Cellular Signalling 23 (2011) 1869–1875

Contents lists available at ScienceDirect

Cellular Signalling

j ourna l homepage: www.e lsev ie r.com/ locate /ce l l s ig

IGF-1 induced HIF-1α-TLR9 cross talk regulates inflammatory responses in glioma

Sanchari Sinha, Nitin Koul, Deobrat Dixit, Vivek Sharma, Ellora Sen ⁎National Brain Research Centre, Manesar, Haryana 122 050, India

⁎ Corresponding author. Tel.: +91 124 2338921x235E-mail addresses: sancharisinha79@gmail.com (S. Si

(N. Koul), dev25@nbrc.ac.in (D. Dixit), viveksharmabt@ellora@nbrc.ac.in (E. Sen).

0898-6568/$ – see front matter © 2011 Elsevier Inc. Aldoi:10.1016/j.cellsig.2011.06.024

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 June 2011Accepted 27 June 2011Available online 3 July 2011

Keywords:IGF-1HIF-1αTLR9Glioblastoma

The insulin-like growth factor (IGF-1) induces hypoxia inducible factor (HIF-1α) regulated genes inglioblastoma multiforme (GBM). As HIF-1α links inflammatory and oncogenic pathways in GBM, weinvestigated whether IGF-1 affects HIF-1α to regulate inflammatory response in glioma cells under normoxia.IGF-1 induced Ras and Calmodulin-dependent kinase II (CaMKII) regulated HIF-1α transcriptional activity inglioma cells. Increase in HIF-1α was concurrent with decreased Toll-like receptor (TLR9) and CXCR4expression and elevated suppressor of cytokine signaling (SOCS3) levels. Interestingly, while synthetic CpGcontaining oligodeoxynucleotide TLR9 agonist (CpG DNA) decreased IGF-1 mediated increase in HIF-1αactivity, siRNA mediated knockdown of HIF-1α decreased TLR9 levels. This suggested that IGF-1 induced HIF-1α-TLR9 axis is regulated by both positive and negative feedback loops. Importantly, TLR9 agonist reversedthe effect of IGF-1 on CXCR4 and SOCS3 expression. While knockdown of HIF-1α abrogated IGF-1 mediatedincrease in SOCS3 it elevated IGF-1 induced decrease in CXCR4 levels. Thus HIF-1α positively and negativelyregulates SOCS3 and CXCR4 expression respectively, in glioma cells. Though TLR9 agonist had no additiveeffect on IGF-1 mediated increase in pro-inflammatory cytokines IL-1β, IL-6 and IL-8, treatment with TLR9agonist alone elevated expression of these pro-inflammatory cytokines. Our studies indicate that a complexHIF-1α-TLR9 cross-talk sustains a self-regulating cycle of inflammatory response through intrinsic negativeand positive feedback mechanisms.

; fax: +91 124 2338910/28.nha), nitinkoul11@gmail.comgmail.com (V. Sharma),

l rights reserved.

© 2011 Elsevier Inc. All rights reserved.

1. Introduction

Signaling pathways emanating from the insulin-like growthfactor-I receptor (IGF-IR) — a transmembrane receptor tyrosinekinase activated by IGF, is associated with the growth and prolifer-ation of several tumor types including glioblastoma [1–3]. AlthoughHIF-1α is a key regulator of cellular response to hypoxia [4] it can alsobe activated under normoxia in response to IGF [5,6]. HIF-1αinfluences tumor growth [7] by regulating expression of genesimplicated in angiogenesis, metabolism and cell survival [8]. IGF-1 isknown to induce HIF-1α accumulation in Kaposi Sarcoma [9] andneuroblastoma [10].

HIF-1α-regulated genes are induced by IGF-1 in glioblastoma [11].Suppression of IGF-1R inhibits tumorigenic potential of rat glioblas-toma cells [12]. We have demonstrated IL-1β induced Ras/Akt/Erkmediated HIF-1α activation in glioma cells [13]. Activated Ras elevatesIGF-I mediated induction of the HIF-1α target VEGF [14]. OncogenicRas activation occurs in GBMs [15] and inhibition of Ras down-regulates HIF-1α activity in GBM [16]. CaMKII which translatesintracellular changes in calcium is associated with glioma migration

[17], and regulates Ras mediated Erk activation [18]. CaMKII regulatesHIF-1α transcriptional activity under intermittent hypoxia [19] aswell as under inflammatory conditions [20].

By demonstrating that Ras rewires HIF1α-driven IL-1β autocrineloop [13,21], we have emphasized the pivotal role of HIF-1α in linkinginflammation and tumorigenesis [22]. HIF-1α is involved in hostimmune response during bacterial infection [23], and LPS is a potentinducer of HIF-1α [24]. Also, HIF-1α accumulation and target geneexpression are impaired upon induction of endotoxin tolerance [25].Activation of toll like receptors (TLRs) which recognize pathogen-associated molecular patterns triggers signaling events that initiateinnate immunity and inflammatory response [26]. Interestingly, HIF-1αregulates hypoxia induced TLR4 expression inmacrophages [27]. WhileTLR9 increases metastatic potential of cancer cells through CXCR4expression [28], the later modulates TLR9 mediated signaling [29]. TLRactivation is modulated by negative regulators such as suppressor ofcytokine signaling (SOCS) that feed backupon and inhibit TLR activation[30]. Moreover, SOCS3 which is constitutively expressed in GBM is notonly involved in inducing radioresistance [31], but it also regulatesCXCR4 function [32]. Given that TLR9 is expressed in GBM [33], weinvestigated whether IGF-1 affect HIF-1α to modulate TLR9 mediatedsignaling in glioma.

Our study to understand the involvement of IGF-1 in HIF-1αregulated inflammatory responses in GBM indicates that IGF-1 inducedRas/CaMKII enhances HIF-1α activation, under normoxia. Increase in

Fig. 1. IGF-1 induces HIF-1α activity in glioma cells in a Ras dependent manner (a)Western blot demonstrates a dose dependent increase in HIF-1α expression in glioma cells treatedwith IGF-1 for 24 h. The figure is a representative of three independent experiments. Blots were reprobed for C23 to establish equivalent loading. (b) Treatment with IGF-1 increasesHIF-1α transcriptional activity in glioma cells. Following transfection of glioma cells with HIF-1α reporter constructs, cells were treated with different concentration of IGF-1 for 24 hand reporter assay was performed to determine HIF-1α activity. The graph represents fold change in activity over control. (c) IGF-1 increases Ras activity in glioma cells. Ras activityin IGF-1 treated glioma cells was determined by the ability of Ras-GTP to bind to a specific protein domain of Raf in the form of a GST-fusion protein. An increase in Ras activity wasobserved in cells treated with IGF-1 for 30 mins. The figure is representative from three independent experiments with similar results. (d) DN-RasN17 inhibits the ability of IGF-1 toinduce HIF-1α activity. Cells co-transfected with RasN17 and HIF-1α luciferase reporter constructs were treated with IGF-1. After 24 h luciferase reporter assay was performed todetermine HIF-1α activity. The graph represents fold change in activity over control. Values in b and d represent the means±SEM from 3 independent experiments. * Significantincrease from control, # Significant decrease from IGF-1 treated cells (Pb0.05).

1870 S. Sinha et al. / Cellular Signalling 23 (2011) 1869–1875

IGF-1 induced HIF-1α is concurrent with decrease TLR9 and CXCR4levels and elevation in SOCS3 expression. Interestingly, IGF-1 inducesbothnegative andpositiveHIF-1α-TLR9 loops in glioma. IGF-1mediatedchanges in SOCS3 and CXCR4 is dependent on both HIF-1α and TLR9.This work forges the first link between IGF-1 induced HIF-1α inregulating TLR9 dependent inflammatory responses in glioma.

2. Materials and methods

2.1. Cell culture

Glioblastoma cell lines U87MG [13] and T98G [13] obtained fromAmerican Type Culture Collection (ATCC) were cultured in DMEMsupplemented with 10% FBS. On attaining semi-confluence, cells wereswitched to serum free media (SFM) and after 6 h, cells were treatedwith different concentration of IGF-1 in SFM. All reagents werepurchased from Sigma unless otherwise stated. The HIF-1α luciferasereporter was a gift by Chinmay Mukhopadhyay (JNU, India) and hasbeen described [34]. DN-RasN17 was purchased from Clontech.

2.2. Western blot analysis

Whole cell lysates and nuclear extract were prepared from cellstreated with or without IGF-1, in the presence or absence of TLR9agonist CpG ODN 2006 (Imgenex) or CaMKII inhibitor KN93 asdescribed previously [13], and protein concentration was determinedby the BCA method. Lysates were electrophoresed on 7% to 12%polyacrylamide gel and Western blot analysis was performed asdescribed [13] using the following antibodies — HIF-1α (ΒD bio-sciences), C23 (Santa Cruz), CaMKII, CXCR4, SOCS3 and TLR9(Imgenex). Antibodies were purchased from Cell Signaling unlessotherwise mentioned.

2.3. Measurement of Ras activity

The Ras activity was performed using a commercially available Rasactivation assay kit purchased fromUpstate Biotechnology, as describedpreviously [13].

2.4. Luciferase assay

Transfection and reporter assay was performed as described [13].Briefly, cells at ~60–70% confluence in 24-well plates were transientlytransfected with 0.3 μg of HIF-1α and 10 ng of the Renilla luciferaseexpression vector pRL-TK as a transfection control using Lipofecta-mine 2000 (Life Technologies-Invitrogen). Briefly, 18 h after trans-fection, cells were serum starved for 4 h followed by treatment with20 ng/ml IGF-1 for 24 h. Luciferase activity was measured using Dualluciferase assay kit according to the manufacturer's protocol (Pro-mega) with GloMax 96 microplate luminometer. The results areexpressed as fold change in activity over control. HIF-1α transcrip-tional activity in the presence of dominant negative Ras constructswere performed by co-transfecting cells with 0.3 μg of RasN17 andHIF-1α reporter constructs. For co-transfection experiments involvingDN construct, control transfection using the appropriate emptyvectors for each construct was used as described [13].

2.5. Small interfering RNA (siRNA) transfection

Eighteen hours prior to transfection, 3 x104 cells were seeded onto24-well plates in medium without antibiotics and transfection with50 nmol/l duplex HIF-1α siRNA (Thermo Fischer Scientific) was carriedusing Lipofectamine 2000 (Life Technologies-Invitrogen) as described[13]. Non-specific siRNA was purchased from Proligo (Singapore).

2.6. Cytokine bead array

Cytometric bead array kit (CBA kit; BD Biosciences) was used toquantitatively measure cytokine levels in the supernatant collectedfrom control and glioma cells treated with different combinations ofIGF-1 and TLR9 agonist CpG ODN 2006 (Imgenex). The assay wasperformed and analyzed on FACS Calibur (Becton Dickinson) asdescribed previously [35].

2.7. Statistical analysis

All comparisons between groups were performed using two-tailedPaired student's t-test. A value of Pb0.05 was considered significant.

1871S. Sinha et al. / Cellular Signalling 23 (2011) 1869–1875

3. Results

3.1. IGF-1 induces HIF-1α expression and transcriptional activity inglioma cells

As IGF-1 induces HIF-1α-regulated genes in glioblastoma [11], wedetermined whether IGF-1 effects HIF-1α expression in glioma. IGF-1elevated HIF-1α expression in glioma cells in a dose dependentmanner as compared to control (Fig. 1a). While treatment with 5 and10 ng/ml of IGF-1 slightly elevated HIF-1α expression, maximalincrease was observed upon treatment with 20 ng/ml of IGF-1 for24 h (Fig. 1a). A significant ~2.3-fold increase in HIF-1α transcrip-tional activity over control was observed in glioma cells treated with10 and 20 ng/ml of IGF-1 for 24 h (Fig. 1b). As increase in HIF-1αactivation was maximal in cells treated with 20 ng/ml of IGF-1 allsubsequent experiments were performed with this dose of IGF-1.

3.2. IGF-1 mediated increase in Ras activity regulates HIF-1α activity

IGF signaling promotes cellular proliferation by stimulatingRas/ERK pathway necessary for tumor growth [6]. Since Ras activationoccurs in GBM [15], we determined whether IGF-1 regulates Ras

Fig. 2. CaMKII regulate IGF-1 mediated HIF-1α activity in a Ras independent manner.(a) IGF-1 increases CaMKII phosphorylation in glioma cells as determined by Westernblot analysis. The figure is a representative from three independent experiments withsimilar results. (b) IGF-1 mediated increase in CaMKII expression is unaltered in cellstransfected with DN-RasN17. Mock and DN-RasN17 transfected cells were treated withIGF-1 for 24 h and CaMKII expression was determined by Western blotting.Representative blot is shown from three independent experiments with identicalresults. Blots were reprobed for β-actin to establish equivalent loading. (c) IGF-1induced HIF-1α activity is dependent on CaMKII activation. Cells transfected with HIF-1α reporter construct were treated with IGF-1 in the presence or absence of 10 μM ofCaMKII inhibitor KN93. After 24 h luciferase reporter assay was performed to determineHIF-1α activity. The graph represents fold change in activity over control. Valuesrepresent the means±SEM from 3 independent experiments. * Significant increasefrom control, # Significant decrease from IGF-1 treated cells (Pb0.05).

activity in glioma cells. IGF-1 elevated Ras activity in glioma cells(Fig. 1c). We have demonstrated the existence of an IL-1β-Ras- HIF-1αaxis in GBM [13]. To investigate whether elevated Ras plays a rolein IGF-1 induced HIF-1α activation, we determined HIF-1α transcrip-tional activity in cells transfected with DN-RasN17 in the presenceand absence of IGF-1. Increase in HIF-1α activation induced by IGF-1was abrogated in cells transfected with DN-RasN17 (Fig. 1d). Thissuggested the involvement of IGF-1 induced Ras in HIF-1α activation.

3.3. IGF-1 induced CaMKII regulates HIF-1α activity independent of Ras

CaMKII regulates HIF-1α transcriptional activity [19] and plays arole in glioma cell migration [17]. As IGF-1 regulates calcium channelsthrough CaMKII signaling [36], we investigated whether CaMKIIregulates HIF-1α transcriptional activity in IGF-1 treated cells. Anincrease in CaMKII phosphorylation was observed upon IGF-1treatment (Fig. 2a). As relation between Ras and CaMKII has beenreported [18] we investigated whether Ras contributes to increasedCaMKII expression in IGF-1 treated cells. The increase in CaMKII is Rasindependent as its expression was not affected in cells transfectedwith DN-Ras (N17) (Fig. 2b). Importantly, IGF-1 induced HIF-1αtranscriptional activity was abrogated in cells treated with CaMKIIinhibitor KN93 (Fig. 2c). Thus, IGF-1 induced Ras and CaMKIIindependently regulate HIF-1α activation.

Fig. 3. Existence of TLR9-CaMKII axis in IGF-1 treated glioma cells. (a) IGF-1 decreasesTLR9 expression in glioma cells. Western blot demonstrates TLR9 expression in gliomacells treated with 20 ng/ml of IGF-1 for 24 h. (b) CaMKII has no effect on TLR9expression in IGF-1 treated glioma cells. Cells were treated with IGF-1 in the presenceand absence of CaMKII inhibitor KN93 andWestern blot was performed to demonstrateTLR9 levels. (c) TLR9 agonist regulates CaMKII expression in glioma cells. Cells weretreated with IGF-1 in the presence and absence of TLR9 agonist CpG ODN 2006(5 μg/ml) and Western blot was performed to demonstrate CaMKII levels. Panels (a, b,c) are representative blots from at least three independent experiments with similarresults. Blots were reprobed for β-actin to establish equivalent loading.

1872 S. Sinha et al. / Cellular Signalling 23 (2011) 1869–1875

3.4. Decreased TLR9 levels in IGF-1 treated cells regulates CaMKII

As TLR9 signaling activates CaMKII [37] and TLR9 is expressed inGBM [33], we determined TLR9 levels in IGF-1 treated cells withelevated CaMKII levels. A decrease in TLR9 expressionwas observed inIGF-1 treated cells (Fig. 3a). As inhibition of CaMKII significantlysuppresses TLR9 induced inflammatory responses [37], we nextdetermined the co-relation between CaMKII and TLR9 in IGF-1 treatedcells. Inhibition of CaMKII has no effect on TLR9 expression, as it levelin control and IGF-1 treated cells were comparable (Fig. 3b).Treatment with TLR9 agonist inhibited IGF-1 mediated increase inCaMKII expression (Fig. 3c).

3.5. Existence of negative and positive TLR9-HIF-1α feed-back loop inglioma cells

As decrease in TLR9 expression was concurrent with increasedHIF-1α activation, we investigated whether TLR9 activation couldaffect HIF-1α in IGF-1 treated cells. IGF-1 induced HIF-1α wasabrogated to control levels in the presence of TLR9 agonist (Fig. 4a).Treatment with TLR9 agonist alone also resulted in a significantdecrease in HIF-1α activation (Fig. 4a). As siRNA mediated knock-down of HIF-1α inhibits hypoxia-induced TLR4 expression inmacrophages [27], we next investigated the role of HIF-1α in TLR9induction. Contrary to our expectation, knockdown of HIF-1αabrogated TLR9 level in both control and IGF-1 treated cells

Fig. 4. Complex cross-talk between HIF-1α and TLR9 in glioma cells. (a) TLR9 activationdecreases IGF-1 induced HIF-1α activation in glioma cells. Cells transfected with HIF-1αreporter construct were treated with IGF-1 in the presence or absence of 5 μg/ml ofTLR9 agonist CpG ODN 2006. After 24 h luciferase reporter assay was performed todetermine HIF-1α activity. The graph represents fold change in activity over control.Values represent the means±SEM from 3 independent experiments. * Significantincrease from control, # Significant decrease from IGF-1 treated cells (Pb0.05). (b)siRNA mediated knock-down of HIF-1α reduces TLR9 levels in both control and IGF-1treated cells. Mock and HIF-1α siRNA transfected cells were treated with IGF-1 for 24 hand TLR9 expression was determined by Western blotting. The figure is representativeblot from at least three independent experiments with similar results. Blots werereprobed for β-actin to establish equivalent loading. Inset shows HIF-1α levels in cellstransfected with HIF-1α siRNA.

(Fig. 4b). This finding highlights the complexity of TLR9 signaling inHIF-1α regulation and vice versa, as both negative and positive TLR9-HIF-1α feedback loops exists in glioma cells.

3.6. IGF-1 mediated decrease in CXCR4 is HIF-1α and TLR9 dependent

HIF-1α regulates CXCR4 expression in GBM [38] and CXCR4 isimportant for TLR9 agonist to enhance metastasis [39]. We thereforedetermined CXCR4 expression in IGF-1 treated cells with elevatedHIF-1α and decreased TLR9 levels. A decrease in CXCR4 expressionwas observed upon IGF-1 treatment (Fig. 5a). As hypoxia has beenshown to increase CXCR4 in glioma cells [38], we determined thewhether HIF-1α regulates CXCR4 expression in these cells. Knock-down of HIF-1α elevated IGF-1 mediated decrease in CXCR4 levels(Fig. 5b). Since decrease in CXCR4 was concurrent with TLR9expression, correlation between the two was also investigated. Thedecrease in CXCR4 levels observed upon IGF-1 treatmentwas revertedto levels comparable to control in the presence of TLR9 agonist(Fig. 5c). Thus, HIF-1α inhibition and TLR9 activation prevents IGF-1mediated decrease in CXCR4 levels.

3.7. Increase in SOCS3 expression in IGF-1 treated cells is HIF-1α andTLR9 dependent

Functional inactivation of CXCR4 mediated response involvesupregulation of SOCS3 [32]. As CXCR4 levels were down-regulated inIGF-1 treated cells, we determined SOCS3 expression in these cells.Treatment with IGF-1 elevated SOCS-3 expression (Fig. 6a). Thoughno direct link between SOCS3 and HIF-1α is known, SOCS3 is inducedby hypoxia in pulmonary arterial smooth muscle cell [40]. Wetherefore investigated the relationship between SOCS3 and HIF-1α bydetermining SOCS3 expression in glioma cells transfected with HIF-

Fig. 5. IGF-1mediated decrease in CXCR4 levels is HIF-1α and TLR9 dependent. (a) IGF-1decrease CXCR4 levels in glioma cells. Western blot demonstrates decrease in CXCR4levels in glioma cells treated with IGF-1 for 24 h. (b) siRNA mediated knock-down ofHIF-1α reverses IGF-1 mediated decrease in CXCR4 expression. Cells transfected withHIF-1α siRNA were treated with IGF-1 for 24 h and CXCR4 expression was determinedby Western blotting. (c) TLR9 activation increases IGF-1 induced decrease in CXCR4expression in glioma cells. Cells were treated with IGF-1 in the presence and absence ofTLR9 agonist for 24 h, andWestern blot was performed to determine CXCR4 expression.The panels (a–c) are representative of three independent experiments. Blots werereprobed with β-actin to establish equivalent loading.

Fig. 6. IGF-1 increases expression of SOCS3 and pro-inflammatory cytokines. (a) IGF-1 increases SOCS3 expression in glioma cells. Cells were treated with IGF-1 for 24 h, andWesternblot was performed to determine SOCS3 expression. (b) IGF-1 increases SOCS3 expression in a HIF-1α dependent manner. siRNA mediated knock-down of HIF-1α attenuates IGF-1induced increase in SOCS3 expression. Mock and HIF-1α siRNA transfected cells were treated with IGF-1 for 24 h and SOCS3 expression was determined by Western blotting. (c)TLR9 activation abrogates IGF-1 induced increase in SOCS3 expression in glioma cells. Cells were treated with IGF-1 in the presence and absence of TLR9 agonist for 24 h, andWesternblot was performed to determine SOCS3 expression. Panels (a, b and c) are representative blots from at least three independent experiments with similar results. Blots were probedwith β-actin to establish equivalent loading. (d) TLR9 activation induces expression of pro-inflammatory cytokines in glioma cells both in the presence and absence of IGF-1. Cellswere treated with IGF-1 in the presence and absence of TLR9 agonist for 24 h, and CBA was performed to determine levels of pro-inflammatory cytokines. Values represent themeans±SEM from 3 individual experiments. * Significant change from control Pb0.05. (e) Proposed model for IGF-1 triggered HIF-1α dependent regulation of inflammatoryresponse in glioma cells.

1873S. Sinha et al. / Cellular Signalling 23 (2011) 1869–1875

1α siRNA. Interestingly, siRNA mediated knock-down of HIF-1αabrogated IGF-1 induced increase in SOCS3 expression (Fig. 6b).Moreover, treatment with TLR9 agonist decreased IGF-1 inducedincrease in SOCS3 expression (Fig. 6c).

3.8. IGF-1 regulates pro-inflammatory response in glioma cells

Sincewehave reported that an IL-1β-HIF-1α autocrine loop sustainsinflammatory response in glioma cells [13], and SOCS3 is known toprotect pancreatic β-cells from IL-1β mediated cytotoxicity [41], weinvestigated the status of IL-1β in IGF-1 treated cells with elevated HIF-1α and SOCS3 expression. A significant increase in IL-1β levels wasobserved upon IGF-1 treatment (Fig. 6d). Treatment with TLR9 agonisteither alone or in presence of IGF-1 elevated IL-1β to levels similar tothat observed upon IGF-1 treatment. Presence of TLR9 agonist had noeffect on the ability of IGF-1 to maintain elevated IL-1β levels (Fig. 6d).Similar trend in the expression of other pro-inflammatory cytokinessuch as IL-6 and IL-8 was observed in glioma cells treated with IGF-1 inthe presence and absence of TLR9 agonist (Fig. 6d). Taken together our

findings indicate that a complex cross-talk between TLR9 and HIF-1αregulates inflammatory responses in IGF-1 treated glioma cells (Fig.6e).

4. Discussion

IGF signaling plays an important role in malignant transformationand protection from apoptosis in a broad range of human cancer[3,42]. Glioma produces IGFs and express elevated levels of IGFreceptors compared with normal brain tissue [43,44]. Besides,inhibition of IGF receptor disrupts a pro-survival and pro-angiogenicIGF-HIF pathway in glioblastoma cells [45]. Our finding suggests thatIGF-1 induced Ras and CaMKII act independently to elevate HIF-1αactivation in glioma cells.

IGF-1 induced increase in HIF-1α activation is concomitant withdecreased TLR9 and CXCR4 levels and elevated SOCS3 expression. Themost important finding of this studywas the identification of a complexcrosstalk between TLR9 and HIF-1α in response to IGF-1, undernormoxia. TLR9 and HIF-1α seem to regulate each other through twoopposite feedback loops.While knockdownofHIF-1αdecreases TLR9bybehaving asmutual regulator of each other in a positive regulatory loop,

1874 S. Sinha et al. / Cellular Signalling 23 (2011) 1869–1875

activation of TLR9 signaling abrogates HIF-1α activation through anegative regulatory loop. This finding highlights the complexity of TLR9signaling in HIF-1α activation and vice versa in glioma cell, as bothnegative and positive TLR9-HIF-1α feedback loops act in tandem toregulate inflammatory response. It is possible that thenegative feedbackoccurs beneath a given threshold of the two proteins and that thepositive loop starts when this threshold is crossed.

As TLR9 signaling regulates both CaMKII and HIF-1α in IGF-1treated cells and since CaMKII triggered by TLR ligands promoteinflammatory responses [37], it is likely that TLR9 act as a sensor tomaintain the inflammatory millieu through regulation of CaMKII andHIF-1α. We have reported that HIF-1α maintains persistently highlevel of IL-1β through an HIF-1α-IL1β autocrine loop in glioma cells[13]. Pro-inflammatory cytokines in the tumor microenvironmentplays a major factor in tumorigenesis [46]. The simultaneous increasein SOCS3 and decrease in TLR9 was concurrent with heightened levelof IL-1β and pro-inflammatory cytokines. It is known that TLR9promotes inflammatory response [37] and SOCS3 prevents IL-1βmediated cytotoxicity [41]. As TLR9 activation decreases IGF-1mediated increase in SOCS3 expression, it is tempting to speculatethat low TLR9 levels maintains elevated SOCS3 expression in IGF-1treated cells to prevent pro-inflammatory cytokines from crossing thethreshold beyond levels required to sustain a TLR9-HIF-1α axis. It ispossible that negative TLR9-HIF-1α axis is initiated to maintainelevated HIF-1α levels in presence of IGF-1. Another importantfinding of this study is the elucidation of the involvement of HIF-1α inthe regulation of SOCS3.

While CXCR4 modulates TLR9 mediated signaling [29], TLR9increasesmetastatic potential of cancer cells through CXCR4 expression[28]. Although TLR9 agonist has no contribution towards sustaining theelevated pro-inflammatory cytokines levels triggered by IGF-1, treat-ment with TLR9 agonist alone elevates pro-inflammatory cytokines tolevels similar to IGF-1 treated cells. TLR9 agonists stimulate innate andadaptive anti-tumor immune responses and have demonstratedpotential for the treatment of cancer [47,48]. This along with the abilityof TLR9 agonist to elevate CXCR4 levelswarrants investigation regardingthe use of TLR9 agonist as an effective anti-glioma target. Thoughpharmacologic inhibition of HIF-1α or the SDF-1/CXCR4 interactionabrogates regrowth of GBM [49], inhibition of HIF-1α reversed IGF-1mediated decrease in CXCR4 expression. These findings suggested thatHIF-1α regulates CXCR4 in different ways depending on the context ofmicroenvironment.

5. Conclusion

Our findings have not only highlighted (i) a previously unrecog-nized function of HIF-1α as an important regulator of SOCS3 and TLR9expression, but has also (ii) established HIF-1α as a link between twoapparently unrelated but crucial component of glioma tumormicroenvironment- growth factor and inflammation. Taken together,our data underline the complexity of HIF-1α in its crosstalk with TLR9under normoxia. This study prompts further investigation intomechanisms governing the dialog between HIF-1α and TLR9 to revealnew molecular components that participates in regulating inflamma-tory responses in glioma.

Conflict of interest

There are no competing financial interests in relation to the workand we have nothing to disclose.

Acknowledgments

The work was supported by a research grant from the Department ofBiotechnology (DBT)Government of India (BT/PR12924/Med/30/235/09)

and partly by the Innovative Young Biotechnologist Award (IYBA) by DBTto ES.

References

[1] H.N. Antoniades, T. Galanopoulos, J. Neville-Golden, M. Maxwell, Int. J. Cancer50 (2) (1992) 215–222.

[2] D. LeRoith, C.T. Roberts Jr., Cancer Lett. 195 (2) (2003) 127–137.[3] M.N. Pollak, E.S. Schernhammer, S.E. Hankinson, Nat. Rev. Cancer 4 (7) (2004)

505–518.[4] N.V. Iyer, L.E. Kotch, F. Agani, S.W. Leung, E. Laughner, R.H. Wenger, M. Gassmann,

J.D. Gearhart, A.M. Lawler, A.Y. Yu, G.L. Semenza, Genes Dev. 12 (2) (1998)149–162.

[5] R. Fukuda, K. Hirota, F. Fan, Y.D. Jung, L.M. Ellis, G.L. Semenza, J. Biol. Chem.277 (41) (2002) 38205–38211.

[6] K.M. Sutton, S. Hayat, N.M. Chau, S. Cook, J. Pouyssegur, A. Ahmed, N. Perusinghe,R. Le Floch, J. Yang, M. Ashcroft, Oncogene 26 (27) (2007) 3920–3929.

[7] P.H. Maxwell, G.U. Dachs, J.M. Gleadle, L.G. Nicholls, A.L. Harris, I.J. Stratford, O.Hankinson, C.W. Pugh, P.J. Ratcliffe, Proc. Natl. Acad. Sci. U S A 94 (15) (1997)8104–8109.

[8] G.L. Semenza, Trends Mol. Med. 8 (4 Suppl.) (2002) S62–S67.[9] S.B. Catrina, I.R. Botusan, A. Rantanen, A.I. Catrina, P. Pyakurel, O. Savu, M. Axelson,

P. Biberfeld, L. Poellinger, K. Brismar, Clin. Cancer Res. 12 (15) (2006) 4506–4514.[10] K. Beppu, K. Nakamura, W.M. Linehan, A. Rapisarda, C.J. Thiele, Cancer Res. 65 (11)

(2005) 4775–4781.[11] W. Zundel, C. Schindler, D. Haas-Kogan, A. Koong, F. Kaper, E. Chen, A.R.

Gottschalk, H.E. Ryan, R.S. Johnson, A.B. Jefferson, D. Stokoe, A.J. Giaccia, GenesDev. 14 (4) (2000) 391–396.

[12] F. Rininsland, T.R. Johnson, C.L. Chernicky, E. Schulze, P. Burfeind, J. Ilan, Proc. Natl.Acad. Sci. U S A 94 (11) (1997) 5854–5859.

[13] V. Sharma, D. Dixit, N. Koul, V.S. Mehta, Sen. E. J. Mol. Med. 89 (2) (2011) 123–136.[14] M. Stearns, J. Tran, M.K. Francis, H. Zhang, C. Sell, Cancer Res. 65 (6) (2005)

2085–2088.[15] A. Guha, N. Lau, I. Huvar, D. Gutmann, J. Provias, T. Pawson, G. Boss, Oncogene

12 (3) (1996) 507–513.[16] R. Blum, J. Jacob-Hirsch, N. Amariglio, G. Rechavi, Y. Kloog, Cancer Res. 65 (3)

(2005) 999–1006.[17] V.A. Cuddapah, H. Sontheimer, J. Biol. Chem. 285 (15) (2010) 11188–11196.[18] M. Illario, A.L. Cavallo, K.U. Bayer, T. Di Matola, G. Fenzi, G. Rossi, M. Vitale, J. Biol.

Chem. 278 (46) (2003) 45101–45108.[19] G. Yuan, J. Nanduri, C.R. Bhasker, G.L. Semenza, N.R. Prabhakar, J. Biol. Chem.

280 (6) (2005) 4321–4328.[20] J. Westra, E. Brouwer, I.A. van Roosmalen, B. Doornbos-van der Meer, M.A. van

Leeuwen,M.D. Posthumus, C.G. Kallenberg, BMCMusculoskeletDisord11 (2010) 61.[21] S. Kaluz, E.G. Van Meir, J. Mol. Med. 89 (2) (2011) 91–94.[22] Y.J. Jung, J.S. Isaacs, S. Lee, J. Trepel, L. Neckers, FASEB J. 17 (14) (2003) 2115–2117.[23] T. Cramer, Y. Yamanishi, B.E. Clausen, I. Forster, R. Pawlinski, N. Mackman, V.H.

Haase, R. Jaenisch, M. Corr, V. Nizet, G.S. Firestein, H.P. Gerber, N. Ferrara, R.S.Johnson, Cell 112 (5) (2003) 645–657.

[24] S. Frede, C. Stockmann, P. Freitag, J. Fandrey, Biochem. J. 396 (3) (2006) 517–527.[25] S. Frede, C. Stockmann, S. Winning, P. Freitag, J. Fandrey, J. Immunol. 182 (10)

(2009) 6470–6476.[26] S. Akira, K. Takeda, Nat. Rev. Immunol. 4 (7) (2004) 499–511.[27] S.Y. Kim, Y.J. Choi, S.M. Joung, B.H. Lee, Y.S. Jung, J.Y. Lee, Immunology 129 (4)

(2010) 516–524.[28] T. Ren, Z.K.Wen, Z.M. Liu, Y.J. Liang, Z.L. Guo, L. Xu, Cancer Biol. Ther. 6 (11) (2007)

1704–1709.[29] M. Ishii, C.M. Hogaboam, A. Joshi, T. Ito, D.J. Fong, S.L. Kunkel, Eur. J. Immunol.

38 (8) (2008) 2290–2302.[30] F.Y. Liew, D. Xu, E.K. Brint, L.A. O'Neill, Nat. Rev. Immunol. 5 (6) (2005) 446–458.[31] H. Zhou, R. Miki, M. Eeva, F.M. Fike, D. Seligson, L. Yang, A. Yoshimura, M.A. Teitell,

C.A. Jamieson, N.A. Cacalano, Clin. Cancer Res. 13 (8) (2007) 2344–2353.[32] S.F. Soriano, P. Hernanz-Falcon, J.M. Rodriguez-Frade, A.M. De Ana, R. Garzon, C.

Carvalho-Pinto, A.J. Vila-Coro, A. Zaballos, D. Balomenos, A.C. Martinez, M.Mellado, J. Exp. Med. 196 (3) (2002) 311–321.

[33] Y. Meng, M. Kujas, Y. Marie, S. Paris, J. Thillet, J.Y. Delattre, A.F. Carpentier, J.Neurooncol. 88 (1) (2008) 19–25.

[34] S. Biswas, M.K. Gupta, D. Chattopadhyay, C.K. Mukhopadhyay, Am. J. Physiol. HeartCirc. Physiol. 292 (2) (2007) H758–766.

[35] V. Sharma, C. Joseph, S. Ghosh, A. Agarwal, M.K. Mishra, E. Sen, Mol. Cancer Ther.6 (9) (2007) 2544–2553.

[36] L. Gao, L.A. Blair, G.D. Salinas, L.A. Needleman, J. Marshall, J. Neurosci. 26 (23)(2006) 6259–6268.

[37] X. Liu, M. Yao, N. Li, C. Wang, Y. Zheng, X. Cao, Blood 112 (13) (2008) 4961–4970.[38] D. Zagzag, Y. Lukyanov, L. Lan, M.A. Ali, M. Esencay, O. Mendez, H. Yee, E.B. Voura,

E.W. Newcomb, Lab. Invest. 86 (12) (2006) 1221–1232.[39] L. Xu, Y. Zhou, Q. Liu, J.M. Luo, M. Qing, X.Y. Tang, X.S. Yao, C.H. Wang, Z.K. Wen,

Biochem. Biophys. Res. Commun. 382 (3) (2009) 571–576.[40] L. Bai, Z. Yu, G. Qian, P. Qian, J. Jiang, G. Wang, C. Bai, Respir. Physiol. Neurobiol.

152 (1) (2006) 83–91.[41] A.E. Karlsen, S.G. Ronn, K. Lindberg, J. Johannesen, E.D. Galsgaard, F. Pociot, J.H.

Nielsen, T. Mandrup-Poulsen, J. Nerup, N. Billestrup, Proc. Natl. Acad. Sci. U S A98 (21) (2001) 12191–12196.

[42] E. Foulstone, S. Prince, O. Zaccheo, J.L. Burns, J. Harper, C. Jacobs, D. Church, A.B.Hassan, J. Pathol. 205 (2) (2005) 145–153.

1875S. Sinha et al. / Cellular Signalling 23 (2011) 1869–1875

[43] R.P. Glick, T.G. Unterman, M. Van der Woude, L.Z. Blaydes, J. Neurosurg. 77 (3)(1992) 445–450.

[44] A.C. Sandberg-Nordqvist, P.A. Stahlbom, M. Reinecke, V.P. Collins, H. von Holst, V.Sara, Cancer Res. 53 (11) (1993) 2475–2478.

[45] M.B. Gariboldi, R. Ravizza, E. Monti, Biochem. Pharmacol. 80 (4) (2010) 455–462.

[46] L.M. Coussens, Z. Werb, Nature 420 (6917) (2002) 860–867.[47] A.M. Krieg, J. Clin. Invest. 117 (5) (2007) 1184–1194.[48] A.M. Krieg, Oncogene 27 (2) (2008) 161–167.[49] M. Kioi, H. Vogel, G. Schultz, R.M. Hoffman, G.R. Harsh, J.M. Brown, J. Clin. Invest.

120 (3) (2010) 694–705.