This study was approved by the Institutional Review Boards of the First Affiliated Hospital of Zhengzhou University. Written informed consent was obtained from all participants for the use of their blood or tissue samples. Full consent was obtained from the Institutional Review Boards of the First Affiliated hospital of Zhengzhou University. Animal care and experiments were approved by the Institute Animal Care and Use Committee of the Cancer Institute and Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College.

ESCC cell lines EC9706, EC109 and NEC were gifts from the Department of Etiology and Carcinogenesis, Cancer Institute and Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College. All the cells were cultured in RMPI-1640 medium with 10% fetal bovine serum (FBS), in 5% CO₂ at 37°C.

Eighty human ESCC tissues used for immunohistochemical staining were collected from the First Affiliated Hospital of Zhengzhou University with the exact post-surgical pathological examination. Eighteen normal tumor-adjacent tissues were acquired from these patients at least 3 cm away from the primary tumor. These patients consisted of 52 females and 28 males. The mean age was 60 years. According to histological grade, 22 patients were in grade I, 37 patients were in grade II, and 21 patients were in grade III. According to clinical stage 6 patients were in stage I, 35 patients were in stage II, 32 patients were in stage III and 7 patients were in stage IV. There were 26 patients without lymph node metastasis and 53 patients with lymph node metastasis.

Serum samples used for the ELISA study were collected from 120 patients. These patients were diagnosed with ESCC between 2007 to 2009 at The First Affiliated Hospital of Zhengzhou University, Henan Cancer Hospital and Zhengzhou Central Hospital. None of these patients received any therapy before the two courses of chemotherapy treatment. The pre-therapy serum was collected 1–3 days before treatment, and the post-therapy serum was collected at the end of the treatments. These patients consisted of 51 females and 121 males. The mean age was 60 years. According to clinical stage, 43 patients were in stage IIB, 70 patients were in stage III and 59 patients were in stage IV. The chemotherapy effect was divided into 4 grades according to the Response Evaluation Criteria in Solid Tumors (RECIST): complete response (CR), all lesions disappeared lasting for at least 4 weeks; partial response (PR), the lesion volume shrank >50% lasting for at least 4 weeks; stable disease (SD), no significant change, shrinking <50% and enlargement <25%; progressive disease (PD), the lesion enlarged >25%.

The microarray data GSE 33426 and GSE 23400 were downloaded from the GEO database: (http://www.ncbi.nlm.nih.gov/gds/). The differentially expressed genes were identified using GEO2R tools on GEO. The heatmap was generated using Genesis 1.0 (Graz, Austria). The biofunction and pathway analysis was performed with IPA^® Software (http://www.ingenuity.com).

Fresh tumor tissues were fixed with formalin, embedded and finally cut into 4-μm sections. The sections were deparaffinized by dimethylbenzene and graded ethanol. Then the sections were incubated with 3% H₂O₂, goat serum, anti-IGF-1R antibody, biotin-labeled secondary antibody, peroxidaze-conjugated streptavidin in order before DAB staining. All the reagents were purchased from ZSGB-Bio (Beijing, China). The staining intensity of IGF-1R was quantified into 3 groups according to the percentage of stained cells: negative, 0–5% cells were stained; weak positive, >5–50% cells were stained; strong positive, >50–100% cells were stained.

Protein was extracted from the cells using lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100). The protein concentration was measured using the the Bradford method. The protein was separated by SDS-PAGE and transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore, Billerica, MA, USA). After incubation with 10% milk, the primary IGF-1R antibody, and the secondary antibody, the blot signals were detected by chemiluminescence using the ECL kit (Applygen, Beijing, China). All the antibodies were purchased from Santa Cruz (Dallas, TX, USA).

pGCsilencer™ U6/Neo/GFP/RNAi-IGF-1R plasmid used for the knockdown of IGF-1R and the pGCsilencer™U6/Neo/GFP negative control were obtained from Genechem Co. (Shanghai, China). Both plasmids were transfected into EC9706 cells using Lipofectamine™ 2000 (Invitrogen, Carlsbad, CA, USA). The cells were digested and seeded into a new 24-well plate, diluted 5 times, 24 h after the transfection. Then, the regular medium was changed to medium containing 600 μg/ml G418 for stable transfection colony screening. After all of the cells without transfection died 4 days later, the remained cells were sustained by medium containing 200 μg/ml G418 and accepted as the stable transfection cell lines: IGF-1R-knockdown cell line and blank transfection control cell line. Western blotting and RT-PCR verified the knockdown of IGF-1R in the cell lines.

Cells were seeded in a 96-well plate at a density of 4,000 cells/well. The next day, 10 μl of 5mg/ml MTT was added to the medium for a 3-h incubation, and the time point was set as zero then detected every 24 h for 6 days. Then the medium was removed, and 150 μl DMSO was added. Finally the OD value was read at 490 nm on a Model 680 microplate reader (Bio-Rad, Hercules, CA, USA) after a 10-min incubation.

Cells were digested by trypsin, washed twice in cold PBS, fixed with ice cold 70% methanol, and incubated at 4°C overnight. Cells were then washed in PBS and incubated with 25 μg/ml propidium iodide containing 30 μg/ml ribonuclease for 30 min at room temperature. Cells were analyzed on the FACSCalibur Cell Sorting System (Becton Dickinson, Franklin Lakes, NJ, USA) using BD FACSuite™ Software in triplicate.

Cells were seeded onto 6-well plates at a density of 300 cells/well. Following 12 days, visible colonies were fixed with methanol and stained with 0.4% crystal violet solution. Colonies were counted, and the number of visible colonies in each well was determined. Each assay was performed in triplicate.

Male BALB/c nude mice 5–6 weeks old were purchased from the Institute of Laboratory Animal Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College. The mouse were divided into three groups and were inoculated with stable IGF-1R-knockdown cells, empty vector control cells or wild-type EC9706 cells, respectively (six mice in each group). A total of 2×10⁶ cells were inoculated subcutaneously into the right dorsal flank. The size of the tumor was measured twice per week using a digital vernier caliper. The tumor volume was determined from the orthogonal dimensions (d1, d2, d3) using the formula (d1 × d2 × d3) × π/6. Five weeks after inoculation, the mice were euthanized by cervical dislocation, and xenograft tumors were collected.

Collected venous blood was centrifuged for 10 min at 2,000 rpm. The upper tube of serum was then taken for cryopreservation and stored at −40°C for ELISA assay. IGF-1 and IGFBF-3 ELISA kits were obtained from Senxiong Biotech (Shanghai, China). The standard curve was established according to the manufacturer’s instructions. Serum samples (100 μl) were incubated in each well for 120 min, followed by 50 μl primary antibody buffer for 60 min, 100 μl secondary primary antibody buffer for 60 min, 100 μl substrate buffer for 10 min and 1 drop stop solution. Then the plate was read at 490 nm on the Model 680 microplate reader. The serum concentrations of IGF-1 and IGFBP3 were calculated by the standard curves.

All statistical analyses were performed using the SPSS 13.0 package. Pearson’s χ² test was used to detect the expression difference in the cancer and normal tissues. All tests of statistical significance were two-sided, and P<0.05 was considered to indicate a statistically significant difference. All quantitative data are presented as means ± SD. Independent samples t-test and one-way ANOVA were adopted to compare the means.

In order to study the expression and potential function of IGF-1R in ESCC, we took advantage of publicly available gene expression data (GSE33426 and GSE23400). In the referenced studies, tumor cluster and normal cells separated by laser-capture microscope or tumor tissues and adjacent normal tissues were compared by cDNA microarray. IGF-1R was identified as a differentially expressed gene in both studies. Moreover, we found 85 and 50 genes, respectively, related with the IGF-1R signaling pathway (Fig. 1A and B). Twenty-three genes in the IGF-1R signaling pathway were identified in both studies (Fig. 1C and D). After IPA analysis, we found that these genes were mainly mapped to AKT and MEK cascade and were related to cell growth, proliferation and survival (Fig. 1E). The bioinformatic analysis supported our hypothesis that IGF-1R plays an important role in ESCC.

Firstly, we used immunohistochemical (IHC) staining to detect the expression of IGF-1R in ESCC tumor tissues (Fig. 2A and C) and adjacent normal tissues (Fig. 2B and D). After IHC staining, IGF-1R showed strong staining in the cancer cell membrane. Of the 80 cancer specimens, 11 cases were negative, 28 cases were weakly positive and 41 cases were strongly positive. The total positive rate was 86.25% and the strong positive rate was 51.25%. Of the 18 normal tumor-adjacent tissues, 7 cases were negative, 9 cases were weakly positive and 2 cases were strongly positive. The total positive rate was 61.11% and the strong positive rate was 11.11%. Both the total positive rate and the strong positive rate were significantly higher in the ESCC tissues than in the normal tumor-adjacent esophageal tissues (total positive rate, χ²=4.630, P<0.05; strong positive rate, χ²=9.614, P<0.01). In addition, we found a relationship between the expression of IGF-1R and cancer patient characteristics (Table I). We found that high expression of IGF-1R was associated with more aggressive lymph node metastasis, lower histological grade and advanced clinical stage. Both positive and strong positive rates of IGF-1R expression were also higher in patients with lymph node metastasis than those without lymph node metastasis (P<0.01), and were also higher in patients with advanced clinical stage (III–IV stage) than those with early clinical stage (I–II stage). The expression of IGF-1R was also increased from a low differentiation grade to a moderate differentiation stage and high differentiation grade (P<0.05 between each of the two groups). No significant difference was noted in relation to age or gender (P>0.05). These results showed that IGF-1R was highly expressed in ESCC tissues, and the high expression of IGF-1R in the ESCC tissues was related to tumor progression.

We determined the expression of IGF-1R in EC9706, EC109 and NEC cells and found that IGF-1R was relatively highly expressed in these three ESCC cell lines. We established stable IGF-1R-knockdown cells by transfection of pGCsilencer™U6/Neo/GFP/RNAi IGF-1R in EC9706 cells. Western blotting and RT-PCR verified the successful knockdown of IGF-1R in these cells (Fig. 3A and B). Compared with the empty vector control, the cell proliferation was inhibited and the doubling time was extended significantly in the IGF-1R-knockdown cells (Fig. 3C). Significant cell cycle arrest occured in the IGF-1R-knockdown cells with an increased percentage of G1 phase cells (P<0.05) and a decreased percentage of S phase cells (P<0.01) (Fig. 3D). In addition, a higher apoptosis rate (P<0.05) and weaker clonogenesis ability (P<0.05) were found after IGF-IR-knockdown (Fig. 3E). All of these results showed that IGF-1R acts as an oncogene in ESCC in vitro. We also used a mouse ESCC xenograft model to detect the tumor growth ability of IGF-1R in vivo. Ten days after cell inoculation, subcutaneous nodules were noted and finally developed into tumors. Although all three cell lines induced visible tumors in the nude mouse, the size of the tumors induced by the stable IGF-1R-knockdown cells was significantly smaller than those formed by the empty control (P<0.05) and wild-type cells (P<0.05) (Fig. 4A and B). The animal assay showed that IGF-1R can induce tumor growth in vivo.

We compared the sensitivity to 5-FU and cisplatin between the IGF-1R-knockdown cells and the empty vector control. We found that the growth inhibition rate of 5-FU and cisplatin in the pGCsilencer™U6-IGF-1R-transfected cells was higher than that in the empty vector-transfected and the non-transfected cells (P<0.05) (Fig. 5A). The high expression of IGF-1R facilitates cancer cell resistance to chemotherapy.

We determined the serum concentrations of IGF-1 and IGFBP-3 in 172 patients both before and after chemotherapy treatment (Fig. 5B). The mean serum concentrations of IGF-1 pre-therapy and post-therapy in all patients were 268.87±61.66 and 266.42±49.98 ng/ml, respectively, without significant difference. The mean serum concentrations of IGFBP-3 pre-therapy and post-therapy in all patients were 2523.2±469.83 and 2598.8±563.56 ng/ml, respectively, without significant difference. In contrast, the serum concentration of IGFBP-3 increased (P<0.01) and the IGF-1/IGFBP-3 ratio decreased (P=0.01) significantly after chemotherapy in the chemotherapy responsive group (CR and PR). The serum concentration of IGF-1 decreased after chemotherapy in the responsive groups but without significance (P>0.05). Accordingly, in the chemotherapy unresponsive group (SD and PD) the IGF-1/IGFBP-3 ratio increased significantly (P<0.05) after chemotherapy; the concentration of IGFBP-3 decreased but without significance (P>0.05). There was no variation in IGF-1 (Table II). These results showed that the IGF-1/IGFBP-3 ratio can be used to predict the chemotherapeutic effect; a decreased IGF-1/IGFBP-3 molar ratio presents a better chemotherapeutic effect. We also evaluated the chemotherapy response rate in 4 groups divided according to the quartiles of plasma IGF-1 and IGFBP-3. Chemotherapy response rate increased from low concentrations of IGF-1 and IGFBP-3 to high concentrations of IGF-1 and IGFBP-3. These results showed that increased serum levels of IGF-1 and IGFBP-3 are associated with significantly higher rates of tumor response (P<0.05).