Metformin Enhances the Therapy Effects of Anti-IGF-1R mAb Figitumumab to NSCLC

The insulin-like growth factor (IGF) signaling system plays a critical role in tumorigenesis, highlighting the potential of targeting IGF-1R as an anti-cancer therapy. Although multiple anti-IGF-1R monoclonal antibody (mAb) drugs have been developed, challenges remain in the validation of the therapeutic effects and understanding the molecular mechanism of these mAbs. Herein, we conducted a study to validate the effect of Figitumumab (CP), an anti-IGF-1R mAb, in a panel of non-small cell lung cancer (NSCLC) cell lines. We found all tested cell lines were sensitive to CP, and CP could block IGF-1R and the downstream PI3K/AKT pathway activation. Unexpectedly, we found CP could activate ERK signaling pathway in IGF-1R kinase independent manner, which we further verified was mainly mediated by β-arrestin2. We also investigated the anti-tumor effect of metformin alone as well as its combination with CP to target NSCLC. Metformin could target IGF-1R signaling pathway by attenuating PI3K/AKT and MEK/ERK signaling pathways and down-regulating IGF-1R. Finally, we found that combining metformin with CP could further induce IGF-1R down-regulation and was more effective to target NSCLC cells. Our data suggests the combining of metformin with CP has additive therapeutic value against NSCLC.

Since many cancer cells are characterized with a constitutive high glucose uptake rate 27 , the "calorie restriction mimetic" activities of metformin are deemed to among the factors that contribute to its inhibitory effects on cancer growth and development [28][29][30] . Therefore, metformin is incorporated into current studies for cancer cell metabolism therapeutic approaches. At the molecular level, the activation of LKB1/AMPK/mTOR pathway and the inhibition of insulin-induced bio-cellular activities are investigated to exert its anti-neoplastic effects 31 .
Beyond the above, the emerging effects of metformin on IGF system capture our attention. Werner et al. has demonstrated that metformin could down-regulate the IGF-1R signaling pathway in a panel of uterine serous carcinoma cell lines, leading to the inhibition of cell proliferation and migration regardless of the p53 status 32 . In another study, metformin is reported to diminish the phosphorylation of a series of receptor tyrosine kinase (RTKs) and the down-stream PI3/AKT/mTOR and Ras/Raf/MEK/ERK pathways in lung tissues in an AMPK independent pattern in liver IGF-1-deficient (LID) mice model with NNK-induced lung carcinogenesis 33 .
β -arrestins are documented to function as E3 ubiquitin ligase adaptors thus mediate ubiquitylation of seven-transmembrane receptors (7TMRs), which result in receptor endocytosis, recycling or degradation 34,35 . Recent studies have broadened the functional scope of β -arrestins in receptor internalization to the field of RTKs, such as TGF-β and IGF-1R 36,37 . The concept of "biased agonist" has cut a figure in the field of 7TMRs, and biased ligands and receptors have been identified that preferentially signal through either G protein-or β -arrestin-biased pathways 34 . Biased agonists give intriguing possibility of directing cellular signaling with unprecedented precision and specificity and facilitate the development of new types of drugs with fewer side effects 35,38 . Our group aims at testing the sensitivity of NSCLC cells to CP and further testing the effect of CP on signaling pathways and receptor down-regulation. In particular, we discussed the role of β -arrestins in regulating cell signaling and therapeutic effect. Furthermore, we tested the therapeutic enhancement effect of metformin when combining with CP to treat NSCLC. We summarized the theoretical basis that support our hypothesis as following: (1) the caloric restriction effect and other anti-neoplastic mechanisms make metformin good choice in cancer therapy (2) anti-IGF-1R mAbs targeting therapy is promising approach against cancer, including NSCLC (3) inhibition effects on IGF-1R signaling system highlighting metformin as a candidate as an adjuvant in anti-IGF-1R system therapy (4) anti-IGF-1R mAb targeting therapy may benefit from metformin for its hypoglycemic property by reducing side effects. In summary our results help to better understand the mechanism and side effects of CP when using it to target NSCLC and we demonstrated the potential of combining CP with metformin as a novel NSCLC therapeutic strategy.

Results
Sensitivity to CP treatment in NSCLC cell lines. Firstly, we investigated the IGF-1R expression level in a panel of NSCLC cell lines, including A549, SK-MES-1, H520, SPC-A-1, H1975 and PC-9 (Fig. 1a). IGF-1R was expressed in all tested NSCLC cell lines, although the exact expression level was variable, with A549 and PC-9 displaying the highest and lowest levels, respectively. We then stimulated these cells with IGF-1, a complete (a) NSCLC cell lysates were prepared and IGF-1R were detected by WB. (b) Cells were starved for 12h and stimulated with 50 ng/ml IGF-1 for 10 min. p-IGF-1R, p-ERK, p-AKT and GAPDH were detected by WB. (c) Cells were treated with 100 ng/ml of CP for 48 h with presence or absence of serum. Cell viability was tested through MTT. The number of viable cells following CP treatment was presented as percentage of untreated cells. Data were presented as mean ± S.E.M. IGF-1R agonist. As shown in Fig. 1b, IGF-1 could induce IGF-1R phosphorylation and activation of downstream RAS/RAF/MEK/ERK and PI3K/AKT signaling pathways in most cell lines, and the level of activation was generally correlated with the amount of IGF-1R in the cell. These results suggested that IGF-1R signaling pathway was responsive in all the cell lines we used.
Subsequently, we investigated the sensitivity of these NSCLC cells to anti-IGF-1R mAb CP (100 ng/mL CP, whose molar concentration is 10-fold less than 50 ng/mL IGF-1). As demonstrated in Fig. 1c, cell viability decreased in all NSCLC cell lines after CP treatment, ranging from 64.7% in SK-MES-1 to 85.0% in PC-9 under serum added medium conditions, and from 45.0% in H1975 and 70.1% in PC-9 under serum free medium (SFM) conditions. CP was originally designed to compete with ligands for binding of IGF-1R, thus preventing the ligand-induced signaling pathway activation. This idea was supported by the inhibitory effect of CP seen in NSCLC cell lines in serum-added medium (Fig. 1c). However, it was surprising to see that CP had more inhibitory effects in cells under SFM conditions. It is possible that besides being an IGF-1 competitor, CP might have another IGF-1R dependent mechanism to inhibit cell proliferation and/or promote cell death.
β-Arrestins regulate IGF-1 but not CP-induced IGF-1R down-regulation. Multiple RTK-targeting mAbs were documented to have a delayed effect on the down-regulation of corresponding receptors. We processed a series of experiments to verify the receptor down-regulation property of CP and investigated its possible mechanisms. In order to get the pure effects of CP without the competition of IGF-1 or other growth factors that are normally presented in serum, all experiments were performed in SFM condition. As shown in Fig. 2a,b, both IGF-1 and CP could induce IGF-1R down-regulation in a time-dependent manner. The IGF-1R down-regulation effect was stronger and more sustained after CP treatment in most cell lines, especially at later time points, while in multiple IGF-1 treated cells (A549, H1975 and SPC-A-1) the expression of IGF-1R rebounded at around 48h. Since the molar concentration of CP was 10-fold less than that of IGF-1, our data suggested that CP is more effective in inducing IGF-1R down-regulation.

CP could induce β-arrestin-dependent ERK activation and inhibiting ERK activation could increase the therapeutic effect of CP against NSCLC. We further investigated the effects of CP on
IGF-1R signaling pathways in NSCLC cells and compared it with IGF-1 stimulation (Fig. 3a). As a full agonist, IGF-1 could induce IGF-1R phosphorylation and downstream signaling pathways activation within 5 min. The activation effects of IGF-1 showed in a time depended pattern, peaking around 10 to 30 min and lasting for at least 60 min. In contrast, p-IGF-1R and p-AKT were undetectable after CP treatment. To our surprise, CP could induce ERK phosphorylation in most NSCLC cell lines, although the ERK activation level was still lower than that induced by IGF-1. Since IGF-1R was kept un-phosphorylated (under the sensitivity of Western Blot) under CP treatment, it suggested that there might be another IGF-1R kinase independent mechanism to mediate the phosphorylation of ERK.
There have been several studies demonstrating the novel role of β -arrestins to regulate the G proteincoupled receptors (GPCRs) signaling pathway 39 . We investigated the hypothesis that β -arrestins play a role in IGF-1R-independent ERK activation. As shown in Fig. 3b, knockdown of either β -arrestin 1 or β -arrestin 2 did not interfere the IGF-1 induced ERK phosphorylation, suggesting different mechanisms of downstream pathways activation between IGF-1R and GPCRs. Intriguingly, knockdown of β -arrestin 1 moderately decreased CP-induced ERK phosphorylation while knockdown of β -arrestin 2 almost completely inhibited the CP-stimulated activation of ERK (Fig. 3b). Our results suggested that CP could induce IGF-1R phosphorylation independent activation of ERK, and this process is mediated through β -arrestins, mainly β -arrestin 2.
Due to the surprising ERK activation from CP treatment, we suspected that it might compromise the therapeutic effect of CP to target NSCLC. We then tried to use U0126, a MEK inhibitor, to block ERK activation. U0126 was added 60 min prior to CP treatment in A549 and SPC-A-1 cell lines. As shown in Fig. 3c, U0126 treatment could further inhibit the proliferation of CP-treated NSCLC cells, with higher U0126 having more significant inhibition effect. These results suggested that accidental ERK activation might be one of the main side effects of CP, and combing CP with ERK inhibitors could potentially enhance the efficacy of CP therapy to target NSCLC.

Anti-proliferation effect of metformin alone and in combination with CP in NSCLC cell lines.
Having shown that CP could be potentially used as a drug to target NSCLC and its side effect of accidental ERK activation, we then tried to investigate if CP could coordinate with metformin to treat NSCLC. Firstly, we investigated the inhibition effects of metformin alone on a panel of NSCLC cell lines. As shown in Fig. 4a, cell viability decreased in all NSCLC cell lines that were tested, ranging from 50.6% in SK-MES-1 to 90.5% in SPC-A-1 cell lines. In addition, in most cell lines IGF-1 could partially rescue the growth inhibitory effects of metformin, implying that simultaneously inhibiting IGF-1R might further improve the anti-tumor effect of metformin. We further tested the effects of metformin on the IGF-1R signaling pathway. As demonstrated in Fig. 4b, metformin could significantly attenuate the IGF-1 induced activation of IGF-1R, AKT and ERK in a panel of NSCLC cell lines, which could at least partially explain the antitumor effects of metformin in Fig. 4a.   Having demonstrated that both metformin and CP target IGF-1R signaling pathway, the next question is whether combining CP and metformin could have additive effects on NSCLC cell lines. Consistent with previous results, CP could effectively inhibit cell proliferation in NSCLC cell lines and CP is more effective in SFM condition than serum-added conditions (Fig. 4c). In addition, the combination of metformin with CP could further decrease the cell survival, with cell viability ranging from 50.6% in SK-MES-1 to 71.5% in SPC-A-1 under serum added conditions while ranging from 34.8% in H1975 to 62.3% in H520 under the SFM conditions. Our result suggested that CP and metformin have additive therapeutic effects to target NSCLC cell lines. Even though both CP and metformin target IGF-1R signaling pathway, the combination might further inhibit IGF-1R activation, and/or either metformin or CP have additional anti-tumor effects that are not overlapped with each other.

Metformin could promote the CP-induced IGF-1R down-regulation.
To further elucidate the additive therapeutic effects of CP and metformin, we tried to check the IGF-1R expression level after drug treatment in NSCLC cell lines. As shown in Fig. 5a, a time-dependent down-regulation of IGF-1R was seen in all the tested cell lines, which was also seen under CP treatment (Fig. 2a). We then assessed the receptor down-regulation after combination metformin and CP treatment. As demonstrated in Fig. 5b, CP plus metformin could further down-regulate IGF-1R compared with CP only group. Therefore, the additive IGF-1R down-regulation could at least partially explain the additive therapeutic effect of combining CP and metformin to target NSCLC.
IGF -1R and β-arrestin 2 are highly expressed in NSCLC tumor tissues. Based on the above results, we recognized the importance of IGF -1R and the corresponding signaling pathway in NSCLC. And the role of β -arrestins in receptor down-regulation and signaling transduction implied the potential of β -arrestins as a new target. We tested mRNA levels of IGF -1R and β -arrestin2 in tumor tissues and the corresponding normal tissues of 35 patients from Shandong Provincial Hospital Affiliated to Shandong University by qRT-PCR. The clinical characters of the patients are described in Table 1. As shown in Fig. 6a, the expression of IGF -1R mRNA was higher in NSCLC tumors than the normal tissue (p = 0.024), confirming the critical role of IGF -1R in carcinogenesis. And β -arrestin2 mRNA level was higher in tumor tissues than normal tissues (p < 0.001) (Fig. 6b), suggesting the active of β -arrestin2 in NSCLC.

Discussion
In the current study, we tested the effects of CP on a panel of NSCLC cell lines and all tested cell lines were sensitive to CP treatment in both serum-added and serum-free medium. On one hand, the inhibition of cell proliferation in serum-added medium verified that CP could block IGF-1R signaling by competing with IGF-1. On the other hand, the proliferation inhibition effects observed in serum-free medium implied that there might have IGF-1R phosphorylation independent mechanisms that account for the therapeutic effect of CP.
We verified IGF-1R down-regulation effect of CP in NSCLC cells and our data showed the roles of β -arrestins in IGF-1 and CP induced IGF-1R down-regulation are different. β -arrestins serves as E3 ligase adaptors to mediate ubiquitylation of 7TMRs and thus scaffold receptor endocytosis and traffick intracellular itinerary of the receptors, and this pattern was also observed in IGF-1R that β-arrestin1 participated in IGF-1R ubiquitylation and down-regulation 39,40 . The exact role of β -arrestin1 and β -arrestin2 seems to be different depending on different cell types and receptors 34 . Also, there's study illustrating that there are different ubiquitylation sites on β -arrestin2 which would govern its interaction with different receptors and subsequent internalization and downstream signaling 41 .
We investigated the effects of CP on IGF-1R signaling pathway and unexpectedly found CP could provoke ERK activation without IGF-1R phosphorylation, which might complain some of the side effects observed in CP clinical trials. Furthermore, we found that CP-induced ERK activation was mediated through β -arrestin2. Based on above, we would like to consider CP as a β -arrestin-biased agonist to IGF-1R. β -arrestins have been shown to scaffold with c-Src and lead to subsequent ERK1/2 activation of in β 2AR 42 , and massive in vitro and in vivo data support β -arrestins as a signaling transducer 34,[43][44][45] . Moreover, the β -arrestins-dependent process was reported to be temporally slower onset, which is consistent with our data that CP-induced p-ERK was both weaker and slower than that of IGF-1. Signaling of the two biased arms is pharmacologically distinct, that is, one biased signaling pathway may translate into favorable physiological effects whereas the other appears to result in unbeneficial results 46,47 . This notion highlights the potential to improve therapeutic outcomes by preferring or avoiding specific signaling arms. When it comes to CP-induced signal activation in NSCLC, the β -arrestin2-biased ERK signaling could be inhibited by U0126, suggesting the potential of controlling the biased signaling to enhance therapeutic effect of CP in the future. A related experiment was done by Leong et al., who demonstrated that the inhibition of MEK signaling with U0126 might help to overcome IGF-1R/IR resistance in vitro and in vivo in colorectal cancer 48 . Although several studies have demonstrated that β -arrestins served as scaffolding proteins to regulate specific components of the MAPK cascade, their exact molecular functions vary depending on specific agonists, receptors and cell types 49,50 . The details of the CP-induced ERK signaling need to be further investigated.
We then assessed the anti-proliferation effects of metformin on a panel of NSCLC cell lines, and our results suggested that the anti-tumor effect of metformin was at least partially though the inhibition of IGF-1R signaling pathway. We also found that metformin could down-regulate IGF-1R and attenuate IGF-1R phosphorylation and the downstream PI3K/AKT and Ras/Raf/MEK/ERK pathway activation after IGF-1 stimulation, which further echoed our hypothesis. Our finding is agreed with previous studies that treatment with metformin inhibit cell viability via inhibition of AKT signaling in glioblastoma 51 . Meanwhile, there are studies demonstrating the effect of metformin on ERK signaling pathway 33,52 . Because recent studies on metformin mainly focus on AMPK signaling pathway 25,31 , our findings help to strengthen the role of metformin on IGF-1R signaling pathway, making metformin a promising candidate in anti-cancer therapy. Actually, there have been studies reporting the use of metformin to improve the clinic prognosis in patients who underwent regular chemotherapy with esophageal adenocarcinoma, rectal cancer and other cancers 53,54 . Some other studies also investigated the synergistic anti-cancer effects of metformin when combining with BRAF or other tyrosine kinase inhibitors in cell lines and animal models 15,55 . Herein, we tested the combination of metformin with CP treatment and we observed an additive anti-cancer effects when targeting NSCLC cell lines, which we think is partially mediated through IGF-1R down-regulation and signaling attenuation.

Conclusion
Although CP can block IGF-1R downstream PI3K/AKT signaling pathway through ligand competition, CP could also provoke an IGF-1R phosphorylation independent ERK activation mediated via β -arrestin2, which may probably influence its therapeutic effect. CP can down-regulate IGF-1R, which would contribute to the therapeutic benefits. The adding of U0126 and metformin may enhance the therapeutic effect of CP to target NSCLC.
Cell Culture. A549, PC-9, H520, SPC-A-1, NCI-H1975 and SK-MES-1 cell lines were purchased from ATCC, SK-MES-1 was cultured in EMEM medium, the other cell lines were cultured in RIPM H1640 medium supplemented with 10% fetal bovine serum (FBS), and grown in an atmosphere of 5% CO 2 /95% humidified air at 37 °C. Cell culture media was from Thermo Scientific Hyclone and FBS was obtained from Life Technologies Gibco.
Western Blot Analysis. Protein samples were dissolved in lithium dodecyl sulfate (LDS) sample buffer (Invitrogen) and equal amounts of samples were separated by SDS/PAGE. We have described the detail protocol previously 56 .
Densitometry Analysis. Band intensity was measured by ImageJ 1.63 (US National Institutes of Health, Bethesda, Maryland, USA,) and displayed relative to band intensity of the stated loading control.
Cell Viability Assay/Cell proliferation assay. Cell proliferation was measured with MTT. Cells were incubated in 96-well tissue culture plates, and after appropriate treatment time, MTT was added to a final concentration of 1 mg/mL, and the reaction mixture was incubated for 4 hours at 37 °C. The resulting crystals were dissolved in DMSO and the absorbance was read at 560 nm. Each experiment was repeated thrice, with cells without any stimulation as controls.
Patient tissues and qRT-PCR. The frozen tissues were from patients with primary NSCLC who had undergone a surgical resection in Shandong Provincial Hospital. This study was reviewed and approved by the Ethical Committee of Shandong Provincial Hospital, and written informed consent was given by participants. All experiments were performed in accordance with relevant guidelines and regulations. Total RNA was obtained from NSCLC frozen tumor tissues and the corresponding pare-neoplastic normal tissues, pericarcinous tissues of lung cancer and cell lines using TRIzol reagent (Invitrogen), according to the manufacturer's instructions. RNA concentration was determined spectrophoto metrically, and integrity was checked by gel electrophoresis. mRNA was converted into cDNA using PrimeScript RT reagent Kit with gDNA Eraser (Takara, Japan). Afterwards, each cDNA was amplified with SYBR Premix Ex Taq ( TliRNaseH Plus, Takara, Japan) and analyzed by Roche LightCycler 480II.
Statistical Analysis. Statistical analysis was done using the Student t test or two-way ANOVA to correct for multiple comparisons, as appropriated. All statistical tests were two-sided and were done using SPSS software 19.0 (SPSS Inc., Chicago, IL).