Proliferation of Colo-357 Pancreatic Carcinoma Cells and Survival of Patients With Pancreatic Carcinoma Are Not Altered by Insulin Glargine

Human pancreatic cancer cells (Colo-357) were incubated for 72 h in RPMI culture medium and treated with either insulin glargine (Sanofi-Aventis, Frankfurt, Germany) or regular human insulin (Novo Nordisk, Bagsværd, Denmark) at the following concentrations: 0, 0.01, 0.1, 1, 10, and 100 nmol/l. After static incubation, Colo-357 cells were either analyzed by flow cytometry or immunoblotting. Colo-357 cells were maintained in RPMI culture medium supplemented with 10% fetal bovine serum, penicillin G (100 units/ml), and streptomycin (100 μg/ml) at 37°C in humidified air containing 5% CO₂. Colo-357 cells are characterized by a high expression level of insulin receptor substrate (IRS) 2, which is implicated in mitogenic signaling after activation of the insulin or IGF-I receptor. Prior reports showed that Colo-357 cells are insulin-responsive cells (11). In contrast to this and other studies, we did not use serum-free medium for preconditioning/cell cycle synchronization of the cells because serum starvation does not represent a condition that occurs physiologically to cancer cells in vivo.

To corroborate the in vitro data, 125 patients were retrospectively examined, in whom total or partial pancreatectomy had been performed due to pancreatic carcinoma. Patients were divided into three groups: 1) control group without diabetes after pancreatectomy, 2) patients with diabetes and subcutaneous insulin glargine, and 3) patients with diabetes and subcutaneous insulin without glargine. All data were updated by contacting each patient's primary care physician. Survival in the groups was compared by Kaplan-Meier survival analysis.

After insulin treatment for 72 h, Colo-357 cells were harvested and washed with PBS. Cells (2.5 × 10⁵ cells) from each sample were used for the measurement of replication (Ki-67) or apoptosis (annexin-V). For detection of replication, cells were permeabilized with a Fix&Perm Cell Permeabilization Kit (An der Grub, Kaumberg, Austria) followed by an immunofluorescent staining with an antibody to Ki-67 antigen (Ki-67 R-phycoerythrin–conjugated mouse anti-human monoclonal antibody set; BD Pharmingen, Heidelberg, Germany). Apoptotic cells were quantified using annexin-V–fluorescein isothiocyanate (Invitrogen, Karlsruhe, Germany). In addition, every sample was stained with propidium iodide (BD Pharmingen) to detect nonviable cells. Antibody labeling as well as annexin-V and propidium iodide staining were done according to the manufacturers directions. Each sample was processed using a standard FACScan flow cytometer (Becton Dickinson, San Jose, CA).

After appropriate incubations, Colo-357 cells were lysed with mammalian protein extraction reagent combined with protease and phosphatase inhibitors (Pierce, Rockford, IL). Cell lysates (30 μg) were subjected to 10% SDS-PAGE and electro-transferred to immobilon-P membranes (Millipore, Billerica, MA). Membranes were incubated with antibodies against the insulin receptor, IGF-I receptor, and pan actin as a loading control (Cell Signaling Technology, Danvers, MA) at dilutions of 1:2,000. IRS2 and phosphoinsulin receptor β-antibodies (Cell Signaling Technology) were used at a dilution of 1:1,000. Corresponding secondary horseradish-conjugated antibodies were diluted to 1:5,000. Bound antibodies were visualized using chemiluminescence (ECL Western blotting detection reagents; Amersham, Buckinghamshire, U.K.).

Changes in the mean rates of replication, apoptosis, and the expression levels of insulin receptor, IGF-I receptor, and IRS2 after each treatment were determined by the Student's t test. Changes at different concentrations of both insulin treatments were examined by ANOVA. P < 0.05 was considered to denote significant differences. Survival analysis (Kaplan-Meier) was performed as a three-sample comparison with the software Statistica (version 6.1).

After 72 h static incubation of Colo-357 cells with either insulin glargine or regular human insulin, flow cytometric analysis was performed to determine proliferation and apoptosis (Fig. 1). Colo-357 cells are characterized by a homogenous autofluorescence, confirming that this cell type is an appropriate model for flow cytometric analysis (Fig. 1A). The fraction of propidium iodide–positive (nonviable) cells was always <5% (data not shown). Cells positive for Ki-67 or annexin-V separated clearly from the unstained cell population (Fig. 1B and C). The replication rate (percentage of Ki-67–positive cells) was 0.55–3.11%, and the number of apoptotic cells (percentage of annexin-V–positive cells) was in the range of 0.94–4.84% (data not shown). There was no concentration-dependent change in cell proliferation after incubation with human insulin or insulin glargine (P = 0.3 and P = 0.4 by ANOVA, respectively) (Fig. 1D). Furthermore, there was no significant difference in cell replication after incubation with human insulin or insulin glargine at any of the individual concentrations tested (P = NS). Similarly, there is no interaction between insulin concentrations and apoptosis for both insulins (P = 0.9 for insulin glargine, P = 0.2 for human insulin by ANOVA) (Fig. 1E), and there were no differences comparing individual concentrations (P = NS).

To demonstrate that Colo-357 cells respond to exogenous insulin, the phosphorylation level of the insulin receptor was measured by immunoblotting using a monoclonal phospho-specific antibody to the insulin receptor β-subunit. After incubation for 30 min, the phosphorylation level of the insulin receptor was increased to a similar degree with regular human insulin or insulin glargine (Fig. 2A and B). Expression levels of insulin receptor, IGF-I receptor, and IRS2 were characterized in additional experiments by immunoblotting using specific rabbit polyclonal antibodies against the β-subunit of the IGF-I receptor and total IRS2 and a monoclonal antibody against the insulin receptor β-subunit (Fig. 2C). The pattern of the expression levels of the IGF-I receptor and IRS2 were similar with both types of insulin (Fig. 2D). There were no differences of the IGF-I receptor expression across the wide spectrum of insulin concentrations. IRS2 expression was downregulated at high concentrations of regular human insulin. Increasing insulin glargine concentrations did not induce changes in the expression levels of total IRS2. Expression of the insulin receptor was downregulated at high concentrations of insulin glargine as well as regular human insulin (Fig. 2D). However, comparing both insulins, the expression levels did not show a significant difference at any of the concentrations tested.

To extend the analysis of the long-acting insulin analogue glargine into the clinical context of pancreatic cancer, we examined the survival of patients after surgery due to pancreatic carcinoma. Subjects were analyzed over a median follow-up period of 22 months after surgery (total n = 125). Strategies of insulin therapy were equally distributed among both groups with diabetic patients (Table 1). There were no differences regarding sex, age, tumor stage, and histomorphology of the tumors between the subject groups. Similarly, the surgical techniques and status of resection, which have an important impact on patients’ survival, did not differ in the three groups. The median survival time was 15 months (Fig. 3A). Survival analysis (Kaplan-Meier) showed that the time-dependent proportion of patients who survived was not altered by insulin glargine (P = 0.4, three-sample comparison) (Fig. 3B) compared with treatment with regular human insulin and subjects without diabetes after pancreatectomy.

In the present study, we sought to investigate the influence of insulin glargine on proliferation of pancreatic cancer cells in comparison with regular human insulin and whether treatment with insulin glargine has an impact on the survival of patients with pancreatic carcinoma. These questions were addressed using a human pancreatic cancer cell line for the in vitro studies and additionally analyzing patients after pancreatectomy due to pancreatic carcinoma. Short-term incubation studies confirmed that the cells used in the present projects respond to exogenous insulin. We report that the turnover of human pancreatic cancer cells is not altered by insulin glargine. There were also no significant differences of the expression patterns of the insulin and IGF-I receptor compared with regular human insulin. Unchanged cell proliferation after prolonged incubation might be partially due to downregulation of the insulin receptor. With low (0.01–1 nmol/l) and high (10–100 nmol/l) insulin concentrations, we reproduced physiological and supraphysiological conditions. This was necessary because insulin selectively activates the insulin receptor at physiological concentrations, whereas at high concentrations it cross-activates the IGF-I receptor (8). Survival after surgery because of pancreatic carcinoma was not influenced by the type of insulin used for therapy of diabetes.

Prior receptor-binding studies using different types of insulin demonstrated that structural modifications of the insulin molecule at its C-terminus are responsible for an altered affinity to the IGF-I receptor (12). Accordingly, an about sixfold higher affinity of insulin glargine to the IGF-I receptor was observed (7). It was hypothesized that mitogenic effects of insulin are mainly mediated through IGF-I receptor stimulation at high insulin concentrations. In an experimental model, the proliferation of human osteosarcoma cells directly correlated with the increased affinity of insulin glargine to the IGF-I receptor (7). From these studies it was inferred that treatment with insulin glargine might induce cell proliferation in various tissues. However, an increased mitogenic potential has not been supported in other studies. In comparison to regular human insulin, the proliferation of insulin glargine–treated cells was not significantly different in human coronary artery endothelial and smooth muscle cells (13), skeletal muscle cells (14), rat fibroblasts (15), cardiac myoblasts, and adult rat ventricular cardiomyocytes (16). One in vivo study (17) in rats and mice reported no significant difference between insulin glargine and NPH insulin with regard to the development of mammary tumors.

In the present study, we extended the experimental evidence into the context of pancreatic cancer and report that insulin glargine does not induce proliferation of pancreatic cancer cells compared with regular human insulin. It is established that the development of pancreatic cancer is associated with diabetes. However, an exact mechanism has not been identified. On the one hand, type 2 diabetes is considered a risk factor for several types of cancer (e.g., of the pancreas [18], breast [19], endometrium [20], liver [21], bladder [22], colon, and rectum [23–25]). Therefore, it was postulated that the increased risk is mainly constituted by elevated concentrations of circulating insulin (26) compared with insulin levels in lean nondiabetic humans. On the other hand, there is also evidence from case-control studies to suggest that pancreatic cancer triggers the development of diabetes because there is a significant association between the two conditions if diabetes is diagnosed concomitantly or within 2 years before the diagnosis of cancer (27). These studies allow the conclusion that glucose metabolism might be influenced by the tumor itself or a secreted product from the tumor with effects on insulin sensitivity (28). Consistent with this hypothesis, it has been reported that insulin sensitivity may be improving after resection of pancreatic tumors (29). Pancreatic ductal adenocarcinoma, which is the most common form of pancreatic cancer in general and in our group of patients, is believed to originate from pancreatic progenitor or stem cells (30,31) and not from pancreatic β-cells (32), which are exposed to the highest local insulin concentrations. When pancreatic progenitor or stem cells can be readily identified and are available for experimental applications, it would be interesting to analyze the concentration-dependent effects of insulin in these cells with respect to tumor formation. Clinically, this would be relevant for patients with diabetes before development of pancreatic cancer. The present result that survival of patients with pancreatic carcinoma is not influenced by diabetes is consistent with other reports showing that survival in patients with pancreatic carcinoma and a palliative therapeutic concept (33) or after surgical resection is not associated with metabolic status (34). However, median survival is short in the present group of patients. Therefore, it might be interesting to perform similar analyses with patients undergoing surgery at earlier tumor stages than T3 and with longer survival.

In summary, we demonstrated that regular human insulin as well as insulin glargine did not influence proliferation and apoptosis in human pancreatic cancer cells. Moreover, therapy with both insulins was not associated with altered survival of patients after pancreatectomy due to pancreatic carcinoma. Based on these results, regular human insulin and insulin glargine may be safely used to treat patients with pancreatic cancer and diabetes.

These studies were funded by the Deutsche Forschungsgemeinschaft, DFG (Ri 1055/3-1). P.P.N. and M.W.B. were supported by the Lautenschläger Foundation.