Oxytocin receptor induces mammary tumorigenesis through prolactin/p-STAT5 pathway

Oxytocin receptor (OXTR) is involved in social behaviors, thermoregulation, and milk ejection, yet little is known about its role in breast cancer. To investigate the role of OXTR in mammary gland development and tumorigenesis, a transgenic mouse model of OXTR overexpression (++Oxtr) was used. Overexpression of OXTR-induced progressive mammary hyperplasia, unexpected milk production, and tumorigenesis in females. OXTR-induced mammary tumors showed ERBB2 upregulation and mixed histological subtypes with predomination of papillary and medullary carcinomas. OXTR overexpression led to an activation of prolactin (PRL)/p-STAT5 pathway and created a microenvironment that promotes mammary-specific tumorigenesis. PRL inhibitor bromocriptine (Br) could mitigate OXTR-driven mammary tumor growth. The study demonstrates Oxtr is an oncogene and a potential drug target for HER2-type breast cancer.


Introduction
Breast cancer is the most common cancer with highest morbidity among females worldwide 1 . It is genetically classified into four subtypes, HER2 + , Luminal A, Luminal B, and triple negative (basal-like subtype) 2 . These subtypes differ significantly in prognosis and responsiveness to various therapeutic options 3 . ERBB2 receptor tyrosine kinase (or HER2) is a family member of epidermal growth factor receptors (EGFRs). Overexpression of HER2 induces tumorigenesis 4 . HER2 is overexpressed in 20-30% of breast tumors 5 and correlates with poor patient outcome 6 . Lapatinib, a tyrosine kinase inhibitor, mitigates mammary tumor growth by blocking HER2 tyrosine kinase activity 7 . Understanding breast cancer development is critical for effective treatments. Mouse models have been developed to mimic clinic phenotypes 8 . MMTV-PyMT mice exhibit papillary and medullary carcinomas 9 similar to HER2 + breast cancer 10 . Brca1 −/− and p53 −/− mice grow basallike tumors 11,12 .
G-protein coupled receptors (GPCRs) regulate a variety of physiologic functions, ranging from blood pressure control, kidney function, allergic response, hormonal disorders to neurologic diseases 21 . Oxytocin receptor (OXTR), a member of GPCRs, is the receptor for neurotransmitter oxytocin (OXT) 22 known to regulate sexual and social behaviors, thermoregulation, and milk ejection 23 . OXTR has been found highly expressed in pathological breast, breast carcinomas, neuroblastomas, and astrocytomas [24][25][26] . OXTR overexpression has been reported in endometrial adenocarcinomas 27 . Connections between breast cancer and OXT/OXTR have been suggested 28 . However, whether and how OXTR regulates mammary gland development and carcinogenesis remains unknown. Our previous study indicates OXTR overexpression disrupts hormonal environment, induces early mammary gland maturation, early involution, and lactation failure 29 . Here we investigated OXTR's role in mammary tumorigenesis using the ++ Oxtr mouse model.

OXTR overexpression induces mammary tumorigenesis
A mouse model with transgenic overexpression of OXTR under β-actin promoter ( ++ Oxtr) was used in this study 29 . OXTR overexpression in mammary gland and brain was confirmed (Fig. S1). Seventeen out of 30 ++ Oxtr females (56.6%) developed tumors in mammary gland from age of 5 to 15 months (Fig. 1A). Among the females, two (11.8%) developed more than one primary neoplasm (Fig. 1B). Tumor growth continued in size (Fig. 1C). They were removed and weighed on 30th day of onset (Fig. 1D). Tumors showed bulging surface, fleshy appearance, areas of hemorrhage, and necrosis (Fig. 1E). However, no visible lung metastases were observed (Fig. S2). No tumors were found in males.
Histological analysis was performed to classify the tumors 30 . H&E staining exhibited multifocal areas and mixed phenotypes of different morphological characteristics. Dominant phenotype for each tumor is noted. Among them, 42.1% have typical papillary patterns with small, finger-like projections. Tumors having epithelial structures with uniform and multilayered nuclei, thin stromal axes, and frond-like branching jutting into larger lumens are classified as papillary carcinomas (Fig. 1F, a, b). Squamous pathology and accumulation of keratin pearls were visible (Fig. 1F, a). Lipids in cells and mucus were detected (Fig. 1F, b). 36.8% of tumors with cord-like structures separated by thin stroma were classified as medullary carcinomas (Fig. 1F, c). 15.8% of tumors were defined as glandular carcinomas. These tumor cells have dark and irregular nuclei, form glandular-like foci or colonies of solid nests. Some tumors showed ductal areas with local invasion of tumor cells (Fig. 1F, d). In addition, some areas contain cells morphologically different from normal mammary gland cells with disorganized and irregular patterns. These are defined as poorly differentiated forms of tumors with malignant potential (Fig.  1F, e). 5.3% of tumors developed small cysts with hobnail- Fig. 1 Tumor onset and histology analysis. A Kaplan-Meier plot of tumor-free survival for wild-type (WT) (n = 29) and ++ Oxtr females (n = 30). B Tumor incidence analysis, WT, n = 29 and ++ Oxtr, n = 30. C Tumor growth measurement by volume (n = 9). D Tumor weights 30 days after palpation (n = 8). Data were represented as mean ± SD. ***P < 0.001, calculated using two-tailed unpaired t-test and Log-rank (Mantel-Cox) test. E Representative macroscopic view of tumors in ++ Oxtr females. shaped cells or clear cytoplasm. These are classified as clear cell carcinoma (Fig. 1F, f). All results support these tumors being mammary carcinoma.

OXTR overexpression induces mammary hyperplasia and unexpected milk production
To explore the process of mammary tumorigenesis, preneoplastic mammary gland morphology was assessed. Whole-mount staining of ++ Oxtr mammary gland revealed enlarged ducts and accelerated alveoli development from age of 3 months ( Fig. 2A). H&E staining showed that ducts were distended and filled with proteinaceous liquid material (Fig. 2B). Macroscopic morphology analysis revealed that the ducts were filled with milk and lasted with age (Fig. 2C). Statistically, 96% of nonpregnant ++ Oxtr females showed milk accumulation (Fig. 2D). Correspondingly, qPCR results confirmed that expression of major milk protein genes Csn2 and Wap were dramatically increased (221 and 157 folds at age of 3 months, and thousands of folds by 9 months, and ++ Oxtr females (n = 24). E Gene expression of major milk proteins Csn2 and Wap, n = 6. F Immunochemistry staining of Ki67 in mammary gland. Nuclei were stained blue with hematoxylin. Scale bar: 100 μm. Quantitative immunostaining of mammary gland using Image Pro Plus, n = 5. G Ki67 immunostaining of ++ Oxtr mammary tumor and WT mammary gland. Scale bar: 100 μm. Quantification of immunostaining using Image Pro Plus, n = 5. H Whole-mount staining of ++ Oxtr mammary gland at tumorigenesis and corresponding WT mammary gland. Scale bar: 500 μm. Data were represented as mean ± SD. ***P < 0.001, calculated using two-tailed unpaired t-test. Fig. 2E). This phenotype is consistent with nipple galactorrhea in early stage of clinical breast cancer.

OXTR overexpression leads to constitutive activation of PRL/p-STAT5 pathway
To determine whether hormonal environment plays a role in OXTR-induced tumorigenesis, serum PRL, P4, estradiol, and OXT at different stages were measured. PRL levels were found to increase with age, peaked at tumorigenesis, and stayed high in ++ Oxtr females (Fig.  4A). However, P4 was lower than WT (Fig. 4B). No changes were found in serum estradiol and OXT (Fig. S5). Expression of OXTR, RANKL, STAT5, and p-STAT5 in mammary gland and tumors were examined by immunoblotting. OXTR in ++ Oxtr mammary gland was constantly high with highest level in tumors (Fig. 4C).
To assess whether OXTR-induced microenvironment can drive metastasis as well, E0771 and B16 cells were injected through tail vein. Larger numbers of visible metastases were readily detectable in ++ Oxtr lungs from E0771 (Fig. S8A, B) with marked increase of metastatic , and ++ Oxtr with Br treatment (n = 12). F p-STAT5 immunostaining of tumors. Nuclei were stained blue with hematoxylin. Scale bar: 100 μm. G Immunoblotting analysis of p-STAT5 of tumors. H Gene expression of Erbb2, Akt1, and Tgfα in tumors by qPCR, n = 6. Data were represented as mean ± SD. **P < 0.01, ***P < 0.001, calculated with one-way analysis of variance (ANOVA). foci (Fig. S8C). However, no significant difference was detected from B16-injected mice (Fig. S8D). Results suggest that OXTR-induced microenvironment can promote mammary-specific tumor growth and metastasis but not melanoma tumors.

Discussion
OXTR overexpression induces dramatic PRL secretion and STAT5 phosphorylation. Nuclear translocation of p-STAT5 leads to increased transcription of mammary epithelial proliferation-related genes, accelerated mammary gland development (unexpected milk secretion), and tumorigenesis (Fig. 7). OXTR induces hormonal changes and creates a mammary gland-specific environment that promotes mammary tumor growth.
Mouse models of mammary tumorigenesis have been established to mimic various subtypes of human breast cancers 8 . Overexpression of HER2 is associated with  5). B Representative photos of mammary tumors from E0771 cells, Scale bar: 1 cm. Quantitative analysis of mammary tumor weights (n = 5 from WT and n = 8 from ++ Oxtr) and volumes (n = 5 from WT and n = 8 from ++ Oxtr). C Representative photos of melanoma tumors from B16 cells, Scale bar: 1 cm. Quantitative analysis of melanoma tumor weights (n = 5) and volumes (n = 5). D Representative photos of cervical tumors from U14 cells, Scale bar: 1 cm. Quantitative analysis of cervical tumor weights (n = 5) and volumes (n = 5). Data were represented as mean ± SD. **P < 0.01, ***P < 0.001, calculated using two-tailed unpaired t-test. metastasis and poor prognosis 35,36 . Our study shows that 57% of ++ Oxtr females develop ERBB2 + mammary tumors with change of PI3K-AKT, MAPK, Jak-STAT, and NF-kappa B pathways, similar to HER2 + breast cancer. High HER2 is accompanied by activation of PI3K/AKT and MAPK pathways, promoting cellular proliferation and survival 37 . ++ Oxtr tumors are morphologically mixed with papillary and medullary carcinoma that are invasive and highly malignant. OXTR-induced hyperprolactinemia, unexpected milk production (nipple discharge), and mammary hyperplasia are all early characteristics of human breast cancer. ++ Oxtr mouse should be an ideal model for HER2 + drug screening and testing.
OXTR overexpression induced high PRL. The excessive PRL secretion leads to accelerated mammogenesis and tumorigenesis. Studies using mouse models lacking either PRL (Prl −/− ) or activated STAT5 have confirmed the role of PRL/p-STAT5 signaling in mammary gland development 38,39 . In response to PRL, p-STAT5 translocates to nucleus and activates target gene transcription 40 . We have identified that 72% of upregulated genes in ++ Oxtr tumors are targets of STAT5 and function in mammary gland development and epithelium cell proliferation. STAT5-targeted genes Erbb2, Akt1, Tgfα, Csn2, and Wap were all upregulated in ++ Oxtr tumors. These rationalize the early symptoms of preneoplasia including mammary hyperplasia, unexpected milk production, and Erbb2 + mammary tumorigenesis in ++ Oxtr females. Our study demonstrates that PRL/p-STAT5 signaling mediates OXTR-induced mammary tumorigenesis. PRL stimulates breast cancer cell proliferation through HER2 expression 41 . Hyperprolactinemia increases risk of breast cancer 42 . High blood PRL is associated with poor prognosis and low survival with metastatic breast cancer 43 . Prl overexpression in mouse mammary gland or transplanted pituitary glands induces mammary carcinomas in aged females 44,45 . Incidence of neoplasms in these females with moderate latency is similar to that of ++ Oxtr females. These reports all support our hypothesis that OXTRinduced ERBB2 + mammary tumors through increased PRL secretion. Br, an inhibitor of PRL/p-STAT5 pathway, can effectively block OXTR-induced PRL secretion, ERBB2 expression, hyper-mammogenesis, and tumorigenesis. The result confirms the role of OXTR through PRL/p-STAT5. The relationship of OXTR, hyperprolactinemia, and ERBB2 expression in breast cancer is established in this study. Moreover, study has shown that metastatic disease-related hyperprolactinemia is significantly more frequent in HER2 + patients 46 , suggesting PRL may stimulate HER2 expression. PRL may be a potential marker for diagnosis of HER2 + breast cancer. Lapatinib, the inhibitor of ERBB2, cannot compromise ++ Oxtr tumor growth. This result suggests that ERBB2 may not be the sole mediator of PRL/p-STAT5-stimulated breast cancer cell proliferation. Study has shown attempts to interfere with HER2 alone have failed to yield an effective treatment 47 .
RANKL is regulated by both P4 and PRL 48,49 . OXTR overexpression leads to increased PRL secretion but downregulates P4. PRL/p-STAT5 axis increased RANKL expression despite low P4, suggesting limited role of P4 in OXTR-induced tumorigenesis. Our results show that OXTR overexpression induces mammary hyperplasia and tumorigenesis by activation of PRL/p-STAT5/ RANKL axis.
Our results indicate that OXTR-induced hormonal environment promotes mammary tumorigenesis exclusively. OXTR expression was detected in human breast cancer cells T47D, MCF7, ZR-75-30, and MDA-MB-231 (Fig. S9). Whether these tumor cells can secrete PRL needs to be determined. In addition, contribution of mammary/brain OXTR overexpression on PRL secretion requires evaluation. We previously reported that mammary OXTR is not a major player in abnormal mammary gland development 29 . OXTR expression in brain can respond to exogenous OXT and stimulates PRL release from pituitary lactotroph 50,51 . We assume that OXTR overexpression in brain may be the major source of PRL secretion in ++ Oxtr mice. Neuron-and mammary glandspecific overexpression of OXTR may further shed light on mammary tumorigenesis.
In conclusion, we have found OXTR overexpression induces ERBB2 + mammary tumors through activation of PRL/p-STAT5 pathway. The activation creates an environment that promotes mammary gland-specific tumor growth. Oxtr is a novel oncogene and a potential new drug target for HER2 + breast cancer. PRL is an important marker for HER2-tumor diagnosis and drug target for HER2 + breast cancer. In addition, Br is an effective antitumor drug for OXTR/PRL-driven HER2 + breast cancer.

Materials and reagents
All general chemicals and reagents were purchased from Sigma, USA and Takara, China.

Animals
All animal studies were performed in accordance with Guide for Care and Use of Laboratory Animals from National Institutes of Health and approved by the Ethics Committee of Shenyang Medical College (SYYXY2018030101). Generation of β-actin-Oxtr ( + + Oxtr) mice (RRID: MGI: 6314370) by us was described previously 29 . Age-matched WT littermates were used as controls. All animals were switched to C57/BL6J background and maintained under pathogen-free conditions at 21 ± 1°C, 50 ± 20% relative humidity, with free access to food and water, and 12:12 h light/dark cycle. Mice were anesthetized with 1% pentobarbital natrium (10 mg/kg) intraperitoneally before euthanizing.

Whole-mount staining
The fourth inguinal mammary glands were dissected and spread on glass slides. After fixation with Carnoy solution, glands were rehydrated gradually through a series of diluted ethanol and immersed in carmine aluminum solution 57 at room temperature overnight. Glands were dehydrated through serial ethanol baths and cleared in xylene.

RNAseq and analysis
Libraries were constructed from WT mammary glands and ++ Oxtr tumors, and sequenced using an Illumina Hiseq platform. Low-quality reads were removed 58 . Clean reads were mapped to mouse genome sequence using TopHat2 59 . Results were presented as fragments per kilobase of transcript per million of mapped reads (FPKMs) 60 . Q value < 0.05 and |log2 (fold change)| >1 were used as threshold for significantly different expression by Cuff diff version 2.0.0 61 . Gene ontology (GO) analysis was performed using Gorilla and estimated by hypergeometric test using custom R scripts. Significance (p value) was adjusted by false discovery rate (FDR) 62 . GO terms with q value < 0.05 were regarded as significantly enriched. The enrichment scores were calculated using Gene Set Enrichment Analysis (GSEA) as described before 63 . OmicShare small tools2 was used to obtain heatmaps. Threshold parameters were set as no rows and column clusters. The GeneVenn online tool was used to create Venn diagrams of gene lists.

Quantitative real-time PCR (qPCR)
Total RNA was purified from mouse tissues using Trizol reagent (Takara). One microgram RNA was reverse transcribed with Prime Script cDNA Synthesis Kit (Takara). The cDNAs were used for PCR with SYBR Green Mix (Takara) following the manufacturer's instruction. Relative expression level was normalized to 18 S ribosomal RNA and calculated using 2 −ΔΔCT value method. PCR primers are listed in Table S1.
Tumor tissue/cell transplantation A 1 mm 3 piece of tumor fragment, E0771, B16, or U14 cells (5 × 10 6 ) were orthotopically transplanted into fourth mammary gland of 3-month-old WT or ++ Oxtr virgin females. Tumor sizes were measured using digital calipers and volume was calculated as ½ (length × width 2 ). Tumors from E0771, B16, and U14 cells were removed after 15 days of injection. For tail vein injection, E0771 and B16 cells (1 × 10 6 ) were suspended in PBS before injection. Metastatic lesions from E0771 or B16 were examined in 4 weeks or day 15 after injection.

Bromocriptine and Lapatinib treatment
Br (Sigma, USA) was dissolved in sterile saline (0.9% NaCl) to a final concentration of 1 mg/ml. After tumor cells injection, mice were treated daily with 200 μg Br subcutaneously for 15 days. Control WT and ++ Oxtr females were treated in parallel with saline.

Statistical analysis
All data were presented as means ± SD. P value was calculated with unpaired two-tailed Student's t-tests to compare two groups, one-way ANOVA to compare more than three groups, and log-rank (Mantel-Cox) test for survival analysis. Asterisks denote statistically significant differences (*P < 0.05; **P < 0.01; ***P < 0.001).

Data availability
Data reported has been deposited in Gene Expression Omnibus (GEO) database (accession number: PRJNA542227).

Ethics Statement
All experiments and protocols were approved by the Ethics Committee of Shenyang Medical College (SYYXY2018030101) and adhered to ethical standards outlined in Guide for Care and Use of Laboratory Animals from National Institutes of Health.

Conflict of interest
The authors declare no competing interests.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.