Nanofiber-mediated inhibition of focal adhesion ... - Oxford Journals

Neuro-Oncology 15(3):319 – 329, 2013. doi:10.1093/neuonc/nos316 Advance Access publication January 17, 2013

N E U RO - O N CO LO GY

Nanofiber-mediated inhibition of focal adhesion kinase sensitizes glioma stemlike cells to epidermal growth factor receptor inhibition Maya Srikanth, Sunit Das, Eric J. Berns, Juno Kim, Samuel I. Stupp, and John A. Kessler Department of Neurology (M.S., J.K., J.A.K.); Department of Neurosurgery (S.D.), Northwestern University, Chicago, Illinois; Department of Biomedical Engineering (E.J.B.); Department of Chemistry and Department of Material Science and Engineering, Northwestern University, Evanston, Illinois (S.I.S.)

Background. Glioblastoma multiforme is the most common glioma in adults and carries a poor prognosis, due to tumor recurrence despite aggressive treatment. Such relapse has been attributed to the persistence of glioma stemlike cells (GSCs), a subpopulation of glioma cells with stem cell properties. Thus, targeting these cells will be critical to achieving meaningful improvement in glioblastoma multiforme survival. We investigated the role of b1-integrin signaling as one such potential target. Methods. We used GSCs isolated from primary human gliomas and maintained in stem cell conditions. We manipulated b1-integrin signaling using a self-assembling peptide amphiphile (PA) displaying the IKVAV (isoleucine-lysine-valine-alanine-valine) epitope as well as lentiviral overexpression, and we assayed the effects on downstream effectors and apoptosis using immunofluorescence. Results. We show that b1-integrin expression correlates with decreased survival in glioma patients and that b1integrin is highly expressed by GSCs. The IKVAV PA potently increases immobilized b1-integrin at the GSC membrane, activating integrin-linked kinase while inhibiting focal adhesion kinase (FAK). The IKVAV PA induces striking apoptosis in GSCs via this FAK inhibition, which is enhanced in combination with inhibition of epidermal growth factor receptor (EGFR). Conversely, lentiviral overexpression of b1-integrin renders GSCs resistant to EGFR inhibition, which was overcome by FAK inhibition.

Received July 22, 2012; accepted November 19, 2012. Corresponding Author: Maya Srikanth, PhD, 303 E. Chicago Avenue, Ward 10-233, Chicago, IL 60611 ([email protected]).

Conclusions. These observations reveal a role for b1-integrin signaling through FAK in GSC treatment resistance and introduce self-assembling PAs as a novel new therapeutic approach for overcoming this resistance. Keywords: beta1-integrin, focal adhesion kinase, glioma stem-like cells, nanotechnology, treatment resistance.

G

lioblastoma multiforme (GBM) is the most common and most deadly adult glioma, with a median survival of just 15 months.1 Tumors inevitably recur, implying that current treatments cannot eradicate the cells capable of tumor formation. Recent studies suggest that this population of cells shares many characteristics with neural stem cells (NSCs), earning the name glioma stemlike cells (GSCs).2 GSCs have demonstrated resistance to a wide variety of chemotherapy drugs,3 due in part to increased expression of both drug efflux pumps4 and the DNA repair protein O6-methyl guanine methyl transferase.5 GSCs also exhibit prominent resistance to ionizing radiation via augmented activation of the DNA damage checkpoint.6 Thus, novel therapeutic approaches targeting GSCs directly will be critical to achieving meaningful improvement in GBM survival. GSCs are dynamically regulated by the specialized perivascular microenvironment in which they reside: the GSC niche. GSCs are actively supported by factors secreted by adjacent tumor endothelial cells,7 and this interaction contributes to the treatment resistance of GSCs.8 In addition, the specialized architecture of extracellular matrix (ECM) present in the perivascular niche critically influences stem cell behavior.9 Cellular adhesion to ECM molecules is largely

# The Author(s) 2013. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: [email protected].

Srikanth et al.: FAK and EGFR inhibition in GSCs

mediated by the integrin family of cell surface receptors.10 As such, the integrins have emerged as a promising new prospective target of GSC-directed therapeutics. We have previously used a self-assembling peptide amphiphile (PA) with an IKVAV (isoleucine-lysinevaline-alanine-valine) epitope (IKVAV PA) to create a surrogate ECM for NSCs, which potently altered their fate via interaction with laminin receptors including b1-integrin.11 b1-integrin forms heterodimers with various a-integrin subunits to mediate binding to multiple different ECM molecules including laminin-1.12 b1-integrin is a marker of mammary stem cells13 and NSCs14 and contributes to the maintenance of the stem cell compartment in these tissues.15,16 Most recently, a6b1-integrin has been identified as a functionally relevant receptor in GSCs, and interference with a6b1 binding to laminin drastically diminishes tumor formation,17 highlighting the potential of GSC niche disruption as a therapeutic approach. b1-integrin has been implicated in tumor initiation and progression in breast and lung cancer,18 where it is a negative prognostic factor.19,20 Importantly, b1-integrin contributes to chemotherapeutic and radiation resistance in these tumors, including resistance to inhibitors of epidermal growth factor receptor (EGFR).21,22 Here, we explored the role of b1-integrin signaling in GSC biology and demonstrate a novel nanotechnology that overcomes GSC resistance to EGFR inhibitors by modulating this signaling.

Materials and Methods Glioma Stemlike Cell Isolation and Culture GSCs were isolated from GBM primary surgical specimens in keeping with protocols approved by the Northwestern University Institutional Review Board and grown as spheres as previously described.23 EGFR activity was inhibited in certain cultures using the EGFR kinase inhibitor PD153035 (Tocris) at 100 nM, while activity of focal adhesion kinase (FAK) was inhibited using PF537228 (Tocris) at 3 mm. Lipid rafts were disrupted using the cholesterol-depleting drug methyl-b-cyclodextrin at 10 mM (Sigma). GSC spheres were passaged by mechanical chopping every 7 – 10 days. Cells were used at passage 12 or less for all experiments, and all experiments were carried out with 2 cell lines. Cell plating in the different PAs was as previously described.11 Briefly, cells in GSC media were mixed 1:1 with a 1 weight percent PA solution in sterile deionized water on glass coverslips to a final concentration of 2 × 105 cells/mL. The PA was allowed to gel at 378C for 2 h before GSC media was added to submerge the gel culture. For controls, the same cell suspension was mixed with scrambled (VVIAK) PA solution and treated as above.

320

NEURO-ONCOLOGY

†

MARCH 2013

Fluorescence In situ Hybridization Fluorescence in situ hybridization (FISH) analysis of EGFR gene dosage was performed as previously described24 using a probe targeting the EGFR locus and a probe for the centromeric region of chromosome 7 (Spectrum Orange dye and Spectrum Green dye, respectively; Abbott Molecular). Nuclei were counterstained with Hoechst 33342. Samples were imaged using an Axiovert 200 standard fluorescence microscope (Zeiss) and analyzed using Axiovision 4.5 software (Zeiss). Immunocytochemistry Cells were fixed in 4% paraformaldehyde (Sigma) in 1× phosphate buffered saline (PBS) for 20 min, washed 3 times in PBS, and incubated with primary antibodies overnight at 48C in 1× PBS containing 1% bovine serum albumin and 0.25% Triton X-100. Following 3 more PBS washes, cells were incubated with the appropriate secondary antibody (Molecular Probes, Invitrogen) at 1:500 in 1× PBS for 1 h at room temperature. Nuclei were counterstained with Hoechst dye (1:5000 in 1× PBS); coverslips were mounted using Prolong Gold antifade reagent (Invitrogen) and imaged on a Leica SP-5 confocal microscope. National Institutes of Health ImageJ software was used to quantify images. The following primary antibodies were used: microtubule associated protein 2 (mouse immunoglobulin [Ig]G1, 1:1000; Abcam), glial fibrillary acidic protein (GFAP; rabbit polyclonal, 1:1000; DakoCytomation), sox2 (rabbit polyclonal 1:500; Millipore), nestin (mouse IgG1, 1:500; BD Biosciences), b1-integrin (mouse IgG1, 1:500; Millipore), active conformations of b1-integrin, clone HUTS-4 (mouse IgG2b, 1:500; Millipore), cleaved caspase 3 (rabbit polyclonal, 1:1000; Cell Signaling), phospho-FAK Tyr397 (rabbit polyclonal, 1:500; Cell Signaling), and phospho–integrin linked kinase (ILK) Ser343 (rabbit polyclonal, 1:250; Abgent). Protein Isolation, Immunoprecipitation, and Western Blotting GSCs were lysed in M-PER protein extraction reagent (Pierce) with 1× Halt Protease + Phosphatase inhibitor cocktail (Thermo Scientific). For immunoprecipitation, a protein A immunoprecipitation kit was used per manufacturer’s instructions (Roche). For Western blot analysis, samples were boiled for 10 min in strong denaturing conditions (nondenaturing conditions were used for samples probed with HUTS-4 antibody), and 5 mg of protein per sample was loaded on 4%–20% sodium dodecyl sulfate–polyacrylamide gels. Proteins were transferred onto polyvinylidene difluoride membranes at 48C for 1 h, which were then blocked in Tris-buffered saline with 0.05% Tween-20) with 5% nonfat dry milk for 1 h at room temperature. Primary antibodies were diluted in blocking solution and incubated with membranes overnight at 48C. After washing, membranes


were incubated with the appropriate horseradish peroxidase–conjugated secondary antibodies (1:2000 in blocking solution; Santa Cruz Biotechnology) and developed using SuperSignal West Pico enhanced chemiluminescent reagent (Thermo Scientific). The following primary antibodies were used: b1-integrin (mouse IgG1, 1:1000; BD Biosciences), active conformations of b1-integrin, clone HUTS-4 (mouse IgG2b, 1:1000; Millipore), phospho-EGFR Tyr1068 and total EGFR (rabbit polyclonal, 1:1000; Cell Signaling), phospho-FAK Tyr397 and total FAK (rabbit polyclonal, 1:1000; Cell Signaling), and glyceraldehyde 3-phosphate dehydrogenase (mouse IgG, 1:5000; Millipore). Flow Cytometry GSCs were dissociated with Accutase, blocked in sterilefiltered 10% fetal bovine serum (FBS) in 1× PBS, and incubated with either CD133/2-allophycocyanin (APC) antibody or mouse IgG1-APC isotype control antibody (Miltenyi Biotec) in blocking solution on ice for 30 min. For b1-integrin staining, phycoerythrin (PE)-conjugated mouse anti-human CD29 or PE-conjugated mouse IgG1 k isotype control (BD Pharmingen) was used. For CD44 staining, APC-conjugated mouse anti-human CD44 or APC-conjugated mouse IgG2b k isotype control (BD Pharmingen) was used. Cells were subjected to flow cytometric analysis with a CyAn machine (Beckman Coulter). RNA Extraction and Quantitative PCR Total RNA from cultured cells was extracted using the RNAqueous-4PCR kit (Ambion), and cDNA was made using Thermoscript reverse transcriptase and oligo-dT primers (Invitrogen). Quantitative PCR was performed with SybrGreen master mix (Applied Biosystems) and a Realplex2 Mastercycler (Eppendorf) using the following cycling parameters: 958C, 15 s; 608C, 60 s for 40 cycles. Lentivirus Production and Infection b1-integrin cDNA (Open Biosystems) was cloned into the elongation factor 1a promoter–internal ribosome entry site 2–enhanced green fluoresecent protein (pEF1aIRES2-EGFP) lentiviral vector,25 generously provided by Linzhao Cheng, Johns Hopkins University. The same vector with no coding sequence cloned upstream of the IRES2 sequence was used as a control. Viral production was as previously described26 using a cytomegalovirus promoter–vesicular stomatitis virus G protein envelope and pPAX2 (gag-pol) packaging plasmids. Seventy-two hours after infection, GSCs were subjected to fluorescence-activated cell sorting (FACS) for green fluorescent protein (GFP) using a MoFlo machine (Beckman Coulter). Mouse Intracranial Xenografts Six- to eight-week-old NOD-SCID-IL2Rg – /2 mice were obtained from The Jackson Laboratory and

subsequently bred in the barrier animal facility at Northwestern University. All animal procedures adhered to Public Health Service Policy on Humane Care and Use of Laboratory Animals. Mice were injected with 1 × 105 GSCs in 2 mL of media in the right striatum as previously described.27 Mice were monitored daily for up to 6 months for the development of symptoms. Tumor formation was assayed by hematoxylin and eosin staining of frozen sections. Statistical Analysis For immunofluorescence experiments, at least 200 cells per condition were counted. All experiments were performed independently 2 –3 times and significance was assessed by 2-way ANOVA. Log-rank analysis was performed for survival data. P , .05 was considered significant.

Results GSCs Express High Levels of b1-Integrin, a Negative Prognostic Marker in Gliomas While b1-integrin is known to be overexpressed in gliomas,28 the relationship between this overexpression and patient survival has not yet been evaluated. We analyzed the Repository for Molecular Brain Neoplasia Data and found that glioma patients with a modest 2-fold upregulation of b1-integrin had significantly shorter survival compared with other patients (Fig. 1A). This survival effect appeared to be dose dependent, as patients with 3-fold upregulation of b1-integrin displayed even shorter overall survival (Fig. 1A). Conversely, no such difference in survival was seen with a5-integrin or a6-integrin expression (data not shown), demonstrating the specificity of the b1-integrin survival effect. Thus b1-integrin confers poor prognosis in glioma, suggesting that cells expressing b1-integrin may contribute to tumor progression and recurrence. As GSCs are thought to be critical to these phenomena, we investigated the expression of b1-integrin in GSCs. GSCs were derived from human glioma surgical specimens using a modified neurosphere assay, as previously described.23 We verified that these cells were capable of self-renewal through serial passaging and that they expressed markers of NSCs, such as nestin, sox2, and CD133 (Supplementary Fig. S1A, B). They showed multilineage differentiation (Supplementary Fig. S1C) and gave rise to gliomas when injected intracranially into immunocompromised mice, demonstrating invasive growth (Supplementary Fig. S1D), microvascular proliferation (Supplementary Fig. S1E), and pseudopalisading necrosis (Supplementary Fig. S1F). We analyzed integrin subunit expression in a panel of GSCs using quantitative PCR. GSCs demonstrated expression of several a-integrin subunits, including a6, confirming previous reports.17 Importantly, b1-integrin was the only b-integrin subunit detectable, demonstrating high

NEURO-ONCOLOGY

†

MARCH 2013

321


Fig. 1. b1-integrin confers poor prognosis in glioma and is the only b-integrin expressed by GSCs. (A) Kaplan–Meier curves of overall survival of glioma patients stratified by b1-integrin expression. Data from Repository for Molecular Brain Neoplasia Data, NCI. (B) Quantitative reverse transcriptase– PCR for integrin subunits in a panel of GSCs, relative to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Inset: immunofluorescence for b1-integrin in GSCs (red), with DAPI (4’,6-diamidino-2-phenylindole) nuclear counterstain (blue). Scale bar: 50 mm. (C) Representative dual flow cytometry plot of b1-integrin and CD133 expression in GSCs (performed in GSC 83 and GSC 92). (D) Representative dual flow cytometry plot of b1-integrin and CD44 expression in GSCs (performed in GSC 83 and GSC 92).

expression in all GSCs assayed (Fig. 1B). We confirmed b1-integrin protein expression by immunofluorescence, with GSCs displaying prominent membranous localization of b1-integrin (Fig. 1B, inset). As CD133+ glioma cells may be enriched for GSCs,27 we examined b1-integrin expression in the subset of CD133+ GSCs by flow cytometry. Virtually every CD133+ GSC also expressed b1-integrin (99.53% + 0.38% CD133 +

322

NEURO-ONCOLOGY

†

MARCH 2013

b1+ vs 0.47% + 0.38% CD133 + b1 – ; Fig. 1C). Similarly, b1-integrin was expressed by almost every CD44+ GSC (97.48% + 0.03% CD44 + b1+ vs 2.52% + 0.03% CD44 + b1 – ; Fig. 1D), another putative GSC marker.29 Thus, b1-integrin is strongly expressed by GSCs and is the only b-integrin subunit detected, implying that it mediates the majority of, if not all, integrin signaling in these cells.


The IKVAV PA Modulates b1-Integrin in GSCs We have previously shown that a self-assembling PA displaying the IKVAV sequence potently alters NSC biology,11 and this bioactivity is likely mediated by interaction with laminin receptors that include b1-integrin.30 We thus explored the effect of the IKVAV PA on GSCs, with a focus on b1-integrin expression. GSCs were cultured in either the IKVAV PA (Fig. 2A) or a scrambled control (VVIAK PA) that has the same chemical properties but lacks bioactivity due to the arrangement of the amino acids in its epitope.31 GSCs encapsulated in the IKVAV PA exhibited strong membranous expression of b1-integrin (Fig. 2B). By contrast, b1-integrin was barely detectable in GSCs grown in the scrambled PA when using the same confocal settings. To determine whether the IKVAV PA could activate b1-integrin, we performed immunofluorescence using an antibody

specific for activated conformations of b1-integrin (HUTS-4).32 GSCs grown in the IKVAV PA demonstrated a striking increase in HUTS-4 immunoreactivity (Fig. 2C). We quantified this increase in activated b1-integrin by Western blot and found that it was over 100-fold higher than levels in the scrambled PA (Fig. 2D). Thus, the IKVAV PA is able to potently upregulate and activate b1-integrin at the GSC membrane.

The IKVAV PA Activates ILK but Inhibits FAK, Possibly via Impaired Lipid Raft Trafficking We next sought to determine whether the IKVAV PA was capable of activating the 2 main downstream effectors of b1-integrin signaling, ILK and FAK. We performed immunofluorescence for activated ILK on GSCs cultured in the IKVAV PA or scrambled PA. The

Fig. 2. The IKVAV PA modulates b1-integrin in GSCs. (A) 3D rendering of IKVAV PA self-assembling into nanofibers in the presence of cations, forming a gel with an extremely high density of epitope presented on its surface. (B) GSCs were cultured in control scrambled PA or bioactive IKVAV PA for 6 days and assayed by immunofluorescence for expression of b1-integrin, with DAPI (4’,6-diamidino-2-phenylindole) nuclear counterstain. Representative confocal z-stacks shown, scale bars: 50 mm. Performed in GSC 59 and GSC 46. (C) GSCs were cultured as in (B) and immunostained for activated b1-integrin (HUTS-4), with DAPI nuclear counterstain. Representative confocal z-stacks shown, scale bars: 50 mm. Performed in GSC 59 and GSC 46. (D) Western blot analysis of activated b1-integrin (HUTS-4) in GSCs cultured as above, with GAPDH (glyceraldehyde 3-phosphate dehydrogenase) loading control; scr: scrambled PA. Performed in GSC 59 and GSC 46.

NEURO-ONCOLOGY

†

MARCH 2013

323


IKVAV PA upregulated ILK autophosphorylated on S343 (pILKS343; Fig. 3A) in GSCs compared with the scrambled PA. Thus, the IKVAV PA is able to activate ILK. We next investigated levels of FAK activation by immunofluorescence for FAK autophosphorylated on Y397 (pFAKY397). While pFAKY397 was detectable in the cytoplasm of GSCs in the scrambled PA, there was a striking decrease in pFAKY397 staining in the IKVAV PA (Fig. 3B). To quantify this difference, we immunoprecipitated total FAK protein from GSCs in the IKVAV or scrambled PA, followed by Western blotting for pFAKY397. This confirmed our immunofluorescence findings, as GSCs in the IKVAV PA demonstrated a 66% decrease in pFAK levels as normalized to total FAK (Fig. 3C). Thus, the IKVAV PA is a potent inhibitor of FAK activation.

Such differential regulation of downstream b1-integrin signaling has been observed in other contexts. For example, FAK signaling is inhibited when trafficking into lipid rafts is impaired,33 whereas ILK signaling is regulated conversely.34 To determine whether disruption of lipid rafts in GSCs mimics the effects of the IKVAV PA, we treated cells with the cholesteroldepleting drug methyl-b-cyclodextrin. This lipid raft disruption resulted in a similar increase in pILKS343 (Supplementary Fig. S2A) with a concomitant decrease in pFAKY397 (Supplementary Fig. S2B). Thus, the effect of lipid raft disruption mirrors the effect of the IKVAV PA on b1-integrin signaling, suggesting impaired localization of b1-integrin to lipid rafts as a potential mechanism underlying the bioactivity of the IKVAV PA (Fig. 4).

Fig. 3. IKVAV PA activates ILK but inhibits FAK. (A) GSCs were cultured in scrambled PA or IKVAV PA and immunostained for phospho-ILK-S343 (pILKS343). Representative confocal z-stacks are shown, scale bar: 50 mm. Performed in GSC 59 and GSC 46. (B) GSCs were cultured as in (A) and immunostained for phospho-FAK-Y397 (pFAKY397). Representative confocal slices are shown, scale bar: 50 mm. Performed in GSC 59 and GSC 46. (C) Total FAK protein was immunoprecipitated from GSCs cultured as in (B) and Western blot analysis was performed for phospho-FAK-Y397 (pFAKY397), with total FAK for reference; scr: scrambled PA. Relative pFAKY397 expression levels were calculated using densitometry and are indicated in blue. Performed in GSC 59 and GSC 46.

324

NEURO-ONCOLOGY

†

MARCH 2013


Fig. 4. Model of b1-integrin and EGFR interaction in GSCs, as modified by IKVAV PA. In the presence of normal ECM (left), activated b1-integrin localizes to lipid rafts in close proximity to EGFR, leading to inhibition of ILK activation and stimulation of FAK activation, ultimately promoting GSC survival. When cultured in IKVAV PA (right), trafficking of activated b1-integrin to lipid rafts is impaired, likely via the presentation of a high density of immobilized ligand. This leads to activation of ILK and inhibition of FAK, which causes apoptosis that can be enhanced by EGFR inhibition (see Fig. 5).

The IKVAV PA Sensitizes GSCs to EGFR Inhibition via Inhibition of FAK We next explored the biological effect of the IKVAV PA on GSCs. As b1-integrin interacts with the EGFR pathway and confers resistance to EGFR inhibitors in breast cancer,21,35 we examined the effect of the IKVAV PA in combination with an EGFR kinase inhibitor. Such inhibitors are a focus of clinical trials in GBM, as the EGFR gene on chromosome 7 is commonly amplified in these tumors. Thus, we first determined whether the EGFR gene was amplified in a panel of GSC lines using FISH. GSCs fell into 3 categories: chromosome 7 polyploidy, extrachromosomal amplification of EGFR, and normal chromosome 7 and EGFR (Fig. 6A). Subsequent studies were performed with both EGFRamplified and EGFRnormal GSCs to discern whether EGFR genetic status influenced responses to b1-integrin modulation and/or EGFR inhibition. We next established that 100 nM PD15303536 was sufficient to inhibit EGFR autophosphorylation on Y1068 (pEGFRY1068) in both EGFRamplified and EGFRnormal GSCs (Fig. 6B). Interestingly, EGFR autophosphorylation was present in EGFRamplified GSCs without epidermal growth factor in the media, suggesting that these cells have ligand-independent EGFR activation, which was nonetheless inhibited by the EGFR kinase inhibitor. To investigate the biological effect of the IKVAV PA in this system, EGFRamplified GSCs were encapsulated in PA nanogels for 3 days and then treated with either 100 nM PD153035 or vehicle (dimethyl sulfoxide [DMSO]) for another 3 days. We used immunofluorescence for cleaved caspase 3 to assay apoptosis. GSCs in the scrambled PA exhibited a 2-fold increase in apoptosis when treated with the EGFR inhibitor (Fig. 5A, also

Fig. 5. IKVAV PA sensitizes GSCs to EGFR inhibition by inhibition of FAK. (A) GSCs were cultured in scrambled PA or IKVAV PA and treated with DMSO or EGFR inhibitor PD153035. Apoptosis was assayed by cleaved caspase 3 (CC3) staining. Mean + SEM of 3 experiments, P , .01 by 2-way ANOVA vs *3 or **2 conditions. Performed in GSC 59 and GSC 46. (B) GSCs were plated in scrambled PA with DMSO (control) or FAK inhibitor, then treated with DMSO or EGFR inhibitor, and apoptosis was assayed by CC3 staining. Mean + SD for 2 experiments, P , .01 by 2-way ANOVA vs *3 or **2 conditions. Performed in GSC 59 and GSC 46.

Supplementary Fig. S3A). The IKVAV PA alone induced an identical level of apoptosis without drug treatment, and in combination with the EGFR inhibitor significantly increased apoptosis almost 3-fold over baseline

NEURO-ONCOLOGY

†

MARCH 2013

325


Fig. 6. The effect of b1-integrin modulation on EGFR inhibition is independent of EGFR status. (A) FISH for EGFR locus and chromosome 7 centromere in GSCs, with Hoechst nuclear counterstain. Representative images of EGFRnormal, EGFRamplified, and EGFRpolyploid GSCs shown. Magnification: 100×. Performed in GSC 16, 18, 31, 42, 46, 57, 59, and 67. (B) Western blot analysis of EGFR autophosphorylation (pEGFRY1068) and total EGFR in EGFRnormal and EGFRamplified GSC lines, with GAPDH (glyceraldehyde 3-phosphate dehydrogenase) loading control. GSCs grown without EGF, with EGF or with EGF + EGFR kinase inhibitor PD153035 (+PD). Performed in GSC 59 and GSC 46. (C) EGFRnormal GSCs were encapsulated in scrambled (scr.) or IKVAV PA, treated with DMSO or EGFR inhibitor, and apoptosis was assayed by cleaved caspase 3 (CC3) staining. Mean + SD for 2 independent experiments, P , .02 by 2-way ANOVA vs *3 or **2 conditions. Performed in GSC 46. (D) EGFRnormal GSCs were infected with control EGFP or b1-integrin lentivirus, FAC-sorted, and treated with DMSO or EGFR inhibitor, and apoptosis was assayed by CC3. Mean + SEM for 3 independent experiments, ***P , .01 by 2-way ANOVA. Performed in GSC 46.

levels (Fig. 5A, also Supplementary Fig. S3A). This powerful induction of apoptosis was independent of EGFR status, as a comparable effect was observed in EGFRnormal GSCs (Fig. 6C). Thus, the ability of the IKVAV PA to trigger apoptosis in GSCs, as well as to enhance sensitivity to EGFR inhibition, is not influenced by EGFR status. FAK is critical for the integration of EGFR and b1-integrin signaling in other systems,37 suggesting that the potent inhibition of FAK by the IKVAV PA may mediate its effects on survival. To investigate this possibility, we used an alternative means of inhibiting FAK, the selective kinase inhibitor PF573228.38 First, we established that 3 mM PF573228 was sufficient to inhibit FAK autophosphorylation on Y397 (pFAKY397) by .90% in GSCs (Supplementary Fig. S3B). We then plated GSCs in the scrambled PA and treated them with either 3 mM PF573228 or vehicle (DMSO). After 3 days, cells were treated with either 100 nM PD153035 or vehicle (DMSO). Levels of apoptosis were assessed 3 days later by immunofluorescence for cleaved caspase 3. EGFR inhibition alone increased apoptosis 1.7-fold (Fig. 5B). FAK inhibitor alone also increased apoptosis about 1.7-fold, but in combination with the EGFR inhibitor was able to augment levels of

326

NEURO-ONCOLOGY

†

MARCH 2013

apoptosis to almost 4-fold the basal level (Fig. 5B). Thus, FAK inhibition with a small molecule mimics the effects of the IKVAV PA, strongly suggesting such inhibition as the mechanism of action of the IKVAV PA.

b1-Integrin Blocks the Effects of EGFR Inhibition in GSCs via FAK Activation We next used lentiviral vectors to overexpress b1-integrin. EGFRamplified GSCs were infected with either pEF-EGFP (control) or pEF-b1-IRES2-EGFP lentivirus. After 3 days, GSCs were sorted for GFP by FACS, and b1-integrin levels were assessed. pEF-b1-IRES2-eGFP elevated b1-integrin levels more than 3-fold (Fig. 7A). To determine the effect of EGFR inhibition in the context of lentivirally mediated b1-integrin overexpression, GSCs were infected and FACS performed as above, then GSCs were plated in EGFR inhibitor and assayed for apoptosis. GSCs infected with control virus showed a robust 2-fold increase in apoptosis when treated with the EGFR inhibitor, almost identical to that observed in the scrambled PA (Fig. 7B). Overexpression of b1-integrin alone did not change the level of apoptosis, but GSCs overexpressing


even GSCs with upregulated b1-integrin sensitive to EGFR inhibition.

Discussion

Fig. 7. b1-integrin blocks apoptosis mediated by EGFR inhibition via FAK. (A) Western blot for b1-integrin in lentivirus-infected GSCs, with GAPDH (glyceraldehyde 3-phosphate dehydrogenase) loading control. (B) GSCs were infected with control EGFP or b1-integrin lentivirus, FAC-sorted, and treated with DMSO or EGFR inhibitor PD153035. Performed in GSC 59 and GSC 46. (C) GSCs were prepared as in (B) but treated with FAK inhibitor PF573228. Performed in GSC 59 and GSC 46. (D) GSCs were prepared as in (B) but treated with EGFR inhibitor PD153035 and FAK inhibitor PF573228 in combination. In (B– D), apoptosis was assayed by cleaved caspase 3 (CC3) staining, shown as mean + SD of 2 experiments, *P , .03 or **P , .01 by 2-way ANOVA. Performed in GSC 59 and GSC 46.

b1-integrin showed no change in apoptosis when treated with the EGFR inhibitor (Fig. 7B), indicating that b1-integrin overexpression completely abrogated the effects of EGFR inhibition in GSCs. This is consistent with previous studies showing b1-integrin overexpression as a clinically relevant mechanism for protecting breast cancer cells from inhibition of the effects of human epidermal growth factor receptor 2, an EGFR family member.21 This effect was also observed in EGFRnormal GSCs (Fig. 6D). Thus, increased b1-integrin blocked the effect of EGFR inhibition independently of EGFR status. This is in keeping with our finding that the effect of the IKVAV PA is identical in EGFRamplified and EGFRnormal GSCs. As FAK was central to the effect of the IKVAV PA, we examined the effect of FAK inhibition in GSCs overexpressing b1-integrin. We treated lentivirally infected GSCs with 3 mM PF573228 to inhibit FAK and assayed for apoptosis 3 days later. The FAK inhibitor increased apoptosis of GSCs infected with control virus nearly 2-fold (Fig. 7C). Importantly, GSCs overexpressing b1-integrin showed the same degree of apoptosis induction with FAK inhibition (Fig. 7C). Thus, b1-integrin upregulation can be overcome by inhibition of FAK activation. Given these results, we investigated the effect of combined inhibition of EGFR and FAK in GSCs. Such joint inhibition yielded a .3-fold increase in apoptotic cells in both control and b1-integrin – overexpressing GSCs (Fig. 7D). Therefore, FAK inhibition renders

EGFR activation is a hallmark of GBM, yet smallmolecule inhibitors of EGFR have yielded disappointing results in clinical trials.39 This suggests that tumorpromoting cells such as GSCs may be resistant to these treatments. We reveal FAK activation by b1-integrin as a mechanism underlying resistance to EGFR inhibition in GSCs and demonstrate that combined inhibition of FAK and EGFR overcomes this resistance. This work highlights the significant role of the GSC niche in treatment resistance. b1-integrin also promotes resistance to EGFR family inhibitors in breast and lung cancer,21,22 and FAK has very recently been implicated as a key mediator in this process.35 Interestingly, b1-integrin signaling can directly activate EGFR in a ligand-independent manner,40 and abrogating this interaction renders lung cancer cells more sensitive to EGFR inhibition.41 Our work also reveals a previously unexplored role for lipid rafts in shaping GSC signaling. Lipid rafts have been characterized as signaling hubs in the cell membrane, facilitating interactions between key regulatory molecules in multiple signaling pathways.42 Such facilitation is critical to the development and maintenance of NSCs.43 In this work, the huge density of IKVAV epitope on the IKVAV PA seems to immobilize b1-integrin at the cell surface, likely disrupting b1-integrin recruitment to lipid rafts. This inhibits FAK activation33 but promotes ILK activation,34 with profound consequences on GSC survival (Fig. 4). Lipid raft localization of EGFR promotes resistance to EGFR inhibitors in breast cancer, and disruption of this localization sensitizes cells to EGFR inhibition,44 similarly to our results with the IKVAV PA. Further characterization of lipid rafts in GSCs may yield additional targets for combating treatment resistance. Previous studies of integrin signaling in GSCs demonstrated drastically diminished tumor formation when a6-integrin signaling was blocked but did not identify the downstream signaling proteins responsible.17 By identifying FAK as the crucial molecular mediator of integrin signaling in GSCs, our work highlights FAK as a promising therapeutic target in glioma. FAK activation has been validated as an independent negative prognostic factor in glioma,45 and FAK inhibition sensitizes tumors to chemotherapy in other systems,46 further validating it as a target for combination therapy in glioma. FAK also plays critical roles in cancer cell invasion47 and radioresistance,48 suggesting that FAK inhibition may also inhibit these phenomena in glioma. One FAK inhibitor has recently completed a phase I trial with promising indications of disease control,49 but further studies involving glioma patients would be required to assess efficacy in this setting. Our work also introduces the IKVAV PA, and more broadly peptide amphiphile nanotechnology,50 as a novel tool in the study and treatment of gliomas.

NEURO-ONCOLOGY

†

MARCH 2013

327


Given the demonstrated importance of niche interactions in GSC regulation, the ability of PAs to generate a surrogate ECM is of obvious relevance to glioma biology. In addition, these unique molecules remain in aqueous solution until they self-assemble into nanofibers in physiologic cation concentrations, allowing for easy administration at the tumor site. The IKVAV PA can be safely and easily injected into the CNS in vivo, where it self-assembles and facilitates recovery in a mouse model of spinal cord injury.51 Thus, it is likely that the IKVAV PA would maintain the biological activity demonstrated here in vivo, even in a diverse cellular environment that included non-GSCs. Finally, PA nanotechnology is infinitely tunable. In addition to intrinsic bioactivity as explored in this work, it can be combined with enzyme cleavage sites and small-molecule inhibitors for controlled drug release and combination therapy.52 Overall, our findings reveal b1-integrin signaling through FAK as a targetable mechanism of GSC resistance to EGFR inhibitors and highlight the IKVAV PA as an innovative technology for translating this discovery to the clinic.

Supplementary Material Supplementary material is available online at NeuroOncology (http://neuro-oncology.oxfordjournals.org/).

Funding This work was supported by the National Institutes of Health (grant nos. F30NS065590 to M. S., R01NS20013 and R01NS21778 to J. A. K., and EB003806 to J. A. K. and S. I. S.) and Northwestern Memorial Hospital (Dixon Translational Research Award to J. A. K. and S. D., Auxiliary Board Award to S. D.).

Acknowledgments We thank Jeffrey Nelson and Paul Mehl of the Robert H. Lurie Flow Cytometry Core Facility for flow cytometry assistance, Ljuba Lyass of the Human Embryonic and Induced Pluripotent Stem Cell Facility for cell culture assistance, Jackie Renfrow (formerly of the Markus Bredel Laboratory) for FISH assistance, and Kessler Lab members for technical assistance and critical review of this manuscript.

Conflict of interest statement. None declared.

References 1.

2.

Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus con-

progenitor

2005;352(10):987 –996.

2004;303(5662):1352 –1355.

Das S, Srikanth M, Kessler JA. Cancer stem cells and glioma. Nat Clin

12. Belkin AM, Stepp MA. Integrins as receptors for laminins. Microsc Res

84 –88. 14. Hall PE, Lathia JD, Miller NG, Caldwell MA, ffrench-Constant C.

spheres identify a population of tumor stem-like cells with angiogenic

Integrins are markers of human neural stem cells. Stem Cells. 2006;24(9):2078 –2084. 15. Taddei I, Deugnier MA, Faraldo MM, et al. Beta1 integrin deletion from

Sciuscio D, Diserens AC, van Dommelen K, et al. Extent and patterns spheres.

Clin

Cancer

Res.

2011;17(2):

255 –266.

the basal compartment of the mammary epithelium affects stem cells. Nature Cell Biol. 2008;10(6):716 –722. 16. Campos LS, Leone DP, Relvas JB, et al. Beta1 integrins activate a MAPK signalling pathway in neural stem cells that contributes to their

Bao S, Wu Q, McLendon RE, et al. Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature. 2006;444(7120):756 –760.

maintenance. Development. 2004;131(14):3433 –3444. 17. Lathia JD, Gallagher J, Heddleston JM, et al. Integrin alpha 6 regulates glioblastoma stem cells. Cell Stem Cell. 2010;6(5):421 –432.

Calabrese C, Poppleton H, Kocak M, et al. A perivascular niche for brain tumor stem cells. Cancer Cell. 2007;11(1):69 – 82.

9.

Science.

Salmaggi A, Boiardi A, Gelati M, et al. Glioblastoma-derived tumoro-

glioblastoma-derived

8.

nanofibers.

mammary gland from a single stem cell. Nature. 2006;439(7072):

of MGMT promoter methylation in glioblastoma- and respective

7.

density

glioblastoma stem cells. Cell Death and Differentiation. 2006;

2006;54(8):850 –860.

6.

high-epitope

13. Shackleton M, Vaillant F, Simpson KJ, et al. Generation of a functional

potential and enhanced multidrug resistance phenotype. Glia. 5.

by

Eramo A, Ricci-Vitiani L, Zeuner A, et al. Chemotherapy resistance of 13(7):1238– 1241.

4.

cells

Tech. 2000;51(3):280 –301.

Pract Neurol. 2008;4(8):427 – 435. 3.

11. Silva GA, Czeisler C, Niece KL, et al. Selective differentiation of neural

comitant and adjuvant temozolomide for glioblastoma. N Engl J Med.

18. Moschos SJ, Drogowski LM, Reppert SL, Kirkwood JM. Integrins and cancer. Oncology (Williston Park). 2007;21(9 Suppl 3):13 – 20.

Hovinga KE, Shimizu F, Wang R, et al. Inhibition of notch signaling in

19. Oshita F, Kameda Y, Hamanaka N, et al. High expression of integrin

glioblastoma targets cancer stem cells via an endothelial cell intermedi-

beta1 and p53 is a greater poor prognostic factor than clinical

ate. Stem Cells (Dayton, Ohio). 2010;28(6):1019 –1029.

stage in small-cell lung cancer. Am J Clin Oncol. 2004;27(3):

Hynes RO. The extracellular matrix: not just pretty fibrils. Science. 2009;326(5957):1216 – 1219.

10. Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell.

NEURO-ONCOLOGY

with decreased survival in invasive breast cancer. Cancer Research. 2007;67(2):659 –664.

2002;110(6):673 – 687.

328

215 – 219. 20. Yao ES, Zhang H, Chen YY, et al. Increased beta1 integrin is associated

†

MARCH 2013


21. Lesniak D, Xu Y, Deschenes J, et al. Beta1-integrin circumvents the

37. Sieg DJ, Hauck CR, Ilic D, et al. FAK integrates growth-factor and integ-

antiproliferative effects of trastuzumab in human epidermal growth

rin

factor receptor-2-positive breast cancer. Cancer Research. 2009;

2000;2(5):249 –256.

69(22):8620 –8628.

signals

to

promote

cell

migration.

Nature

Cell

Biol.

38. Slack-Davis JK, Martin KH, Tilghman RW, et al. Cellular characterization

22. Ju L, Zhou C, Li W, Yan L. Integrin beta1 over-expression associates with resistance to tyrosine kinase inhibitor gefitinib in non-small cell lung cancer. J Cell Biochem. 2010;111(6):1565 –1574.

of

a

novel

focal

adhesion

kinase

inhibitor.

J

Biol

Chem.

2007;282(20):14845 – 14852. 39. Rich JN, Reardon DA, Peery T, et al. Phase II trial of gefitinib in recurrent

23. Galli R, Binda E, Orfanelli U, et al. Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res. 2004;64(19):7011 –7021.

glioblastoma. J Clin Oncol. 2004;22(1):133 –142. 40. Moro L, Dolce L, Cabodi S, et al. Integrin-induced epidermal growth factor (EGF) receptor activation requires c-Src and p130Cas and leads

24. Pfisterer WK, Hank NC, Preul MC, et al. Diagnostic and prognostic significance of genetic regional heterogeneity in meningiomas. Neuro Oncol. 2004;6(4):290 –299.

to phosphorylation of specific EGF receptor tyrosines. J Biol Chem. 2002;277(11):9405 –9414. 41. Morello V, Cabodi S, Sigismund S, et al. Beta1 integrin controls EGFR

25. Pomper MG, Hammond H, Yu X, et al. Serial imaging of human embryonic stem-cell engraftment and teratoma formation in live mouse models. Cell Res. 2009;19(3):370 –379.

signaling and tumorigenic properties of lung cancer cells. Oncogene. 2011;30(39):4087 –4096. 42. Decker L, Baron W, ffrench-Constant C. Lipid rafts: microenvironments

26. Bonaguidi MA, McGuire T, Hu M, Kan L, Samanta J, Kessler JA. LIF and BMP signaling generate separate and discrete types of GFAP-expressing cells. Development. 2005;132(24):5503 –5514.

for integrin-growth factor interactions in neural development. Biochem Soc Trans. 2004;32(Pt3):426 –430. 43. Campos LS, Decker L, Taylor V, Skarnes W. Notch, epidermal growth

27. Singh SK, Hawkins C, Clarke ID, et al. Identification of human brain

factor receptor, and beta1-integrin pathways are coordinated in neural stem cells. J Biol Chem. 2006;281(8):5300–5309.

tumour initiating cells. Nature. 2004;432(7015):396 –401. 28. Paulus W, Baur I, Schuppan D, Roggendorf W. Characterization of

44. Irwin ME, Mueller KL, Bohin N, Ge Y, Boerner JL. Lipid raft localization

integrin receptors in normal and neoplastic human brain. Am J

of EGFR alters the response of cancer cells to the EGFR tyrosine kinase

Pathol. 1993;143(1):154 – 163.

inhibitor gefitinib. J Cell Physiol. 2011;226(9):2316 –2328.

29. Son MJ, Woolard K, Nam DH, Lee J, Fine HA. SSEA-1 is an enrichment

45. Ding L, Sun X, You Y, Liu N, Fu Z. Expression of focal adhesion

marker for tumor-initiating cells in human glioblastoma. Cell Stem Cell.

kinase and phosphorylated focal adhesion kinase in human gliomas is

2009;4(5):440 –452.

associated with unfavorable overall survival. Transl Res. 2010;

30. Tomaselli KJ, Doherty P, Emmett CJ, Damsky CH, Walsh FS, Reichardt

156(1):45–52.

LF. Expression of beta 1 integrins in sensory neurons of the dorsal

46. Halder J, Kamat AA, Landen CN, Jr., et al. Focal adhesion kinase

root ganglion and their functions in neurite outgrowth on two laminin

targeting using in vivo short interfering RNA delivery in neutral lipo-

isoforms. J Neurosci. 1993;13(11):4880 –4888.

somes for ovarian carcinoma therapy. Clin Cancer Res. 2006;

31. Nomizu M, Weeks BS, Weston CA, Kim WH, Kleinman HK, Yamada Y. Structure-activity study of a laminin alpha 1 chain active peptide segment Ile-Lys-Val-Ala-Val (IKVAV). FEBS Lett. 1995;365(2–3):227–231. 32. Luque A, Gomez M, Puzon W, Takada Y, Sanchez-Madrid F, Cabanas

12(16):4916 –4924. 47. Mitra SK, Hanson DA, Schlaepfer DD. Focal adhesion kinase: in command and control of cell motility. Nat Rev Mol Cell Biol. 2005;6(1):56 –68.

C. Activated conformations of very late activation integrins detected

48. Kasahara T, Koguchi E, Funakoshi M, Aizu-Yokota E, Sonoda Y.

by a group of antibodies (HUTS) specific for a novel regulatory

Antiapoptotic action of focal adhesion kinase (FAK) against ionizing ra-

region (355 –425) of the common beta 1 chain. J Biol Chem. 1996;271(19):11067 –11075.

diation. Antioxid Redox Signal. 2002;4(3):491 –499. 49. Infante JR, Camidge DR, Mileshkin LR, et al. Safety, pharmacokinetic,

33. Seong J, Ouyang M, Kim T, et al. Detection of focal adhesion kinase ac-

and pharmacodynamic phase I dose-escalation trial of PF-00562271,

tivation at membrane microdomains by fluorescence resonance energy

an inhibitor of focal adhesion kinase, in advanced solid tumors. J Clin

transfer. Nat Commun. 2011;2:406.

Oncol. 2012;30(13):1527 –1533.

34. Chun J, Hyun S, Kwon T, Lee EJ, Hong SK, Kang SS. The subcellular lo-

50. Hartgerink JD, Beniash E, Stupp SI. Self-assembly and mineraliza-

calization control of integrin linked kinase 1 through its protein-protein

tion of peptide-amphiphile nanofibers. Science. 2001;294(5547):

interaction with caveolin-1. Cell Signal. 2005;17(6):751 –760.

1684 – 1688.

35. Huang C, Park CC, Hilsenbeck SG, et al. Beta1 integrin mediates an al-

51. Tysseling-Mattiace VM, Sahni V, Niece KL, et al. Self-assembling nano-

ternative survival pathway in breast cancer cells resistant to lapatinib.

fibers inhibit glial scar formation and promote axon elongation after

Breast Cancer Res. 2011;13(4):R84.

spinal cord injury. J Neurosci. 2008;28(14):3814 –3823.

36. Fry DW, Kraker AJ, McMichael A, et al. A specific inhibitor of the epidermal

growth

factor

1994;265(5175):1093 – 1095.

receptor

tyrosine

kinase.

Science.

52. Kim JK, Anderson J, Jun HW, Repka MA, Jo S. Self-assembling peptide amphiphile-based nanofiber gel for bioresponsive cisplatin delivery. Mol Pharm. 2009;6(3):978 –985.

NEURO-ONCOLOGY

†

MARCH 2013

329