brain-mimetic 3d culture platforms allow investigation of ... · cd168 (also known as hmmr or...

Translational Science

Brain-Mimetic 3D Culture Platforms AllowInvestigation of Cooperative Effects ofExtracellular Matrix Features on TherapeuticResistance in GlioblastomaWeikun Xiao1, Rongyu Zhang1, Alireza Sohrabi1, Arshia Ehsanipour1, Songping Sun1,Jesse Liang1, Christopher M.Walthers1, Lisa Ta2, David A. Nathanson2,3,4, andStephanie K. Seidlits1,3,4,5

Abstract

Glioblastoma (GBM) tumors exhibit potentially actionablegenetic alterations against which targeted therapies have beeneffective in treatment of other cancers. However, these therapieshave largely failed in GBM patients. A notable example is kinaseinhibitors of EGFR, which display poor clinical efficacy despiteoverexpression and/or mutation of EGFR in >50% of GBM. Inaddressing this issue, preclinical models may be limited by theinability to accurately replicate pathophysiologic interactions ofGBM cells with unique aspects of the brain extracellular matrix(ECM), which is relatively enriched in hyaluronic acid (HA) andflexible. In this study, we present a brain-mimetic biomaterialECMplatform for 3D culturing of patient-derived GBM cells, withimproved pathophysiologic properties as an experimentalmodel.

Compared with orthotopic xenograft assays, the novel biomate-rial cultures we developed better preserved the physiology andkinetics of acquired resistance to the EGFR inhibition than glio-masphere cultures. Orthogonal modulation of both HA contentand mechanical properties of biomaterial scaffolds was requiredto achieve this result. Overall, our findings show how specificinteractions betweenGBMcell receptors and scaffold componentscontribute significantly to resistance to the cytotoxic effects ofEGFR inhibition.

Significance: Three-dimensional culture scaffolds of glioblas-toma provide a better physiological representation over currentmethods of patient-derived cell culture and xenograft models.Cancer Res; 78(5); 1358–70. �2017 AACR.

IntroductionGlioblastoma (GBM) is the most common and aggressive

cancer originating in the central nervous system (CNS; ref. 1).High lethality is largely attributed to poor therapeutic responseto current treatments (2). Similar to many peripheral cancers,EGFR, a receptor tyrosine kinase (RTK), is overexpressed and/ormutated inmore than 50%ofGBM tumors (2). EGFR stimulationactivates the PI3K–AKT and MAPK–ERK pathways to promoteprogression, invasion, survival, and drug resistance (3). Targetedinhibition of EGFR has been largely successful in treatment ofnon-CNS malignancies with EGFR amplification and/or muta-tion, including breast cancer and non–small cell lung carcinoma

(4, 5). However, EGFR inhibition has largely been ineffectiveagainst GBM in clinical trials (6).

We contend that the unique brain microenvironment substan-tially contributes to the failure of targeted EGFR therapies inGBM.In contrast to malignant cancers of peripheral origin, GBM rarelymetastasizes beyond the brain (7), indicating a strong preferencefor the brain microenvironment. The long chain polysaccharidehyaluronic acid (HA) is abundant in brain extracellular matrix(ECM). In GBM, overexpression of HA is associated with manyphenotypic changes associated with cancer progression includinginitial tumor development, cancer cell proliferation, invasion,resistance to therapeutic agents and posttreatment recurrence (8).Likewise, cell surface receptors for HA, including CD44 andCD168 (also known as HMMR or RHAMM), are often upregu-lated on GBM cells (9–11) and increased CD44 expression inclinical tumors is a poor prognostic indicator (12).

Increased deposition of ECM proteins—which interact withintegrins through the RGD sequence—directly correlates withpoor GBM prognosis (13–16). Treatment of GBM with Cilengi-tide—a cyclic RGD peptide designed to inhibit RGD interactionswith integrin avb3—has shown modest success as an adjuncttherapy to prevent integrin-mediated protection from drug-induced apoptosis in clinical trials (17). However, "inside-out"activation of integrin b1 by HA-bound CD44 can bypassdirect integrin-ECM interactions to promote survival of gastroin-testinal, breast, and ovarian cancer cells (18–20). Mechanicalsignaling from the GBM microenvironment, which is stifferthan normal brain, has also been reported to influence drug

1Department of Bioengineering, University of California, Los Angeles, California.2Department of Molecular and Medical Pharmacology, University of California,Los Angeles, California. 3Jonsson Comprehensive Cancer Center, University ofCalifornia, Los Angeles, California. 4Brain Research Institute, University ofCalifornia, Los Angeles, California. 5Broad Stem Cell Research Center, Universityof California, Los Angeles, California.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corrected online February 25, 2019.

Corresponding Author: Stephanie K. Seidlits, University of California LosAngeles, 420 Westwood Plaza, Room 5121H, Engineering V, Los Angeles, CA90095. Phone: 310-267-5244; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-17-2429

�2017 American Association for Cancer Research.

CancerResearch

Cancer Res; 78(5) March 1, 20181358

on August 5, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst February 28, 2018; DOI: 10.1158/0008-5472.CAN-17-2429





http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-17-2429&domain=pdf&date_stamp=2019-2-25

http://cancerres.aacrjournals.org/



resistance (16, 21). Integrins, CD44, and RTKs are mechanore-sponsive (15, 22–24).

Taken together, previous reports indicate that interactionsamong various ECM receptors act to protect cancer cells fromdrug-induced apoptosis and facilitate resistance. Thus, we positthat GBM tumors acquire resistance to EGFR inhibition throughinteractions with the local ECM, alternatively activating oncogenicpathways through both cooperation with and compensation forEGFR. However, the influence of the local microenvironment ontherapeutic resistance in GBM has been challenging to study, ascurrently available experimental models fail to account for thecomplex array of ECM components surrounding GBM tumors anddo not adequately reflect clinical outcomes. To address this limi-tation, we have developed a novel biomaterial platform for three-dimensional (3D) culture of patient-derived GBM cells in whichHA content, integrin-binding peptide concentration and compres-sive modulus can be orthogonally tuned to mimic GBM tumors.

Using this hydrogel platform, 3D-cultured GBM cells devel-oped rapid resistance to erlotinib—a small molecule for targetedinhibition of EGFR currently used in clinical treatment of severalcancers (4, 5). Unlike patient-matched gliomasphere cultures,biomaterial-cultured GBM cells maintained expression of ECMreceptors and resistance kinetics similar to orthotopic xenograftsin mice. Orthogonal tuning of both HA content and scaffoldstiffness was required to achieve these results—where a soft,HA-rich microenvironment that mimicked native brain wasnecessary to protect GBM from cytostatic and cytotoxic effects oferlotinib treatment. Furthermore, results demonstrate that theseeffects are at least partially mediated by HA-CD44 interactions.Finally, biomaterial cultures demonstrated the cooperative effectsof integrin and CD44 engagement, where inclusion of RGDpeptides into soft, HA-rich hydrogels significantly increased apo-ptotic resistance to erlotinib treatment. In future studies, theseex vivo culture platforms can provide (i) a controlled experimentalspace in which to quantify the independent and combined effectsof individual ECM components on drug resistance and (ii) moreaccurate predictions of in vivo tumor physiology.

Materials and MethodsAll reagents and materials were purchased from ThermoFisher

Scientific, unless otherwise specified. More details on procedurescan be found in Supplementary Methods.

Mouse xenograftsAll studies were approved by the UCLAOffice of Animal Rights

Oversight. For intracranial xenografts, GBM39 or HK301 cellswith constitutive expression of Gaussia Luciferase were injected(2 � 105 cells) into the right striatum (2 mm lateral and 1 mmposterior to bregma, at a depth of 2 mm) of female nod-SCID-gmice (6–8weeks old). Tumor burdenwasmonitored semi-weeklyvia bioluminescence imaging using an IVIS 200 instrument. Twoweeks following injection, mice were randomized into two treat-ment arms—vehicle or erlotinib. Mice were euthanized whenmoribund, which was defined by a loss of 25% to 30% bodyweight from start of treatment in addition to symptoms such asneurological defects, paralysis, hydrocephalus, and hunching. Forsubcutaneous xenograft studies, 1 � 106 cells/100 mL wereinjected into subcutaneously into the right flank of mice. Treat-ment was initiated once tumors had reached 1 mm3 (approxi-mately 2–3 weeks). Animals were euthanized once subcutaneous

tumors grew large enough to impede movement. Erlotinib (50 or75 mg/kg, Cayman Chemicals) was administered through oralgavage. Tissues frommice used for survival studies were extracted,paraffin-embedded, sectioned (5 mm) and analyzed usingimmunohistochemistry.

Hydrogel fabricationHA (�700 kDa, LifeCore Biomedical) disaccharides (�5%)

were modified with thiol groups via the carboxlic acid. RGDpeptide (GenScript) or free cysteine Sigma-Aldrich) were conju-gated to approximately 25% of maleimide groups on 20 kDa,4-arm PEG-maleimide (Laysan Bio, Inc.). Hydrogels were cross-linked viaMichael-type Addition bymixing thiolated HA, 20 kDaPEG-thiol (Laysan Bio), and PEG-maleimide dissolved in HEPESbuffer (pH 6.8). Linear compressive testing was performed usingan Instron 5564 material testing device (Instron).

GBM cell culturePrimaryGBM cell lines GBM39,HK301, andHK423were used.

HK301 and HK423 cells were generously provided by Dr. HarleyKornblum at UCLA. HK301 and HK423 lines were collected inApril 2010 and October 2013, respectively, with strict adherenceto UCLA Institutional Review Board protocol 10-000655 (25).GBM39 was collected in span of 1999–2006 (26). HK301 andHK423 cells were used between passages 15 and 25. GBM39 cellswere used at fewer than 15 passages. All cell lines were authen-ticated previously by short-tandem repeat analysis (27). Cellcultures routinely tested negative for mycoplasma contamination(Life Technologies, C7028). Cells (50,000/mL) were cultured inDMEM/F12 with 1xG21 (Gemini Bio), 1% penicillin/streptomy-cin, 50 ng/ml EGF (Peprotech), 20 ng/mL FGF-2 (Peprotech), and25 mg/mL heparin (Sigma-Aldrich).When gliomaspheres reachedaround 200 mmindiameter, theywere dissociated into single cellsin 1mLof TrypLE Express andfiltered through 70-mmcell strainer.For hydrogel cultures, dissociated cells were resuspended inpeptide-modified PEG-maleimide at 1 million cells/mL prior tomixing the HA-thiol/PEG-thiol to initiate crosslinking. Mediumwas replaced 4 days later. In some cases, 24 hours after encapsu-lation, hydrogel cultures were soaked in live/dead assay solution(Life Technologies L3224) for 30 minutes at room temperature.Hydrogel cultures were then placed on coverglass and imagedusing confocal microscopy (Leica SP5).

Drug treatmentEncapsulated single cells were cultured in hydrogels for 1 week

before treatment. Gliomasphere cultures were treated right afterdissociation, as previously described (28). Erlotinib was recon-stituted as a 10 mmol/L stock solution in dimethylsulfoxide(DMSO). Erlotinib was then diluted to 1 mmol/L in culturemedium. DMSO alone was used as a vehicle (i.e., negativecontrol). Cyclo-RGD was dissolved in PBS as 10 mmol/L stockthen dissolved in media as 50 mmol/L. Culture medium and drugwere replenished every third day.

Quantification of apoptosisCryopreserved hydrogel blocks were prepared and stained in

parallel for each experimental repeat (n ¼ 3 individual repeats)using an antibody against cleaved poly ADPpolymerase (c-PARP)andHoechst 33342 as a nuclear counterstain. At least four imagesfrom randomly chosen locations per sectionwere taken from least2 sections. Data were analyzed using ImageProPlus software. The

Glioblastoma Drug Resistance in 3D Cultures

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1359




area fraction of c-PARPþ toHoeschtþwas defined as percentage ofapoptotic cells. Only cells with nuclear colocalization of c-PARPand Hoescht were considered to be apoptotic.

For detailed procedures of hydrogel fabrication, hydrogelcryopreservation, Lentivirus preparation and transduction,mechanical characterization, Western blotting, immunohis-tochemistry and imaging, and flow cytometry, please refer toSupplementary Methods.

ResultsBrain microenvironment facilitates resistance to EGFRinhibition

To investigate how the unique brain microenvironment influ-ences physiology and treatment response of xenografted tumors,we transplanted patient-derivedGBM cells at either intracranial orsubcutaneous (dorsal flank) sites in nudemice.Once tumorswereestablished, mice were treated with either erlotinib or vehicle.Orthotopic transplants of both primary GBM cell lines (GBM39,

HK301) responded poorly to erlotinib (Fig. 1A; SupplementaryFig. S1A–S1C) despite its effectiveness in gliomasphere cultures(Supplementary Fig. S1D). Erlotinib treatment suppressed growthof intracranially xenografted GBM39 tumors for only 10 days,after which time the tumors failed to respond (Fig. 1A; Supple-mentary Fig. S1C). Inmice with orthotopically implanted HK301tumors, erlotinib had no detectable effect on tumor growth orsurvival (Supplementary Fig. S1A and S1B). In contrast, erlotinibtreatment inhibited growth of subcutaneously xenograftedGBM39 tumors for more than 200 days before tumors exhibitedacquired resistance (Fig. 1B). Tumors of HK301 cells could beestablished at orthotopic, but not subcutaneous, transplantationsite, further indicating that the subcutaneous tissue microenvi-ronment may not be amenable to GBM tumor growth.

Immunohistochemistry revealed HA presence surroundingand within intracranially xenografted tumors (Fig. 1C). Incontrast, HA deposition was only found within tumors, andnot surrounding tissue, in subcutaneous xenografts (Fig. 1C).Expression of phosphorylated EGFR (p-EGFR) and CD44

Figure 1.

Glioblastoma xenografts acquireresistance to erlotinib at intracranialsites with faster kinetics than atsubcutaneous sites. A,Bioluminescence imaging oforthotopic xenografts of GBM39 cells(normalized to day 0 before treatmentwith 50 mg/kg erlotinib). Error bars,SD (n ¼ 3). B, Volume ofsubcutaneously xenografted tumors ofGBM39 cells (normalized to day 0before treatment with 50 mg/kgerlotinib). Error bars, SD (n ¼ 4). C,Representative images ofimmunohistochemical staining for HA(brown, positive stain; purple,hematoxylin) of intracranial andsubcutaneous xenografts of GBM39cells. White dashed lines delineatetumor and surrounding tissues. Scalebars, 100 mm. Images of negativecontrols can be found inSupplementary Fig. S10A. D, Left,representative images ofimmunofluorescence staining forCD44 (red), p-EGFR (Tyr1068; green),and Hoechst 33342 (blue) ofintracranial and subcutaneousxenografts of GBM39 cells. Arrows,cells expressing both p-EGFR andCD44. Right, hematoxylin and eosin(H&E) images of adjacent sectionsfrom the same tissues. Images ofnegative controls can be found inSupplementary Fig. S10B. Scale bars,200 mm.

Xiao et al.

Cancer Res; 78(5) March 1, 2018 Cancer Research1360




remained high in intracranial xenografts regardless of treatment(Fig. 1D; Supplementary Fig. S2A). However, minimal expressionwas observed in subcutaneous xenografts (SupplementaryFig. S2A). Furthermore, after erlotinib treatment, CD44 andp-EGFR expression were nearly undetectable (Fig. 1D). Based onthese results, we posited that the ubiquitous abundance of HA inthe brain ECMmight contribute to faster acquisition of resistanceto EGFR inhibition. This hypothesis was further supported by theobservation that CD44 and p-EGFR were often coexpressed byGBM cells (Fig. 1D; Supplementary Fig. S2A and S2B).

Brain-mimetic hydrogels preserve in vivo phenotype ofGBM cells

Although it is clear that the brain ECM contributes to charac-teristic multi-drug resistance in GBM (8, 13, 21), specificmechan-isms have been difficult to uncover using conventional methods.To address current experimental limitations, we developed brain-mimetic biomaterials as 3D culture platforms to maintain thephysiology of patient-derived GBM cells. Hydrogel biomaterialswere fabricated so that multiple aspects of the local ECM—

including HA concentration, mechanical properties, and integ-rin-binding peptides—could be varied independently (Fig. 2A).

Given its potential importance for maintaining CD44expression and resistance to EGFR inhibition (Fig. 1), HA waschosen as a base for hydrogel fabrication. High molecular weightHA (�700 kDa), which induces CD44 clustering to achievedistinct biological effects from its low molecular weight counter-parts (29), was used to best mimic the native brain ECM. HA wasmodified with thiol groups to enable crosslinking via maleimidegroups on polyethylene glycol (PEG) macromers. Thiol groupswere conjugated to approximately 5% of HA disaccharidesthrough carboxylates on N-glucuronic acid (Fig. 2B).

Hydrogel mechanical properties were also selected to bestmimic native brain, which exhibits a linear compressive modulusaround 1 kPa and a shear elastic modulus around 200 Pa (16, 30,31). The concentration of HA was altered independently ofstiffness by substituting bioinert PEG-thiol for HA-thiol to main-tain a constant total polymer content and molar ratio of thiols tomaleimide groups for crosslinking (Fig. 2A). Figure 2C demon-strates that increasing total PEG concentration from0.5% (w/v) to1% (w/v) yields hydrogels that are twice as stiff—1 kPa (1173 �77 Pa) and 2 kPa (2160 � 48 Pa) linear compressive moduli,respectively—while keeping HA concentration constant at 0.5%(w/v). In addition, HA concentration was lowered to 0.1% (w/v)while maintaining a linear compressive modulus of 1 kPa (981�141 Pa). Hereafter, the mechanical properties of these hydrogelsare referred to as 1 kPa or 2 kPa.

As higher compressivemoduliwere achievedby increasing totalpolymer content, there remained apossibility that diffusionof keymolecules, such as erlotinib and growth factors, would varybetween softer and stiffer hydrogels and skew results. To evaluatethis possibility, diffusion of fluorescent dye-labeled dextrans ofvarying molecular weights through hydrogels was quantified(Fig. 2D). For 20 kDa and 70 kDa dextrans, effective diffusionthrough the 1 kPa and 2 kPa hydrogels were statistically equiv-alent. For all hydrogels formulations, diffusive equilibrium (i.e.,Mt/Minf) was reached by around 7 hours for 20 kDa and around11 hours for 70 kDa dextrans. An upper size limit was found at150 kDa, which did not diffuse into any of the hydrogels. Thisresult indicates that the hydrodynamic radius of 150 kDa dextranis larger than hydrogel mesh size. Results confirm that availability

of nutrients, FGF-2 and EGF (<15 kDa), and erlotinib (�300Da),was equivalent for all hydrogel cultures investigated.

GBM cell lines (HK301, GBM39, and HK423), isolated fromthree individual patients, were investigated. All three lineswere highly susceptible to erlotinib treatment when cultured asgliomaspheres (Supplementary Fig. S1D). Gliomaspheres weredissociated into single cells and suspended in hydrogel precursorsolution immediately prior to crosslinking. A live/dead assayperformed 24 hours after cell encapsulation confirmed thatthe majority of cells remained viable in all hydrogel conditions(Fig. 3A; Supplementary Fig. S3A). Using an EDU-based assay toquantify numbers of proliferating cells, we found that approxi-mately 20%more HK301 cells entered S-phase within a 2.5 hoursperiod when cultured in 3D thanwhen cultured as gliomaspheres(P < 0.05; Fig. 3B). Variations in HA content or compressivemodulus had no significant effects on cell proliferation.

Basal levels of CD44 were cell line dependent. For example,CD44was observedonHK423 cells in all conditions, butwas onlydetectable on HK301 cells when cultured in high HA-contenthydrogels. For all cell lines evaluated, culture in hydrogels withhigh HA (0.5% w/v) induced increased CD44 expression com-pared with culture in hydrogels with low HA (0.1% w/v) orgliomaspheres (Fig. 3C; Supplementary Fig. S4A–S4C). Thesefindings agree with our in vivo results, where murine xenograftsrobustly express CD44 when seeded in the HA-rich brain, but notat subcutaneous sites. (Fig. 1D; Supplementary Fig. S2). Hydrogelmodulus had no effects on CD44 expression (Fig. 3C).

GBM cells in brain-mimetic hydrogels rapidly acquire drugresistance

Effects of erlotinib treatment on GBM cells cultured in hydro-gels or gliomaspheres were investigated. To mimic brain tissue,hydrogels with 0.5% (w/v) HA and 1 kPa compressive moduluswere used (30, 32). To characterize the cytotoxic effects of erlo-tinib, numbers of cells positive for nuclear cleaved poly ADPpolymerase (c-PARP) were counted (Fig. 4A). By the 6th day oftreatment,GBMcells culturedwithinhydrogels displayed levels ofapoptosis indistinguishable from nontreated controls (Fig. 4B).Gliomasphere cultures had significantly more apoptotic cellswhen treated with erlotinib (Fig. 4B). In contrast to gliomaspherecultures, after 12 days of treatment, GBM cells cultured withinhydrogels also had acquired resistance to the cytostatic effects oferlotinib (Fig. 4C). In addition to erlotinib, acquisition of resis-tance to EGFR inhibition in hydrogels was observed when cellswere treated with lapatinib (Supplementary Fig. S5A).

Further investigation confirmed nearly complete inhibition ofwild-type EGFR phosphorylation (p-wtEGFR) in gliomaspherecultures by day 3 of erlotinib treatment (Fig. 5A and B). In contrast,erlotinib only partially inhibited p-wtEGFR in hydrogel-culturedcells. Furthermore, erlotinib treatment increased total wtEGFRexpression in both hydrogel and gliomasphere cultures (Fig. 5AandB).Given thehigh rateof cell death in erlotinib-treated, lowHA(0.1%w/v) hydrogel cultures, we were not able to obtain sufficientprotein lysate toperformWesternblots.However, immunostainingdemonstrated thathigherHAcontent inhydrogels corresponded toincreased p-EGFR expression in cultured cells (Supplementary Fig.S4). CD44 expression also increased with HA content and, likexenografted tumors (Fig. 1D, Supplementary Fig. S2), generallyoverlapped areas of p-EGFR expression (Supplementary Fig. S3).

While HK423 cells express only wtEGFR and not the truncatedand constitutively activate mutant, EGFRvIII, HK301 cells express






Figure 2.

Fabrication and characterization of biomaterial platforms. A, Schematic of hydrogel encapsulation of GBM cells for 3D culture. Cysteine-bearing peptideswere first conjugated to PEG maleimide. Single GBM cells were resuspended in PEG maleimide-peptide before mixing with PEG thiol and 5% thiolated HA.Molar ratio of thiol to maleimide was maintained at approximately 1.1 to 1. B, Representative H1-NMR spectrum of thiolation, indicating that approximately5% of HA glucuronic acid groups have been modified with a thiol. C, Linear compressive moduli of hydrogels. Percentages indicate weight to volume ratios (w/v).Error bars, SEM (n ¼ 3). One-way ANOVA with Tukey test for multiple comparisons was performed (��, P < 0.001). D, Left, effective diffusion coefficients (cm2/s)for 20 kDa and 70 kDa dextrans, respectively, through hydrogels. Two-way ANOVA with Tukey test for multiple comparisons was performed (�� , P < 0.001).Error bars, SEM (n ¼ 3). Right, diffusion over time of dextran through HA hydrogels. Mt/Minf is defined as the ratio of dextran released at a specific time (Mt)to the total amount of dextran released at infinite time (Minf). Mt/Minf is plotted against the square root of time (s1/2) so that the slope indicates diffusion rate.HA percentage indicates volume to weight percentage (% w/v). ns, nonsignificant.

Xiao et al.





bothwtEGFRandEGFRvIII andGBM39 cells express only EGFRvIII(Supplementary Fig. S5B). Although erlotinib-treated HK301 cellscultured in gliomaspheres or hydrogels upregulated total expres-sion of wtEGFR (Fig. 5A and B), only in hydrogel cultures was p-EGFRvIII increased. Likewise for HK423 cells, erlotinib treatmentinduced higher total wt-EGFR levels in hydrogel and gliomasphere.While erlotinib treatment did attenuate p-wtEGFR in HK423 cellscultured in hydrogels or gliomaspheres, this effect was only partialin hydrogel cultures. As with HK301 cells, GBM39 cells cultured inHAhydrogels increased p-EGFRvIII (Fig. 5A andB).While erlotinibtreatment reduced p-EGFRvIII, levels remained higher than treatedgliomaspheres. In all three cell lines when treated, p-EGFR levelswere always significantly higherwhencultured inHAhydrogel thangliomasphere cultures (Fig. 5A and B).

Downstream pathways of EGFR include PI3K–AKT andMAPK–ERK, both of which many studies have reported main-tain survival and growth potential of GBM tumors (3). Thus, wecharacterized the effects of erlotinib treatment on phosphory-lation of AKT and ERK1/2 (Fig. 5A and C). For all threecell lines, culture in HA hydrogels increased p-AKT levelscompared with gliomaspheres. While p-AKT levels were notaltered significantly in erlotinib-treated gliomaspheres, p-AKTlevels significantly increased in hydrogel-cultured HK301 andHK423 cells (Fig. 5A and C). In HK301 gliomaspheres, erloti-nib treatment significantly decreased p-ERK levels. Althoughnot statistically significant, a similar trend was observed forHK423 and GBM39 gliomaspheres (Fig. 5A and C). In allhydrogel cultures, erlotinib treatment had no significant effects

Figure 3.

Characterization of patient-derived GBM cells in 3D hydrogel culture. A, Representative confocal microscopy images showing live (green) and dead (red)cells 24 hours after hydrogel encapsulation of HK301 cells. Scale bar, 100 mm. Quantification of the percentage of viable cells (HK301) 24 hours afterencapsulation is shown on the lower right. Error bar, SEM (n¼ 3). One-way ANOVAwith Tukeymultiplicity test was performed. NS, nonsignificant.B, Proliferation ofHK301 cells in hydrogel and gliomasphere cultures. On the fourth day of culture, encapsulated cells were pulsed with EdU (1 mmol/L) for 2.5 hours. Cellswere removed from hydrogels and percentage of cells that had proliferated (EdUþ), assessed using a flow cytometer. One-way ANOVAwith Tukey test for multiplecomparisons was performed (� , P < 0.05). C, Representative images of immunofluorescence staining for CD44 (red) in cryosectioned hydrogel andgliomasphere cultures of HK301 cells. Scale bar, 100 mm. HA percentage indicates volume to weight percentage (% w/v).






on p-ERK, and p-ERK levels were significantly higher than inerlotinib-treated gliomaspheres (Fig. 5A and C).

Biomaterials to quantify effects of ECM cues on drug resistanceFigures 1 to 4 suggest that CD44 expression and HA content

support the ability of GBM cells to gain erlotinib resistance.Others have reported that the mechanical microenvironmentcontributes to GBM tumor progression (16, 24). Unlike glioma-spheres, biomaterial platforms also provide a defined, 3Dmechanical environment to cultured cells. Thus, we explored thecooperative influence of hydrogel compressive modulus and HAcontent on acquisition of resistance to EGFR inhibition viaerlotinib. Importantly, mechanical modulus was varied indepen-dently of HA concentration, so that the individual and combinedcontributions of each were experimentally decoupled.

Patient-derivedGBMcells (HK301)were cultured inHAhydro-gels to characterize the independent effects of HA concentrationand compressive modulus on response to erlotinib. Total num-bers of cells were tracked using bioluminescence imaging of livecultures transduced to constitutively express firefly luciferase.Results demonstrate that cells cultured in 3D hydrogels with ahigher HA concentration (0.5% w/v) and lower compressivemodulus (1 kPa) gained resistance to erlotinib by the ninth dayof treatment (Fig. 6A). By the 15th day of treatment, there weremore total cells in erlotinib-treated than untreated cultures,indicating that erlotinib-resistant cells proliferate faster. GBMcells cultured in hydrogels with high HA content (0.5% w/v),but with a stiffer modulus (2 kPa), also acquired some resistanceto erlotinib; however, not until the 12th day of treatment. Inaddition, cell numbers in treated cultures remained only approx-imately 50%of those in nontreated cultures after 15 days. Finally,GBM cells cultured in soft hydrogels (1 kPa) with a low HAconcentration (0.1% w/v) did not acquire erlotinib resistancewithin 15 days. Instead their response was comparable with thatof gliomasphere cultures, with bioluminescence signals close to

background on the 15th day of treatment. Furthermore, minimalHA was observed in cultured gliomaspheres (Supplementary Fig.S6A). Together, these results indicate that high HA content wasrequired for acquisition of resistance.

Cytotoxic and cytostatic effects of erlotinib treatment onGBM cells cultured in hydrogels were evaluated. Erlotinib-treatedcells cultured in soft hydrogels (1 kPa) with high HA content(0.5% w/v) proliferated significantly faster than their untreatedcontrols at the 3- and 6-day time points (Fig. 6B). This increase inproliferation correlated to the increased total numbers of viablecells observed by the 12th day of treatment (Fig. 6A). Erlotinibtreatment also induced a slight increase in cell proliferation on thethird day in other hydrogel conditions (Fig. 6B). While prolifer-ation had decreased by the 6th day of treatment in all conditions,it remained elevated in 3D hydrogel cultures compared withgliomaspheres. Finally, only GBM cells cultured in soft, high HAhydrogels had escaped the cytotoxic effects of erlotinib on the 6thday of treatment (Fig. 6C). Notably, the kinetics of resistanceacquisition to erlotinib of GBM cells cultured in soft, high HAhydrogels (Fig. 6) were comparable with those observed inpatient-matched orthotopic xenografts in mice (Fig. 1A).

To further investigate the role of CD44, we used shRNA len-tivirus to knockdown CD44 expression (Supplementary Fig. S6B)and repeated erlotinib-treatment experiments. Lack of CD44mitigated both cytotoxic and cytostatic resistance to erlotinib(Fig. 6D and E). Despite restoration of erlotinib efficacy for thefirst 6 days of treatment, even cells lacking CD44 expressiongained resistance to the cytostatic effects of erlotinib by the 12thday of treatment (Fig. 6E; Supplementary Fig. S7A and S7B).Although CD44 is a major receptor for HA, other HA receptors,such as CD168, may act to compensate for lost CD44 activity andfacilitate delayed acquisition of erlotinib resistance. In soft, highHA hydrogel cultures, we observed a unique pattern of CD168expression around the edges of cell masses resembling that ofCD44 (Supplementary Fig. S8A). In low HA hydrogels, CD168

Figure 4.

GBM cells in 3D, HA hydrogel culturesacquire cytotoxic and cytostaticresistance to erlotinib. A,Representative images ofimmunofluorescence staining ofc-PARP in HK301 cells after 6 days oferlotinib treatment. Scale bar, 200 mm.B, Quantification of apoptotic cells(c-PARPþ) after 6 days of erlotinibtreatment. Error bars, SEM (n ¼ 3).Student t tests were performed (� , P <0.05; �� , P < 0.01). C, Proliferation rateof cells (EdU incorporation over 2.5hours) after 12 days of erlotinibtreatment. Erlotinib-treated sampleswere normalized to nontreatedsamples for each condition. Student ttest was performed (�� , P <0.01). Errorbars, SEM (n ¼ 3).

Xiao et al.





Figure 5.

Signaling analysis for hydrogel- and gliomasphere (GS)-cultured cells. A, Representative Western blot images of 72 hours after erlotinib treatment. HApercentage indicates volume to weight (%w/v). B and C,Quantification (integrated intensity signals) of phospho-EGFR (B), phospho-AKT and phospho-ERK1/2 (C)forWestern blots. Error bars, SD from independent repeats [HK301 (n¼ 5), HK423 (n¼4), andGBM39 (n¼ 3)]. One-wayANOVAandTukeymultiple comparison testwere used (� , P < 0.05; �� , P < 0.01; �� , P < 0.001; �� , P < 0.0001). ns, nonsignificant.






Figure 6.

HA content and modulus contribute to kinetics of acquisition of erlotinib resistance. A, Chemiluminescent signal measured 2 hours after addition ofD-luciferin (1 mmol/L). Signals of erlotinib-treated HK301 cells were normalized to nontreated samples and signal before treatment (day 0) for each condition.Two-way ANOVA (culture condition, time) with Šid�ak test for multiple comparisons of hydrogel condition against gliomasphere culture was performed. Error bars,SD (n ¼ 3). B, Proliferation rate of HK301 cells (EdU incorporation over 2.5 hours) during erlotinib treatment. Erlotinib-treated samples were normalized tonontreated samples for each condition. Error bars, SEM (n ¼ 3). Two-way ANOVA (culture condition, time) with Šid�ak test for multiple comparisons of hydrogelcondition against gliomasphere culture were performed. C and D, Percentage of apoptotic cells (c-PARPþ HK301 cells) after 6 days of erlotinib treatment.Error bars, SEM (n ¼ 3). One-way ANOVA with Tukey test for multiple comparisons was performed. Representative images of c-PARP staining can be found inSupplementary Fig. S11A. E, Proliferation rate of HK301 cells (EdU incorporation over 2.5 hours) during erlotinib treatment. Erlotinib-treated samples werenormalized to nontreated samples for each condition. Error bars, SEM (n ¼ 3). Two-way ANOVA (cell type, time) with Šid�ak test for multiple comparisons wereperformed. HA percentage indicates volume to weight percentage (% w/v). �, P < 0.05; �� , P < 0.01; �� , P < 0.0001.

Xiao et al.





expression was confined to the nucleus, where it participates information of mitotic spindles (33).

RGD and HA cooperate to evade erlotinib-induced apoptosisLike CD44, integrins can relay both biochemical and mechan-

ical cues through activation of FAK–PI3K–AKT and MAPK-ERKpathways–all previously implicated as mediators of resistance totreatment in GBM (16, 23, 24). To investigate cooperative inter-actions with HA-bound CD44, the ubiquitous integrin-bindingsequence RGD was incorporated into hydrogel platforms(Fig. 2A). First, incorporation of RGD peptides into in softhydrogels (1 kPa) with high HA content (0.5% w/v) facilitatedGBMcell spreading out of spheroidalmasses into the surroundinghydrogel (Fig. 7A; Supplementary Fig. S9A). Next, we investigatedhow RGD incorporation into hydrogels affected cytotoxic effectsof erlotinib treatment. On the third day of treatment, GBM cellscultured in high HA (0.5% w/v) hydrogels with RGD providedsignificant protection against erlotinib-induced apoptosis com-pared with those in high HA hydrogels without RGD or low HA(0.1%w/v) hydrogelswith RGD(Fig. 7B). These results imply thatengagement of integrin and HA receptors cooperate to amplifyresistance to EGFR inhibition.

To investigate downstream effects of integrin engagement, weinvestigated phosphorylation of zyxin and FAK – prominentsignaling proteins associated with integrin activation. When cul-tured in high HA hydrogels containing RGD, GBM cells upregu-lated p-zyxin (Fig. 7C). This result is not unexpected given the roleof zyxin in integrin-mediated cell spreading and migration in 3Dculture (34) and the invasive morphology of cells cultured in3D hydrogels (Fig. 7A). However, erlotinib treatment did notaffect p-zyxin levels in HK301 or HK423 cells (Fig 7C; Supple-mentary Fig. S9B).

Integrin activation of FAK is thought to facilitate cancer cellresistance to drug-induced apoptosis (15). Levels of p-FAK weresimilar in untreated cultures in HA hydrogels with or withoutRGD. However, when cultured in hydrogels containing RGD,erlotinib treatment increased p-FAK activity (Fig. 7C; Supplemen-tary Fig. S7B). To further confirm that apoptotic resistance wasmediated by cell-RGD interactions, cyclo-RGD was used as aninhibitor (17). Addition of cyclo-RGD effectively reversed cellspreading (Fig. 7D) and reduced p-zyxin (Fig. 7E; SupplementaryFig. S9B). When treated with erlotinib and cyclo-RGD, p-FAKlevels were comparable with nontreated cells (Fig. 7E; Supple-mentary Fig. S9B). Moreover, this combined treatment reversedthe ability of RGD to rescue cells from erlotinib-induced apopto-sis (Fig. 7F).

DiscussionTherapeutic resistance plays a critical role in GBM lethality.

However, preclinical studies have inadequately accounted for theinfluence of the unique properties of brain tissue. Here, we reporta biomaterial platform for 3D culture of primary GBM cells thatmimics native brain tissue. In agreement with previous reports,our investigations found that the ECM surrounding brain tumorsis enriched inHA (35).Moreover, our data demonstrate that GBMtumors xenografted at non-CNS anatomical sites, which containless HA in their ECM, acquire resistance to RTK inhibition on asignificantly longer time scale (Fig. 1). Orthogonal tuning ofbiomaterial features revealed that an HA-rich, mechanically softculture environment is required for GBM cells to acquire resis-

tance to RTK inhibition (Fig. 6A). Furthermore, 3D cultures ofpatient-derived GBM cells in biomaterials with defined HA con-tent and mechanical properties rapidly developed resistance toerlotinib in a manner consistent with patient-matched mousexenografts (Figs. 1 and 6; Supplementary Fig. S1A) and clinicalreports (6). Specifically, in both orthotopic animal and high HAhydrogel experimental settings, GBM cells acquired erlotinibresistance between 9 and 12 days of treatment. Taken together,results demonstrate the utility of these biomaterial cultures asex vivo models of GBM that better recapitulate the brain micro-environment than standard culture methods yet are easier, moreaffordable, less time consuming (days vs. weeks to establishtumors) and provide a more controlled experimental contextthan animal models.

For all primary GBM cell lines evaluated, changes in cytotoxicand cytostatic effects of erlotinib over time were consistent withacquisition of resistance—where despite an initial inhibition ofp-EGFR, treatment reduced apoptosis while increasing prolif-eration (Figs. 4 and 5). 3D culture in HA-rich hydrogels aloneincreased p-AKT, while erlotinib treatment further upregulatedp-AKT levels (Fig. 5A and C). This finding indicates that GBMcells resistant to EGFR inhibition may be more aggressive—apossible explanation of increased proliferation in HK301 aftertreatment in high HA hydrogel (Fig. 5A and C and Fig. 6B). Inaddition to EGFR inhibition via the small molecule erlotinib,hydrogel-cultured GBM cells gained resistance to lapatinib—another RTK inhibitor—providing evidence that culture ofprimary GBM cells in 3D brain-mimetic biomaterials representa clinically relevant method for evaluate drug response (Sup-plementary Fig. S5A).

Here, we implemented several key improvements over previ-ously reportedHA-based biomaterials for 3D cell culture (24, 36–40) that enabled development of culture platforms representing acompelling new preclinical model for studying mechanisms ofdrug resistance in brain cancers. First, HAwasminimallymodified(�5% of disaccharides contain a thiol substitution) to maintainthe native ability of high molecular weight HA to interact withCD44 receptors. In contrast, other methods often modify up to70% of HA disaccharides.

Second, while previous methods have relied on HA polysac-charides with molecular weights at or below 200 kDa, we incor-poratedHAwith a range ofmolecular weights from500–750 kDa.This size difference has significant effects on HA bioactivity(29, 41). For example, while high molecular weight HA (500–1,000 kDa), which is found in abundance in native brain, sup-presses the immune system, 200 kDa HA stimulates cytokineproduction (29). High molecular weight HA has also beenreported to activate RTKs more efficiently than smaller HA chains(41). Although someHA is detected in subcutaneously implantedGBM tumors, it is thought that tumors contain high amounts oflow molecular weight of HA, which contribute to tumor growthand angiogenesis (42).

Third, the biomaterial platforms reported here achieved effec-tive decoupling of HA content, stiffness, availability of integrin-binding sites and diffusion. Previously reported methods for HAhydrogel fabrication typically increased hydrogel stiffness byincreasing HA concentration, and thus total polymer content(30, 37). Another common method substitutes gelatin to varyHA concentration without altering total polymer content; how-ever, gelatin contains RGD sites (39, 40). In our system, orthog-onal control of these variables enabled systematic investigation of






Figure 7.

Interactions of integrins and CD44 with the scaffold protect glioblastoma cells from erlotinib-induced apoptosis (HK301 cells). A, Representative phasecontrast images of hydrogel-cultured cells 8 days after encapsulation. Scale bar, 200 mm. B, Percentage of apoptotic cells (c-PARPþ) after 3 days of erlotinibtreatment. Error bars, SEM (n ¼ 3). Two-way ANOVA (hydrogel condition, treatment) with Šid�ak test for multiple comparisons was performed (�� , P < 0.001). C,Right, representative Western blot images 6 days after erlotinib treatment. HK301 was encapsulated in different gel types. Left, normalized integrated intensitysignals of phospho-FAK, total-FAK, and phospho-zyxin. Error bars, SD across individual experimental repeats (n ¼ 4). D, Representative phase contrast images 3days after treatment with cyclo-RGD (50 mmol/L). E, Right, representative Western blot image (HK301 cells after 6 days of treatment). Left, normalized integratedintensity signals of phospho-FAK and phospho-zyxin. Error bar, SD across individual experimental repeats (n ¼ 4). F, Percentage of apoptotic cells (c-PARPþ)3 days after cyclo-RGD treatment. Error bars, SEM (n¼ 3). One-way ANOVA with Tukey test for multiple comparisons was performed (� , P < 0.05; �� , P < 0.01). HApercentage indicates volume to weight (% w/v). E, 1 mmol/L erlotinib; ER, 1 mmol/L erlotinib and 50 mmol/L cyclo-RGD; R, 50 mmol/L cyclo-RGD; V, vehicle(DMSO or PBS). Representative images of c-PARP staining can be found in Supplementary Fig. S11B. ns, nonsignificant.

Xiao et al.





how individual features interact to facilitate acquisition of treat-ment resistance.

Finally, the majority of previous studies of GBM in 3D bio-materialmodels have explored only immortalized cells lines, suchas U87 cells, which likely have significant phenotypic deviationsfrom primary, patient-isolated GBM cells (43). Although logisti-cally more challenging, primary GBM cells—as used here—aremore likely to yield clinically translatable findings. Moreover,compatibility of an ex vivo experimental platform with primaryGBM cells isolated from multiple patients will facilitate futureapplication to personalized medicine.

The lack of CD44 expression in hydrogel cultures with lowerHA content (Fig. 3C; Supplementary Fig. S4) indicates HA mayinduce upregulation of CD44 receptors that can then respondto mechanical cues. Previous studies have shown that mechan-ical cues are transduced through the PI3K–AKT pathway,despite EGFR inhibition (22). This may explain why EGFRresistance is more pronounced in GBM cultures with compa-rable HA levels and CD44 expression, but different mechanicalmoduli. As knockdown of CD44 restored the cytostatic andcytotoxic effects of erlotinib, we are confident that HA-CD44interactions contributed to acquisition of erlotinib resistance(Fig. 6D and E). However, cytostatic effects were lost overtime—implying that GBM cells eventually acquired resistancethrough a CD44-independent mechanism. The observationthat the CD168 receptor was highly expressed at the cellmembrane in these cultures (Supplementary Fig. S8A) suggeststhat CD168-HA interactions may compensate for the loss ofCD44 to permit acquisition of erlotinib resistance.

Our results are in agreement with previous reports thatHA-bound CD44 facilitates activation of wtEGFR, and thusresistance to EGFR inhibition (Fig. 5A and B; SupplementaryFig. S4; refs. 22, 41, 44). The observations that areas of CD44and p-EGFR expression overlap in GBM cells cultured inHA-rich hydrogels (Fig. 3C; Supplementary Fig. S4) and tumorsxenografted in HA-rich brain (Fig. 1D; Supplementary Fig. S2)indicate that HA-bound CD44 may increase activation of EGFRthrough physical interactions at the cell membrane and facil-itate resistance to EGFR inhibition. This has been previouslyreported to occur in orthotopic xenografts and clinical samples(22, 44). CD44 may also affect activation of EGFRvIII, acommon variant in clinical tumors associated with resistanceto EGFR inhibition and worse patient outcomes (2, 28, 45).Notably, RGD peptides acted synergistically with HA in hydro-gels to induce cell spreading and protect GBM cells fromerlotinib-induced apoptosis (Fig. 7B). These results imply thata combination therapy of integrin, CD44 and EGFR inhibitionmay have clinical potential.

Given the complexity of GBM tumors in vivo—including pow-erful cooperative mechanisms and the presence of confoundingvariables such as the blood-brain barrier—it has been challengingto isolate the contributions of individual ECM features usinganimal models. On the other hand, standard in vitro culturemethods do not account for key features of the brain ECM thatare crucial to preserving tumor physiology and obtaining exper-imental results with clinical relevance. Here, we describe bioma-terial platforms that recapitulate the brain microenvironmentto produce ex vivo cultures of primary GBM cell lines withunique genetic and phenotypic profiles that are physiologicallyrepresentative of clinical tumors. Specifically, mechanisms and

kinetics of acquisition of resistance to EGFR inhibition werepreserved in biomaterial, but not in standard gliomasphere cul-tures. Compared with animal models, these biomaterial scaffoldsprovide researchers with a platform in which to perform highlycontrolled experiments faster, cheaper andmore reproducibly. Inaddition, scaffolds are optically transparent—permitting imagingof 3D cultures—and compatible with standard techniques fortissue processing—including sectioning and histological staining.The ability to independently vary individual parameters withinthe ECM enables characterization of how multiple ECM cues acttogether to facilitate acquisition of treatment resistance andamplify aggressive characteristics. Here, this function was usedto demonstrate how mechanical modulus, HA content and RGDpeptides mediate acquisition of resistance to RTK inhibitionthrough cooperative interactions among HA, CD44, integrins,and EGFR. In conclusion, these biomimetic scaffolds with orthog-onal control over ECM parameters provide a unique tool forresearchers to better understand how the complex microenviron-ment in GBM tumors fuels treatment resistance and cancerprogression.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: W. Xiao, D.A. Nathanson, S.K. SeidlitsDevelopment of methodology: W. Xiao, R. Zhang, A. Sohrabi, A. Ehsanipour,S. Sun, J. Liang, C.M. Walthers, S.K. SeidlitsAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.):W. Xiao, R. Zhang, A. Sohrabi, A. Ehsanipour, S. Sun,L. Ta, S.K. SeidlitsAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis):W. Xiao, R. Zhang, A. Sohrabi, L. Ta, D.A. Nathanson,S.K. SeidlitsWriting, review, and/or revision of themanuscript:W. Xiao, C.M.Walthers, L.Ta, D.A. Nathanson, S.K. SeidlitsAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): W. Xiao, A. Ehsanipour, L. Ta, S.K. SeidlitsStudy supervision: C.M. Walthers, S.K. Seidlits

AcknowledgmentsThis work was supported with funding from the NIH (R21NS093199)

and the UCLA ARC 3R's Award. Our sincerest thanks go to the lab of Dr.Harley Kornblum for provision of the HK301 and HK423 cell lines and thelab of Dr. Benjamin Wu for use of the Instron instrument for linearmechanical testing. We also thank UCLA Tissue Pathology Core Laboratory(TPCL) for cryosectioning, Advanced Light Microscopy/Spectroscopy corefacility (ALMS) in California Nanosystems Institute (CNSI) at UCLA for useof the confocal microscope, UCLA Crump Institute for Molecular Imagingfor using IVIS imaging system, UCLA Molecular Instrumentation Center(MIC) for providing magnetic resonance spectroscopy, and Flow CytometryCore in Jonsson Comprehensive Cancer Center (JCCC) at UCLA for pro-viding instrumentation for flow cytometry. The monoclonal antibodyagainst CD44 (clone H4C4, developed by August, J.T./Hildreth, J.E.K.) wasobtained from the Developmental Studies Hybridoma Bank, created byNICHD of the NIH and maintained at the University of Iowa, Departmentof Biology, Iowa City, IA 52242.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 12, 2017; revised November 15, 2017; accepted December19, 2017; published OnlineFirst December 27, 2017.






References1. Ostrom QT, Gittleman H, Farah P, Ondracek A, Chen Y, Wolinsky Y, et al.

CBTRUS statistical report: primary brain and central nervous systemtumors diagnosed in the United States in 2006–2010. J Neurooncol2013;15:788–96.

2. Furnari FB, Cloughesy TF, Cavenee WK, Mischel PS. Heterogeneity ofepidermal growth factor receptor signalling networks in glioblastoma.Nat Rev Cancer 2015;15:302–10.

3. Taylor TE, Furnari FB, Cavenee WK. Targeting EGFR for treatment ofglioblastoma: molecular basis to overcome resistance. Curr Cancer DrugTargets 2012;12:197–209.

4. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, Tan EH, Hirsh V,Thongprasert S, et al. Erlotinib in previously treated non–small-cell lungcancer. N Engl J Med 2005;353:123–32.

5. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, et al.Use of chemotherapy plus a monoclonal antibody against HER2 formetastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344:783–92.

6. Brandes AA, Franceschi E, Tosoni A, Hegi ME, Stupp R. Epidermal growthfactor receptor inhibitors in neuro-oncology: hopes and disappointments.Clin Cancer Res 2008;14:957–60.

7. Bernstein JJ,WoodardCA.Glioblastoma cells donot intravasate into bloodvessels. Neurosurgery 1995;36:124–32.

8. Gilg AG, Tye SL, Tolliver LB, Wheeler WG, Visconti RP, Duncan JD, et al.Targeting hyaluronan interactions in malignant gliomas and their drug-resistant multipotent progenitors. Clin Cancer Res 2008;14:1804–13.

9. Tilghman J, Wu H, Sang Y, Shi X, Guerrero-Cazares H, Quinones-HinojosaA, et al.HMMRmaintains the stemness and tumorigenicity of glioblastomastem-like cells. Cancer Res 2014;74:3168–79.

10. Pietras A, Katz AM, Ekstr€om EJ, Wee B, Halliday JJ, Pitter KL, et al.Osteopontin-CD44 signaling in the glioma perivascular niche enhancescancer stem cell phenotypes and promotes aggressive tumor growth. CellStem Cell 2014;14:357–69.

11. Lathia JD, Mack SC, Mulkearns-Hubert EE, Valentim CLL, Rich JN. Cancerstem cells in glioblastoma. Genes Dev 2015;29:1203–17.

12. Ranuncolo SM, Ladeda V, Specterman S, Varela M, Lastiri J, Morandi A,et al. CD44 expression in human gliomas. J Surg Oncol 2002;79:30–6.

13. Gladson CL. The extracellular matrix of gliomas: modulation of cellfunction. J Neuropathol Exp Neurol 1999;58:1029–40.

14. Mahesparan R, Read T-A, Lund-Johansen M, Skaftnesmo K, Bjerkvig R,Engebraaten O. Expression of extracellular matrix components in a highlyinfiltrative in vivo glioma model. Acta Neuropathol 2003;105:49–57.

15. Guo W, Giancotti FG. Integrin signalling during tumour progression. NatRev Mol Cell Biol 2004;5:816.

16. Miroshnikova YA,Mouw JK, Barnes JM, PickupMW, Lakins JN, KimY, et al.Tissuemechanics promote IDH1-dependentHIF1a-tenascin C feedback toregulate glioblastoma aggression. Nat Cell Biol 2016;18:1336–45.

17. ReardonDA, Neyns B, Weller M, Tonn JC, Nabors LB, Stupp R. Cilengitide:an RGDpentapeptideanb3 andanb5 integrin inhibitor in development forglioblastoma and other malignancies. Futur Oncol 2011;7:339–54.

18. Wang SJ, Bourguignon LYW. Hyaluronan and the interaction betweenCD44 and epidermal growth factor receptor in oncogenic signaling andchemotherapy resistance in head and neck cancer. Arch Otolaryngol NeckSurg 2006;132:771–8.

19. Lee J-L, Wang M-J, Sudhir P-R, Chen J-Y. CD44 engagement promotesmatrix-derived survival through the CD44-SRC-integrin axis in lipid rafts.Mol Cell Biol 2008;28:5710–23.

20. Lee J-L, WangM-J, Sudhir P-R, Chen G-D, Chi C-W, Chen J-Y. Osteopontinpromotes integrin activation through outside-in and inside-out mechan-isms: OPN-CD44V interaction enhances survival in gastrointestinal cancercells. Cancer Res 2007;67:2089–97.

21. Zustiak SP, Dadhwal S,Medina C, Steczina S, Chehreghanianzabi Y, AshrafA, et al. Threedimensional matrix stiffness and adhesive ligands affectcancer cell response to toxins. Biotechnol Bioeng 2016;113:443–52.

22. Herishanu Y, Gibellini F, Njuguna N, Hazan-Halevy I, Keyvanfar K, Lee E,et al. CD44 signaling via PI3K/AKT andMAPK/ERK pathways protects CLLcells from spontaneous and drug induced apoptosis through MCL-1. LeukLymphoma 2011;52:1758.

23. Chopra A, Murray ME, Byfield FJ, Mendez MG, Halleluyan R, Restle DJ,et al. Augmentation of integrin-mediated mechanotransduction by hya-luronic acid. Biomaterials 2014;35:71–82.

24. Kim Y, Kumar S. CD44-mediated adhesion to hyaluronic acid contributesto mechanosensing and invasive motility. Mol Cancer Res 2014;12:1416–29.

25. Laks DR, Crisman TJ, Shih M, Mottahedeh J, Gao F, Sperry J, et al. Large-scale assessment of the gliomasphere model system. Neuro Oncol2016;18:1367–78.

26. Sarkaria JN, Carlson BL, Schroeder MA, Grogan P, Brown PD, Giannini C,et al. Use of an orthotopic xenograft model for assessing the effect ofepidermal growth factor receptor amplification on glioblastoma radiationresponse cancer therapy : preclinical epidermal growth factor receptoramplification on glioblastoma radiation respon. Clin Cancer Res 2006;12:2264–71.

27. Mai WX, Gosa L, Daniels VW, Ta L, Tsang JE, Higgins B, et al. Cytoplasmicp53 couples oncogene-driven glucose metabolism to apoptosis and is atherapeutic target in glioblastoma. Nat Med 2017;1342–51.

28. Nathanson DA, Gini B, Mottahedeh J, Visnyei K, Koga T, Gomez G, et al.Targeted therapy resistance mediated by dynamic regulation of extrachro-mosomal mutant EGFR DNA. Science (80-) 2014;343:72–6.

29. Stern R, Asari AA, Sugahara KN. Hyaluronan fragments: an information-rich system. Eur J Cell Biol 2006;85:699–715.

30. Seidlits SK, Khaing ZZ, Petersen RR, Nickels JD, Vanscoy JE, Shear JB, et al.The effects of hyaluronic acid hydrogels with tunable mechanical proper-ties on neural progenitor cell differentiation. Biomaterials 2010;31:3930–40.

31. Georges PC, Miller WJ, Meaney DF, Sawyer ES, Janmey PA. Matrices withcompliance comparable to that of brain tissue select neuronal over glialgrowth in mixed cortical cultures. Biophys J 2006;90:3012–8.

32. Armstrong SE, Bell DR. Measurement of high-molecular-weight hyalur-onan in solid tissue using agarose gel electrophoresis. Anal Biochem2002;308:255–64.

33. Telmer PG, Tolg C, McCarthy JB, Turley EA. How does a protein with dualmitotic spindle and extracellular matrix receptor functions affect tumorsusceptibility and progression? Commun Integr Biol 2011;4:182–5.

34. Fraley SI, Feng Y, Giri A, Longmore GD, Wirtz D. Dimensional andtemporal controls of three-dimensional cell migration by zyxin and bind-ing partners. Nat Commun 2012;3:719.

35. Jadin L, Pastorino S, Symons R, Nomura N, Jiang P, Juarez T, et al.Hyaluronan expression in primary and secondary brain tumors. AnnTranslMed 2015;3:80.

36. Ananthanarayanan B, Kim Y, Kumar S. Elucidating themechanobiology ofmalignant brain tumors using a brain matrix-mimetic hyaluronic acidhydrogel platform. Biomaterials 2011;32:7913–23.

37. Heffernan JM,OverstreetDJ, Le LD, VernonBL, Sirianni RW.Bioengineeredscaffolds for 3D analysis of glioblastoma proliferation and invasion. AnnBiomed Eng 2015;43:1965–77.

38. Wang C, Tong X, Yang F. Bioengineered 3Dbrain tumormodel to elucidatethe effects of matrix stiffness on glioblastoma cell behavior using PEG-based hydrogels. Mol Pharm 2014;11:2115–25.

39. Pedron S, Harley BAC. Impact of the biophysical features of a 3D gelatinmicroenvironment on glioblastoma malignancy. J Biomed Mater Res PartA 2013;101:3404–15.

40. Chen JE, Pedron S, Harley BAC. The combined influence of hydrogelstiffness and matrixbound hyaluronic acid content on glioblastoma inva-sion. Macromol Biosci 2017. doi: 10.1002/mabi.201700018.

41. Lokeshwar VB, Selzer MG. Differences in hyaluronic acid-mediated func-tions and signaling in arterial, microvessel, and vein-derived humanendothelial cells. J Biol Chem 2000;275:27641–9.

42. Misra S, Hascall VC, Markwald RR, Ghatak S. Interactions between hyalur-onan and its receptors (CD44, RHAMM) regulate the activities of inflam-mation and cancer. Front Immunol 2015;6:201.

43. Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM, et al. Tumor stemcells derived from glioblastomas cultured in bFGF and EGF more closelymirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell 2006;9:391–403.

44. Tsatas D, Kanagasundaram V, Kaye A, Novak U. EGF receptor modifiescellular responses to hyaluronan in glioblastoma cell lines. J Clin Neurosci2002;9:282–8.

45. Shinojima N, Tada K, Shiraishi S, Kamiryo T, Kochi M, Nakamura H, et al.Prognostic value of epidermal growth factor receptor in patients withglioblastoma multiforme. Cancer Res 2003;63:6962–70.


Xiao et al.




Correction

Correction: Brain-Mimetic 3D CulturePlatforms Allow Investigation of CooperativeEffects of Extracellular Matrix Features onTherapeutic Resistance in GlioblastomaWeikun Xiao, Rongyu Zhang, Alireza Sohrabi,Arshia Ehsanipour, Songping Sun, Jesse Liang,Christopher M.Walthers, Lisa Ta, David A. Nathanson, andStephanie K. Seidlits

In the original version of this article (1), in Fig. 2C, the axis label of the last bar shouldread "0.1%HA1%PEG" insteadof "0.1%HA0.5%PEG." Also, in Fig. 4B, the names ofthe cell lines were accidentally switched; the top panel should be HK301 and thebottom panel should be HK423. The errors have been corrected in the latest onlineHTML and PDF versions of the article. The authors regret these errors.

Reference1. XiaoW, Zhang R, Sohrabi A, Ehsanipour A, Sun S, Liang J, et al. Brain-mimetic 3D culture platforms

allow investigation of cooperative effects of extracellularmatrix features on therapeutic resistance inglioblastoma. Cancer Res 2018;78:1358–70.

Published online March 15, 2019.doi: 10.1158/0008-5472.CAN-19-0265�2019 American Association for Cancer Research.

CancerResearch

Cancer Res; 79(6) March 15, 20191260

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-0265&domain=pdf&date_stamp=2019-2-26

2018;78:1358-1370. Published OnlineFirst February 28, 2018.Cancer Res Weikun Xiao, Rongyu Zhang, Alireza Sohrabi, et al. Resistance in Glioblastoma

TherapeuticCooperative Effects of Extracellular Matrix Features on Brain-Mimetic 3D Culture Platforms Allow Investigation of

Updated version

10.1158/0008-5472.CAN-17-2429doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2018/01/09/0008-5472.CAN-17-2429.DC2

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/78/5/1358.full#ref-list-1

This article cites 43 articles, 11 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/78/5/1358.full#related-urls

This article has been cited by 5 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/78/5/1358To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-17-2429

http://cancerres.aacrjournals.org/content/suppl/2018/01/09/0008-5472.CAN-17-2429.DC2

http://cancerres.aacrjournals.org/content/78/5/1358.full#ref-list-1

http://cancerres.aacrjournals.org/content/78/5/1358.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/78/5/1358


brain-mimetic 3d culture platforms allow investigation of ... · cd168 (also known as hmmr or...

Documents