With Oncomine™ Research Premium Edition, you can:
Design better experiments...Answer more questions with fewer experiments, select the most promising gene or cell line, and test your hypothesis.
Gain more biological insights...Discover novel targets for therapeutic development, interrogate gene expression profiles, and identify drug and biological interactions.
Prepare to publish faster...Validate your results faster, visualize data easier and make connections to clinical significance.
The Oncomine™ Platform—from web applications to translational bioinformatics services—provides solutions for individual researchers and multinational companies, with robust, peer-reviewed analysis methods and a powerful set of analysis functions that compute gene expression signatures, clusters and gene-set modules, automatically extracting biological insights from the data. It has become an industry-standard tool cited in more than 1,100 peer-reviewed journal articles. The Oncomine Platform has been used as a foundation for ground-breaking discoveries with unique features that include:
- Scalability — with 700+ independent datasets
- High quality — with expertly curated data
- Consistency — with a rich, extensive and controlled ontology of terms
- Standardized analysis — with conventions that assure clear and consistent interpretations of results
Oncomine Research Edition remains free to the academic and nonprofit cancer research communities.
Oncomine Research Premium Edition is a subscription version that includes compelling features such as multi-gene search, expanded analyses, custom concept upload, export capability and enhanced support. Click here for more information.
Breast cancer (BC) is the most frequently diagnosed cancer among women and it is the leading cause of female cancer-related death . In 2017, 255,180 new BC cases and 41,070 BC deaths are estimated to occur in United States . Although the mechanism of BC is poorly understood, it has been widely accepted that environmental factors, genetic factors, and their interactions could contribute to the risk of BC . It has been demonstrated that breast cancer associated gene 1 (BRCA1) gene mutations were associated with the high incidence of BC .
BC cells could escape immune surveillance by suppressing the T-cell response . T-cell activation is dependent on the balance of costimulation and coinhibition . Cluster of differentiation 28 (CD28), constitutively expressed on the majority of T cells, is the primary T-cell costimulatory molecule and enhances T-cell activation and proliferation . In addition, B7 family members were identified as the coinhibitory molecular, such as cytotoxic T-lymphocyte antigen 4 (CTLA-4) and programmed cell death 1 ligand 1 (PD-L1) [8,9]. Isitmangil et al.  found that the levels of CD28 were significantly higher in BC patients than in the control group. Coinhibitory molecular CTLA-4 blockade has been demonstrated to enhance antitumor immunity . CD28 is encoded at the same chromosomal loci as CTLA4 and thus has 31% amino acid homology, and interacts closely with CTLA4 . Based on these observations, CD28 was associated with the development of BC by affecting the T-cell function.
Recently, two studies investigated the association between CD28 gene polymorphism and BC risk [10,13]. Chen et al.  found that CD28 rs3116496 polymorphism increased the risk of BC in a Chinese population. Isitmangil et al.  conducted another case–control study in a Turkish population and failed to detect significant association between rs3116496 polymorphism and BC risk. Different genetic backgrounds may contribute to the divergence. Therefore, we conducted a hospital-based case–control study with 312 BC patients and 312 controls to validate the conclusion in a Chinese Han population of Zhejiang province.
Patients and methods
We obtained approval of the study protocol from the Ethics Committee of Wenzhou Central Hospital (Zhejiang, China). All subjects provided written informed consent before participating in the study.
BC patients (312) were recruited from the Wenzhou Central Hospital (Zhejiang, China) between September, 2013 and January, 2016. Individuals with previous cancer history and metastasized cancer from other organs were not included in the present study. Cancer-free controls, frequency-matched to the cases on age were randomly selected from the same institutions during the same time period. Each patient was interviewed by trained personnel using a pretested questionnaire to obtain information on demographic data and related risk factors for BC. After the interview, 5 ml of peripheral blood was collected from each subject. The estrogen receptor (ER)/progesterone receptor (PR) information was obtained from the medical records of the hospitals.
Oncomine (www.oncomine.org), a cancer microarray database and web-based data mining platform, aims to stimulate new discovery from genome-wide expression analyses and compare the transcriptome data in most major types of cancer with respective normal tissues [14,15]. The filters were as follows: Gene: CD28; Analysis Type: Cancer vs. Normal Analysis; Cancer: Breast Cancer; Sample Type: Clinical Specimen. We got CD28 expression data in Breast from TCGA breast statistics.
The Kaplan–Meier plotter
The prognostic significance of the mRNA expression of CD28 gene in BC was evaluated using the Kaplan–Meier plotter (www.kmplot.com) . With the purpose to assess prognostic value of a specific gene, the patient samples were divided into two cohorts according to the median expression of the gene (high vs. low expression). We analyzed the overall survival (OS) of BC patients by using a Kaplan–Meier survival plot. Log rank P-value and hazard ratio (HR) with 95% confidence intervals were calculated and displayed on the figure.
Isolation of DNA and genotyping
Blood samples were collected using vacutainers and transferred to test tubes containing ethylenediaminetetraacetic acid (EDTA). Genomic DNA was isolated from whole blood using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany). Genotyping was done by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) using the Mass ARRAY system (Sequenom, San Diego, CA, U.S.A.).
The demographic and clinical characteristics of study participants were evaluated by using the chi-squared test. Odds ratios (ORs) and 95% confidence intervals (CIs) were used to estimate the association between CD28 gene rs3116496 polymorphism and risk of BC by logistic regression analyses. Hardy–Weinberg equilibrium (HWE) for CD28 genotype distributions in controls was tested by a goodness-of-fit chi-squared test. All statistical analyses were performed using the SAS software package (var. 9.1.3; SAS Institute, Cary, NC, U.S.A.). P<0.05 was considered to indicate a significant difference.
Oncomine analysis and prognosis analysis
Using Oncomine analysis, we investigated mRNA levels of CD28 gene in BC. As is shown in Figure 1, 0 for no value, 1 for the control group, and 2–9 for different types of BC. Three types of BC with larger sample size could present more reliable data. They were invasive BC, invasive ductal breast cancer (IDC), and invasive lobular breast cancer (ILC). The number of down-regulation datasets was found more than the number of up-regulation datasets. Overall, CD28 gene was down-regulated in BC compared with normal breast tissue.
The Student’s t-test was conducted using the Oncomine software (www.oncomine.org). The boxes represent the 25th through 75th percentiles. The horizontal lines represent the medians.
Subsequently, we investigated the association between CD28 gene expression and BC prognosis using a Kaplan–Meier plotter. Survival curves were plotted for all BC patients (Figure 2). CD28 high expression was found to be associated with better OS for all BC patients (HR, 0.8; 95%CI, 0.64–0.99; P=0.041). We speculated that CD28 is a tumor suppressor gene, but more convincing studies are needed to validate the conclusions because these clinical data are from the database.
Kaplan–Meier plots of overall survival: comparison of patients with high vs. low expression of CD28 in BC patients. The Kaplan–Meier plots were generated by Kaplan–Meier Plotter (http://www.kmplot.com).
Characteristics of the study population
The demographic and clinical characteristics of all subjects are listed in Table 1. A total of 312 BC patients and 312 healthy controls participated in the present study. Subjects were adequately matched for age and family history of cancer (P=0.882 and 0.650 respectively). The types of BC were mainly concentrated on invasive BC (7.7%), IDC (64.1%), and ILC (25.3%).
The genotypes of CD28 rs3116496 in two groups were summarized in Table 2. The genotype distribution in the controls conformed to HWE (P=0.083), suggesting these subjects could represent the total population. Carriers with genotype CC were at 2.15 times higher risk of BC than those with genotype TT (OR = 2.15, 95% CI = 1.02-4.51, P=0.044). Moreover, CD28 rs3116496 polymorphism achieved significant difference in the association with BC in the recessive model (P=0.040), but not in the dominant model (P=0.504). The C allele of rs3116496 polymorphism exhibited no significant association with the risk of BC.
In the present study, we found that CD28 gene expression is down-regulated in BC patients, whereas CD28 high expression is associated with better OS for BC by bioinformatics analysis. In addition, CC genotype of rs3116496 polymorphism showed positively correlation with the increased risk of BC.
Tumor-specific T-cell response was beneficial to limiting the development of cancer, which was influenced by costimulatory and coinhibitory signals . As one of the best characterized costimulatory molecules, CD28 competes with CLTA-4 (coinhibitory molecules) for B7 binding to enhance T-cell proliferation . Therefore, CD28 gene mutations may break the balance between costimulatory and coinhibitory molecules and changed the susceptibility of cancer.
Recently, the association between CD28 rs3116496 polymorphism and cancer risk has been widely investigated, such as cervical cancer , non-small-cell lung cancer , colorectal cancer , BC , and renal cell carcinoma . Only two studies evaluated the role of CD28 rs3116496 polymorphism in BC risk [10,13]. In 2012, Chen et al. conducted a case–control study with 565 BC patients and 605 controls to investigate the association between CD28 rs3116496 polymorphism and BC risk . The results indicated that C allele of rs3116496 polymorphism increased the risk of BC in the northern Chinese population . Isitmangil et al.  performed a case–control study in a Turkish population (79 cases and 76 controls) and failed to detect significant association between rs3116496 polymorphism and BC risk. In our study, CC genotype of rs3116496 polymorphism showed positive correlation with the increased risk of BC in the eastern Chinese population (312 cases and 312 controls). Obviously, the findings of studies from China (Asians) and Turkey (Caucasians) were different. We hypothesized that the following possible reasons may explain the different results of rs3116496 polymorphism between Asians and Caucasians. First, genetic heterogeneity for BC may exist in different populations. Second, the discrepancy may be explained by clinical heterogeneity between the different populations. Third, the sample size of the Caucasian population (79 cases and 76 controls) might not have been sufficiently large to reach a convincing conclusion when compared with Asian populations. Additionally, the different genotyping methods and random errors may also explain the different findings between Asians and Caucasians. It is noteworthy that we found CC genotype increased the risk of BC, while the study conducted by Chen et al. indicated C allele was associated with the risk of BC .
We attempted to make explanations for our positive results. First, the functional significance of rs3116496 polymorphism is unclear because it is located in the intron region. The significant association between CD28 rs3116496 polymorphism and BC risk may reflect linkage disequilibrium with another potential functional variant or closely linked susceptibility gene. Second, the expression of CD28 was down-regulated in BC patients, showing worse OS for BC. Therefore, we guessed that CD28 gene mutants may alter the expression of CD28, thus contributing the risk of BC by reducing T-cell response.
Several limitations in the present study need to be addressed. First, expression data and survival data of CD28 are from the database, and we did not verify its validity by the present study. Second, the sample size of the present study is not large, which may get false-positive results. Third, other exposure information was missing, such as smoking and diet. Fourth, selection bias could not be avoided because the present study is a hospital-based case–control study. Fifth, we cannot add the experimental results in vivo and in vitro in the present study. Finally, we cannot make sure whether the location of rs3116496 polymorphism would affect the process of gene splicing.
In conclusion, the present study provides evidence that CD28 gene was down-regulated in the BC patients, which was harmful to the survival of BC patients. In addition, we find that CD28 rs3116496 polymorphism increases the risk of BC. However, our results were obtained from a moderate-sized sample, and larger, well-designed studies with ethnically diverse populations are warranted to further validate these findings.
Abbreviations: BC, breast cancer; BRCA1, CD28; CD28, breast cancer associated gene 1; CTLA-4, cytotoxic T-lymphocyte antigen 4; ER, estrogen receptor; HR, hazard ratio; HWE, Hardy–Weinberg equilibrium; IDC, invasive ductal breast cancer; ILC, invasive lobular breast cancer; MALDI-TOF MS, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; OS, overall survival; PD-L1, programmed cell death 1 ligand 1; PR, progesterone receptor; SNP, single nucleotide polymorphism