MAM domain containing 2 is a potential breast cancer biomarker that exhibits tumour‐suppressive activity

Abstract Objectives The aim of this study was to discover new potential biomarkers of breast cancer and investigate their cellular functions. Materials and methods We analysed the gene expression profiles of matched pairs of breast tumour and normal tissues from 24 breast cancer patients. Tetracycline‐inducible MAMDC2 expression system was established and used to evaluate cell proliferation in vitro and in vivo. MAMDC2‐mediated signalling was determined using immunoblot analysis. Results We identified MAMDC2 as a down‐regulated gene showing significant prognostic capability. Overexpression of MAMDC2 or treatment with MAMDC2‐containing culture medium significantly inhibited the cell proliferation of T‐47D cells. Furthermore, MAMDC2 expression reduced in vivo growth of T‐47D xenograft tumours. MAMDC2 may exert its growth‐inhibitory functions by attenuating the MAPK signalling pathway. Conclusion We report that MAMDC2 has a tumour‐suppressive role and, as a secretory protein, it might be useful as a biomarker for breast cancer treatment.

studies reported that MAMDC2 gene expression is differentially regulated in certain human cancer types, including CML, head and neck squamous cell carcinoma and breast cancer. [9][10][11] A gene expression analysis reported MAMDC2 as one of three genes that are correlated with disease-free survival of breast cancer patients. 12 Although previous reports have shown that the MAMDC2 expression is associated with various human cancer types, its exact molecular function has not been defined. In the present study, we have demonstrated that MAMDC2 has a growth-inhibitory function by regulating MAPK signalling pathway. The differentially regulated genes were selected by statistical analysis using GEO database (http://www.ncbi.nlm.njh.gov/geo): GSE22035, GSE5764, GSE26910 and GSE21422. Gene expression analysis was performed using the Nanostring nCounter system (Nanostring Technologies, USA) according to the manufacturer's instructions. The quantified probe counts were normalized to β-actin gene. All statistical analyses were performed with Medcalc software (Belgium). P-values less than .05 were considered statistically significant. media. RT-PCR amplification was performed using a monoplex RT-PCR with 2X TOPsimple™ DyeMix-multi HOT premix (Enzynomics, Korea).

MAMDC2 antibody from Abcam Inc
For the subcellular localization of MAMDC2, HeLa cells were plated on a glass cover slide and transfected with MAMDC2-EGFP.
To visualize mitochondria, pDsRed2-Mito vector (Takara, Japan) was co-transfected. For ER, the cells were incubated with monoclonal anti-KDEL antibody (Enzo Life Sciences), followed by incubation with Alexa Fluor 594-labelled anti-mouse antibody (Thermo Fisher Scientific).

| In vivo xenograft experiment
Six-week-old female NSG mice (The Jackson Laboratory, USA) were acclimated for one week and then ovariectomized. Then, the mice were implanted with 90 days slow-release oestradiol pel-

| Apoptosis assays
For cell cycle analysis, cells were fixed in 70% ethanol overnight.
After treatment with RNase I, cells were stained with 5 μg/mL propidium iodide (PI). Annexin V staining was performed using

| MAMDC2 is down-regulated in breast cancer cells
To discover new biomarkers of breast cancer, we obtained matched pairs of breast tissue from tumour and non-tumour region of 24 female patients diagnosed with IDC (average age; 50.5 ± 14.1 years; Table S1). Then, we selected potential breast cancer biomarkers by analysing microarray data from GEO database and functional annotation through Gene Ontology database. Finally, 24 putative cell surface and secreted proteins were chosen and a total 27 genes, including 3 controls, had their expression profiles analysed using the Nanostring nCounter system ( Figure S1). A total of 22 genes were differentially regulated more than 2-fold, with 9 genes up-regulated, while 13 were down-regulated (Table S2).  Figure 1F).
Next, in order to evaluate MAMDC2 as a useful biomarker, we tried to detect MAMDC2 protein in human serum or breast tissues. Unfortunately, however, the MAMDC2 protein was not monitored in the human sample due to the lack of specific antibody. Instead, we used the Kaplan-Meier (KM) plotter database and found a close relation between the MAMDC2 expression level and survival rates of the breast cancer patients. High MAMDC2 expression group showed clearly increased survival curve in total ( Figure 1G) and ER-positive ( Figure 1H) patients.
Interestingly, however, we could not find any differences in the triple-negative cases ( Figure 1I). Overall, these clinical data may indicate that MAMDC2 is a reliable prognostic biomarker for breast cancer.

| MAMDC2 is an N-glycosylated secretory protein
Although it has not been documented, MAMDC2 was predicted to be a secretory protein, based on the presence of a short N-terminal signal sequence and absence of any transmembrane domain ( Figure 2A). To determine whether this is the case, we transiently transfected pcDNA3-MAMDC2-FLAG plasmid into T-47D cells and monitored MAMDC2 expression by immunoblotting.
Interestingly, we observed high amounts of MAMDC2 protein in both cell lysate and the culture supernatant, indicating that it is secreted into the extracellular space ( Figure 2B). We next examined the glycosylation status of MAMDC2 protein by treating it with PNGase F. The molecular weight of MAMDC2-FLAG protein collected from both cell lysate and culture supernatant was reduced after incubation with PNGase F, demonstrating that both cellular and secreted MAMDC2 proteins are N-glycosylated ( Figure 2C).
In order to determine the intracellular localization of MAMDC2 proteins, we transfected pEGFP-MAMDC2 construct into HeLa cells. In contrast to EGFP which is dispersed throughout the cytosol, EGFP-conjugated MAMDC2 substantially co-localized with ER-Tracker, but not with Mito-Tracker ( Figure 2D). This suggests that MAMDC2 protein is synthesized in the ER and released from the cells via a secretory pathway. is not enough to inhibit cell growth, and therefore, it is likely that MAMDC2 functions as an extracellular regulator of cell proliferation.

| The N-terminal region of MAMDC2 is important for its cell growth-inhibitory activity
To determine which regions are critical for the cell growth-suppressive function, we generated several deletion mutant constructs of MAMDC2 using pcDNA3-MAMDC2-FLAG plasmid. Because this protein contains four sequential MAM domains, that is D1, D2, D3 and D4 from the N-terminus, we made three truncated versions of MAMDC2 each of which contained two MAM domains, that is D1-2, D2-3 and D3-4 ( Figure 4A). Then, their expression was confirmed in both cell lysate and culture supernatant ( Figure 4B). When transfected into T-47D cells, two MAMDC2 variants, D1-2 and D2-3, exhibited inhibitory effects on the growth of T-47D cells that were similar to the wild-type MAMDC2 ( Figure 4C,D). Taken together, we propose that the second MAM domain from the N-terminus, D2, may have a critical role in mediating cell growth inhibition.

| In vivo tumour growth in MAMDC2 xenograft model
To evaluate the effect of MAMDC2 on tumour growth in vivo, Tet-On-MAMDC2 cells or Tet-On-Ctrl cells were injected into female NOD scid IL2 receptor gamma null (NSG) mice (n = 5). As T-47D is an ER+ cell line, the xenografted mice also received a slow-release oestrogen pellet to promote in vivo tumour formation. In addition, MAMDC2 expression was induced by providing food containing doxycycline from day zero. Although the mean body weight of each group did not change significantly, the Tet-On-MAMDC2 xenograft mice exhibited tumour regression ( Figure 6A,B). Seventy days after tumour cell injection, all mice were sacrificed and tumour xenografts were excised. The average tumour weight of MAMDC2-expressing xenografts was ~62% of controls (P < .05; Figure 6C,D), indicating that the MAMDC2 expression can attenuate in vivo tumour cell proliferation.
Next, we examined the effect of MAMDC2 knockdown on cell proliferation. Because we could not detect endogenous MAMDC2 protein in most cancer cell lines, we performed the experiment using Tet-On-MAMDC2 cells. Two siRNAs were selected to repress MAMDC2 expression to about half the level of control siRNA-treated cells ( Figure 6E). While no significant difference was observed in 24-hours cultures, a clear recovery of growth was observed in siMAMDC2-treated cells after 72 hours ( Figure 6F,G), providing further strong evidence that the expression of MAMDC2 is closely linked with cell viability.

| D ISCUSS I ON
Breast cancer is a highly heterogeneous disease exhibiting diverse clinical features. 14 Although different studies classify breast cancer cell lines into different categories, the most general subtyping is based on the expression of three immunohistochemistry markers, ER, progesterone receptor (PR) and the human epidermal growth factor receptor 2 (HER2). 15 Although ER-positive tumours are very common, accounting for 70 to 80% of all breast cancer cases, only about 30% of cell lines are ER+, because ER-negative cells are more likely to be established. 16 Despites the broad phenotypic spectrum of breast cancers, three cell lines (MCF7, T-47D and MDA-MB-231) account for more than two-thirds of breast cancer studies. 14 T-47D, which is often categorized as 'ER+ luminal subtype', exhibits low-levels of genomic abnormalities. 17  While T-47D is considered as non-aggressive cell line, MDA-MB-231 have a highly aggressive and invasive phenotype that can easily invade extracellular matrix with high metalloproteinase activities. 18 In fact, the MAMDC2 protein was quickly degraded in the MDA-MB-231 culture media ( Figure 7F). Therefore, the resistance of MDA-MB-231 cells to the MAMDC2-mediated growth control is related to the rapid degradation of extracellular MAMDC2 protein by metalloproteinases.
The KM plotter analysis revealed that MAMDC2 expression level is closely associated with the survival rates of the breast cancer patients in the ER+ group ( Figure 1H), which was completely abrogated in the triple-negative subtype ( Figure 1I). Furthermore, when we analysed, the MAMDC2 expression using two GEO profiles, that in aggressive breast tumour subtypes, was relatively higher than non-aggressive or less aggressive subtypes ( Figure S4).
These data may demonstrate that the highly aggressive breast tumour cells are no longer influenced by MAMDC2 expression. Unlike in MDA-BD-231 cells, ER signalling appears to negatively regulate MAMDC2 gene expression in T-47D and MCF-7 cells ( Figure 1E).
Since MAMDC2 exerts strong growth-inhibitory effects on these ER+ cells, ER activation could be more important in these cells to suppress MAMDC2 gene expression. On the contrary, we found that addition of oestradiol did not affect the growth-inhibitory function of MAMDC2 even in ER+ cells ( Figure S5), suggesting that ER signalling is not closely associated with the MAMDC2-mediated growth control. Collectively, we report that MAMDC2 is a novel tumour inhibitor gene that can be used a prognostic marker for the ER-positive breast tumour.

ACK N OWLED G EM ENTS
This work is supported by the National Research Foundation of Korea grant (NRF-2020R1A2C201128811) and funded by the Korean government (MSIT).

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflicts of interest with the contents presented in this study.

AUTH O R CO NTR I B UTI O N S
All authors participated in study design, data interpretation, and analysis and manuscript review. PJM conceptualized and acquired funding. HL involved in investigation and validation. YC visualized the study. BCP provided the resources. SJ and PJM supervised the study. SL wrote the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Soojin Lee
https://orcid.org/0000-0002-0383-5025 F I G U R E 7 MAMDC2 may inhibit RAS-ERK pathway. A-F, The culture media of Tet-On-MAMDC2 cells were collected and concentrated 10-fold. Then, concentrated culture media (10x) was added to cells to final concentrations (0×, 2×, and 4×) for 3 days. Immunoblotting was performed with the culture media of T-47D (C) and MDA-MB-231 cells (F). **P < .001. Scale bar = 200 μm. G, Phosphorylation levels of MAPK components in the normal growth condition (H) Tet-On-MAMDC2 or Tet-On-Ctrl cells were cultured with tetracycline for 48 h. I, Conditioned media collected from Tet-On-MAMDC2 or Tet-On-Ctrl cells were added to the culture media (1x) of MDA-MB-231 cells for 48 h