ZEB1 (a positive enhancer) and KLF5 (a negative silencer) affect transcription factors and play inherently conserved roles in tumorigenesis and multidrug resistance. In humans, the rs2295080T-allele at the mTOR promoter locus has been associated with human cancer risk; however, the 63 bp spacing of another SNP rs2295079 has not been identified. Here, we discovered, for the first time, that rs2295079 (-78C/G) and rs2295080 (-141G/T) formed linkage haplotypes, with Ht1 (-78C/-141G) and Ht2 (-78G/-141T) being dominant, which were associated with distinct susceptibility to breast cancer, response to paclitaxel, and clinical outcomes in breast cancer. At the cellular level, compared with Ht1, Ht2 exhibits a much stronger effect on promoting mTOR expression, leading to enhanced tumor cell growth and strengthened resistance to PTX treatment. Mechanistically, the -141T allele of Ht2 creates a novel ZEB1-binding site; meanwhile, the -78C allele of Ht1 exists as an emerging KLF5-binding site, which synergistically induces promote/inhibit mTOR expression, cell proliferation, and excretion of cytotoxic drugs through the ZEB1/KLF5–mTOR–CCND1/ABCB1 cascade, thereby affecting the response to paclitaxel treatment in vivo and in vitro. Our results suggest the existence of a ZEB1/KLF5–mTOR–CCND1/ABCB1 axis in human cells that could be involved in paclitaxel response pathways and functionally regulate interindividualized breast cancer susceptibility and prognosis.

Implications:

This study highlights the function of haplotypes of mTOR -78C/-141G and -78G/-141T, in affecting breast cancer susceptibility and paclitaxel response regulated by ZEB1/KLF5–mTOR–CCND1/ABCB1 axis.

This article is featured in Highlights of This Issue, p. 2143

Breast cancer is a leading cause of cancer-related death in women, with an estimate of more than two million new cases worldwide in 2018, and it accounts for 30% of all new cancer cases in women in the United States (1, 2). Accumulating evidence supported that the hereditary genetic variations of candidate genes, such as SNPs and their haplotypes, induced interindividualized cancer susceptibility and diverse responses to chemotherapeutic agents (3, 4). Compared with conventional analysis of individual single SNPs, recently developed multiallelic haplotype-based association analysis based on a cluster of two to three SNP markers significantly improved the power and robustness of detecting cancer susceptibility loci and responsiveness variability (5, 6). However, it is less well understood why SNPs form haplotypes, or “block-like” linkage disequilibrium (LD), and the functional advantages the linked SNPs have, and what the underlying molecular mechanisms are, need to be further examined.

The mTOR gene is located at chromosome 1q36.2 (7). mTOR acts as a central switch that determines cell fate in tumorigenesis, including cell proliferation (8–12), migration (13), and multidrug resistance (14, 15). The SNP rs2295080 (-141G/T), located in the promoter of the mTOR gene, has been found to be associated with an increased risk of various cancers, such as gastric cancer (16, 17), esophageal carcinoma (18), bladder cancer (19), etc. (20–22). However, previous studies have consistently focused on the SNP but have failed to study any molecules involved in tumorigenesis and progression. Given the functional importance of rs2295080, this SNP warrants further investigation to determine how it is controlled by key transcriptional factors, whether SNP rs2295080 gives rise to LD, and the potential functional impacts of the relevant haplotypes on cell proliferation, progression, and drug response in breast cancer.

ZEB1 and KLF5 are both prime elements of a network of transcription factors that play important roles in tumorigenesis and multidrug resistance (23–26). Recently, Liu and colleagues found that mTOR is positively regulated by ZEB1 because of multiple consensus E-box–binding sites for ZEB1 in mouse embryonic fibroblasts (27). KLF5, another key transcription factor, plays important roles in the regulation of cell cycle and protection against cell apoptosis (28). Furthermore, KLF5 is capable of protecting cells from a variety of harmful stimuli (29, 30).

In this study, a haplotype-based case–control study was performed on the basis of the HapMap Project Database (http://hapmap.ncbi.nlm.nih.gov/), and SNPs in the 5′ untranslated region (UTR) of mTOR gene were genotyped and sequenced. We discovered that rs2295080 (-141G/T) and rs2295079 (-78C/G) constitute linkage haplotypes, with Ht1 (-78C/-141G) and Ht2 (-78G/-141T) being dominant, which are associated with distinct susceptibility to breast cancer, response to paclitaxel, and clinical prognosis. We further found that Ht1 and Ht2 have silencer/enhancer properties, and a novel ZEB1-binding site created by the -141T allele of Ht2 and an emerging KLF5 binding site that exists on the -78C allele of Ht1 synergistically affect expression of mTOR in response to diverse paclitaxel and clinical outcomes. Molecular analyses were performed in cell lines, and xenograft mouse models revealed that ZEB1/KLF5 mediates the occurrence of functional interactions between the 5′ UTR of mTOR promoter haplotypes through the ZEB1/KLF5–mTOR–CCND1/ABCB1 axis, thereby affecting interindividualized breast cancer susceptibility, progression, and paclitaxel response.

Study population

The study was subject to approval by the Institutional Review Board of China Medical University, and all subjects gave their written informed consent prior to inclusion in the ongoing study. Briefly, 547 female patients with histopathologically confirmed breast cancer and 504 female healthy volunteers were consecutively recruited from First Hospital of China Medical University between October 2008 and June 2013. After an interview, 5 mL of peripheral venous blood was collected for further SNP genotyping. Clinical data were obtained from the interviewer-administered health risk questionnaires and medical records. The healthy controls were recruited from individuals who visited the same hospitals for a health checkup, had no known medical illness or hereditary disorders, and were not taking any medications.

SNP genotyping

Genomic DNA was isolated from a leukocyte cell pellet of each blood sample by using previously described methods (31). The TaqMan allelic discrimination method was used to genotype the selected SNPs. The SNP allele–specific probes were labeled with the fluorescent dyes HEX and FAM by using the TaqMan SNP Genotyping Assays of the ABI 7500 Fast Real-Time PCR platform (Applied Biosystems, Life Technologies Corporation). For quality control, 10% of the samples were confirmed by repeated DNA-sequencing analysis, and the genotypes were 100% concordant.

IHC

The samples were incubated overnight in 4°C with the primary antibody (p-mTOR, 1:100, Cell Signaling Technology, #2976; mTOR, 1:100, Cell Signaling Technology, #2983; ABCB1, 1:200, Boster, A00049; and CCND1, 1:200, Boster, BA0770). The immunostaining was examined under a light microscope by two pathologists blinded to the experimental conditions. The intensity of immunoreactivity was scored as follows: 0 for no staining, 1 for weak staining, 2 for moderate staining, and 3 for strong staining. The percentage of stained cells was scored by using 5% increments (0%, 5%, 10%, … 100%). The final immunoreactive score was determined by multiplying the intensity score with the percentage of positively stained cells.

Cell culture

Human breast cancer MCF7 cell and 293T cell were purchased from Peking Union Medical College Cell Resource Center (PUMCCRC, Beijing, China). Cell lines were authenticated using short tandem repeat analysis by Shanghai GeneChem Co., Ltd. The cells were demonstrated to be free of contamination with mycoplasma by PCR. The MCF7 cells were cultured in DMEM culture medium (Invitrogen) containing 10% FBS (HyClone), 100 U/mL penicillin (Invitrogen), and 100 U/mL streptomycin (Invitrogen) with culture environment of 5% CO2 and 37°C. Cells between third and eighth passage were used for the experiments. The cells were maintained in a logarithmic growth phase by routine passages every 4 to 5 days at 1:3 split ratio. Paclitaxel-resistant MCF7 cell (MCF7/PTX) was induced by our laboratory. MCF7/PTX was added PTX on time to maintain the drug resistance phenotype.

Cell proliferation and viability assay

MCF7 or MCF7/PTX cells were seeded into 96-well plates (3,000 cells/well). Twenty-four hours after transfection, paclitaxel (0.03, 0.11, 0.46, 1.83, 7.3, 29, 117, and 468 nmol/L) was added into the culture medium. Forty-eight hours later, the cells were incubated with 10 μL of CCK-8 (Cell Counting Kit-8, Beyotime) at 37°C for 4 hours. The absorbance value was measured at 450 nm using an Anthos 2010 microplate reader (Anthos Labtec Instruments). The inhibition rate = (normal control − experience)/(normal control − negative control) × 100%. There was no paclitaxel in the normal control well, and there were no cells in the negative control well. The experiment was repeated three times.

Western blot analysis

The total protein of MCF7 or MCF7/paclitaxel cells was extracted and quantified. Sixty micrograms of proteins from each cell lysate were separated by 10% SDS-PAGE electrophoresis and transferred to a nitrocellulose membrane. Primary antibodies and the final antibody concentrations are as follows: p-mTOR (1:1,000, 430.9 ng/mL, Cell Signaling Technology, #5536), mTOR (1:1,000, 89.0 ng/mL, Cell Signaling Technology, #2983), ABCB1 (1:1,000, 200 ng/mL, Boster, A00049), CCND1 (1:1,000, 200 ng/mL, Boster, BM0771), and GAPDH (1:1,000, 200 ng/mL, Boster, BM3876) were incubated overnight at 4°C. The secondary antibody (Boster, BA1054, 1:1,000, 1 μg/mL, BA1050, 1:1,000, 1 μg/mL) was incubated at room temperature for 1 hour. Chemiluminescence was detected by ECL (Amersham). Data were statistically analyzed by FluorChem V2.0 (Alpha Innotech Corp).

Immunofluorescence

MCF7 cells were seeded at a density of 5 × 104 cells/well in a chamber slide. Twenty-four hours after transfection, cells were fixed with 4% (w/v) paraformaldehyde in PBS for 10 minutes. Primary antibodies and the final antibody concentrations are as follows: p-mTOR (1:100, 4.309 μg/mL, Cell Signaling Technology, #5536) and mTOR (1:100, 0.890 μg/mL, Cell Signaling Technology, #2983) were incubated overnight at 4°C. The nuclei were located with DAPI (diamidino-2-phenylindole). Fluorescence was detected using a confocal laser scanning microscope (FV1000S-SIM/IX81, Olympus).

Cell-cycle analysis

MCF7 and MCF7/paclitaxel cells were cultured in 6-well plates (5 × 104 cells/well). After transfection, the cells were cultured with 7.3 nmol/L paclitaxel for another 24 hours. Cells were harvested, and cell density was adjusted to 107 cells/mL. The suspension (100 μL) was incubated in the dark at room temperature for 30 minutes with a solution of PI (propidium iodide). The cell cycle was analyzed by flow cytometry using a FACSCalibur instrument. The percentage of cells in S phase in the different groups was compared. The experiment was repeated at least three times.

In vivo xenograft experiments

The transplanted tumor model of BALB/c nude mice (female, 4- to 6-week-old) was prepared by subcutaneous injection at the scapula of MCF7 cells. The cells (8 × 106) were suspended in 200 μL PBS and Matrigel (1:1; BD Biosciences). Every 3 days, the weights of the xenograft mice were recorded until the end of the experiment. Ten to 12 days after cell injection, tumor formation was measured at intervals of 3 days, and the tumor volume was then calculated by the length (L) and width (W) according to the following formula: V = (L × W2) × 0.5. At the same time, paclitaxel (40.45 mg/kg) was injected through the caudal vein at the first time (day 17) in the groups of paclitaxel alone, pcmTOR-Ht1 and pcmTOR-Ht2 treated (injected intratumorally), once every 3 days a total of 6 times. Simultaneously, the NC group was established (n = 5). The effects on the weight of the xenograft mice and in vivo tumor growth were observed. The plasma was collected 4 hours after paclitaxel injection at 20, 26, and 32 days. Mice were sacrificed and tumors were weighed and harvested for histologic examination and UPLC-MS/MS. All mice were bred in pathogen-free conditions at the Animal Center of China Medical University. The Animal Ethics Committee of China Medical University approved this study. All animal studies were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals. The assay was performed as in our previous publications, with some modifications (32, 33).

UPLC-MS/MS assay

Tissue from tumors was weighed and then homogenized with a high-speed homogenizer (Shanghai Kingdom Biochemical Instrument). Cells were cultured in 6-well plates (5 × 104 cells/well). pcmTOR-Ht1, pcmTOR-Ht2, and NC were transfected into each group. Twenty-four hours later, the cells were cultured with 7.3 nmol/L paclitaxel for another 24 hours. Then, cells were lysed by repeated freezing and thawing. After centrifugation at 10,000 × g for 5 minutes, supernatant was collected for UPLC-MS analysis. The detailed procedure was shown in the Supplementary Methods.

Luciferase reporter assay

The cells were transferred to 96-well plates (2 × 104/well). The pGL3-Ht1, pGL3-Ht2, and pRL-TK plasmids were transfected into 293T or MCF7 cells. FLuc luminescence was measured by using Dual-Luciferase Reporter Assay System (Promega). The value of Fluc luciferase activity was corrected by Rluc luciferase activity. Relative luciferase activity was normalized to the negative control. Assays were conducted in triplicate in a single experiment and then as three independent experiments.

Fusion protein–NanoLUC luciferase reporter analysis

NLF1-C[CMV/Hygro] Vector was purchased from the Promega Corporation. The pNluc-ZEB1 and pNluc-KLF5 could generate an in-frame Nluc fusion protein. The pGL3-Ht1, pGL3-Ht2, pNLuc-ZEB1, and pNLuc-KLF5 plasmids were transfected at a concentration of 1 μg/mL. Cells were harvested at the following times: 4, 8, 12, 16, 20, and 24 hours, and then luciferase activity was detected by using Nano-Glo Dual-Luciferase Reporter (NanoDLR) Assay System (Promega).

Chromatin immunoprecipitation

The plasmids pcmTOR-Ht1 and pcmTOR-Ht2 were transfected to MCF7 cells. Samples were collected and subjected to immunoprecipitation with ZEB1 antibody (GENE Tex, GTX105278) and KLF5 antibody (MILLIPORE, 07-1580) overnight at 4°C. DNA fragments were purified by MiniBest Universal Genomic DNA Extraction Kit Ver.5.0 (Takara). A 2 μL of chromatin immunoprecipitation (ChIP) extraction solution was subjected to 40 cycles of amplification for qPCR assay. PCR products were sequenced by Sangon Biotech.

Statistical analysis

All data analysis was carried out using SPSS 16.0 software packages (SPSS Inc.). Associations between SNPs and breast cancer and clinical variables were analyzed by OR and 95% confidence intervals (CI) using unconditional logistic regression models. Overall survival (OS) was defined as the time between the surgery and death or last known follow-up. Disease-free survival (DFS) was measured from surgery until an occurrence of recurrence, death, or last known follow-up. The survival curves were generated by using the Kaplan–Meier method, and log-rank test was used to estimate the associations of the DFS or OS with SNPs. Multivariate Cox proportional hazards regression models were applied to obtain the adjusted HR and 95% CI for evaluating the independent prognostic value of genotypes and clinical variables. Data from qRT-PCR, reporter assays, cell proliferation assays, UPLC-MS/MS assays, and in vivo xenograft experiments were analyzed using two-tailed Student t test. All statistical tests were two-sided, and a P value of <0.05 was considered statistically significant.

Chemotherapy response- and prognosis-associated polymorphism and haplotype identification in the mTOR promoter region

We explored the connection between breast cancer risk and genetic variations in the mTOR promoter region. On the basis of the HapMap Project Database (http://hapmap.ncbi.nlm.nih.gov/), -1,100 base pairs upstream from the transcription start site of mTOR gene were genotyped and sequenced in a relatively large number of samples in a case–control study (Supplementary Table S1). Three SNPs at loci -78 (rs2295079, C/G), -141 (rs2295080, G/T), and -422 (rs17027478, G/T) were identified in the mTOR promoter region (the minor allele frequency >10%). Interestingly, SNPs at loci -78 (C/G) and -141 (G/T), but not at locus -422 (G/T), constitute a perfect LD block with r2 = 0.860 (Fig. 1A). Ht1 (-78C/-141G) and Ht2 (-78G/-141T) account for more than 97.0% among the genotypes (Fig. 1B). The distribution frequency of Ht1 and Ht2 was 27% and 71% in controls, and 22% and 77% in breast cancer, respectively (Fig. 1C). Further cancer risk analysis revealed that Ht2 was significantly associated with an increased susceptibility of breast cancer [adjusted OR (95% CI): 1.339 (1.096–1.637), P = 0.004; Supplementary Table S2]. The distribution frequency of Ht2/Ht2 homozygous variation was a high-risk haplotype compared with those carrying Ht1/Ht1 (Fig. 1D and E).

Figure 1.

The association of Ht1 or Ht2 with susceptibility and clinical outcome in patients with breast cancer. A, Schematic graph representing the distribution of candidate SNPs and their haplotypes in the promoter region of mTOR gene. B, The percentage distribution of Ht1 and Ht2 in the healthy controls and the patients with breast cancer. C, Table listing the frequency distribution of haplotypes in the patients with breast cancer and controls. D and E, Pie chart illustrating and table listing mTOR haplotypes percentage distribution in both controls and patients. F, RECIST criteria and logistic regression analysis evaluated the response variation to paclitaxel-based NCT in patients with breast cancer with haplotypes of the mTOR gene. G, Kaplan–Meier plots of overall survival probability comparison between Ht1 and other haplotypes, between Ht2 and other haplotypes, and between Ht1 and Ht2 in the patients with breast cancer who received paclitaxel (PTX)-based chemotherapy. H, Comparison of the median survival time between Ht1 and other haplotypes, between Ht2 and other haplotypes, and between Ht1 and Ht2 in the patients with breast cancer who received paclitaxel-based chemotherapy.

Figure 1.

The association of Ht1 or Ht2 with susceptibility and clinical outcome in patients with breast cancer. A, Schematic graph representing the distribution of candidate SNPs and their haplotypes in the promoter region of mTOR gene. B, The percentage distribution of Ht1 and Ht2 in the healthy controls and the patients with breast cancer. C, Table listing the frequency distribution of haplotypes in the patients with breast cancer and controls. D and E, Pie chart illustrating and table listing mTOR haplotypes percentage distribution in both controls and patients. F, RECIST criteria and logistic regression analysis evaluated the response variation to paclitaxel-based NCT in patients with breast cancer with haplotypes of the mTOR gene. G, Kaplan–Meier plots of overall survival probability comparison between Ht1 and other haplotypes, between Ht2 and other haplotypes, and between Ht1 and Ht2 in the patients with breast cancer who received paclitaxel (PTX)-based chemotherapy. H, Comparison of the median survival time between Ht1 and other haplotypes, between Ht2 and other haplotypes, and between Ht1 and Ht2 in the patients with breast cancer who received paclitaxel-based chemotherapy.

Close modal

The correlation of Ht1 and Ht2 with clinicopathologic parameters in patients with breast cancer was summarized in Supplementary Table S3. The distribution frequencies of Ht1 and Ht2 were significantly associated with tumor size (P = 0.019), clinical stages (P = 0.004), and lymph node metastasis status (P = 0.015). For Ht2 carriers, a significantly increased frequency was observed in tumors of >3.0 cm [adjusted OR (95% CI): 1.438 (1.063–1.947)], clinical stage III or IV [adjusted OR (95% CI): 1.607 (1.159–2.228)], and higher in lymph node-positive patients [adjusted OR (95% CI): 1.431 (1.072–1.911)]. We therefore modeled the associations with each SNP on cancer risk and clinicopathologic parameters. These conditional risk estimates and significance levels are presented in Supplementary Tables S3 and S4. These data suggest that Ht1 and Ht2 were the most predominant genotypes and potential biomarkers for evaluating individualized cancer susceptibility and progression.

For further analysis, we selected the direct haplotype of Ht1 and Ht2 to assess the response individuation of neoadjuvant chemotherapy (NCT). The response evaluation criteria in solid tumors (RECIST) were used to evaluate the response to paclitaxel-based chemotherapy. The response ratio in the patients carrying Ht2 (65.3%) was significantly lower (P = 0.012) than in the patients carrying Ht1 (81.8%), indicating a strong association between Ht2 and chemotherapy resistance in patients with breast cancer (Fig. 1F; Supplementary Table S5).

Next, we explored the effects of the two representative haplotypes on prognostic impacts using the Kaplan–Meier analysis and multivariate Cox regression analysis in patients with breast cancer. Specifically, with paclitaxel-based chemotherapy, the low-risk haplotype Ht1/Ht1 (-78CC/-141GG) carriers had significantly longer OS and median survival time than others (Fig. 1G and H). Multivariate Cox regression analysis repeated and verified that the haplotypes (Ht2 vs. Ht1) were independent prognostic indicators for both OS and DFS in patients with breast cancer, and Ht2 is associated with a significantly shorter OS [P = 0.008, adjusted HR (95% CI): 2.535 (2.339–3.847)] and DFS [P = 0.010, 2.224 (2.072–3.695)] than Ht1 (Table 1). Notably, the prognostic impacts of each SNP (i.e., -78GG vs. -78CC; -144TT vs. -144GG) in the above patient cohorts were much more limited than the haplotypes (Ht2 vs. Ht1), suggesting that the linked SNPs contributed synergistically to the overall prognostic impacts of the corresponding haplotypes.

Table 1.

Multivariate COX regression analysis of mTOR genetic polymorphisms in association with DFS and OS in patients with breast cancer

OSDFS
VariableAdjusted HR (95% CI)aPaAdjusted HR (95% CI)aPa
mTOR-78C/G 
 CC 1.00  1.00  
 GC 1.211 (0.795–1.845) 0.374 0.661 (0.299–1.459) 0.306 
 GG 1.855 (1.029–3.3440.039 0.671 (0.211–2.128) 0.498 
mTOR-141G/T 
 TT 1.00  1.00  
 TG 2.222 (0.970–5.102) 0.059 0.896 (0.651–1.232) 0.499 
 GG 3.145 (1.901–8.264) 0.075 0.905 (0.738–1.110) 0.338 
Combined genotype 
 -78GG/-141TT 1.00  1.00  
 -78GC/-141GT 0.741 (0.441–1.245) 0.258 0.815 (0.510–1.302) 0.392 
 -78CC/-141GG 1.613 (0.981–2.650) 0.049 1.380 (0.844–2.255) 0.199 
mTOR haplotypes 
mTOR-Ht1 1.00  1.00  
mTOR-Ht2 2.535 (2.339–3.847) 0.008 2.224 (2.072–3.695) 0.010 
Age (years) 
 <50 1.00  1.00  
 ≥50 1.031 (0.592–1.794) 0.914 1.010 (0.580–1.759) 0.971 
Menopausal status 
 Premenopausal 1.00  1.00  
 Postmenopausal 1.262 (0.729–2.187) 0.406 1.419 (0.703–2.867) 0.329 
Family history of cancer 
 No 1.00  1.00  
 Yes 1.790 (1.424–2.250) <0.001 2.566 (1.293–5.094) 0.007 
Tumor size (cm) 
 ≤3.0 1.00  1.00  
 >3.0 4.397 (3.131–6.175) <0.001 0.941 (0.640–1.382) 0.756 
Lymph node metastasis status 
 Node-negative 1.00  1.00  
 Node-positive 2.942 (2.009–4.308) <0.001 2.326 (1.189–4.550) 0.014 
OSDFS
VariableAdjusted HR (95% CI)aPaAdjusted HR (95% CI)aPa
mTOR-78C/G 
 CC 1.00  1.00  
 GC 1.211 (0.795–1.845) 0.374 0.661 (0.299–1.459) 0.306 
 GG 1.855 (1.029–3.3440.039 0.671 (0.211–2.128) 0.498 
mTOR-141G/T 
 TT 1.00  1.00  
 TG 2.222 (0.970–5.102) 0.059 0.896 (0.651–1.232) 0.499 
 GG 3.145 (1.901–8.264) 0.075 0.905 (0.738–1.110) 0.338 
Combined genotype 
 -78GG/-141TT 1.00  1.00  
 -78GC/-141GT 0.741 (0.441–1.245) 0.258 0.815 (0.510–1.302) 0.392 
 -78CC/-141GG 1.613 (0.981–2.650) 0.049 1.380 (0.844–2.255) 0.199 
mTOR haplotypes 
mTOR-Ht1 1.00  1.00  
mTOR-Ht2 2.535 (2.339–3.847) 0.008 2.224 (2.072–3.695) 0.010 
Age (years) 
 <50 1.00  1.00  
 ≥50 1.031 (0.592–1.794) 0.914 1.010 (0.580–1.759) 0.971 
Menopausal status 
 Premenopausal 1.00  1.00  
 Postmenopausal 1.262 (0.729–2.187) 0.406 1.419 (0.703–2.867) 0.329 
Family history of cancer 
 No 1.00  1.00  
 Yes 1.790 (1.424–2.250) <0.001 2.566 (1.293–5.094) 0.007 
Tumor size (cm) 
 ≤3.0 1.00  1.00  
 >3.0 4.397 (3.131–6.175) <0.001 0.941 (0.640–1.382) 0.756 
Lymph node metastasis status 
 Node-negative 1.00  1.00  
 Node-positive 2.942 (2.009–4.308) <0.001 2.326 (1.189–4.550) 0.014 

aP value and adjusted HR (95% CI) were assessed using multivariate Cox regression analysis. P < 0.05 marked in bold.

mTOR haplotypes are associated with the expression of mTOR in tissues of patient with breast cancer

Although a large number of SNPs of the mTOR gene have been found to be associated with an increased risk of various cancers, the effect of mTOR haplotypes has not yet been studied. To identify the potential haplotypes to be selected for functional analysis, we used different techniques to assess associations between these two haplotypes and local gene expression. (i) mRNA expression from the RNA-sequencing dataset platform (http://www.cbioportal.org/) of The Cancer Genome Atlas (TCGA; n = 825) was compared and showed elevated mRNA expression of mTOR, ABCB1, and CCND1 in breast tumors compared with adjacent normal breast tissue (Fig. 2A and B). (ii) Protein expression detected by IHC in breast cancer tissues based on 76 patients who received paclitacel-based NCT identified a significant correlation of the risk haplotype Ht2/Ht2 carriers with elevated p-mTOR, mTOR, ABCB1, and CCND1 levels (Fig. 2C and D). (iii) Further protein correlation analysis revealed that the elevated mTOR protein expression showed a strong positive correlation with ABCB1 (P < 0.0001) and CCND1 (P < 0.0001; Fig. 2E). Consistent with the imbalance in protein expression, prognostic analysis revealed that Ht2/Ht2 carriers had shorter OS (Fig. 2F). Using our in-house data, we analyzed the association of mTOR expression with the prognosis of patients with breast cancer who received NCT (n = 884) in the Kaplan–Meier Plotter (http://kmplot.com/analysis/), and we found that high expression of mTOR was significantly associated with shorter DFS (P = 0.013) and OS (P = 0.0033; Fig. 2G). Such imbalanced expression and prognosis data indicate that Ht2 is associated with expression differences in local genes and exhibits a strong prognostic impact on survival in patients with breast cancer who received NCT, likely through altering the expression of mTOR.

Figure 2.

The correlation of mTOR haplotypes with protein expression and prognosis in patients with breast cancer. A, TCGA data illustrated elevated levels of mTOR mRNA expression in breast cancer by RNA sequencing. B, TCGA data illustrated elevated expression levels of mTOR, ABCB1, and CCND1 in breast cancer tissues compared with those of normal tissues. C and D, IHC assays represent the protein expression of p-mTOR, mTOR, CCND1, and ABCB1 in patients with breast cancer (20 ×); Ht2/Ht2 carriers are associated with higher expression of proteins. E, mTOR expression in the patients showed a positive correlation with expression of ABCB1 and CCND1. F, Kaplan–Meier plot showed Ht2 carriers correlated with a shorter OS in NCT-treated patients. G, NCT patients with higher expression of mTOR were associated with shorter DFS and OS using the available data (n = 884; http://kmplot.com/analysis). **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 2.

The correlation of mTOR haplotypes with protein expression and prognosis in patients with breast cancer. A, TCGA data illustrated elevated levels of mTOR mRNA expression in breast cancer by RNA sequencing. B, TCGA data illustrated elevated expression levels of mTOR, ABCB1, and CCND1 in breast cancer tissues compared with those of normal tissues. C and D, IHC assays represent the protein expression of p-mTOR, mTOR, CCND1, and ABCB1 in patients with breast cancer (20 ×); Ht2/Ht2 carriers are associated with higher expression of proteins. E, mTOR expression in the patients showed a positive correlation with expression of ABCB1 and CCND1. F, Kaplan–Meier plot showed Ht2 carriers correlated with a shorter OS in NCT-treated patients. G, NCT patients with higher expression of mTOR were associated with shorter DFS and OS using the available data (n = 884; http://kmplot.com/analysis). **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Close modal

Prioritizing mTOR haplotypes for functional assays in human breast cancer cells

To examine the haplotype-specific biological effects of the mTOR promoter, we utilized paclitaxel-sensitive MCF7 cells and PTX-resistant MCF7/paclitaxel cells transfected with pcmTOR-Ht1 (containing Ht1 genotypes of the mTOR promoter and mTOR-cDNA), pcmTOR-Ht2 (containing Ht2 genotypes of the mTOR promoter and mTOR-cDNA), and NC vector (Fig. 3A). Assays of cell viability in MCF7 cells showed that the proliferation of time course pcmTOR-Ht2 was much higher than pcmTOR-Ht1 in MCF7 but not in MCF7/PTX cells (Fig. 3B; Supplementary Fig. S1A). This finding was further confirmed at the mRNA and protein levels of p-mTOR, mTOR, ABCB1, and CCND1 by qRT-PCR, Western blot, and immunofluorescence assays (Fig. 3C–E; Supplementary Fig. S1B).

Figure 3.

Elevated Ht2 expression is highly correlated with promoting cell proliferation, cycle, and chemoresistance to paclitaxel (PTX). A, Schematic graph representing the recombinant constructs containing Ht1 and Ht2, respectively. B, Cell proliferation and viability assay in MCF7 cells transfected with the haplotype pcmTOR-Ht1, pcmTOR-Ht2, or NC vector. C, qRT-PCR assays of mTOR mRNA expression. D, Western blot assays of p-mTOR and mTOR protein expression, accompanied with ABCB1 and CCND1. E, Immunofluorescence assays of p-mTOR and mTOR protein expression. F, IC50 values were determined in MCF7 cells transfected with pcmTOR-Ht1, pcmTOR-Ht2, or empty vector combined with treatment with paclitaxel. G, UPLC-MS/MS assays determined the intracellular concentration of paclitaxel in MCF7 cells in different treatment groups. H, Representative histograms depicting cell-cycle profiles and S-phase fraction (SPF) in MCF7 cells transfected with pcmTOR-Ht1, pcmTOR-Ht2, and empty vector, respectively. Cells were stained with PI and analyzed by flow cytometry assays. The results are means of triplicate independent experiments.**, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 3.

Elevated Ht2 expression is highly correlated with promoting cell proliferation, cycle, and chemoresistance to paclitaxel (PTX). A, Schematic graph representing the recombinant constructs containing Ht1 and Ht2, respectively. B, Cell proliferation and viability assay in MCF7 cells transfected with the haplotype pcmTOR-Ht1, pcmTOR-Ht2, or NC vector. C, qRT-PCR assays of mTOR mRNA expression. D, Western blot assays of p-mTOR and mTOR protein expression, accompanied with ABCB1 and CCND1. E, Immunofluorescence assays of p-mTOR and mTOR protein expression. F, IC50 values were determined in MCF7 cells transfected with pcmTOR-Ht1, pcmTOR-Ht2, or empty vector combined with treatment with paclitaxel. G, UPLC-MS/MS assays determined the intracellular concentration of paclitaxel in MCF7 cells in different treatment groups. H, Representative histograms depicting cell-cycle profiles and S-phase fraction (SPF) in MCF7 cells transfected with pcmTOR-Ht1, pcmTOR-Ht2, and empty vector, respectively. Cells were stained with PI and analyzed by flow cytometry assays. The results are means of triplicate independent experiments.**, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Close modal

To further confirm the cellular mechanisms underlying haplotype-specific effects in breast cancer following paclitaxel treatment, we compared the effects between pcmTOR-Ht1 and pcmTOR-Ht2-mediated mTOR overexpression. In paclitaxel-sensitive MCF7 cells, forced expression of pcmTOR-Ht2 substantially increased PTX IC50 from 4.17 nmol/L (control) to 61.61 nmol/L (Fig. 3F), but there was no significant difference in MCF7/paclitaxel cells (Supplementary Fig. S1C). Furthermore, the intracellular paclitaxel concentration was significantly reduced in MCF7 cells transfected with pcmTOR-Ht2 (Fig. 3G). However, the intracellular paclitaxel concentration was not significantly different in MCF7/paclitaxel cells transfected with different plasmids (Supplementary Fig. S1D). To further confirm these findings, we also employed cell-cycle assays. We observed that the percentage of cells in S-phase among MCF7 cells transfected with pcmTOR-Ht2 was significantly elevated compared with those transfected with pcmTOR-Ht1 or NC, but there was no significant difference in MCF7/paclitaxel cells (Fig. 3H; Supplementary Fig. S1E). These data suggested that overexpression of Ht2 promoted cell viability and chemoresistance to paclitaxel in MCF7 cells.

In vivo effects of mTOR haplotypes on tumor growth and chemoresistance of human breast cancer in xenograft mice

To investigate the in vivo effects of Ht1 and Ht2 on the chemosensitivity of paclitaxel, MCF7 cell lines transfected with pcmTOR-Ht1, pcmTOR-Ht2, or NC were subcutaneously implanted into nude mice, followed by paclitaxel treatment (Fig. 4A). The mice were sacrificed, and the tumors were removed during the seventh week after implantation. There were no significant differences in body weight of xenograft mice of different groups (Fig. 4B). Although paclitaxel treatment dramatically inhibited tumor growth, forced expression of pcmTOR-Ht2 almost completely reversed this inhibition. In contrast, forced expression of pcmTOR-Ht1 showed only a moderate effect on rescuing the inhibition caused by paclitaxel treatment (Fig. 4C–E). Furthermore, IHC assays showed that the levels of p-mTOR, mTOR, ABCB1, and CCND1 were much higher in tumors with Ht2 overexpression than with Ht1 (Fig. 4F).

Figure 4.

The effect of Ht1 and Ht2 on tumor growth and chemosensitivity to paclitaxel in xenograft mice. A, Diagram showing the time of tumor formation and administration in vivo. B, The body weight of mice in each group (n = 5). C, Growth curves of tumor volumes in each group (n = 5). D, Tumor weight of xenograft mice at sacrifice (n = 5). E, Photographs of tumors of each group removed at day 35 after implantation. F, H&E staining showed the transplantation tumors after transfection with pcmTOR-Ht1 and pcmTOR-Ht2; IHC method showed p-mTOR and mTOR expression in the xenograft mice after transfection of pcmTOR-Ht1 and pcmTOR-Ht2. Comparison of the expression of CCND1 and ABCB1 between pcmTOR-Ht1 and pcmTOR-Ht2. G, Diagram showing the time of tumor formation and blood collection in vivo. H, UPLC-MS/MS detection of docetaxel (internal standard) and paclitaxel in transplantation tumors. I, UPLC-MS/MS assays determined the concentration of paclitaxel in plasma and transplantation tumors at day 20, day 26, and day 32. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 4.

The effect of Ht1 and Ht2 on tumor growth and chemosensitivity to paclitaxel in xenograft mice. A, Diagram showing the time of tumor formation and administration in vivo. B, The body weight of mice in each group (n = 5). C, Growth curves of tumor volumes in each group (n = 5). D, Tumor weight of xenograft mice at sacrifice (n = 5). E, Photographs of tumors of each group removed at day 35 after implantation. F, H&E staining showed the transplantation tumors after transfection with pcmTOR-Ht1 and pcmTOR-Ht2; IHC method showed p-mTOR and mTOR expression in the xenograft mice after transfection of pcmTOR-Ht1 and pcmTOR-Ht2. Comparison of the expression of CCND1 and ABCB1 between pcmTOR-Ht1 and pcmTOR-Ht2. G, Diagram showing the time of tumor formation and blood collection in vivo. H, UPLC-MS/MS detection of docetaxel (internal standard) and paclitaxel in transplantation tumors. I, UPLC-MS/MS assays determined the concentration of paclitaxel in plasma and transplantation tumors at day 20, day 26, and day 32. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

Next, to further investigate whether the concentration of paclitaxel in plasma and in transplantation tumors was different among the group of pcmTOR-Ht1 + paclitaxel, pcmTOR-Ht2 + paclitaxel, and NC + paclitaxel, similar xenograft mice models were constructed according to the above method (Fig. 4G). The transplanted tumors and the blood samples of 3 mice from each group were taken at day 20, day 26, and day 32, and the paclitaxel concentrations were determined by the UPLC-MS/MS method (Supplementary Fig. S2). Notably, the cellular paclitaxel concentration of the pcmTOR-Ht2 + paclitaxel group was significantly decreased in comparison to the pcmTOR-Ht1 + paclitaxel and NC + paclitaxel groups at day 20 (P = 0.0258 and P = 0.0095, respectively), day 26 (P = 0.0420 and P = 0.0434, respectively), and day 32 (P = 0.0070 and P = 0.0188, respectively). However, there was no significant difference observed in the paclitaxel concentration of the plasma samples (Fig. 4H and I). These above data indicate that, compared with Ht1, Ht2 exhibits a stronger effect on promoting expression of mTOR, which in turn results in a much more robust effect on promoting tumor growth and conferring resistance to paclitaxel treatment in vivo.

Transcription factor–binding analyses

We used both bioinformatic analyses and functional studies to examine DNA transcription factor interactions in these two prioritized haplotypes. The Jaspar (http://jaspar.binf.ku.dk/) and Gene Regulation (http://gene-regulation.com/pub/databases.html) prediction tools were utilized to predict the haplotypes altering transcription factor binding (Fig. 5A and B). Subsequently, analysis of RNA fold secondary structure using RNAfolder (http://rna.tbi.univie.ac.at) showed evidence of distinct secondary RNA structures between Ht1 and Ht2 (Fig. 5C). Specifically, motif analysis showed that the -78C allele falls in the middle of a consensus KLF5 motif (relative score of 86%), whereas the -78G allele is predicted to have no favorable binding. In contrast, ZEB1 can specifically bind to the -141T allele of high-risk Ht2 (relative score of 81%) but not the -141G allele of Ht1 (Fig. 5C). Next, analyses of a further RNA-seq dataset of breast tumor samples from TCGA showed that elevated mTOR expression is strongly positively correlated with ZEB1 expression (P < 0.0001) and negatively correlated with KLF5 expression (P = 0.0203; Fig. 5D).

Figure 5.

Transcription factors KLF5 and ZEB1 binding analyses. A and B, Jaspar bioinformatics predicting allele-specific binding sites of Ht1 with KLF5 (left) and Ht2 with ZEB1 (right). C, RNAfolder bioinformatics predicting the diversity of secondary structure with the variations of Ht1 and Ht2. The module of specific binding affinity between KLF5 and -78C within Ht1 and ZEB1 and -141T within Ht2. D, TCGA dataset showed that elevated mTOR mRNA expression was positively correlated with ZEB1 and negatively correlated with KLF5 using RNA-sequencing. E and F, Schematic graph representing the construction of pGL3-Ht1 and pGL3-Ht2 recombinant plasmids; luciferase reporter gene assay depicting the relative mTOR gene transcription activity after transfection of pGL3-Ht1 and pGL3-Ht2 to 293T and MCF7 cells. G and H, Protein–luciferase fusion reporter assays depicting the transcriptional activities from 0 to 24 hours in each group. I and J, qPCR and ChIP assays using antibodies against KLF5 or ZEB1 in MCF7 cells. Normal rabbit IgG was used as a nonspecific antibody control. Graphs present the results of two biological replicates. ****, P < 0.0001.

Figure 5.

Transcription factors KLF5 and ZEB1 binding analyses. A and B, Jaspar bioinformatics predicting allele-specific binding sites of Ht1 with KLF5 (left) and Ht2 with ZEB1 (right). C, RNAfolder bioinformatics predicting the diversity of secondary structure with the variations of Ht1 and Ht2. The module of specific binding affinity between KLF5 and -78C within Ht1 and ZEB1 and -141T within Ht2. D, TCGA dataset showed that elevated mTOR mRNA expression was positively correlated with ZEB1 and negatively correlated with KLF5 using RNA-sequencing. E and F, Schematic graph representing the construction of pGL3-Ht1 and pGL3-Ht2 recombinant plasmids; luciferase reporter gene assay depicting the relative mTOR gene transcription activity after transfection of pGL3-Ht1 and pGL3-Ht2 to 293T and MCF7 cells. G and H, Protein–luciferase fusion reporter assays depicting the transcriptional activities from 0 to 24 hours in each group. I and J, qPCR and ChIP assays using antibodies against KLF5 or ZEB1 in MCF7 cells. Normal rabbit IgG was used as a nonspecific antibody control. Graphs present the results of two biological replicates. ****, P < 0.0001.

Close modal

To further verify whether the motifs display competitive binding to any prioritized haplotype sites, luciferase activities were assessed in 293T cells and MCF7 cells transfected with promoter–reporter constructs pGL3-Ht1 and pGL3-Ht2 (Fig. 5E and F). Assays of luciferase activities showed that the transcriptional activities of Ht2 sequence were much higher than those of Ht1 sequence in both 293T and MCF7 cell lines (P < 0.0001).

Then, we cloned pNluc-KLF5 constructs (that express Nluc and KLF5 fusion proteins) and pNluc-ZEB1 constructs (that express Nluc and ZEB1 fusion proteins), which were cotransfected with pGL3-Ht1 or pGL3-Ht2 plasmids, respectively. We verified that transcriptional activities of pGL3-Ht1 can be continuously and time-dependently decreased by KLF5 (P < 0.0001), indicating that KLF5 acts as an orientation-independent transcriptional silencer (Fig. 5G). In contrast, pGL3-Ht2 with ZEB1 had relatively higher target gene promoter activity than without ZEB1 (P < 0.0001), suggesting that ZEB1 acts as an orientation-dependent enhancer (Fig. 5H). To further assess the allelic binding specificity of the mTOR locus, we performed the KLF5 and ZEB1 ChIP, followed by qPCR and sequencing. ChIP assays confirmed enrichment of KLF5 binding to Ht1 and ZEB1 binding to Ht2 (Fig. 5I and J).

Prioritizing candidate haplotypes for bioinformatics and functional assays

We considered two putative transcription factors that play a role in mTOR regulation. This finding was further confirmed at the cell viability, mRNA, and protein expression levels, as detected by CCK8, qRT-PCR, and Western blot, respectively. The cell viability of MCR7 cells transfected with pcmTOR-Ht1 and KLF5 was much lower than that of cells transfected with pcmTOR-Ht2 and KLF5. This finding is because KLF5 can inhibit the overexpression of pcmTOR-Ht1 but cannot inhibit pcmTOR-Ht2. Remarkably, the cell proliferation was strongly elevated in MCF7 cells transfected with pcmTOR-Ht2 combined with ZEB1, as compared with cells transfected with pcmTOR-Ht1 combined with ZEB1. This result is because ZEB1 could elevate the overexpression of Ht2 but not Ht1 (Fig. 6A). Moreover, similar results were observed in mRNA and protein expression data (Fig. 6B–D). Taken together, the competition with transcription factor–binding sites suggested that the haplotypes synergistically contribute to substantially promoting the expression of mTOR by facilitating the binding of ZEB1 (an enhancer) and inhibiting the binding of KLF5 (a silencer; Fig. 6E).

Figure 6.

Bioinformatics and functional assays of KLF5 and ZEB1. A, Cell proliferation and viability assay in MCF7 cells transfected with KLF5 or ZEB1 vectors with pcmTOR-Ht1 and pcmTOR-Ht2. B, qRT-PCR assays of mTOR mRNA expression. C and D, Western blot assays of p-mTOR and mTOR protein expression, accompanied with ABCB1 and CCND1. E, A graphical abstract of mTOR haplotypes regulated by KLF5 and ZEB1 in breast cells. KLF5 binds to the low-risk allele of Ht1 (-78C/-141G), whereas ZEB1 binds to the high-risk allele of Ht2 (-78G/-141T) to regulate gene transcription, expression, cell proliferation, and chemoresistance to paclitaxel. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Figure 6.

Bioinformatics and functional assays of KLF5 and ZEB1. A, Cell proliferation and viability assay in MCF7 cells transfected with KLF5 or ZEB1 vectors with pcmTOR-Ht1 and pcmTOR-Ht2. B, qRT-PCR assays of mTOR mRNA expression. C and D, Western blot assays of p-mTOR and mTOR protein expression, accompanied with ABCB1 and CCND1. E, A graphical abstract of mTOR haplotypes regulated by KLF5 and ZEB1 in breast cells. KLF5 binds to the low-risk allele of Ht1 (-78C/-141G), whereas ZEB1 binds to the high-risk allele of Ht2 (-78G/-141T) to regulate gene transcription, expression, cell proliferation, and chemoresistance to paclitaxel. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Close modal

Genetic variation of SNPs constitutes the genetic basis for cancer susceptibility and different responses to chemotherapy (34–36). However, most previous reports have focused on the study of single SNPs, ignoring the effect of potential haplotypes of multiple SNPs (37, 38). Through a large-scale case (n = 547)-control (n = 504) study, we showed that, compared with Ht1, Ht2 is associated with increased risk of susceptibility of breast cancer, as well as higher expression level of mTOR and poor prognosis, suggesting that Ht2 may be a key genetic factor that increases cancer risk and promotes progression. Notably, no such large differences could be discovered if the analyses were focused on the individual single SNPs (i.e., comparing -78G with -78C or comparing -141G with -141T). Thus, two linage variations of each haplotype allele may function synergistically or in combination to cause the overall effects or impacts of the given haplotype.

In addition, many previous studies have simply reported the data of correlation analysis of the SNPs and cancer phenotypes, with limited efforts to understand whether and how the SNPs affect target gene expression, cancer pathogenesis, and drug response (39, 40). Here, we show that, compared with Ht1, Ht2 weakens the binding of KLF5 and enhances the binding of ZEB1, thereby resulting in a much stronger effect on promoting the expression of mTOR. Thus, our studies provide an interesting model showing how two linked SNPs in the promoter region of a given gene coordinately fine-tune the expression of the target gene, highlighting the functional importance and necessity of forming haplotypes or LD.

mTOR protein, a serine/threonine protein kinase, can be phosphorylated by itself or by PKB, and thereby regulating the cell growth, survival and autophagy, and so forth (41). PKB exerts its function by phosphorylating Ser2448 of mTOR protein. However, amino acid starvation rapidly attenuated the reactivity of the Ser2448 phospho-specific antibody with mTOR (42). Our additional functional studies demonstrated that Ht2 exhibits a much stronger effect on promoting mTOR, CCND1, and ABCB1 expression, leading to enhanced tumor cell cycle, cell proliferation, and increased resistance to paclitaxel treatment, which is in agreement with the clinical data in patients with breast cancer.

The xenograft mice model of this study was performed by revised method as described in previous publication (43–45) and in this revised manuscript. The MCF7 cells were subcutaneously injected to nude mice with a Matrigel (1:1) mixture instead of estrogen. Previous studies supported that tumor cell lines that are difficult to form tumor in nude mice can easily grow if implanted with Matrigel, and the tumor formation rate is mainly dependent on the concentration of Matrigel and cell numbers (46, 47). Thus, Matrigel could assist tumor formation and then save study process and animal costs for less feeding time. Recent studies showed that the number of MCF7 cells injected into the mice was from 2 × 105 to 1 × 107 (48, 49). In our study, 8 × 106 MCF7 cells were injected per mice with Matrigel (1:1). More importantly, we searched the relevant articles and found that estrogen can upregulate mRNA and protein expression of ABCB1 and CCND1 (50, 51), which is not suitable for our study.

In conclusion, we have identified novel functional haplotypes in the mTOR promoter region in patients with breast cancer. To date, only very few published SNP-related studies have shown data from animal models that assess the biological functions of the SNPs or haplotypes, and our study will undeniably add an interesting new example (52, 53). These novel mTOR haplotypes could serve as potential biomarkers for evaluating cancer susceptibility and drug response. Patients carrying Ht2 should be treated differently from those carrying Ht1, as the former are often resistant to standard NCT (e.g., paclitaxel treatment) due to increased mTOR expression. Therefore, our studies not only provide an excellent model showing how haplotypes coordinately modulate target gene expression and thereby affect cancer susceptibility and drug responses but also suggest that mTOR promoter haplotypes are important biomarkers for breast cancer risk stratifications and interindividualized treatment for patients.

No potential conflicts of interest were disclosed.

Conception and design: Q. Chen, H. Wu

Development of methodology: Q. Chen, M. He, Y. Wang, H. Zhao, O.I. Olapade, H. Wu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Guan, F. Jin, Y. Liu, H. Wu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Hu, J. Zhang

Writing, review, and/or revision of the manuscript: X. Deng, J. Chen, O.I. Olapade, H. Wu, M. Wei

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B. Wei, W. Yao, H. Wu

Study supervision: J. Chen, M. Wei

The authors gratefully appreciate the efforts and contributions of doctors, nurses, and technical staff at the First Hospital of China Medical University. The research was supported by National Natural Science Foundation of China (nos. 31828005, 81872905, 81402948, 81673475, 81501346, 81603149, 81601370), National Natural Science Foundation of China and Liaoning Joint Fund Key Program (nos. U1608281), Shenyang S&T Projects (17-123-9-00, Z18-4-020), the Key Laboratory Foundation from Shenyang S&T Projects (F16-094-1-00), Liaoning Province Scientific Research Foundation (nos. LK201646), and Key Laboratory Foundation from Liaoning Province (nos. LS201617).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Siegel
RL
,
Miller
KD
,
Jemal
A
. 
Cancer statistics, 2018
.
CA Cancer J Clin
2018
;
68
:
7
30
.
2.
Bray
F
,
Ferlay
J
,
Soerjomataram
I
,
Siegel
RL
,
Torre
LA
,
Jemal
A
. 
Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries
.
CA Cancer J Clin
2018
;
68
:
394
424
.
3.
Fleming
ND
,
Agadjanian
H
,
Nassanian
H
,
Miller
CW
,
Orsulic
S
,
Karlan
BY
, et al
Xeroderma pigmentosum complementation group C single-nucleotide polymorphisms in the nucleotide excision repair pathway correlate with prolonged progression-free survival in advanced ovarian cancer
.
Cancer
2012
;
118
:
689
97
.
4.
Takada
M
,
Nagai
S
,
Haruta
M
,
Sugino
RP
,
Tozuka
K
,
Takei
H
, et al
BRCA1 alterations with additional defects in DNA damage response genes may confer chemoresistance to BRCA-like breast cancers treated with neoadjuvant chemotherapy
.
Genes Chromosomes Cancer
2017
;
56
:
405
20
.
5.
Apostolou
P
,
Pertesi
M
,
Aleporou-Marinou
V
,
Dimitrakakis
C
,
Papadimitriou
C
,
Razis
E
, et al
Haplotype analysis reveals that the recurrent BRCA1 deletion of exons 23 and 24 is a Greek founder mutation
.
Clin Genet
2016
;
91
:
482
7
.
6.
Sarwar
R
,
Mahjabeen
I
,
Bashir
K
,
Saeed
S
,
Kayani
MA
. 
Haplotype based analysis of XRCC3 gene polymorphisms in thyroid cancer
.
Cell Physiol Biochem
2017
;
42
:
22
33
.
7.
Jiang
J
,
Fridley
BL
,
Feng
Q
,
Abo
RP
,
Brisbin
A
,
Batzler
A
, et al
Genome-wide association study for biomarker identification of rapamycin and everolimus using a lymphoblastoid cell line system
.
Front Genet
2013
;
4
:
1
12
.
8.
Feng
FB
,
Qiu
HY
. 
Effects of artesunate on chondrocyte proliferation, apoptosis and autophagy through the PI3K/AKT/mTOR signaling pathway in rat models with rheumatoid arthritis
.
Biomed Pharmacother
2018
;
102
:
1209
20
.
9.
Zhao
Z
,
Li
J
,
Tan
F
,
Gao
S
,
He
J
. 
mTOR up-regulation of BEX4 promotes lung adenocarcinoma cell proliferation by potentiating OCT4
.
Biochem Biophys Res Commun
2018
;
500
:
302
9
.
10.
Du
L
,
Li
X
,
Zhen
L
,
Chen
W
,
Mu
L
,
Zhang
Y
, et al
Everolimus inhibits breast cancer cell growth through PI3K/AKT/mTOR signaling pathway
.
Mol Med Rep
2018
;
17
:
7163
9
.
11.
Mu
Y
,
Zheng
D
,
Wang
C
,
Yu
W
,
Zhang
X
. 
RagD regulates amino acid mediated-casein synthesis and cell proliferation via mTOR signalling in cow mammary epithelial cells
.
J Dairy Res
2018
;
85
:
204
11
.
12.
Lv
X
,
Li
CY
,
Han
P
,
Xu
XY
. 
MicroRNA-520a-3p inhibits cell growth and metastasis of non-small cell lung cancer through PI3K/AKT/mTOR signaling pathway
.
Eur Rev Med Pharmacol Sci
2018
;
22
:
2321
7
.
13.
Nguyen
JT
,
Ray
C
,
Fox
AL
,
Mendonça
DB
,
Kim
JK
,
Krebsbach
PH
. 
Mammalian EAK-7 activates alternative mTOR signaling to regulate cell proliferation and migration
.
Sci Adv
2018
;
4
:
1
15
.
14.
Milošević
Z
,
Banković
J
,
Dinić
J
,
Tsimplouli
C
,
Sereti
E
,
Dragoj
M
, et al
Potential of the dual mTOR kinase inhibitor AZD2014 to overcome paclitaxel resistance in anaplastic thyroid carcinoma
.
Cell Oncol
2018
;
41
:
409
26
.
15.
Zhang
L
,
Wang
Z
,
Khishignyam
T
,
Chen
T
,
Zhou
C
,
Zhang
Z
, et al
In vitro anti-leukemia activity of dual PI3K/mTOR inhibitor Voxtalisib on HL60 and K562 cells, as well as their multidrug resistance counterparts HL60/ADR and K562/A02 cells
.
Biomed Pharmacother
2018
;
103
:
1069
78
.
16.
Wang
MY
,
Li
QX
,
He
J
,
Qiu
LX
,
Wang
YN
,
Li
J
, et al
Genetic variations in the mTOR gene contribute toward gastric adenocarcinoma susceptibility in an Eastern Chinese population
.
Pharmacogenet Genomics
2015
;
25
:
521
30
.
17.
Qi
L
,
Sun
K
,
Zhuang
Y
,
Yang
J
,
Chen
J
. 
Study on the association between PI3K/AKT/mTOR signaling pathway gene polymorphism and susceptibility to gastric cancer
.
J BUON
2017
;
22
:
1488
93
.
18.
Zhu
J
,
Wang
M
,
Zhu
M
,
He
J
,
Wang
JC
,
Jin
L
, et al
Associations of PI3KR1 and mTOR polymorphisms with esophageal squamous cell carcinoma risk and gene-environment interactions in Eastern Chinese populations
.
Sci Rep
2015
;
5
:
8250
.
19.
Bizhani
F
,
Hashemi
M
,
Danesh
H
,
Nouralizadeh
A
,
Narouie
B
,
Bahari
G
, et al
Association between single nucleotide polymorphisms in the PI3K/AKT/mTOR pathway and bladder cancer risk in a sample of Iranian population
.
EXCLI J
2018
;
17
:
3
13
.
20.
Xu
M
,
Gao
Y
,
Yu
T
,
Wang
J
,
Cheng
L
,
Cheng
L
, et al
Functional promoter rs2295080 T>G variant in MTOR gene is associated with risk of colorectal cancer in a Chinese population
.
Biomed Pharmacother
2015
;
70
:
28
32
.
21.
Cao
Q
,
Ju
X
,
Li
P
,
Meng
X
,
Shao
P
,
Cai
H
, et al
A functional variant in the MTOR promoter modulates its expression and is associated with renal cell cancer risk
.
PLoS One
2012
;
7
:
e50302
.
22.
Huang
L
,
Huang
J
,
Wu
P
,
Li
Q
,
Rong
L
,
Xue
Y
, et al
Association of genetic variations in mTOR with risk of childhood acute lymphoblastic leukemia in a Chinese population
.
Leuk Lymphoma
2012
;
53
:
947
51
.
23.
Liang
W
,
Song
S
,
Xu
Y
,
Li
H
,
Liu
H
. 
Knockdown of ZEB1 suppressed the formation of vasculogenic mimicry and epithelial-mesenchymal transition in the human breast cancer cell line MDA-MB-231
.
Mol Med Rep
2018
;
17
:
6711
6
.
24.
Maturi
V
,
Enroth
S
,
Heldin
CH
,
Moustakas
A
. 
Genome-wide binding of transcription factor ZEB1 in triple-negative breast cancer cells
.
J Cell Physiol
2018
;
233
:
7113
27
.
25.
Lazarova
D
,
Bordonaro
M
. 
ZEB1 mediates drug resistance and EMT in p300-deficient CRC
.
J Cancer
2017
;
8
:
1453
9
.
26.
Farrugia
MK
,
Sharma
SB
,
Lin
CC
,
McLaughlin
SL
,
Vanderbilt
DB
,
Ammer
AG
, et al
Regulation of anti-apoptotic signaling by Kruppel-like factors 4 and 5 mediates lapatinib resistance in breast cancer
.
Cell Death Dis
2015
;
6
:
e1699
.
27.
Liu
Y
,
Sánchez-Tilló
E
,
Lu
X
,
Clem
B
,
Telang
S
,
Jenson
AB
, et al
Rb1 family mutation is sufficient for sarcoma initiation
.
Nat Commun
2013
;
4
:
2650
.
28.
Farrugia
MK
,
Vanderbilt
DB
,
Salkeni
MA
,
Ruppert
JM
. 
Kruppel-like pluripotency factors as modulators of cancer cell therapeutic responses
.
Cancer Res
2016
;
76
:
1677
82
.
29.
Xing
C
,
Ci
X
,
Sun
X
,
Fu
X
,
Zhang
Z
,
Dong
EN
, et al
Klf5 deletion promotes Pten deletion-initiated luminal-type mouse prostate tumors through multiple oncogenic signaling pathways
.
Neoplasia
2014
;
16
:
883
99
.
30.
Ge
F
,
Chen
W
,
Yang
R
,
Zhou
Z
,
Chang
N
,
Chen
C
, et al
WWOX suppresses KLF5 expression and breast cancer cell growth
.
Chin J Cancer Res
2014
;
26
:
511
6
.
31.
Wu
H
,
Kang
H
,
Liu
Y
,
Tong
W
,
Liu
D
,
Yang
X
, et al
Roles of ABCB1 gene polymorphisms and haplotype in susceptibility to breast carcinoma risk and clinical outcomes
.
J Cancer Res Clin Oncol
2012
;
138
:
1449
62
.
32.
He
M
,
Fu
Y
,
Yan
Y
,
Xiao
Q
,
Wu
H
,
Yao
W
, et al
The hedgehog signalling pathway mediates drug response of MCF-7 mammosphere cells in breast cancer patients
.
Clin Sci
2015
;
129
:
809
22
.
33.
Wang
Y
,
Zhao
L
,
Xiao
Q
,
Jiang
L
,
He
M
,
Bai
X
, et al
miR-302a/b/c/d cooperatively inhibit BCRP expression to increase drug sensitivity in breast cancer cells
.
Gynecol Oncol
2016
;
141
:
592
601
.
34.
Xu
JZ
,
Wen
F
,
Wang
XR
. 
The eIF3a Arg803Lys genetic polymorphism is associated with susceptibility to and chemoradiotherapy efficacy in cervical carcinoma
.
Kaohsiung J Med Sci
2017
;
33
:
187
94
.
35.
Tecza
K
,
Pamula-Pilat
J
,
Lanuszewska
J
,
Grzybowska
E
. 
Genetic polymorphisms and response to 5-fluorouracil, doxorubicin and cyclophosphamide chemotherapy in breast cancer patients
.
Oncotarget
2016
;
7
:
66790
808
.
36.
Zhao
XQ
,
Cao
WJ
,
Yang
HP
,
Yang
XW
,
Tang
P
,
Sun
L
, et al
DPYD gene polymorphisms are associated with risk and chemotherapy prognosis in pediatric patients with acute lymphoblastic leukemia
.
Tumour Biol
2016
;
37
:
10393
402
.
37.
Zhang
Z
,
Chen
Q
,
Zhang
J
,
Wang
Y
,
Hu
X
,
Yin
S
, et al
Associations of genetic polymorphisms in pTEN/AKT/mTOR signaling pathway genes with cancer risk: a meta-analysis in Asian population
.
Sci Rep
2017
;
7
:
17844
.
38.
Zining
J
,
Lu
X
,
Caiyun
H
,
Yuan
Y
. 
Genetic polymorphisms of mTOR and cancer risk: a systematic review and updated meta-analysis
.
Oncotarget
2016
;
7
:
57464
80
.
39.
Chen
M
,
Cassidy
A
,
Gu
J
,
Delclos
GL
,
Zhen
F
,
Yang
H
, et al
Genetic variations in PI3K-AKT-mTOR pathway and bladder cancer risk
.
Carcinogenesis
2009
;
30
:
2047
52
.
40.
Zhu
ML
,
Yu
H
,
Shi
TY
,
He
J
,
Wang
MY
,
Li
QX
, et al
Polymorphisms in mTORC1 genes modulate risk of esophageal squamous cell carcinoma in eastern Chinese populations
.
J Thorac Oncol
2013
;
8
:
788
95
.
41.
Vazquez-Martin
A
,
Oliveras-Ferraros
C
,
Bernadó
L
,
López-Bonet
E
,
Menendez
JA
. 
The serine 2481-autophosphorylated form of mammalian target of rapamycin (mTOR) is localized to midzone and midbody in dividing cancer cells
.
Biochem Biophys Res Commun
2009
;
380
:
638
43
.
42.
Navé
BT
,
Ouwens
M
,
Withers
DJ
,
Alessi
DR
,
Shepherd
PR
. 
Mammalian target of rapamycin is a direct target for protein kinase B: identification of a convergence point for opposing effects of insulin and amino-acid deficiency on protein translation
.
Biochem J
1999
;
344
:
427
31
.
43.
Li
J
,
Gao
Z
,
Zhao
D
,
Zhang
L
,
Qiao
X
,
Zhao
Y
, et al
PI-273, a substrate-competitive, specific small-molecule inhibitor of PI4KIIα, inhibits the growth of breast cancer cells
.
Cancer Res
2017
;
77
:
6253
66
.
44.
Jiang
YJ
,
Lee
CL
,
Wang
Q
,
Zhou
ZW
,
Yang
F
,
Jin
C
, et al
Establishment of an orthotopic pancreatic cancer mouse model: cells suspended and injected in Matrigel
.
World J Gastroenterol
2014
;
20
:
9476
85
.
45.
Clark
PA
,
Bhattacharya
S
,
Elmayan
A
,
Darjatmoko
SR
,
Thuro
BA
,
Yan
MB
, et al
Resveratrol targeting of AKT and p53 in glioblastoma and glioblastoma stem-like cells to suppress growth and infiltration
.
J Neurosurg
2017
;
126
:
1448
60
.
46.
Benton
G
,
Arnaoutova
I
,
George
J
,
Kleinman
HK
,
Koblinski
J
. 
Matrigel: from discovery and ECM mimicry to assays and models for cancer research
.
Adv Drug Deliv Rev
2014
;
15
:
3
18
.
47.
Benton
G
,
Kleinman
HK
,
George
J
,
Arnaoutova
I
. 
Multiple uses of basement membrane-like matrix (BME/Matrigel) in vitro and in vivo with cancer cells
.
Int J Cancer
2011
;
128
:
1751
7
.
48.
Ren
P
,
Ren
X
,
Cheng
L
,
Xu
L
. 
Frankincense, pine needle and geranium essential oils suppress tumor progression through the regulation of the AMPK/mTOR pathway in breast cancer
.
Oncol Rep
2018
;
39
:
129
37
.
49.
Ning
Y
,
Wang
C
,
Liu
X
,
Du
Y
,
Liu
S
,
Liu
K
, et al
CK2-mediated CCDC106 phosphorylation is required for p53 degradation in cancer progression
.
J Exp Clin Cancer Res
2019
;
38
:
131
.
50.
Chen
Y
,
Zhang
K
,
Li
Y
,
Guo
R
,
Zhang
K
,
Zhong
G
, et al
Oestrogen-related receptor alpha mediates chemotherapy resistance of osteosarcoma cells via regulation of ABCB1
.
J Cell Mol Med
2019
;
23
:
2115
24
.
51.
Mulligan
WA
,
Wegner
KA
,
Keil
KP
,
Mehta
V
,
Taketo
MM
,
Vezina
CM
. 
Beta-catenin and estrogen signaling collaborate to drive cyclin D1 expression in developing mouse prostate
.
Differentiation
2017
;
93
:
66
71
.
52.
Qiu
F
,
Yang
L
,
Ling
X
,
Yang
R
,
Yang
X
,
Zhang
L
, et al
Sequence variation in mature microRNA-499 confers unfavorable prognosis of lung cancer patients treated with platinum-based chemotherapy
.
Clin Cancer Res
2015
;
21
:
1602
13
.
53.
You
M
,
Wang
D
,
Liu
P
,
Vikis
H
,
James
M
,
Lu
Y
, et al
Fine mapping of chromosome 6q23-25 region in familial lung cancer families reveals RGS17 as a likely candidate gene
.
Clin Cancer Res
2009
;
15
:
2666
74
.