Abstract
LILRB2 is an inhibitory receptor involved in immune cells. A variety of cancer cells have been observed to express LILRB2, which has been related to development of cancers. Recently, ANGPTL2 was found to be bound to LILRB2 as a high affinity ligand. Expression and function of LILRB2 and ANGPTL2 in colorectal cancer (CRC) remain unknown. To explore differential expression, 364 CRCs, 5 adenomas, and 205 normal samples for LILRB2 and 338 CRCs, 5 adenomas, and 232 normal samples for ANGPTL2 were studied in Oncomine and GEO databases. We noted that LILRB2 was significantly increased in CRC compared to adenoma and normal tissues. ANGPTL2 was higher in adenoma than normal tissues and further increased in CRC than adenoma. Copy number of LILRB2 and ANGPTL2 DNA was also more increased in CRC than in normal tissue. Furthermore, immunohistochemistry analysis of 155 pairs of primary CRC and normal tissues verified the positive rates of LILRB2 and ANGPTL2 were 87.10% (135/155) and 97.44% (151/155) in CRC, with almost no expression in normal tissues. LILRB2 and ANGPTL2 were significantly associated with tumor size, worse cell differentiation, lymph node metastasis, and advanced disease stage. Levels of ANGPTL2 were adversely related to survival of CRC patients, consistent with results in GEPIA (TCGA data) database. Moreover, a significant positive correlation was found between LILRB2 and ANGPTL2 in CRC. These findings suggest that ANGPTL2 and LILRB2 play an important role in CRC occurrence and progression. ANGPTL2 and LILRB2 could serve as novel biomarkers for treatment and prognosis of CRC.
Keywords: ANGPTL2, LILRB2, colorectal cancer
Introduction
Colorectal cancer (CRC) is the third most frequent cancer and fourth leading cause of malignant tumor related deaths. According to data from International Agency for Research on Cancer (IARC) in 2012, there are about 1,360,000 new cases and 690,000 deaths annually, worldwide [1]. Fortunately, targeted therapy has significantly prolonged survival and enhanced the quality of life of CRC patients. However, drug resistance to targeted drugs remains a challenge for treatment of CRC. Moreover, targeted drugs can cause adverse reactions to some normal tissues which express the target. Therefore, more specific targets or novel biomarkers need to be identified to improve the curative effect and prognosis of CRC.
Leukocyte immunoglobulin-like receptor subfamily B member 2 (LILRB2) is a member of the leukocyte immunoglobulin-like receptor (LILR) family and is also known as ILT-4 or CD85d. A cell-surface receptor, LILRB2 can negatively regulate the immune response by recruiting tyrosine phosphatases, like SHP1 or SHP2. LILRB2 is predominantly expressed in monocytes, macrophages, and myeloid cells and also be expressed by endothelial cells, placental trophoblasts, and decidual macrophages [2-5], suggesting that LILRB2 may play an important role in various physiological functions. Interestingly, LILRB2 and its mouse homolog, paired immunoglobulin-like receptor (PIRB), were found to be overexpressed in leukemic stem cells (LSCs) and hematopoietic stem cells (HSCs), critical to the maintenance of stemness [2]. Furthermore, LILRB2, as well as LILRB1 and LILRB4, is expressed in many solid tumors [6-10]. High LILRB2 expression was found in non-small cell lung cancer (NSCLC) and breast cancer and has been associated with lymph node metastasis and less number of tumor infiltrating lymphocytes [8,9]. Higher LILRB2 has been found in more aggressive pancreatic ductal carcinomas (PDAC) [10]. These results suggest that LILRB2 could be involved in progression of malignant tumors by non-immunoregulation and immunoregulation. However, expression and function of LILRB2 in CRC remains poorly understood.
Angiopoietin-like proteins (ANGPTLs) are a family constituting seven secreted glycoproteins with the same domain structure as angiopoietin [11]. ANGPTLs are extensive expressed in many tissues such as hematopoietic system and vascular system and liver and play an important role in inflammation, angiogenesis, and lipid metabolism [11,12]. Several ANGPTLs, including ANGPTL2, can stimulate activities of HSCs in humans and mice [13-15]. ANGPTL2 has been found to induce inflammatory carcinogenesis in skin squamous cell carcinoma (SCC) [16]. Significantly increased ANGPTL2 expression also has been observed in various solid tumors, regulating proliferation and metastasis of cells such as NSCLC, ovarian cancer, and sarcoma [17-19]. In addition, overexpressed ANGPTL2 can induce epithelial-to-mesenchymal transition (EMT) in PDAC [20]. These reports suggest that ANGPTL2 may be a critical factor in carcinogenesis and cancer progression. However, the role of ANGPTL2 in CRC remains largely undefined.
Recently, LILRB2 has been identified as a receptor for ANGPTL2 [2]. Co-expression of LILRB2 and ANGPTL2 has been reported in NSCLC to promote the proliferation of cancer cells [21]. To our knowledge, expression and function of LILRB2 and ANGPTL2 in CRC has not been reported. Compared with other ANGPTLs, ANGPTL2 presents the highest affinity for LILRB2 [2]. Thus, we speculated that there may be interaction between LILRB2 and ANGPTL2 and that this interaction may be involved in molecular mechanisms of CRC progression. Our present work was designed to evaluate the expression and role of both LILRB2 and ANGPTL2 in CRC.
Materials and methods
Bioinformatics analysis
We made a differential expression analysis for mRNA expression and DNA copy number variation (CNV) of LILRB2 and ANGPTL2, in the Oncomine Cancer Microarray Database (https://www.oncomine.org), to assess the expression pattern of LILRB2 and ANGPTL2 in CRC. Meta-analysis of LILRB2 and ANGPTL2 expression was performed in datasets. Coexpression gene analysis of LILRB2 and ANGPTL2 was performed and presented in heatmap. Datasets in the database were used in which CRC type, sample sizes, folds change, t-test values, and p values were obtained, showing a significant difference between the two groups. The filter criteria were P<0.05, fold change: all, gene rank: all. Original data of the microarray was downloaded from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). Survival analysis of LILRB2 and ANGPTL2 using TCGA data was performed in Gene Expression Profiling Interactive Analysis (GEPIA) Database (http://gepia.cancer-pku.cn/index.html). Oncomine, GEO, and GEPIA are all free public databases.
Patients and samples
CRC samples from 155 patients were obtained from CRC patients at Tongji Hospital affiliated to Tongji University from 2008 to 2010. None of the patients had received chemotherapy or radiotherapy before surgery and no perioperative mortality was observed. There were 95 males (59.375%) and 65 females (40.625%) with median age of 73 years (range: 33-98 years). Patients were classified according to criteria described in the seventh edition of AJCC/UICC. Main clinicopathological variables of the patients are summarized (Table 1). Samples were contacted by phone to confirm their health status. Median follow up duration was 68 months (range = 1-95 months). This study was approved by the Review Board and Ethics Committee of Tongji Hospital affiliated to Tongji University and written informed consent was obtained from each patient.
Table 1.
Correlation between LILRB2/ANGPTL2 expression and clinicopathological parameters
| Characteristics | N | LILRB2 expression | P value | N | ANGPTL2 expression | P value | ||
|---|---|---|---|---|---|---|---|---|
|
|
|
|||||||
| Low | High | Low | High | |||||
| All case (155) | ||||||||
| Age (years) | ||||||||
| ≤73 | 79 | 38 | 41 | 0.453 | 78 | 18 | 60 | 0.018a |
| >73 | 76 | 32 | 44 | 77 | 7 | 70 | ||
| Gender | ||||||||
| Male | 92 | 40 | 52 | 0.611 | 91 | 20 | 71 | 0.018a |
| Female | 63 | 30 | 33 | 64 | 5 | 59 | ||
| Tumor size (cm) | ||||||||
| ≤4.95 | 77 | 26 | 51 | 0.005b | 81 | 14 | 67 | 0.683 |
| >4.95 | 78 | 44 | 34 | 74 | 11 | 63 | ||
| Differentiation | ||||||||
| Well | 23 | 16 | 7 | 0.011a | 22 | 7 | 15 | 0.031a |
| Moderate/Poor | 132 | 54 | 78 | 133 | 18 | 115 | ||
| Dukes stage | ||||||||
| A/B | 75 | 40 | 35 | 0.048a | 70 | 16 | 54 | 0.039a |
| C/D | 80 | 30 | 50 | 85 | 9 | 76 | ||
| T classification | ||||||||
| T1 | 0 | 0 | 0 | 0.029a | 0 | 0 | 0 | 0.558 |
| T2 | 9 | 6 | 3 | 10 | 2 | 8 | ||
| T3 | 53 | 30 | 23 | 56 | 11 | 45 | ||
| T4 | 93 | 34 | 59 | 89 | 12 | 77 | ||
| N classification | ||||||||
| N0 | 74 | 40 | 34 | 0.033a | 76 | 16 | 60 | 0.102 |
| N1/N2 | 78 | 30 | 51 | 79 | 9 | 70 | ||
| M classification | ||||||||
| M0 | 137 | 61 | 76 | 0.661 | 137 | 20 | 117 | 0.173 |
| Ml | 18 | 9 | 9 | 18 | 5 | 13 | ||
| Tumor location | ||||||||
| Proximal | 59 | 27 | 32 | 0.906 | 59 | 9 | 50 | 0.816 |
| Distal | 96 | 43 | 53 | 96 | 16 | 80 | ||
| Tumor site | ||||||||
| Colon | 101 | 44 | 57 | 0.585 | 101 | 14 | 87 | 0.294 |
| Rectum | 54 | 26 | 28 | 54 | 11 | 43 | ||
P<0.05;
P<0.01.
Tissue microarray and immunohistochemistry
CRC and adjacent normal tissues were constructed into tissue microarray (TAMs) blocks and cut into 4 μm serial sections. After being deparaffinized and rehydrated, the sections were heated in an autoclave for antigen retrieval. Endogenous peroxidases were blocked with 0.3% H2O2 solution and nonspecific background staining was blocked with serum. Immunohistochemical staining for LILRB2 and ANGPTL2 was performed with primary antibodies: anti-LILRB2 (1:150, LSBio) and anti-ANGPTL2 (1:200, Proteintech). Then, sections were incubated with a HRP-conjugated anti-mouse/rabbit secondary antibody for 45 minutes and visualized with 3,3’-diaminobenzidine solution and counterstained with hematoxylin. Sections incubated with rabbit IgG substitute primary antibodies were used as negative controls. All specimens were evaluated by two independent pathologists unaware of the clinical records of patients. Scores were obtained as the sum of intensity score (0 = negative, 1 = weak, 2 = moderate, 3 = strong) multiplied by the proportion score representing the percentage of stained cells (1 = 10-24%, 2 = 25-49%, 3 = 50-74%, 4 = 75-100%). Brown staining of tumor cells >10% was considered as positive. High expression of LILRB2 or ANGPTL2 was defined as score ≥6. Otherwise, scores represented low expression.
Statistical analysis
SPSS software version 20.0 was used for statistical analyses. Data are represented as mean ± standard deviation (SD). Student’s t-test was used to compare data between two groups. Association between expression of LILRB2 or ANGPTL2 and clinicopathological variables was analyzed using χ2 test. Overall survival (OS) curves were constructed using Kaplan-Meier method and analyzed using log-rank test. Spearman’s correlation analysis was used to analyze correlation between LLILRB2 and ANGPTL2 expression. P value <0.05 was considered to indicate a statistically significant difference.
Results
Expression of LILRB2 and ANGPTL2 in CRC and other digestive system cancers
In Oncomine database, we found that LILRB2 and ANGPTL2 were upregulated in 152 and 134 significant unique analyses, respectively. In CRC, LILRB2 was upregulated in 12 datasets which complied with inclusion criteria, including 10 mRNA datasets and 2 DNA datasets. For ANGPTL2, 15 datasets consistent with inclusion criteria were upregulated, including 13 mRNA datasets and 2 DNA datasets. Moreover, LILRB2 and ANGPTL2 also were found to be upregulated in many datasets of digestive system tumors, such as esophageal cancer, gastric cancer, liver cancer, and pancreatic cancer (Table 2). Ten mRNA datasets from 5 microarrays, including 274 CRC and 111 normal samples, were analyzed for mRNA differential expression of LILRB2. The fold change of LILRB2 mRNA in CRC was 1.156 to 2.25, compared with normal tissues (Figure 1A; Table 3). Twelve mRNA datasets from 6 microarrays, including 296 CRC and 138 normal samples, were implicated in the differential expression analysis of ANGPTL2. The fold change of ANGPTL2 mRNA in CRC was 1.213 to 2.557 (Figure 1D; Table 3). Ten LILRB2 mRNA datasets and 12 ANGPTL2 mRNA datasets were analyzed by meta-analysis, with results showing that the median gene rank of LILRB2 and ANGPTL2 in overexpressed genes was 5132.5 and 4020.0, respectively (P = 0.004, P = 0.005). Furthermore, expression of LILRB2 mRNA and ANGPTL2 mRNA was increased in cecum adenocarcinoma, colon adenocarcinoma, rectal adenocarcinoma, rectosigmoid adenocarcinoma, colon mucinous adenocarcinoma, and rectal mucinous adenocarcinoma. Fold change of LILRB2 and ANGPTL2 mRNA in subtypes of CRC was 1.276 to 1.786 and 1.213 to 1.733, compared with normal colon tissues, respectively (Figures 2, 3).
Table 2.
Expression of LILRB2 and ANGPTL2 in various malignant tumors
| Type of cancer (Oncomine database) | LILRB2 (No. of analyses) Cancer vs. Normal | ANGPTL2 (No. of analyses) Cancer vs. Normal |
|---|---|---|
|
| ||
| Up-regulated | Up-regulated | |
| Bladder cancer | 4 | 0 |
| Brain and CNS cancer | 11 | 18 |
| Breast cancer | 25 | 4 |
| Cervical cancer | 5 | 1 |
| Colorectal cancer | 12 | 15 |
| Esophageal cancer | 3 | 8 |
| Gastric cancer | 11 | 12 |
| Head and neck cancer | 9 | 7 |
| Kidney cancer | 14 | 10 |
| Leukemia | 7 | 16 |
| Liver cancer | 4 | 3 |
| Lung cancer | 2 | 2 |
| Lymphoma | 22 | 18 |
| Melanoma | 2 | 2 |
| Myeloma | 1 | 1 |
| Other cancers | 11 | 7 |
| Ovarian cancer | 0 | 0 |
| Pancreatic cancer | 4 | 4 |
| Prostate cancer | 3 | 3 |
| Sarcoma | 4 | 5 |
| Significant unique analyses | 152 | 134 |
CNS: central nervous system.
Figure 1.
LILRB2 and ANGPTL2 expression in CRC and normal tissues at mRNA and DNA (copy number variation) level. A. LILRB2 mRNA (Hong Colorectal microarray). B, C. LILRB2 DNA (Kurashina Colon microarray). D. ANGPTL2 mRNA (Hong Colorectal microarray). E, F. ANGPTL2 DNA (Kurashina Colon microarray).
Table 3.
LILRB2 and ANGPTL2 overexpression datasets in CRC
| Gene | Microarray (mRNA) | Dataset | t-Test | Fold change | P value |
|---|---|---|---|---|---|
| LILRB2 | Kaiser Colon | Rectal Mucinous Adenocarcinoma (n = 4) vs. Normal (n = 5) | 4.279 | 1.426 | 0.002b |
| Rectosigmoid Adenocarcinoma (n = 10) vs. Normal (n = 5) | 3.109 | 1.311 | 0.006b | ||
| Colon Mucinous Adenocarcinoma (n = 13) vs. Normal (n = 5) | 3.106 | 1.729 | 0.004b | ||
| Cecum Adenocarcinoma (n = 17) vs. Normal (n = 5) | 2.620 | 1.310 | 0.010a | ||
| Rectal Adenocarcinoma (n = 8) vs. Normal (n = 5) | 2.262 | 1.786 | 0.027a | ||
| Colon Adenocarcinoma (n = 41) vs. Normal (n = 5) | 2.785 | 1.276 | 0.009b | ||
| Skrzypczak Colorectal | Colon Carcinoma Epithelia (n = 10) vs. Normal (n = 5) | 3.061 | 1.156 | 0.005b | |
| Hong Colorectal | Colorectal Carcinoma (n = 70) vs. Normal (n = 12) | 3.081 | 2.250 | 0.003b | |
| Gaedcke Colorectal | Rectal Adenocarcinoma (n = 65) vs. Normal (n = 65) | 3.604 | 1.376 | 2.30E-4c | |
| Skrzypczak Colorectal | Colorectal Carcinoma (n = 36) vs. Normal (n = 24) | 1.693 | 1.183 | 0.049a | |
| ANGPTL2 | Gaedcke Colorectal | Rectal Adenocarcinoma (n = 65) vs. Normal (n = 65) | 10.903 | 2.374 | 5.70E-20c |
| Skrzypczak Colorectal 2 | Colon Carcinoma (n = 5) vs. Normal (n = 10) | 11.289 | 2.557 | 3.58E-7c | |
| Colon Carcinoma (n = 5) vs. Normal (n = 10) | 8.360 | 2.252 | 4.11E-5c | ||
| Skrzypczak Colorectal | Colorectal Carcinoma (n = 36) vs. Normal (n = 24) | 4.906 | 2.371 | 6.54E-6c | |
| Hong Colorectal | Colorectal Carcinoma (n = 70) vs. Normal (n = 12) | 6.688 | 2.348 | 2.22E-7c | |
| Kaiser Colon | Colon Mucinous Adenocarcinoma (n = 13) vs. Normal (n = 5) | 3.963 | 1.733 | 0.001b | |
| Rectal Adenocarcinoma (n = 8) vs. Normal (n = 5) | 2.808 | 1.607 | 0.009b | ||
| Rectal Mucinous Adenocarcinoma (n = 4) vs. Normal (n = 5) | 2.546 | 1.284 | 0.019a | ||
| Colon Adenocarcinoma (n = 41) vs. Normal (n = 5) | 3.196 | 1.432 | 0.009b | ||
| Rectosigmoid Adenocarcinoma (n = 10) vs. Normal (n = 5) | 1.878 | 1.247 | 0.042a | ||
| Cecum Adenocarcinoma (n = 17) vs. Normal (n = 5) | 2.143 | 1.213 | 0.033a | ||
| TCGA Colorectal | Colon Mucinous Adenocarcinoma (n = 22) vs. Normal (n = 22) | 1.964 | 1.305 | 0.028a |
P<0.05;
P<0.01;
P<0.001.
Figure 2.
Differential expression analysis of LILRB2 in subtype of CRC. A. Colon (n = 5) vs. Cecum Adenocarcinoma (n = 17), fold change = 1.310, P = 0.010. B. Colon (n = 5) vs. Colon Adenocarcinoma (n = 41), fold change = 1.276, P = 0.009. C. Colon (n = 5) vs. Rectal Adenocarcinoma (n = 8), fold change = 1.786, P = 0.027. D. Colon (n = 5) vs. Rectosigmoid Adenocarcinoma (n = 10), fold change = 1.311, P = 0.006. E. Colon (n = 5) vs. Colon Mucinous Adenocarcinoma (n = 13), fold change = 1.729, P = 0.004. F. Colon (n = 5) vs. Rectal Mucinous Adenocarcinoma (n=4), fold change = 1.426, P = 0.002.
Figure 3.
Differential expression analysis of ANGPTL2 in subtype of CRC. A. Colon (n = 5) vs. Cecum Adenocarcinoma (n = 17), fold change = 1.213, P = 0.033. B. Colon (n = 5) vs. Colon Adenocarcinoma (n = 41), fold change = 1.432, P = 0.009. C. Colon (n = 5) vs. Rectal Adenocarcinoma (n = 8), fold change = 1.607, P = 0.009. D. Colon (n = 5) vs. Rectosigmoid Adenocarcinoma (n = 10), fold change = 1.247, P = 0.042. E. Colon (n = 5) vs. Colon Mucinous Adenocarcinoma (n = 13), fold change = 1.733, P = 0.001. F. Colon (n = 5) vs. Rectal Mucinous Adenocarcinoma (n = 4), fold change = 1.284, P = 0.019.
Expression of LILRB2 and ANGPTL2 in normal colons, colon adenomas, and colon cancer tissues
The original data of the Skrzypczak Colorectal 2 microarray were downloaded from GEO database (Table 4). It was noted that there was no significant difference in LILRB2 mRNA expression levels between colon adenoma and normal colon tissues (P = 0.2368). While compared with colon adenoma, LILRB2 mRNA was significantly increased in colon cancer and fold change was 2.218 (P = 0.011). Similarly, expression of ANGPTL2 mRNA showed a significant difference between normal colons and colon adenoma tissues, with fold change of 1.213 (P = 0.017). Compared with colon adenomas, ANGPTL2 mRNA was further increased in colon cancer, with fold change of 1.71 (P<0.001).
Table 4.
Fluorescence values of LILRB2 and ANGPTL2 probe in Skrzypczak Colorectal 2 microarray
| LILRB2 (Reporter: 207697_x_at) | ANGPTL2 (Reporter: 213004_at) | ||||
|---|---|---|---|---|---|
|
| |||||
| Tissue Subtype | GEO ID | ID_REF | Tissue Subtype | GEO ID | ID_REF |
| Colon (normal) | GSM523242 | 3.66491 | Colon Adenoma | GSM523244 | 3.7952 |
| Colon (normal) | GSM523243 | 3.60744 | Colon Adenoma | GSM523247 | 4.19941 |
| Colon (normal) | GSM523246 | 3.70634 | Colon Adenoma | GSM523251 | 4.1066 |
| Colon (normal) | GSM523268 | 3.60522 | Colon Adenoma | GSM523252 | 3.87153 |
| Colon (normal) | GSM523269 | 3.68977 | Colon Adenoma | GSM523281 | 3.83996 |
| Colon (normal) | GSM523270 | 3.689 | Colon Carcinoma | GSM523248 | 4.80783 |
| Colon (normal) | GSM523271 | 4.42752 | Colon Carcinoma | GSM523249 | 5.16897 |
| Colon (normal) | GSM523273 | 4.60052 | Colon Carcinoma | GSM523250 | 5.96317 |
| Colon (normal) | GSM523274 | 3.85642 | Colon Carcinoma | GSM523253 | 5.40926 |
| Colon (normal) | GSM523280 | 4.02841 | Colon Carcinoma | GSM523275 | 4.77843 |
| Colon Adenoma | GSM523244 | 3.59502 | |||
| Colon Adenoma | GSM523247 | 3.81785 | |||
| Colon Adenoma | GSM523251 | 3.79166 | |||
| Colon Adenoma | GSM523252 | 3.60936 | |||
| Colon Adenoma | GSM523281 | 3.59599 | |||
DNA copy number variations of LILRB2 and ANGPTL2
CNV of LILRB2 and ANGPTL2 was analyzed in Kurashina Colon microarray, including 188 samples. Compare to normal tissues, DNA copy number of LILRB2 and ANGPTL2 increased in CRC. Compared with normal tissues, the fold change of LILRB2 in CRC was 1.082, 1.056 (P = 2.07E-6, P = 0.005) (Figure 1B, 1C), and 1.11, 1.045 (P = 0.041, P = 0.001) (Figure 1E, 1F) for ANGPTL2 in datasets, respectively. Meta-analysis results of LILRB2 and ANGPTL2 copy number showed the median gene rank of LILRB2 and ANGPTL2 was 3187.0 and 2415.5 in overexpressed genes, respectively (P = 0.002, P = 0.021). These results indicate that DNA copy number of LILRB2 and ANGPTL2 was more increased in CRC than in normal tissues.
Co-expression analyses of LILRB2 and ANGPTL2 in CRC
Three expression profile microarrays were found containing co-expression data of LILRB2 and ANGPTL2. Results revealed that LILRB2 mRNA and ANGPTL2 mRNA had a similar expression patterns and the correlation coefficients in 3 microarrays were 0.368, 0.293, and 0.238, respectively (Table 5; Figure 4). In addition, correlation between LILRB2 and ANGPTL2, at mRNA level, was examined in GEPIA database using The Cancer Genome Atlas (TCGA) data. These results showed that there was a significant positive correlation between mRNA expression of LILRB2 and ANGPTL2 in colon adenocarcinoma (COAD) and rectal adenocarcinoma (READ). Correlation coefficient was 0.73 and 0.65, respectively (P<0.01, P<0.01) (Figure 5A, 5C). However, there was no correlation between LILRB2 and ANGPTL2 in normal colons and rectum tissues (P = 0.23, P = 0.62) (Figure 5B, 5D). In conclusion, LILRB2 and ANGPTL2 were significantly positive correlated in CRC tissues.
Table 5.
LILRB2 and ANGPTL2 co-expression microarrays
| Microarray | Cancer Type | LILRB2 Reporter | ANGPTL2 Reporter | Sample (n) | Correlation |
|---|---|---|---|---|---|
| Smith Colorectal 2 | Colorectal Adenocarcinoma | 210146_x_at | 219514_at | 55 | 0.368 |
| 213004_at | |||||
| 213001_at | |||||
| Jorissen Colorectal | Colorectal Adenocarcinoma | 210146_x_at | 211148_s_at | 74 | 0.293 |
| Ayers Colorectal | Colon Carcinoma | 207697_x_at | 219514_at | 59 | 0.238 |
Figure 4.
Co-expression analysis of LILRB2 and ANGPTL2 in CRC. A. Smith Colorectal 2 microarray. B. Jorissen Colorectal microarray. C. Ayers Colorectal microarray.
Figure 5.
Correlation analysis of LILRB2 and ANGPTL2 expression using TCGA data in GEPIA database. A. Correlation of LILRB2 and ANGPTL2 in colon adenocarcinoma. B. Correlation of LILRB2 and ANGPTL2 in normal colon. C. Correlation of LILRB2 and ANGPTL2 in rectum adenocarcinoma. D. Correlation of LILRB2 and ANGPTL2 in normal rectum. TPM: transcripts per million.
LILRB2 and ANGPTL2 protein expression in primary CRC tissues
Expression of ANGPTL2 and LILRB2 in 155 primary CRC tissues was examined by immunohistochemistry. LILRB2 was found in membranes, cytoplasm, or both, and ANGPTL2 was observed in the cytoplasm of CRC cells. In contrast, LILRB2 and ANGPTL2 were scarcely stained in adjacent normal colorectal tissue cells (Figure 6). Positive rates of LILRB2 and ANGPTL2 in CRC samples were 87.10% (135/155) and 97.44% (151/155), respectively. High expression of LILRB2 in CRC tissues was associated with tumor size (P = 0.005), worse cell differentiation (P = 0.011), advanced Dukes stage (P = 0.048), T stage (P = 0.029) and lymph node metastasis (P = 0.033) (Table 1). Similarly, high expression of ANGPTL2 in CRC tissues was associated with ages of more than 73 years (P = 0.018), female (P = 0.018), worse cell differentiation (P = 0.031), and advanced Dukes stage (P = 0.039) (Table 1). No significant correlation was observed between LILRB2 and age, gender, and tumor location. ANGPTL2 expression had no relation to tumor size, TNM stage, and tumor location (Table 1). Furthermore, expression of LILRB2 and ANGPTL2 showed significant correlation at the protein level (r = 0.505, P<0.001).
Figure 6.
Expression of LILRB2 and ANGPTL2 in primary CRC and normal tissues. A, E. normal control tissues, - (×200). B, F. well differentiation CRC tissues, + (×200). C, G. middle differentiation CRC tissues, ++ (×200). D, H. poor differentiation CRC tissues, +++ (×200).
Correlation between LILRB2, ANGPTL2 protein expression, and OS of CRC patients
An important question is whether expression of LILRB2 and ANGPTL2 could be used to predict prognosis of CRC patients. In survival analysis in GEPIA database by TCGA data, significant differences were observed in the OS and disease-free survival (DFS) at mRNA level between ANGPTL2 high expression group and low expression group of CRC patients (P = 0.0042, P = 0.0013) (Figure 7A, 7B). No difference of OS and DFS was found in LILRB2 (P = 0.37, P = 0.57) (Figure 7D, 7E). In addition, log-rank test analysis showed that OS of high expression of ANGPTL2 protein group was lower than the low expression group (P = 0.0303) (Figure 7C). Levels of ANGPTL2 negatively correlated with OS of CRC patients. However, no significant difference in OS between LILRB2 high and low protein expression groups was found (P = 0.44) (Figure 7F).
Figure 7.
Relationship between LILRB2 or ANGPTL2 expression and patient survival. A, D. OS/mRNA level, low (n) = 181, high (n) = 181. B, E. DFS/mRNA level, low (n) = 181, high (n) = 181. C, F. OS/protein level, low ANGPTL2 (n) = 134, high ANGPTL2 (n) = 21, low LILRB2 (n) = 88, high LILRB2 (n) = 67.
Discussion
Past studies have shown that expression of LILRB2 is increased in multiple tumor tissues including acute myeloid leukemia cells, breast cancer, and NSCLC [8,9,22]. Overexpression of LILRB2 can promote occurrence and development of tumors and is closely related to prognosis [21,23,24]. We used public microarray databases to evaluate expression of LILRB2 and ANGPTL2 in CRC. Our results showed that both LILRB2 and ANGPTL2 were overexpressed, not only in CRC but also in other digestive system cancers. Expression of LILRB2 and ANGPTL2 was significantly upregulated at mRNA and DNA levels in CRC, indicating that increased mRNA of LILRB2 and ANGPTL2 may be caused by change of DNA copy number and increased LILRB2 and ANGPTL2 may play a role. In order to reduce error and expand the sample size, we performed meta-analysis of LILRB2 and ANGPTL2 at mRNA and DNA levels. These results, likewise, supported the increase of LILRB2 and ANGPTL2 in CRC. In addition, we found that mRNA expression of LILRB2 and ANGPTL2, in each subtype of CRC, was higher than normal tissues. This further suggests that LILRB2 and ANGPTL2 may play a role as oncogene in development of CRC.
Occurrence of CRC is a multistep biological process. About 85% of CRCs originate from adenomatous polyps. It is essential to understand the roles of LILRB2 and ANGPTL2 in the evolution of CRC. Therefore, we examined expression of LILRB2 and ANGPTL2 in normal tissues, adenomas, and CRC tissues. Our results showed that expression of LILRB2 in colon adenomas was not significantly different from that in normal tissues but expression of ANGPTL2 in colon adenomas was higher than in normal tissues. It can be speculated that ANGPTL2 may play a role in the transition from normal tissue to adenoma by binding to LILRB2. In addition, expression of LILRB2 and ANGPTL2 in cancer tissues was significantly higher than in adenomas. High expression of LILRB2 and ANGPTL2 in CRC indicates that LILRB2 and ANGPTL2 interaction plays a role in promoting the transformation of adenoma into carcinoma. In particular, high expression of LILRB2 was not observed in adenomas, proving that LILRB2 is closely related to occurrence and development of cancer. Levels of LILRB2 protein in CRC tissues was further confirmed by immunohistochemistry and was significantly higher than normal tissues. In this study, the positive expression rate of LILRB2 in CRC was 87.10%.
Currently, there are no reports on expression of LILRB2 in CRC tissues. This study takes the lead in verifying the high expression of LILRB2 in CRC. In addition, LILRB2 was found to be correlated with tumor size, worse cell differentiation, and advanced disease stage, suggesting that LILRB2 may be involved in progression of CRC. Moreover, we analyzed correlation between LILRB2 and OS of CRC patients. Unexpectedly, we were unable to find a significant correlation between LILRB2 and OS. We explored mRNA expression levels of LILRB2 using TCGA data in GEPIA database and found that LILRB2 mRNA in CRC was upregulated, compared with corresponding normal tissues. While in survival analysis, there was no significant difference between LILRB2 mRNA high expression group and low expression group. Results of survival analysis at mRNA level in GEPIA database were consistent with the results at protein level. There may be some reasons that significant correlation was found between expression of LILRB2 and OS. First, the results may have been interfered by sampling error for patients from a single area. Second, LILRB2 might not be an independent prognostic factor. As ligand of LILRB2, ANGPTL2 negative correlation with OS and DFS may explain this phenomenon, in part. Therefore, the prognosis significance of LILRB2 still requires much research.
Previously, ANGPTL2 has been considered as an “orphan ligand”, but a recent study has found that ANGPTL2 is a high affinity ligand for LILRB2 and supports development of a variety of malignant tumors. It was reported that ANGPTL2 binds LILRB2 and supports development of LSCs and leukemia [2]. The combination of ANGPTL2 and LILRB2 has also been found to promote growth of tumor cells in NSCLC [24]. Does ANGPTL2 also bind to receptor LILRB2 as a ligand in CRC? No synergistic expression of LILRB2 and ANGPTL2 in CRC has been reported. Therefore, expression of ANGPTL2 in CRC was examined in this study. It was noted that expression of ANGPTL2 in CRC tissues increased significantly. The positive rate of ANGPTL2 in CRC tissues was 97.44%. In addition, the level of ANGPTL2 had a significant correlation with clinicopathological parameters, suggesting that ANGPTL2 was involved in development of CRC. Analysis of the correlation between ANGPTL2 protein and OS showed that survival time of patients with high ANGPTL2 expression was shorter than those with low expression of ANGPTL2. A consistent OS and DFS result also was obtained at the mRNA level using TCGA data in GEPIA database.
The relationship between LILRB2 and ANGPTL2 was comprehended by co-expression analysis, providing evidence for the binding of ANGPTL2 to LILRB2 in CRC. It was found that there was a significant positive correlation between LILRB2 and ANGPTL2 in many microarrays, in Oncomine database. Furthermore, LILRB2 and ANGPTL2 also presented a highly positive correlation by TCGA data. These results suggest that ANGPTL2 plays a biological role acts as a ligand binding to LILRB2 receptors in CRC. In addition, some stromal cells also expressed LILRB2 and ANGPTL2, suggesting that the inhibition function of LILRB2 for immune cells may play a role in the microenvironment around cancer cells. ANGPTL2 may also be secreted by cancer cells and be involved as a soluble ligand of LILRB2. LILRB2 and ANGPTL2 may play immunosuppression and tumor sustaining roles. Application of bioinformatics analysis laid a foundation for our hypothesis. Large sample sizes avoid errors produced by small sample sizes in a single study.
In conclusion, LILRB2 and ANGPTL2 play important roles in CRC, especially in the transformation from normal and precancerous lesions to CRC. They are also correlated with prognosis. This study presents a perspective for occurrence and progression of CRC and provides a basis for exploring LILRB2 and ANGPTL2 in CRC. It also has a guiding significance for clinical treatment and prognosis of CRC. The molecular mechanisms still require further study in the future.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant no.81570053).
Disclosure of conflict of interest
None.
References
- 1.Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer statistics, 2012. CA Cancer J Clin. 2015;65:87–108. doi: 10.3322/caac.21262. [DOI] [PubMed] [Google Scholar]
- 2.Zheng J, Umikawa M, Cui C, Li J, Chen X, Zhang C, Huynh H, Kang X, Silvany R, Wan X, Ye J, Canto AP, Chen SH, Wang HY, Ward ES, Zhang CC. Inhibitory receptors bind ANGPTLs and support blood stem cells and leukaemia development. Nature. 2012;485:656–660. doi: 10.1038/nature11095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Vlad G, Chang CC, Colovai AI, Berloco P, Cortesini R, Suciu-Foca N. Immunoglobulin-like transcript 3: a crucial regulator of dendritic cell function. Hum Immunol. 2009;70:340–344. doi: 10.1016/j.humimm.2009.03.004. [DOI] [PubMed] [Google Scholar]
- 4.Hunt JS, Petroff MG, McIntire RH, Ober C. HLA-G and immune tolerance in pregnancy. FASEB J. 2005;19:681–693. doi: 10.1096/fj.04-2078rev. [DOI] [PubMed] [Google Scholar]
- 5.Petroff MG, Sedlmayr P, Azzola D, Hunt JS. Decidual macrophages are potentially susceptible to inhibition by class Ia and class Ib HLA molecules. J Reprod Immunol. 2002;56:3–17. doi: 10.1016/s0165-0378(02)00024-4. [DOI] [PubMed] [Google Scholar]
- 6.Khan MF, Bahr JM, Yellapa A, Bitterman P, Abramowicz JS, Edassery SL, Basu S, Rotmensch J, Barua A. Expression of leukocyte inhibitory immunoglobulin-like transcript 3 receptors by ovarian tumors in laying hen model of spontaneous ovarian cancer. Transl Oncol. 2012;5:85–91. doi: 10.1593/tlo.11328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhang Y, Lu N, Xue Y, Zhang M, Li Y, Si Y, Bian X, Jia Y, Wang Y. Expression of immunoglobulin-like transcript (ILT)2 and ILT3 in human gastric cancer and its clinical significance. Mol Med Rep. 2012;5:910–916. doi: 10.3892/mmr.2012.744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Liu J, Wang L, Gao W, Li L, Cui X, Yang H, Lin W, Dang Q, Zhang N, Sun Y. Inhibitory receptor immunoglobulin-like transcript 4 was highly expressed in primary ductal and lobular breast cancer and significantly correlated with IL-10. Diagn Pathol. 2014;9:85. doi: 10.1186/1746-1596-9-85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sun Y, Liu J, Gao P, Wang Y, Liu C. Expression of Ig-like transcript 4 inhibitory receptor in human non-small cell lung cancer. Chest. 2008;134:783–788. doi: 10.1378/chest.07-1100. [DOI] [PubMed] [Google Scholar]
- 10.Carbone C, Piro G, Fassan M, Tamburrino A, Mina MM, Zanotto M, Chiao PJ, Bassi C, Scarpa A, Tortora G, Melisi D. An angiopoietin-like protein 2 autocrine signaling promotes EMT during pancreatic ductal carcinogenesis. Oncotarget. 2015;6:13822–13834. doi: 10.18632/oncotarget.2635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hato T, Tabata M, Oike Y. The role of angiopoietin-like proteins in angiogenesis and metabolism. Trends Cardiovasc Med. 2008;18:6–14. doi: 10.1016/j.tcm.2007.10.003. [DOI] [PubMed] [Google Scholar]
- 12.Zhang H, Wong CC, Wei H, Gilkes DM, Korangath P, Chaturvedi P, Schito L, Chen J, Krishnamachary B, Winnard PT Jr, Raman V, Zhen L, Mitzner WA, Sukumar S, Semenza GL. HIF-1-dependent expression of angiopoietin-like 4 and L1CAM mediates vascular metastasis of hypoxic breast cancer cells to the lungs. Oncogene. 2012;31:1757–1770. doi: 10.1038/onc.2011.365. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 13.Zhang CC, Kaba M, Ge G, Xie K, Tong W, Hug C, Lodish HF. Angiopoietin-like proteins stimulate ex vivo expansion of hematopoietic stem cells. Nat Med. 2006;12:240–245. doi: 10.1038/nm1342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zhang CC, Kaba M, Iizuka S, Huynh H, Lodish HF. Angiopoietin-like 5 and IGFBP2 stimulate ex vivo expansion of human cord blood hematopoietic stem cells as assayed by NOD/SCID transplantation. Blood. 2008;111:3415–3423. doi: 10.1182/blood-2007-11-122119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zheng J, Huynh H, Umikawa M, Silvany R, Zhang CC. Angiopoietin-like protein 3 supports the activity of hematopoietic stem cells in the bone marrow niche. Blood. 2011;117:470–479. doi: 10.1182/blood-2010-06-291716. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Aoi J, Endo M, Kadomatsu T, Miyata K, Nakano M, Horiguchi H, Ogata A, Odagiri H, Yano M, Araki K, Jinnin M, Ito T, Hirakawa S, Ihn H, Oike Y. Angiopoietin-like protein 2 is an important facilitator of inflammatory carcinogenesis and metastasis. Cancer Res. 2011;71:7502–7512. doi: 10.1158/0008-5472.CAN-11-1758. [DOI] [PubMed] [Google Scholar]
- 17.Sasaki H, Suzuki A, Shitara M, Hikosaka Y, Okuda K, Moriyama S, Yano M, Fujii Y. Angiopoietin-like protein ANGPTL2 gene expression is correlated with lymph node metastasis in lung cancer. Oncol Lett. 2012;4:1325–1328. doi: 10.3892/ol.2012.924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kikuchi R, Tsuda H, Kozaki K, Kanai Y, Kasamatsu T, Sengoku K, Hirohashi S, Inazawa J, Imoto I. Frequent inactivation of a putative tumor suppressor, angiopoietin-like protein 2, in ovarian cancer. Cancer Res. 2008;68:5067–5075. doi: 10.1158/0008-5472.CAN-08-0062. [DOI] [PubMed] [Google Scholar]
- 19.Teicher BA. Searching for molecular targets in sarcoma. Biochem Pharmacol. 2012;84:1–10. doi: 10.1016/j.bcp.2012.02.009. [DOI] [PubMed] [Google Scholar]
- 20.Carbone C, Moccia T, Zhu C, Paradiso G, Budillon A, Chiao PJ, Abbruzzese JL, Melisi D. Anti-VEGF treatment-resistant pancreatic cancers secrete proinflammatory factors that contribute to malignant progression by inducing an EMT cell phenotype. Clin Cancer Res. 2011;17:5822–5832. doi: 10.1158/1078-0432.CCR-11-1185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang L, Geng T, Guo X, Liu J, Zhang P, Yang D, Li J, Yu S, Sun Y. Co-expression of immunoglobulin-like transcript 4 and angiopoietin-like proteins in human non-small cell lung cancer. Mol Med Rep. 2015;11:2789–2796. doi: 10.3892/mmr.2014.3029. [DOI] [PubMed] [Google Scholar]
- 22.Kang X, Lu Z, Cui C, Deng M, Fan Y, Dong B, Han X, Xie F, Tyner JW, Coligan JE, Collins RH, Xiao X, You MJ, Zhang CC. The ITIM-containing receptor LAIR1 is essential for acute myeloid leukaemia development. Nat Cell Biol. 2015;17:665–677. doi: 10.1038/ncb3158. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhang P, Yu S, Li H, Liu C, Li J, Lin W, Gao A, Wang L, Gao W, Sun Y. ILT4 drives B7-H3 expression via PI3K/AKT/mTOR signalling and ILT4/B7-H3 co-expression correlates with poor prognosis in non-small cell lung cancer. FEBS Lett. 2015;589:2248–2256. doi: 10.1016/j.febslet.2015.06.037. [DOI] [PubMed] [Google Scholar]
- 24.Liu X, Yu X, Xie J, Zhan M, Yu Z, Xie L, Zeng H, Zhang F, Chen G, Yi X, Zheng J. ANGPTL2/LILRB2 signaling promotes the propagation of lung cancer cells. Oncotarget. 2015;6:21004–21015. doi: 10.18632/oncotarget.4217. [DOI] [PMC free article] [PubMed] [Google Scholar]