Hypermethylation of TMEM240 Involved in Expression Deciency Predicts Poor Hormone Therapy Response and Disease Progression in Breast Cancer

Background: Approximately 25% of patients with early-stage breast cancer experience cancer progression throughout the disease course. Alterations in TMEM240 in breast cancer were identied and investigated to monitor treatment response and disease progression. Methods: Circulating methylated TMEM240 in the plasma of breast cancer patients was used to monitor treatment response and relapse events. Illumina methylation arrays were used to identify novel hypermethylated CpG sites and genes related to poor hormone therapy response. Quantitative methylation-specic real-time polymerase chain reaction (QMSP), quantitative real-time reverse transcription PCR, and immunohistochemical analyses were performed to measure DNA methylation, mRNA and protein expression levels in 335 breast samples from Taiwanese and Korean patients. Kaplan–Meier curves, Cox proportional hazards regression and receiver operating characteristic curves were used to analyze 10-year survival and disease progression. The Cancer Genome Atlas (TCGA) dataset was used to investigate TMEM240 alterations in Western countries. Transient transfection and knockdown of TMEM240 were performed to determine its biological functions and its relationship to hormone drug treatment response in breast cancer cells. Results: Aberrant methylated TMEM240 was identied in breast cancer patients with poor hormone therapy response using genome-wide methylation analysis in the Taiwan and TCGA breast cancer cohorts. A cell model showed that TMEM240, which is localized to the cell membrane and cytoplasm, represses breast cancer cell proliferation and cell migration. TMEM240 protein expression was observed in normal breast tissues, but not detected in 88.2% (67/76) of breast tumors and in 90.0% (9/10) of metastatic tumors from breast cancer patients. Almost all triple-negative breast cancer patients (95.7%, 22/23) had decient TMEM240 protein expression. QMSP revealed that in 54.5% (55/101) of Taiwanese breast cancer patients, the methylation level of TMEM240 was at least 2-fold higher in tumor tissues than in the matched normal breast tissues. Patients with hypermethylation of TMEM240 had poor 10-year overall survival (p = 0.003) and poor treatment response, especially hormone therapy response (p < 0.001). Prediction of disease progression based on circulating methylated TMEM240 was found to have 87.5% sensitivity, 93.1% specicity, and 90.2% accuracy, better than the currently used biomarkers CEA and CA-153. Circulating methylated TMEM240 is a potential biomarker for treatment response and disease progression monitoring in breast cancer. glyceraldehyde 3-phosphate dehydrogenase gene; Ki-67, marker of proliferation Ki-67; QMSP, quantitative methylation specic real-time polymerase chain reaction; QPCR, Quantitative real-time reverse transcription polymerase chain reaction; SAC21, spinocerebellar ataxia 21; SRB, sulforhodamine B; TNBC, Triple negative breast cancer; TCGA, The Cancer Genome Atlas; TMEM240, transmembrane Protein 240; tumor suppressor gene;

Genome-wide methylation analysis of 5 paired breast cancer tissues and corresponding noncancerous breast tissues was performed using the Illumina In nium HumanMethylation450 BeadChip array (Illumina, San Diego, CA, USA) for one sample and the In nium MethylationEPIC Kit (Illumina) for the remaining 4 samples, as previously reported [17]. The two arrays contain more than 450,000 and 850,000 methylation sites, respectively, and provide genome-wide coverage of the gene region and CpG island coverage, respectively, including 99% of RefSeq genes. Bisul te conversion of 500 ng of genomic DNA was performed using the EpiTect Fast DNA Bisul te Kit (Qiagen, Bonn, Germany, Cat. No. 59826). Methylation scores for each CpG site were represented as "beta" values ranging from 0 (unmethylated) to 1 (fully methylated) based on determination of the ratios of the methylated signal intensities to the sums of the methylated and unmethylated signal outputs.
Cell lines, cell culture, and drug treatment The MDA-MB-231 and T47D breast cancer cell lines used in this study were obtained from the Bioresource Collection and Research Center (http://www.bcrc. rdi.org.tw/). MDA-MB-231 cells were cultured in DMEM/F12 supplemented with human platelet lysate (hPL, American Red Cross, USA) and 1% penicillin/streptomycin. T47D cells were cultured in DMEM/F12 supplemented with human platelet lysate (hPL, American Red Cross, USA), 1% penicillin/streptomycin and 6 ng/ml insulin. For the TMEM240 demethylation assay, MDA-MB-231 cells were treated with dimethyl sulfoxide (DMSO) or with the demethylation agent decitabine (DAC, Sigma-Aldrich, St. Louis, MO, USA) for 96 hours. DAC was dissolved in DMSO.
After treatment of the cells, DNA and RNA were extracted, and methylation and gene expression levels were analyzed. For the hormone therapy response assay, T47D cells were treated with DMSO or with a series of concentrations of Tamoxifen (0, 10 and 20 μM) for 48 hours (Sigma-Aldrich, St. Louis, MO, USA).

Immunohistochemical assay
Immunohistochemical staining with an antibody against TMEM240 (1:35, Sigma-Aldrich, HPA066721, St. Louis, MO, USA) was performed using an iView DAB detection kit (Ventana, Tucson, AZ, USA) and a BenchMark XT autostainer. The assay included both positive and negative controls. The researchers who evaluated the immunohistochemical staining results were blinded to the clinical follow-up data. The intensity of TMEM240 expression was identi ed semiquantitatively as no expression, low expression (weaker than or equal to the expression intensity observed in normal colon epithelium), or high expression (stronger than the expression intensity observed in normal colon epithelium).
Plasmid extraction, con rmation and puri cation Plasmid DNA was extracted using the Geneaid™ Midi Plasmid Kit (Geneaid Biotech Ltd., Cat. No. PI025) according to the manufacturer's instructions. The extracted DNA was subjected to preliminary length analysis by sequenced to con rm errorless production. The plasmid concentration was measured using a NanoDrop 2000C ultramicrowavelength spectrometer (Thermo Fisher Scienti c, USA), and the plasmid was stored at -20 °C until further use. cDNA expression construct, RNAi, and transfection TMEM240 interference RNA was obtained from Life Technologies Corporation. Transfections were performed using 10 nM si-TMEM240 or nontargeting siRNAs, and Lipofectamine-RNAiMax and Lipofectamine 3000 reagent (Invitrogen) was used to transfect MDA-MB-231 and T47D cells according to the manufacturer's protocol.

Transwell assay
Transwell assays were used to study cell migration. In the transwell assays, the upper and lower chambers of the culture wells were separated by a semipermeable membrane (Falcon) with a pore size of 8 μm. Approximately 2 × 10 4 and 1× 10 5 treated and untreated MDA-MB-231 and T47D cells, respectively, were seeded in the upper chamber. Then, 300 μL of serum-free DMEM/F12 was added as culture medium, and 800 μL of serumcontaining culture medium was added as a chemical attractant in the lower chamber. After 16 hours of incubation, the cells retained over the membrane were washed twice with PBS, xed with 4% formaldehyde and stained with 1% crystal violet/ddH 2 O for 60 minutes at room temperature.
Five randomly chosen areas were photographed using a camera attached to a microscope (Nikon), and ImageJ was used to quantify the number of cells in each area.

Wound healing assay
The wound healing assays were performed using culture inserts (Ibidi, GmbH, Martinsried, Germany). After seeding 1× 10 5 cells overnight, the cells were transfected with siRNA for 48 hours. The culture inserts were then removed, the wounded areas were photographed using a camera attached to a microscope (Nikon), and ImageJ was used to calculate the wound areas.

Sulforhodamine B assay
A sulforhodamine B (SRB) assay was used to measure the cell proliferation rate. MDA-MB-231 and T47D cells were seeded in 96-well plates at densities of 8 × 10 3 and 1 × 10 4 cells/well, respectively, and incubated for 24 h and 48 h, respectively. The cells were then xed with 10% trichloroacetic acid for 10 min. After staining with SRB for 30 min, excess dye was removed by washing the cells ve times with 1% acetic acid. Cell proliferation was assessed using a microplate reader to determine the absorbance of the SRB solution at 515 nm.

Statistical analysis
All statistical analyses were performed using SPSS (SPSS Inc., Chicago, IL, USA). Pearson's chi-squared test was used to compare breast cancer patients with respect to TMEM240 methylation, RNA expression, and clinical data including age, sex, tumor type, TNM tumor stage, race, menopausal state, and ER, PR and HER2 status. The t-test was used to compare cells transfected with or without TMEM240 plasmid or si-TMEM240 and cells with and without drug treatment. Multivariate Cox proportional hazards regression analyses (adjusted for age, sex, race, tumor subtype, and tumor stage) were further used to analyze the correlation between TMEM240 hypermethylation and 10-year overall survival in breast cancer patients. Comparisons of hypermethylation and hypomethylation curves that yielded log-rank test p values of less than 0.05 were considered statistically signi cant. The TMEM240 methylation level and drug treatment response in breast cancer patients whose data were accessed through the TCGA portal were analyzed using the Mann-Whitney test. In addition to accuracy, other commonly used measures of evaluating the classi cation, such as the receiver operating characteristic curve (ROC) and area under the curve (AUC), sensitivity, speci city, false-positive rate and false-negative rate, are also reported.

Results
TMEM240 was identi ed in samples from Taiwanese and Western breast cancer patients by genome-wide methylation analysis To identify a novel potential biomarker in breast cancer patients with poor hormone therapy response, we used ve criteria to screen potential targets: (1) hypermethylation in Taiwanese breast cancer patients; (2) hypermethylation in Western breast cancer patients; (3) hypermethylation in breast cancer patients with poor hormone therapy response compared with complete response; (4) a methylation level in normal tissues that was close to 0; and (5) low expression in breast cancer patients ( Figure 1A). First, to identify critical tumor suppressor genes, the In nium Methylation Assay was applied to 5 breast cancer tissue samples and paired noncancerous breast tissue samples. A total of 2612 genes were hypermethylated according to the criterion ΔAvg_β (βTumor -βNormal) > 0.4. Second, we analyzed the TCGA Illumina In nium HumanMethylation450 BeadChip array data of 87 paired Western breast cancer patients. A total of 6882 genes were found to be hypermethylated according to the criterion ΔAvg_β (βTumor -βNormal) > 0.4. Next, the top 20 genes with the highest methylation levels in 8 breast cancer patients with poor hormone therapy response compared with 21 patients with complete response to hormone therapy were identi ed. Next, we further found that 11940 genes showed much lower DNA methylation levels in breast, colon, rectal, lung, uterine, gastric, esophageal, pancreatic, liver, and prostate normal tissues. Finally, TCGA RNA sequencing data of 38 paired breast cancer samples from Western patients showed that the expression of 2474 genes were decreased by 50% in the breast cancer samples. The TMEM240 gene was identi ed using InteractiVenn ( Figure 1B). Few reports about TMEM240 in women cancer was found. Methylation of TMEM240 was further analyzed in the TCGA cohort, and the gene was found to be highly methylated in breast cancer, endometrial and uterine cancer. The cluster analysis of the TMEM240 methylation pattern was visualized as a heatmap ( Figure 1C). The role of TMEM240 in breast cancer is unclear. Therefore, TMEM240 in breast cancer was selected for further analysis. A comprehensive analysis of its epigenetic alterations, mRNA and protein expression was performed, and the biological role of TMEM240 was further studied. To investigate whether TMEM240 is associated with breast cancer cell migration, MDA-MB-231 cells were transfected with TMEM240 or si-TMEM240 for 24 h. The motility of the cells was then analyzed using transwell assays and wound healing assays. The data revealed that an increase in TMEM240 expression suppressed the migration ability of MDA-MB-231 cells by 75.6% ( Figure 2E). Knockdown of TMEM-240 in MDA-MB-231 cells signi cantly induced cell migration by 63.0% and 62.7% based on the results obtained using transwell assays ( Figure 2F) and wound healing assays ( Figure 2G), respectively.

TMEM240 protein is mainly distributed in cell membranes and in the cytoplasm
Although TMEM240 is predicted to be a membrane protein, to date no study has reported the intracellular distribution of the TMEM240 protein.
According to the structure of TMEM240 protein reported in the UniProtKB/Swiss-Prot database, the protein contains two transmembrane protein regions located between amino acid residues 5~25 and 90~110. Exogenous expression of TMEM240 and immuno uorescent staining were used to examine the distribution of the protein in the DLD-1 cell line. Deconvolution and 3D reconstruction of immuno uorescence images showed that the TMEM240 protein was mainly concentrated in the cytoplasm and cell membranes ( Figure 4).
Low TMEM240 protein expression in Taiwanese and South Korean breast cancer patients TMEM240 reveals the tumor suppressor potential for breast cancer cell growth and migration ( Figure 2). To investigate whether TMEM240 protein expression is altered in cancerous breast tissues, TMEM240 protein expression in 76 breast tumors from 40 Korean breast cancer patients and 36 Taiwanese breast cancer patients was analyzed by immunohistochemistry. TMEM240 was observed to be localized to the cell membrane and cytoplasm in normal breast tissues ( Figure 3A). The protein was expressed at lower-than-normal levels in 88.2% (67/76) of the tumors from breast cancer patients and in 90.0% (9/10) of metastatic tumors from breast cancer patients ( Figure 3B and 3C and Table 1). Almost all triple-negative breast cancer patients (95.7%, 22/23) had de cient TMEM240 protein expression (Table 1).

Promoter hypermethylation and low TMEM240 mRNA expression in Taiwanese breast cancer patients
Low expression of TMEM240 protein was observed in breast cancer patients. We investigated whether TMEM240 mRNA was also expressed at lower levels in breast cancer. We analyzed TMEM240 mRNA expression in 52 paired Taiwanese breast cancer tissues. In 50.0% (26/52) of these tissues, TMEM240 mRNA expression was lower in the breast cancer tumor tissue than in the normal breast tissue ( Figure 3D, Table 1). We further analyzed the methylation patterns of TMEM240 in paired 101 Taiwan breast cancer patients, the methylation level of TMEM240 was at least 2-fold higher in 54.5% (55/101) breast tumor tissues than in the matched normal breast tissues ( Figure 3E, 3F and Table 1). The DNA hypermethylation levels and mRNA expression levels of TMEM240 showed a signi cant negative correlation by Spearman rank correlation coe cient analysis (p = 0.037). To determine whether hypermethylation of TMEM240 is involve in the regulation of mRNA expression, TMEM240 mRNA expression was investigated using administration of the DNA demethylating drug decitabine (DAC) to T47D and MDA-MB-231 breast cancer cells. The cells were treated with DMSO and DAC for 48 h. In the DAC groups for the two cell lines, methylation of TMEM240 decreased to 28.36% and 7.8%, respectively, of that in the DMSO group (p = 0.001, Figure 3G and 3H, left panel), and TMEM240 mRNA expression increased by 150-fold and 337-fold, respectively (p < 0.001, Figure 3G and 3H, right panel), suggesting that hypermethylation of the TMEM240 promoter is the main mechanism through which TMEM240 silencing occurs.
TMEM240 promoter hypermethylation and low mRNA expression in breast cancer tissues from the TCGA dataset To further evaluate alterations in TMEM240 hypermethylation and mRNA expression in Western breast cancer patients, we analyzed the TCGA data of the Illumina In nium HumanMethylation450 BeadChip array for 78 breast cancer tumors, 78 matched normal tissues and 623 breast cancer tumor tissues and displayed the methylation levels in a heatmap. The exon 1 region of TMEM240 was hypermethylated in 40.3% (251/623) of the breast tumor tissues ( Table 2). Analysis of RNA sequencing data from TCGA showed that TMEM240 mRNA expression was reduced by half in 51.4% (37/72) of the breast cancer tumor tissues compared with the matched normal breast tissues (p = 0.019, Figure S3) and in 60.2% (458/761) of tumors from breast cancer patients (Table 2). The DNA hypermethylation levels and mRNA expression levels of TMEM240 showed a signi cant negative correlation on Spearman rank correlation coe cient analysis (p=0.049). Hypermethylation of TMEM240 was associated with Asian, ERnegative, PR-negative and triple-negative breast cancer patients and patients with invasive ductal carcinoma (all p < 0.001, Table 2). In addition, Kaplan-Meier curves indicated that patients with hypermethylation of TMEM240 had a poor survival rate ( Figure 4A, log rank test, p = 0.003). A Cox proportional hazards survival analysis further adjusted for race, age, tumor type, tumor stage and menopausal state showed that TMEM240 promoter hypermethylation was signi cantly and independently associated with 10-year overall survival (Table 3, p = 0.002).
Hypermethylation of TMEM240 in breast cancer was associated with poor treatment response in the TCGA cohort set To further investigate whether hypermethylation of TMEM240 is associated with poor treatment response, we analyzed the relationship between hypermethylation of TMEM240 and clinical treatment response to chemotherapy, hormone therapy and targeted therapy in patients from the TCGA cohort. The results indicated that patients with hypermethylation of TMEM240 had poor chemotherapy response (Table 4, p = 0.012) and poor hormone therapy response (Table 4, p < 0.001). Better hormone therapy response was observed in 85.0% of patients with lower methylation of TMEM240 but in only 28.6% of patients with hypermethylation of TMEM240 (Table 4B, Mann-Whitney U test, p = 0.005). Higher methylation of TMEM240 was associated with poorer response to tamoxifen treatment ( Figure 4C, Mann-Whitney U test, p = 0.041) and with poorer response to aromatase inhibitor treatment ( Figure 4D, Mann-Whitney U test, p = 0.037).
To further investigate whether the expression of TMEM240 may be involved in the response to hormone drug treatment, a cell proliferation assay was performed after overexpression and/or knockdown of TMEM240 and tamoxifen treatment in T47D breast cancer cells. The proliferation of T47D (ER+/PR+) cells was signi cantly decreased by 62.9% when cells transfected with si-control were treated with 20 mM tamoxifen (p = 0.003), but only a 31.1% decrease in proliferation was observed in cancer cells transfected with si-TMEM240 ( Figure 4E). The data indicate that lower expression of TMEM240 is related to resistance to tamoxifen treatment. Overexpression of TMEM240 in T47D cells induced 76.9% cancer cell death when the cells were treated with 20 mM tamoxifen but only a 46.0% decrease in the vector control cancer cells when treated with 20 mM tamoxifen ( Figure 4F).
Circulating methylated TMEM240 predicts disease progression and poor hormone therapy response in Taiwanese breast cancer patients Hypermethylation of TMEM240 was found in breast tumors of patients who displayed poor treatment response, especially in tumors from patients who received hormone therapy. Detection of circulating methylated TMEM240 in the plasma of patients with poor treatment response could provide a potential tool for real-time monitoring of clinical outcomes after medical treatment. Breast cancer patients were recruited from the Taipei Medical University Hospital and Shuang Ho Hospital and were followed up for at least 1 year. After these patients received treatment, circulating methylated DNA was extracted from their plasma at 3-6 month intervals and analyzed by QMSP. The patients with poor prognosis had signi cantly higher circulating methylated TMEM240 levels than other patients but did not display higher levels of CA-15-3 and CEA ( Table 5). The level of circulating methylated TMEM240 dramatically and gradually decreased in breast cancer patients following treatment (Case 1 and Case 2, Figure 5A-B). When patients experienced disease progression, recurrence or metastasis, the levels of circulating methylated TMEM240 increased signi cantly (Case 3 and Case 4, Figure 5C-E, Mann-Whitney U test, p < 0.001). The circulating methylated TMEM240 test for poor prognosis prediction was found to have 87.5% sensitivity (28/32), 93.1% speci city (27/29), and 90.2% accuracy (55/61), values that are better than those obtained using the currently used biomarkers CEA and CA-153 (Table 5 and Figure 5).
Hypermethylation of TMEM240 in breast cancer was associated with poor response to hormone therapy in the TCGA cohort. We further investigated whether patients with an increase in circulating methylated TMEM240 in plasma experienced disease progression, recurrence or metastasis after hormone therapy. The results indicated that the patients with poor response after hormone therapy had signi cantly higher levels of circulating methylated TMEM240 (Table 5). The circulating methylated TMEM240 test for poor hormone therapy response prediction was found to have a sensitivity of 76.9% (10/13), a speci city of 95.7% (22/23), and an accuracy of 88.9% (32/36) ( Figure 5E-F).

Discussion
Aberrant promoter hypermethylation of CpG islands associated with TSGs can cause transcriptional silencing and contribute to tumorigenesis. In the present investigation, hypermethylation of TMEM240 in patients with poor hormone therapy response was identi ed using genome-wide methylation array analysis. QMSP con rmed the presence of TMEM240 hypermethylation in Taiwanese breast cancer tumor tissues compared with normal tissues. In the TCGA cohort, hypermethylation of the promoter region of TMEM240 was found in 40.3% of tumors. It occurs more frequently in Asian patients (54.8%). Results similar to those found in the Asian TCGA cohort were consistently found in Taiwanese breast cancer patients (54.5%). Low expression of TMEM240 protein was found in most Taiwanese and Korean breast cancer patients. Moreover, patients with hypermethylation of TMEM240 had poor 10-year overall survival. In addition, hypermethylation of TMEM240 was observed in patients with progressive disease, especially in patients treated with hormone therapy. The results obtained for Taiwanese breast cancer patients were similar to those obtained for patients in the TCGA cohort. Hypermethylation of TMEM240 in 87.8% of CRC, 80.0% of esophageal cancer and 80.4% of liver cancer patients was reported in our previous study [13]. Here, we further analyzed a TCGA cohort and found high methylation of TMEM240 in breast cancer and in endometrial and uterine cancer. Alterations in TMEM240 in female cancer are less frequent than those that occur in gastrointestinal cancer, but they are associated with poor clinical treatment response and poor prognosis in breast cancer. Circulating methylated TMEM240 dramatically and gradually decreased and then diminished in breast cancer patients without disease progression, whereas it increased in breast cancer patients with recurrence or metastasis. Hypermethylation of TMEM240 leads to low expression of TMEM240 mRNA and low TMEM240 protein expression. TMEM240 expression induces breast cancer cell death and enhances the cellular response to hormone therapy drugs, suggesting that de ciency in TMEM240 expression plays an important role during cancer progression in breast cancer patients.
In addition to its association with progressive disease and poor prognosis, hypermethylated TMEM240 was found to be strongly associated with ER/PR negative breast cancer, TNBC and poor hormone therapy response. Almost all triple-negative breast cancer patients (95.7%, 22/23) displayed de cient TMEM240 protein expression. Patients with hypermethylation of TMEM240 often had poor hormone therapy response (Table 3, p < 0.001). Patients who had circulating hypermethylated TMEM240 also experienced disease progression (Figure 7). The data indicate that TMEM240 de ciency is involved in breast tumorigenesis through pathways other than the ER/PR and HER2 pathways, leading to a poor hormone therapy response. Even patients with hypermethylation of TMEM240 and positive ER/PR expression exhibited poor hormone therapy responses, including treatment with tamoxifen or aromatase inhibitors (AIs) ( Table 4). Only in patients in which TMEM240 was expressed at su cient levels in the cancer cells did hormone therapy produce a good therapeutic response (Figure 2 and Figure 3).
Ki-67 as a parameter and multigene analysis (MGA) have been used to predict the response to hormone therapy [25,26]. Methylation of TMEM240 may play a role in determining resistance to AI drugs such as Ki-67 or MGA. The mechanisms and pathways that are modulated by TMEM240 are worth investigating further. TNBC represents a group of breast cancers with heterogeneous genomic features. There are several different subtypes of TNBC, including the Vanderbilt subtype and the Baylor subtype [27, 28]. Each subtype carries a different set of mutant genes [27,28]. Further study may focus on the relationship between TMEM240 and speci c subtypes of TNBC.
Advances in detection technology have reduced breast cancer death rates in several Western countries [15]. Therefore, the development and use of biomarkers of treatment response can improve patient outcomes. The presence of breast-derived circulating DNA is indicative of residual disease after treatment [29]. Circulating methylated TMEM240 dramatically and gradually decreases and then diminishes in patients with various subtypes of breast cancer who do not show disease progression (Figure 7), suggesting that measurement of circulating methylated TMEM240 could be used to detect the presence of residual disease. In addition, the level of circulating methylated TMEM240 in plasma increased further in breast cancer patients with recurrence or metastasis (Figure 7). In these patients, the concentrations of CEA and CA15-3 in serum remained normal or increased much later than did the circulating methylated TMEM240. The detection of CEA and CA15-3 was incapable of revealing disease progression and poor treatment response in several patients ( Table 5). Measurement of circulating methylated TMEM240 could be used to monitor and detect early disease progression after treatment and during long-term follow-up. Although hypermethylation of TMEM240 also occurs in other types of cancer, its high alteration in cancers may assist the detection of disease progression. Combining measurement of TMEM240 hypermethylation with the measurement of additional breast cancer-speci c methylated DNA biomarkers that are associated with disease progression will improve detection sensitivity and cancer speci city.

Conclusion
De ciency in TMEM240 expression plays an important role during cancer progression in breast cancer patients. Circulating hypermethylated TMEM240 may represent a potential biomarker for disease progression and poor hormone therapy response. The data generated in this study are available from the corresponding author upon reasonable request.

Competing interests
There are no con icts of interest.  Tables   Table 1.   2. For some categories, the number of samples (n) was lower than the overall number analyzed because clinical data were unavailable for those samples. 3. When the TMEM240 expression level in breast tumors was less than half of the mean of TMEM240 expression levels in adjacent normal breast tissues was de ned as low expression. 4. The TMEM240 promoter methylation level in breast tumors being 2-fold higher than in adjacent normal breast tissues was de ned as hypermethylation. 2. When the TMEM240 expression level in breast tumors was less than half of the mean of TMEM240 expression levels in adjacent normal breast tissues was de ned as low expression from TCGA data set using RNA sequencing analysis. 3.*, P < 0.05; **, P < 0.01; ***, P < 0.001.  1. These results were analyzed by the Fisher's exact test. The patients with a treatment and monitoring duration of greater than one year were included in this analysis. When the circulating methylated TMEM240 levels normalized by circulating ACTB in plasma of breast cancer patients was higher than 0.002 was de ned as abnormal. 2. For concentration of CA-153, CEA and Ki-67 expression, the number of samples (n) was lower than the overall number analyzed because clinical data were unavailable for those samples.   overexpression was measured via transwell assays. si-TMEM240 was transfected into MDA-MB-231 cells for 24 h, and the distribution of the cells was then analyzed using transwell assays (F) and wound healing assays (G). The data are presented as the mean ± SD; * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001. A t-test was used to calculate group differences in all experiments. Experiments were performed using at least two biological duplicates and three technical replicates. Localization of the TMEM240 protein was determined by deconvolution and 3D reconstruction.