma and carcinoma are common in dogs, and the immunohistochemical characteristics of myoepithelial cells in complex tumors have been elucidated [2, 5, 8, 17,18,19]. However, the immunohistochemical characteristics of myoepithelial cells in ductal tumors remain unclear. Therefore, the first objective of the present study was to clarify the immunohistochemical characteristics of suprabasal myoepithelial cells in ductal tumors.

In human normal breast tissue, the oxytocin receptor (OXTR) is expressed in myoepithelial cells in the acinus and ducts [20]. The primary function of oxytocin and OXTR in normal breast tissue is to mediate milk ejection by the contraction of myoepithelial cells. Recent studies indicated the involvement of OXTR in the development and progression of human breast tumors [13]. A study using a transgenic mouse model overexpressing OXTR revealed a markedly higher rate of mammary tumors in transgenic mice than in wild-type mice through the promotion of cell proliferation-related gene transcription by OXTR overexpression [12]. Another study showed that human breast tumors highly expressing OXTR were associated with worse prognosis than those with low expression of OXTR [14]. Based on these findings, OXTR has potential as a therapeutic target for breast tumors in human medicine [9, 13, 14].

The histological localization of OXTR in canine normal mammary glands remains unclear. Only one study has examined OXTR expression in canine mammary tumors [3]. Benavente et al. performed immunohistochemistry for OXTR in canine mammary tumors and revealed a relationship between OXTR expression levels and neoplastic malignancy [3]. In that study, the majority of cases were classified as simple tumors, such as adenoma-simple and carcinoma-simple, and neoplastic cells were immunolabeled with OXTR [3]. However, histological subtypes with proliferation of myoepithelial cells were only benign mixed tumors [3]. Since OXTR is expressed in myoepithelial cells in human normal breast tissue [20], it is essential to elucidate the localization of OXTR in canine normal mammary glands and mammary tumors with the proliferation of myoepithelial cells. Therefore, the second objective of the present study was to reveal the immunohistochemical characteristics of OXTR in canine normal mammary glands and ductal tumors.

The proliferating suprabasal myoepithelial cells in ductal tumors are difficult to distinguish from luminal cells on HE-stained sections. One of the goals of our study is to aid in the differentiation of ductal tumors from simple tumors. Although some ductal tumors are accompanied by proliferation of interstitial myoepithelial cells, such cases can be differentiated from simple tumors on HE-stained sections. Therefore, they were excluded from the present study. Canine mammary ductal adenomas (8 cases) and ductal carcinomas (5 cases) without the proliferation of interstitial myoepithelial cells in sections were selected from the archives of the Laboratory of Veterinary Pathology in Obihiro University of Agriculture and Veterinary Medicine. To re-confirm the histological classification, HE-stained sections were re-evaluated based on the latest international histological classification [23]. Briefly, ductal tumors were histologically diagnosed by the presence of neoplastic cells clearly arranged in two layers surrounding a slit-like lumina (Supplementary Fig. 1). In the present study, among neoplastic cells arranged in bilayered cords or tubules, those arranged in the inner layer were described as neoplastic luminal (NL) cells, while those arranged in the outer layer were described as neoplastic suprabasal myoepithelial (NSM) cells.

Formalin-fixed paraffin-embedded sections were subjected to immunohistochemistry using anti-pan-cytokeratin (CK), anti-CK19, anti-vimentin, anti-α smooth muscle actin (SMA), anti-calponin, and anti-p63 antibodies (Supplementary Table 1). We also conducted double immunofluorescence for OXTR/pan-CK and OXTR/SMA using a rabbit polyclonal anti-mouse OXTR antibody (Supplementary Table 1) to examine the localization of positive reactions for OXTR in canine normal mammary glands and ductal tumors. An NCBI BLASTP search using the registered OXTR protein sequences of dog (NP_001185588.1) and mouse (NP_001074616.1) showed 90.59% identity with 96% query coverage, suggesting the appropriateness of its application of this antibody to canine tissue. Among the mammary tissues at the surgical margins resected together with the tumor, areas without apparent histopathological changes, such as mammary lobular hyperplasia, were assessed as normal mammary tissue in the present study. Based on a previous study [21], mammary ducts in normal tissue were classified into three types: interlobular ducts, intralobular ducts, and terminal ductal lobular units, to separately evaluate the results of immunohistochemistry and immunofluorescence examinations. Some sections were also subjected to Periodic acid Schiff (PAS) staining to confirm the location of the basement membrane.

All cases examined in the present study were female. The average age of 8 cases with ductal adenoma was 11.3 years, and 3 were spayed. The average age of 5 cases with ductal carcinoma was 11.6 years, and 4 were intact.

In normal mammary glands, luminal cells in the interlobular ducts, intralobular ducts, and terminal ductal lobular units were immunolabeled with pan-CK and CK19, while myoepithelial cells were immunolabeled with pan-CK, vimentin, SMA, calponin, and p63 (Supplementary Fig. 2). The results of double immunofluorescence for OXTR/SMA showed positive reactions for OXTR in luminal cells in interlobular ducts, intralobular ducts, and terminal ductal lobular units, but not in myoepithelial cells (Fig. 1). No apparent differences in these immunohistochemical staining patterns, including OXTR, were observed between spayed and intact females.

Fig. 1.

Interlobular duct (a–c) and intralobular duct and terminal ductal lobular units (d–f) in a normal mammary gland, dog. Immunofluorescence for α-smooth muscle actin (SMA) (a, d), oxytocin receptor (OXTR) (b, e), and merged figures (c, f). Asterisks indicate the lumens of interlobular or intralobular ducts. Myoepithelial cells are positively stained with SMA (green) in interlobular ducts, intralobular ducts, and terminal ductal lobular units. Positive reactions for OXTR (red) are observed in luminal cells, which are located in the inner layer to myoepithelial cells.

In ductal adenomas and carcinomas, the results of PAS staining indicated that the structure of the basement membrane was retained and NSM cells were connected to the basement membrane (Supplementary Fig. 3). NL cells were immunolabeled with pan-CK and CK19. NSM cells were strongly immunolabeled with p63 and weakly and focally labeled with pan-CK and vimentin. Positive reactions for CK19, SMA, and calponin were not detected in NSM cells (Fig. 2a–f). The results of double immunofluorescence for OXTR/pan-CK indicated that positive reactions for OXTR in ductal adenomas and carcinomas were detected in both NL and NSM cells (Fig. 2g–i). No apparent differences were observed in immunohistochemical staining patterns between spayed and intact females.

Fig. 2.

Ductal adenoma in a mammary gland, dog. Immunohistochemistry (a–f) and immunofluorescence (g–i) (the inset indicates a magnified figure of each). Neoplastic cells are arranged in two layers surrounding a slit-like lumina. Neoplastic luminal cells, which are arranged in the inner layer, are immunopositive for pan-cytokeratin (CK) (a) and CK19 (b). Neoplastic suprabasal myoepithelial cells, which are arranged in the outer layer, are negative for SMA (d) and calponin (e), weakly and focally positive for pan-CK (a) and vimentin (c), and immunopositive for p63 (f). Figure showing immunofluorescence for pan-CK (g) and OXTR (h), and their merged figure (i) indicates that positive reactions for OXTR are strongly detected both in neoplastic luminal cells and neoplastic suprabasal myoepithelial cells.

The results of immunohistochemistry for pan-CK, CK19, vimentin, SMA, calponin, and p63 of normal mammary ducts in the present study were consistent with previous findings [2, 5, 17,18,19]. Since ductal adenomas and carcinomas show differentiation in the interlobular ducts [6, 7, 23], we initially suspected that ductal adenomas and carcinomas have the same immunohistochemical features as normal interlobular ducts, as previously described [23]. The immunohistochemical characteristics of NL cells were consistent with those of luminal cells in normal mammary ducts. On the other hand, the immunohistochemical characteristics of NSM cells differed from those of myoepithelial cells in normal mammary ducts. While NSM cells positively stained for p63, they were negative for SMA and calponin. Furthermore, the immunoreactivity of NSM cells for pan-CK and vimentin was markedly weaker than that of their normal counterparts. The immunohistochemical characteristics of canine mammary tumors with the proliferation of myoepithelial cells, such as complex carcinoma, were previously reported [2, 5, 17,18,19]. SMA-/calponin-/p63+/vimentin+ (weakly)/ pan-CK+ (weakly) myoepithelial cells, which were observed in ductal adenomas and carcinomas, have not been detected in other canine mammary tumor subtypes or normal mammary glands. Since myoepithelial cells may be in various stages of differentiation in mammary tumors, the use of multiple myoepithelial markers is recommended for their identification by immunohistochemistry [23]. To differentiate ductal tumors from simple tumors, it is important to confirm the proliferation of myoepithelial cells. Although SMA, calponin, p63, and vimentin are general myoepithelial markers of canine mammary glands [23], the suprabasal myoepithelial cells in ductal tumors showed strong immunopositivity only for p63, while the other markers were negative or weakly positive. These findings revealed that p63 was the most appropriate myoepithelial marker for detecting suprabasal myoepithelial cells in ductal tumors, and also that the combination of vimentin, SMA, and calponin was insufficient.

The main function of luminal cells in mammary glands is milk production stimulated by prolactin [1], while that of myoepithelial cells is to mediate milk ejection stimulated by oxytocin [13]. OXTR in human normal mammary ducts and the acinus is located in myoepithelial cells [20]. The localization of OXTR in canine normal mammary glands has not been investigated histologically. In the present study, the results of immunofluorescence indicated that OXTR was located in luminal cells. A previous study revealed that the NL cells of canine simple tumors were immunolabeled with OXTR [3]. This finding may support the present results showing OXTR expression in luminal cells in normal canine mammary glands. Although myoepithelial cells in normal mammary ducts did not exhibit positive immunoreactivity for OXTR in the present study, this finding does not indicate the absence of OXTR expression in these cells. It may suggest that OXTR expression in canine normal myoepithelial cells is markedly lower than that in luminal cells. Taken together, differences were noted in the expression pattern of OXTR in normal mammary glands between humans and dogs. This result may indicate the difference in the function of oxytocin and OXTR in normal mammary glands among those species. In lactating rabbits and rats, OXTR is expressed both in myoepithelial and luminal cells. In these species, oxytocin acts through OXTR to not only induce contraction of myoepithelial cells but also promote the intracellular transport of secretory vesicles, such as casein micelles, toward the luminal side in luminal cells [15]. OXTR expressed in luminal cells of the canine mammary gland may likewise have a role in promoting intracellular secretory vesicle transport. However, because the previous study examined only lactating animals, it remains unclear whether a similar mechanism operates in non-lactating animals. Further studies are required to clarify the functions of oxytocin and its receptor in luminal cells of the canine mammary gland.

The majority of human breast tumors originate from luminal cells, not myoepithelial cells [13]. Although OXTR expression was not observed in normal luminal cells, it was detected in NL cells and hyperplastic luminal cells [4, 10]. Li et al. demonstrated in a mouse model that OXTR overexpression increased serum prolactin levels, activated the PRL/p-STAT5 pathway, and upregulated tumor-promoting genes such as Erbb2 and Akt1, thereby promoting cell proliferation and mammary tumorigenesis [12]. Liu et al. further showed that OXTR-overexpressing breast cancer cell lines displayed enhanced responsiveness to epidermal growth factor and increased cell motility via the ERK1/2-RSK-rpS6 pathway, suggesting that OXTR may also be involved in the invasion of neoplastic cells [14]. In addition to these studies, oxytocin-OXTR signalling has been reported to activate multiple downstream cascades in a G protein-dependent manner, and its effects appear to be context-dependent, ranging from tumor-promoting to tumor growth-inhibitory [13]. For example, activation of Gαq/11 has been linked to cell proliferation through increased intracellular calcium signalling [13, 22], whereas activation of Gαi has been suggested to suppress the PI3K/Akt pathway and thereby exert antiproliferation effects [11, 13]. Moreover, the biological effects of oxytocin and OXTR appear to depend not only the type of G protein but also on the characteristics of the tumor cells themselves. Oxytocin has been reported to stimulate proliferation in some human breast cancer cell lines, while showing no effect or even inhibitory effects in others [13]. Few studies have examined the relationship between mammary tumors and OXTR expression in dogs. In the present study, OXTR immunoreactivity in normal canine mammary ducts was restricted to luminal cells, whereas in ductal tumors it was observed both in NL and NSM cells. The altered immunohistochemical expression pattern suggested that NSM cells, at least, may be influenced by oxytocin and OXTR. We could not examine which OXTR-related signalling pathways were activated in the ductal tumors, and it remains unclear whether the OXTR expression in NSM cells is associated with enhanced malignancy, including increased tumor growth or cell migration, or instead reflects antiproliferation. Further studies are needed to determine the precise functions of OXTR in canine ductal tumors and other types of mammary tumors.

A limitation of the present study is that the cross-reactivity of the anti-mouse OXTR antibody in canine tissues was not directly examined. Although an NCBI BLASTP search using the registered OXTR protein sequences of dog and mouse showed 90.59% identity, this sequence similarity does not in itself confirm antibody specificity or cross-reactivity. Therefore, the lack of direct validation, such as by western blotting, should be considered a limitation of the present study.

In conclusion, we revealed the presence of NSM cells with a characteristic immunophenotype in canine ductal tumors (vimentin+ (weakly)/SMA-/calponin-/p63+). The results obtained indicate that p63 is the most appropriate myoepithelial marker for differentiating ductal tumors from simple tumors. We also found that the localization of OXTR in normal canine mammary glands was in luminal cells, which is different from human normal breast tissue. In canine ductal tumors, positive reactions for OXTR were observed in both NL and NSM cells; however, further studies are required to elucidate the role of OXTR in the tumorigenesis of canine ductal tumors.

CONFLICT OF INTEREST

The authors declare no conflicts of interest with respect to the publication of this manuscript.