Adamts18 mRNA is expressed in the myoepithelial cells of mammary ducts

The distribution of Adamts18 mRNA in mouse mammary glands was determined by in situ hybridization [10]. The results showed that Adamts18 was mainly expressed in the myoepithelial cells of mammary ducts at different stages, including pubertal stage (1-month, M), adult stage (2M, 3M, and 12M), and elderly stage (18M) (Fig. 1A). Adamts18 mRNA levels in mammary tissues of female virgin mice aged from 1 to 30 months were also detected by qRT-PCR. The results showed that Adamts18 expression was highest at 2 months of age, and then decreased in adult and maintained at low level in elderly mammary tissues of virgin females (Fig. 1B).

Fig. 1

Expression of Adamts18 mRNA in mouse mammary gland. (A) In situ hybridization analysis of Adamts18 mRNA in mammary gland of mice. In situ hybridization–positive signals appear as pink dots in cells. Scale bar = 50 μm. M, month. Neg, negative probe hybridization of mammary gland as negative control. (B) The expression of Adamts18 mRNA in mammary gland of mice aged from 1 to 30 months was analyzed by qRT-PCR. The relative quantity of Adamts18 mRNA was normalized to that of the housekeeping gene Gapdh using the ΔΔCt method. Each dot represents an individual. Data are expressed as mean ± SD

ADAMTS18 deficiency increases the risk of HER2-positive mammary tumorigenesis and metastasis

The progression of primary mammary epithelial cells to a malignant phenotype involves multiple genetic events, including activation of oncogenes and inactivation of specific tumor suppressor genes. To investigate the association of ADAMTS18 with HER2-positive breast cancer, we developed Her2t/w/Adamts18+/+ and Her2t/w/Adamts18−/− mice with C57BL/6-FVB mixed background for subsequent experiments (Additional file 1: Fig. S1A). Female virgin mice were palpated weekly to detect the presence of spontaneous mammary tumors until 30 months of age. There was no significant difference in mean body weight between Her2t/w/Adamts18+/+ mice and Her2t/w/Adamts18−/− mice during the 30-month observation period (Additional file 1: Fig. S1B). The median survival time of Her2t/w/Adamts18−/− mice was shorter than that of Her2t/w/Adamts18+/+ mice (26 months vs 28.8 months, P = 0.0001) (Additional file 1: Fig. S1C), but the survival curve does not distinguish between natural and tumor-related deaths as mice both with and without visible tumors began to die after 26 months.

Due to halve Her2 expression and with FVB-C57BL/6 mixed background, both Her2t/w/Adamts18−/− mice (10 of 32, ~ 31%) and Her2t/w/Adamts18+/+ mice (2 of 36, ~ 5.6%) showed significantly lower incidence of spontaneous mammary tumors when compared with Her2t/t mice (12 of 12, 100%) (P < 0.001) (Fig. 2A). Nevertheless, the incidence of spontaneous mammary tumors was significantly higher in Her2t/w/Adamts18−/− mice than in Her2t/w/Adamts18+/+ mice (31% vs 5.6%, P = 0.01). Among the 10 tumor-bearing Her2t/w/Adamts18−/− mice, 7 developed more than two mammary tumors, while only one mammary tumor was detected in each of the 2 tumor-bearing Her2t/w/Adamts18+/+ mice (Additional file 1: Fig. S2). In these tumor-bearing mice, spontaneous tumors appeared significantly earlier in Her2t/w/Adamts18−/− mice than in Her2t/w/Adamts18+/+ mice. The mean latency for mammary tumor development was 210 days in Her2t/t mice, 494 days in Her2t/w/Adamts18−/− mice, and 647 days in Her2t/w/Adamts18+/+ mice, respectively (Fig. 2B). In addition, the mammary tumor-bearing Her2t/w/Adamts18−/− mice exhibited significant lung metastases when compared with Her2t/w/Adamts18+/+ mice (-/- vs. + / + : 9 of 32, ~ 28% vs. 0 of 36, 0%; P = 0.0022) (Fig. 2C and D). These tumor-bearing Her2t/w/Adamts18−/− mice also developed metastases in the liver (6 of 32, ~ 19%, P = 0.0219), kidney (5 of 32, ~ 16%, P = 0.0457), and peritoneal cavity (3 of 32, ~ 10%, P = 0.198) when compared with Her2t/w/Adamts18+/+ mice (Additional file 1: Fig. S3). Collectively, the 10 mammary tumor-bearing Her2t/w/Adamts18−/− mice were all accompanied by metastases, while the 2 Her2t/w/Adamts18+/+ mice exhibited only primary mammary tumors (Additional file 1: Table S2), suggesting that ADAMTS18 deficiency increases the risk of HER2-positive spontaneous mammary metastasis.

Fig. 2

ADAMTS18 deficiency increases the risk of HER2-positive mammary tumorigenesis and metastasis. (A) Representative images of spontaneous mammary tumors and the incidence of macroscopic mammary tumor cases in Her2t/t mice with FVB background (12 of 12, 100%), and Her2t/w/Adamts18+/+ mice (2 of 36, ~ 5.6%) and Her2t/w/Adamts18−/− mice (10 of 32, ~ 31%) with FVB-C57BL/6 mixed background. (B) The mean latency of spontaneous mammary tumors in tumor-bearing mice. Red represents the formation of multiple tumors in an individual mouse. Each dot, square, or triangle represents an individual. (C) Representative images and HE staining of mammary tumors and lung metastasis from the indicated genotypes. Brown arrow, lung tumor. Scale bars, 20 and 100 μm, respectively. In Her2t/w/Adamts18-/- mice (middle and lower panels), tumor type in the lungs was consistent with the primary site of the mammary tumors. (D) The incidence of tumors in distant organs of Her2t/w/Adamts18+/+ mice (n = 36) and Her2t/w/Adamts18−/− mice (n = 32). *p < 0.05; **p < 0.01; ***p < 0.001; p values of tumor incidence between two groups were determined through chi-squared analysis

Her2 t/w/Adamts18 −/− mammary tumor cells show increased proliferation, migration, and invasion capacity in vitro

To understand why Her2t/w/Adamts18−/− mice are more likely to form tumors and migrate than Her2t/w/Adamts18+/+ mice, we compared the proliferation, migration, and invasion capacity of primary tumor cells isolated from Her2t/w/Adamts18−/− mammary tumors, Her2t/w/Adamts18+/+ mammary tumors, and Her2t/t mammary tumors. At 12 h, Her2t/t and Her2t/w/Adamts18−/− mammary tumor cells showed significantly increased proliferation when compared with Her2t/w/Adamts18+/+ mammary tumor cells (Fig. 3A). Scratch would healing assays showed that the migration capacity of Her2t/t and Her2t/w/Adamts18−/− mammary tumor cells was significant increased compared with Her2t/w/Adamts18+/+ mammary tumor cells at 6 h and 12 h, respectively (Fig. 3B-D). Transwell assays showed that the invasion of Her2t/t and Her2t/w/Adamts18−/− mammary tumor cells was significantly increased compared with Her2t/w/Adamts18+/+ mammary tumor cells at 6 h (Fig. 3E and F).

Fig. 3

Increased proliferation, migration, and invasion capacity in Her2t/w/Adamts18−/− mammary tumor cells. (A) Counting kit-8 (CCK8) assay for cell proliferation of mammary tumor cells after 12-h (h) culture. (B–D) Migration of mammary tumors by scratch wound healing assays. Representative images of migrated mammary tumors in each group are shown (B). The scratch healing ratio was obtained by comparing the scratch area of 6 h and 12 h with that of 0 h. A smaller ratio indicates faster cell migration (healing) (C and D). Scale bar = 50 μm. (E–F) Transwell assay of invasive ability of mammary tumors. Representative images of migrated mammary tumors in each group are shown (E). Scale bar = 200 μm. Cells at the lower surface of the transwell chamber were counted (F). Data are expressed as mean ± SD from Her2t/w/Adamts18+/+ mammary tumors (n = 2), Her2t/w/Adamts18−/− mammary tumors (n = 3), and Her2t/t mammary tumors (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant; one-way ANOVA

ADAMTS18 deficiency promotes mammary hyperplasia

The development of breast cancer is a multi-step process that includes hyperplasia, precancerous lesions, carcinoma in situ, and metastasis. We then characterized the mammary glands of 30-month-old Her2t/w/Adamts18−/− and Her2t/w/Adamts18+/+ mice without macroscopic tumors to determine whether ADAMTS18 has an effect on mammary hyperplasia. When compared with those of Her2t/w/Adamts18+/+ mice, the mammary ducts of Her2t/w/Adamts18−/− mice were disordered with an increased number of branches (Fig. 4A and B). Histological analysis showed that the transverse section of the lateral ductal branch of Her2t/w/Adamts18−/− mice was thickened (Fig. 4C). The number of Ki-67-positive mammary epithelial cells in the mammary ducts of 30-month-old Her2t/w/Adamts18−/− mice was significantly increased (Fig. 4D and E). We further investigated the earlier proliferation of mammary epithelial cells in mice with different genotypes. The results showed that at 10 months of age, Ki-67 positive signals were significantly increased in the mammary ducts of Her2t/w/Adamts18−/− mice compared with those in Her2t/w/Adamts18+/+ mice (Additional file 1: Figure S4). Together, these results suggest that ADAMTS18 deficiency promotes mammary hyperplasia.

Fig. 4

ADAMTS18 deficiency promotes mammary hyperplasia. (A) Whole-mount preparation of mammary glands from the indicated genotypes at 30 months of age. Scale bars = 500 μm. Arrowheads indicate branches and terminal end buds. LN, lymph node. (B) Quantification of the branch point in the two genotypes indicated in B (n = 3/group). (C) Representative H&E-stained cross-sections of mammary glands from the indicated genotypes. Arrowheads indicate thickened tubules. Scale bars = 100 μm. (D) Representative immunohistochemical staining of cross-sections of mammary glands from the indicated genotypes using Ki-67 antibody. Scale bars = 50 μm. (E) Quantification of Ki-67-positive cells (n = 4/group). Data are expressed as mean ± SD. *p < 0.05; ***p < 0.001; Student’s t-test, two tailed

ADAMTS18 deficiency increases the activity of ERK and PI3K/AKT signaling pathways

HER2 is mainly involved in RAS-RAF-MEK-ERK pathway for cell proliferation and PI3K-AKT-mTOR pathway for cell survival [27, 28]. To determine whether ADAMTS18 deficiency synergistically strengthens the HER2 downstream signaling pathway, we examined the activity of ERK signaling pathway in the mammary epithelium of 30-month-old mice with different genotypes. Immunohistochemical analysis showed that the total ERK levels in Her2t/w/Adamts18+/+ and Her2t/w/Adamts18−/− mammary glands were comparable (Fig. 5A and B). However, the levels of Thr202/Tyr204-phosphorylated ERK were significantly increased in mammary epithelial cells of Her2t/w/Adamts18−/− mice (Fig. 5C and D). Considering that fibronectin is one of the important indicators involved in epithelial-mesenchymal transitions (EMT) process, and its effect on EMT is mediated by ERK kinase pathway, we then detected the expression level of fibronectin in mammary epithelium of mice with different genotypes. The results showed that the expression of fibronectin in the mammary epithelium of Her2t/w/Adamts18−/− mice was significantly higher than that of Her2t/w/Adamts18+/+ mice (Fig. 5E and F). Likewise, Western blot results showed that the levels of p-ERK1/2 (Thr202/Tyr204) and p-AKT (Ser-473) were significantly higher in Her2t/w/Adamts18−/− mammary glands than those in Her2t/w/Adamts18+/+ mammary glands (Additional file 1: Fig. S5). The level of E-cadherin (epithelial markers) in Her2t/w/Adamts18−/− mammary glands was significantly lower than that of Her2t/w/Adamts18+/+ mammary glands, while the levels of N-cadherin and fibronectin (mesenchymal markers) were significantly higher in Her2t/w/Adamts18−/− mammary glands than those in Her2t/w/Adamts18+/+ mammary glands (Additional file 1: Fig. S5).

Fig. 5

Enhanced ERK1/2 activity in mammary epithelium of 30-month-old Her2t/w/Adamts18−/− mice. (A, C, E) Representative immunochemistry staining of total ERK (A), Thr202/Tyr204-phosphorylated ERK (C), and fibronectin (FN) (E) in mammary epithelium. Scale bar, 100 μm. (B, D, F) Quantification of ERK (B), pERK (D), and FN-positive areas (F) was performed with Image Pro Plus. Each dot or square represents an individual. Data are expressed as mean ± SD (n = 4). **p < 0.01; ***p < 0.001; ns, not significant; Student’s t-test, two tailed

ADAMTS18 deficiency causes alterations of mammary ECM

To determine whether increased ERK1/2 and AKT signaling activity is associated with ECM alterations caused by ADAMTS18 deficiency, we examined the expression and distribution of ECM molecules and epithelial cell receptors in the mammary glands of two genotypes of mice at 30 months of age. qRT-PCR results showed that the transcription levels of Fn, LNα5, Col1a1, Integrin (Itg)α3, Itgα5 and Itgβ1 were significantly increased in the mammary glands of Her2t/w/Adamts18−/− mice at 30 months of age (Additional file 1: Table S3). The expression levels of Lnα1, β1, β2, β3, γ1, γ2, Itgβ3, Dag1, Nid1, Vcan, and Ddr1 were comparable between the two genotypes. IHF analysis showed that LN in the Her2t/w/Adamts18+/+ BM was distributed evenly and continuously, while LN in the Her2t/w/Adamts18−/− BM was distributed with local intensification and fragmentation (Fig. 6A). Fibrillar ColI was rarely distributed in the Her2t/w/Adamts18+/+ mammary stroma, while it was widely distributed in the Her2t/w/Adamts18−/− mammary stroma. Compared with the low level of FN distribution in Her2t/w/Adamts18+/+ BM, FN was abundantly distributed in Her2t/w/Adamts18−/− BM. Western blot results showed that the levels of LNα5 were significantly increased in Her2t/w/Adamts18−/− mammary glands (Fig. 6B). Sandwich ELISA results showed that the levels of ColI and LN in Her2t/w/Adamts18−/− mammary glands were significantly higher than those in Her2t/w/Adamts18+/+ mammary glands, and the ColIV level in Her2t/w/Adamts18−/− mammary glands was significantly lower than that in Her2t/w/Adamts18+/+ mammary glands (Fig. 6C–E).

Fig. 6

Expression and distribution of extracellular matrix molecules in mammary glands of 30-month-old mice with different genotypes. (A) Representative immunofluorescent staining of laminin (LN), collagen I (ColI), and fibronectin (FN) in the mammary glands of mice with different genotypes (30-month-old Her2t/w/Adamts18+/+ and Her2t/w/Adamts18−/− mice). Blue, 4',6-diamidino-2-phenylindole (DAPI); white short arrow, fractured laminin; yellow short arrow, local accumulation of laminin; white arrow, Col I distributed in stroma of mammary glands; Green arrow, fibronectin distributed in BM of mammary glands. Scale bars = 50 μm. (B) Western blot analysis of the protein levels of LNα5 in mammary glands of mice (n = 3/group). (C-E) Levels of ColI (C), LN (D), and ColIV (E) measured by sandwich enzyme-linked immunosorbent assay (n = 6/group). Data are expressed as mean ± SD (n = 6). *p < 0.05; Student’s t-test, two tailed

Western blot results showed that the protein levels of ITGα3, ITGα5, ITGβ1, phosphorylated c-JUN, and LOXL2 in Her2t/w/Adamts18−/− mammary glands were significantly higher than those in Her2t/w/Adamts18+/+ mammary glands (Fig. 7A–F). We also examined the expression of MMP2 and MMP9, as they play a key role in the degradation of the BM. The level of MMP9 in mammary glands of Her2t/w/Adamts18−/− mice at 30 months of age was significantly higher than that of Her2t/w/Adamts18+/+ mice (Fig. 7G). There were no significant differences in MMP2 and TIMP-1 levels between Her2t/w/Adamts18+/+ and Her2t/w/Adamts18−/− mammary glands (data not shown).

Fig. 7

Analysis of the protein levels of integrinα3 (ITGA3), α5 (ITGA5), β1(ITGB1), c-JUN, phosphorylated c-JUN (p–c-JUN), LOXL2 and MMP9 in mammary glands of 30-month-old mice. (A) Representative Western blot images of different proteins from the indicated genotypes (30-month-old Her2t/w/Adamts18+/+ and Her2t/w/Adamts18−/− mice). (B-G) Relative expression levels of the proteins are represented as p–c-JUN/c-JUN or protein/GAPDH. Each dot or square represents an individual. Data are expressed as mean ± SD (n = 4). *p < 0.05; **p < 0.01; ***p < 0.001; Student’s t-test, two tailed

Low ADAMTS18 expression in distant metastases of HER2 + tumors relapsed posttrastuzumab treatment

To investigate the clinical association of ADAMTS18 with HER2 + breast cancer, GEO dataset (GSE191230) was used for the analysis (Additional file 1: Table S4). In this dataset, RNA sequencing (RNA-seq) analysis was performed on 13 treatment-naive HER2 + breast tumors and 7 distant metastases of HER2 + tumors relapsed posttrastuzumab treatment (Additional file 1: Fig. S6A). ADAMTS18 expression was significantly downregulated in distant metastatic samples compared with the primary tumors (Additional file 1: Fig. S6B).