Insulin-mediated signaling promotes proliferation and survival of glioblastoma through Akt activation

Abstract

Background: Metabolic complications such as obesity, hyperglycemia, and type 2 diabetes are associated with poor outcomes in patients with glioblastoma. To control peritumoral edema, use of chronic high-dose steroids in glioblastoma patients is common, which can result in de novo diabetic symptoms. These metabolic complications may affect tumors via profound mechanisms, including activation of insulin receptor (InsR) and the related insulin-like growth factor 1 receptor (IGF1R) in malignant cells.

Methods: In the present study, we assessed expression of InsR in glioblastoma surgical specimens and glioblastoma response to insulin at physiologically relevant concentrations. We further determined whether genetic or pharmacological targeting of InsR affected oncogenic functions of glioblastoma in vitro and in vivo.

Results: We showed that InsR was commonly expressed in glioblastoma surgical specimens and xenograft tumor lines, with mitogenic isoform-A predominating. Insulin at physiologically relevant concentrations promoted glioblastoma cell growth and survival, potentially via Akt activation. Depletion of InsR impaired cellular functions and repressed orthotopic tumor growth. The absence of InsR compromised downstream Akt activity, but yet stimulated IGF1R expression. Targeting both InsR and IGF1R with dual kinase inhibitors resulted in effective blockade of downstream signaling, loss of cell viability, and repression of xenograft tumor growth.

Conclusions: Taken together, our work suggests that glioblastoma is sensitive to the mitogenic functions of insulin, thus significant insulin exposure imposes risks to glioblastoma patients. Additionally, dual inhibition of InsR and IGF1R exhibits promise for treating glioblastoma.

Keywords: Akt; IGF1R; glioblastoma; insulin; insulin receptor.

Fig. 1.

Characterization of InsR and IGF1R expression in glioblastoma. (A) Expression of InsR and IGF1R in high-grade glioma surgical specimens (lane 1–8) compared with normal brain tissues (lane 9–10). (B) Immunoblotting of InsR and IGF1R in glioblastoma ex vivo cultures. (C) Representative images (400×) of immunohistochemical staining of InsR and IGF1R in paraffin-embedded sections of a glioblastoma patient specimen, VU11046. The scale bars represent 100 µm.

Fig. 2.

Insulin, IGF1, and IGF2 stimulate glioblastoma cell growth and downstream signaling. (A and C) T4105 and GBM44 cells were cultured in insulin-free medium and treated with insulin, IGF1, or IGF2 with a 3-fold serial dilution for 5 days. Dose response curves were fitted by GraphPad Prism 5 using a 3-parameter nonlinear regression algorithm. Viability readings of cells cultured in insulin-free medium were assigned as 100%. ^#P < .05 by Student's t-test for insulin vs IGF1. (B and D) Following insulin starvation, growth rates of T4105 or GBM44 cells in the presence of 0.5 or 5 ng/mL insulin were determined as described in methods. *P < .01 by Student's t-test for insulin vs insulin-free medium. (E) T4105 and GBM44 cells were starved in insulin-free medium overnight. Cells were stimulated by indicated ligands for 30 min and harvested for immunoblotting with actin as the loading control.

Fig. 3.

Knockdown of InsR impairs glioblastoma proliferation and survival. (A) T4105 cells were infected with lentivirus directing expression of nontargeting shRNA (NT) or shRNA sequences specific to InsR. After selection with 1 µg/mL puromycin for 2 days, relative InsR mRNA levels were determined by qRT-PCR. (B) Cells were then used to determine the growth rate and (C) neurosphere formation. (D) Representative cell cycle distribution and (E) caspase-3/7 activation were assessed 3 days after puromycin selection. Caspase activation was measured by the Caspase-Glo 3/7 kit (Promega) and normalized to the corresponding cell viability. *P < .001 by Student's t-test.

Fig. 4.

IGF1R is also implicated in glioblastoma proliferation and survival. (A) Following completion of puromycin selection, T4105 cells were subject to immunoblotting for indicated proteins and (B) qRT-PCR for expression of InsR and IGF1R. The mRNA levels of InsR and IGF1R were normalized to actin mRNA. *P < .001 by Student's t-test. (C) Growth curves after IGF1R knockdown were determined as described in Fig. 3B. (D) Following lentiviral infection and puromycin selection, 5000 T4105 cells were injected intracranially into athymic nude mice (n = 6). Animals were sacrificed upon development of neurologic signs. Median survival for the NT group was 28 days, and for the InsR-knockdown group and IGF1R-knockdown group 38 days (log-rank P = .0011) and 36 days (log-rank P = .0004), respectively.

Fig. 5.

Dual inhibition of InsR and IGF1R reduces glioblastoma cell viability and represses subcutaneous xenograft growth. (A) InsR/IGF1R dual inhibitors, OSI-906 and BMS-754807, dose-dependently inhibited T4105 viability and (B) neurosphere formation. *P < .01, **P < .001 by Student's t-test. (C) T4105 cells were treated with OSI-906 or BMS-754807 at indicated concentrations for 24 hours, lysed, and subjected to immunoblotting. (D) T4105 cells were infected with a lentivirus directing expression of myristoylated Akt1. Dose-response curves to OSI-906 were assessed for control and Akt-activated cells. (E) T4105 subcutaneous tumors were established in both flanks of athymic nude mice (n = 5) and treated with 50 mg/kg OSI-906 or vehicle once daily, 5 days on, 3 days off. Data are mean ± SE. (F) At the end of the experiment, tumors were dissected and the wet weight was measured. P = .02 by Student's t-test.