Fibroblast-Mediated Collagen Remodeling Within the Tumor ... · Statistical analysis All data were analyzed using Prism 6 software (GraphPad). Differences with P values of 0.05 were

Microenvironment and Immunology

Fibroblast-Mediated Collagen RemodelingWithin the Tumor Microenvironment FacilitatesProgression of Thyroid Cancers Driven byBrafV600E and Pten LossLee Ann Jolly1, Sergey Novitskiy2, Phillip Owens2, Nicole Massoll3,Nikki Cheng4,Wei Fang4, Harold L. Moses2, and Aime T. Franco1

Abstract

Contributions of the tumor microenvironment (TME) to pro-gression in thyroid cancer are largely unexplored and may illu-minate a basis for understanding rarer aggressive cases of thisdisease. In this study,we investigated the relationship between theTME and thyroid cancer progression in a mouse model wherethyroid-specific expression of oncogenic BRAF and loss of Pten(BrafV600E/Pten�/�/TPO-Cre) leads to papillary thyroid cancers(PTC) that rapidly progress to poorly differentiated thyroid cancer(PDTC). We found that fibroblasts were recruited to the TME ofBrafV600E/Pten�/�/TPO-Cre thyroid tumors. Conditioned mediafrom cell lines established from these tumors, but not tumorsdriven by mutant H-ras, induced fibroblast migration andproliferation in vitro. Notably, the extracellular matrix of

BrafV600E/Pten�/�/TPO-Cre tumors was enriched with stromal-derived fibrillar collagen, compared with wild-type or Hras-driven tumors. Further, type I collagen enhanced the motility ofBrafV600E/Pten�/�/TPO-Cre tumor cells in vitro. In clinical speci-mens, we found COL1A1 and LOX to be upregulated in PTC andexpressed at highest levels in PDTC and anaplastic thyroid cancer.Additionally, increased expression levels of COL1A1 and LOXwere associatedwith decreased survival in thyroid cancer patients.Overall, our results identified fibroblast recruitment and remo-deling of the extracellular matrix as pivotal features of the TME inpromoting thyroid cancer progression, illuminating candidatetherapeutic targets and biomarkers in advanced forms of thismalignancy. Cancer Res; 76(7); 1804–13. �2016 AACR.

IntroductionThyroid cancer is themost common endocrinemalignancy and

is predicted to be the fourthmost commonly diagnosed cancer by2030 (1). The BRAFV600E mutation is the most common geneticalteration in thyroid cancer, in particular papillary thyroid cancer(PTC), and is associated with more aggressive disease (2). Poorlydifferentiated (PDTC) and anaplastic thyroid cancers (ATC) oftenhave mutations in BRAF as well as mutations that result inconstitutive PI3K signaling, and are often unresponsive to treat-ments for thyroid cancer, including radiation and chemotherapy(3–5). Additionally, the success of targeted inhibition strategiesfor advanced thyroid cancers has been limited (6, 7). The increas-ing incidences of thyroid cancer, coupled with our relative lack of

understanding of drivers of disease progression, underscore theneed for novel therapeutics as well as identification of biomarkersthat are predictive of aggressive disease.

The tumor microenvironment (TME) is comprised of multiplecellular and non-cellular components, including extracellularmatrix (ECM) proteins, that converge to promote tumorigenesisin a variety of solid malignancies (8, 9). Tumor cells induce themigration of non-malignant cells, such as fibroblasts, endothelialcells, and immune cells, to the TME through direct cell–cellcontact and indirect mechanisms, which collectively support thedevelopment of a primary tumor niche (10). The influence oftumor-stromal cross-talk on tumor progression is recognized inmany different types of cancer. However, the mechanisms bywhich tumor cells establish a permissive niche that promotesthyroid cancer progression remain largely undefined.

Cancer-associated fibroblasts (CAF) represent a heterogeneouspopulation of fibroblasts that are recruited and activated toaugment tumor progression in many different solid tumors(11–13). In addition to stimulating tumor cell proliferation,angiogenesis, invasion, and metastasis, CAFs also drive tumori-genesis by upregulating the production of ECM components,including type 1 collagen (Col 1) and ECM modifying enzymes(14–16). Col 1 is themost abundant ECM scaffolding protein andits increased deposition in the TME is associated with tumorprogression (17–19), increased incidence of metastasis (20) anddrug resistance (21) in human cancers. These observations aresupported by in vivo and in vitro studies demonstrating that Col 1promotes the migration, invasion, and metastasis of tumor cells(22–24).

1Department of Physiology and Biophysics, University of Arkansas forMedical Sciences, Little Rock, Arkansas. 2Department of Cancer Biol-ogy, Vanderbilt University, Nashville, Tennessee. 3Department ofPathology, Winthrop P. Rockefeller Cancer Institute, University ofArkansas for Medical Sciences, Little Rock, Arkansas. 4Departmentof Pathology and Laboratory Medicine, University of Kansas MedicalCenter, Kansas City, Kansas.

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

Corresponding Author: Aime T. Franco, Department of Physiology and Bio-physics, University of Arkansas forMedical Sciences, 4301WestMarkhamStreet,Biomedical Research Building II B238-2, Little Rock, AR 72205. Phone: 501-603-1359; Fax: 501-686-8167; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-15-2351

�2016 American Association for Cancer Research.

CancerResearch

Cancer Res; 76(7) April 1, 20161804

on November 25, 2020. © 2016 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 27, 2016; DOI: 10.1158/0008-5472.CAN-15-2351

http://cancerres.aacrjournals.org/

A thoroughunderstanding of the role of tumor derived signals inestablishing an environment conducive to tumor development andthe effects of stromal derived signals on tumor cell behavior inthyroid cancer is largely unexplored. To identify potential mechan-ismsof thyroid cancerprogression in the contextof theTME, anovelmodel of thyroid cancer progression (BrafV600E/Pten�/�/TPO-Cre)was created and the TME dissected to identify factors that influencethyroid tumorigenesis. Braf activation and Pten loss cooperate inPTC development that rapidly progresses to PDTC characterized bya fibrotic and reactive tumor stroma enriched with CAFs, fibrillarcollagen deposits, and increased expression of lysyl oxidase (Lox),an ECM modifying enzyme that catalyzes collagen fiber cross-linking. We extended these findings to human disease and foundthat increased COL 1 and LOX expression is associated with moreaggressive well-differentiated thyroid cancer subtypes, PDTCs, anda poorer overall survival rate. Based on these observations, wepropose that a regulatory loop exists between thyroid tumor cells,CAFs, collagen, and Lox, which potentiates thyroid cancer progres-sion. These components may serve as therapeutic targets foradvanced thyroid cancers, and future studies will investigate ther-apeutic strategies targeting the TME and ECM in our in vivomodels.

Materials and MethodsExperimental animals

All animal experiments were performed at the University ofArkansas for Medical Sciences and approved by the InstitutionalAnimal Care and Use Committee. The LSL-BrafV600E,Ptenfl/fl, andthyroid peroxidase promoter (TPO)-Cre strains have been previ-ously described (25, 26). Mice were on mixed C57BL6/129SVJgenetic backgrounds. Genotypes were determined by PCR aspreviously described (25, 26).

Histology and immunohistochemistryThyroid tissues were fixed in 10% formalin-buffered acetate

and embedded in paraffin. Five-micrometer sections were pre-pared, and histological diagnosis was performed by a thyroidpathologist (N. Massoll). For further details, see SupplementaryMaterials and Methods.

Cell linesThe Braf, B297T, and B1180T cell lines were established from

BrafV600E/Pten�/� /TPO-Cre thyroid tumors and H340T, andH245T cell lines were established from HrasG12V/Pten�/�/TPO-Cre thyroid tumors, detailed in Supplementary Materials andMethods. Cell lines were authenticated using short tandem repeat(STR) DNA profiling (DDCMedical). Independent murinemam-mary CAF lines (mCAF and 4F)were isolated fromMMTV-PyVmTmodel as previously described (27).

RT-PCR analysisTotal RNA was extracted using the RNeasy Plus Mini Kit (Qia-

gen). Equal amounts of RNA template were reverse transcribedusing the Verso cDNA synthesis kit (Thermo Scientific). Differ-ential mRNA expression of type 1 collagen (Col1a1), lysyl oxidase(Lox) and 18s was measured using TaqMan Mastermix and pre-designed Taqman assays (Applied Biosystems). Four microlitersof cDNA from tumor samples and independent passages of eachcell line were run in triplicate on a Bio-Rad CFX96. Q-Genesoftware (28) was used to determine relative normalized expres-sion to 18s. Data analysis was based on the Ct method.

Migration assaysMigration assays were performed in 24-well plates with Fluor-

oblok inserts (Falcon). Forty thousand mCAF or 4F fibroblastswere seeded on each Fluoroblok insert in 0.5%FBS/F12. Condi-tioned media from tumor lines were added to the bottom cham-ber of the Fluoroblok plate. Eight hours after incubation at 37�C,cells on the Fluoroblok inserts were stainedwith 2mmol/LCalceinAM (Life Technologies). Fluorescent values were obtained at awavelength of 485 ex/520 em on a Synergy H1multimode reader(BioTek) to quantitate migration. The optics position of the platereaderwas set to read from the bottomof the plate in order to onlyimage cells that had migrated through the transwell. Images weretaken on the EVOS FL imaging system.

Proliferation assaysProliferation assays were performed in 96-well plates. One

thousand fibroblasts were seeded in quadruplicate per treatmentcondition in completemedia for attachment. Following overnightattachment, the media were replaced with medium containing0.5% FBS, 10% FBS, or conditioned media from tumor cell lines.Proliferation was assessed using the CellTiter-Glo luminescentcell viability assay (Promega). One well per treatment was incu-bated with a 1 mmol/L Calcein AM for 30 minutes at 37�C andimaged using an EVOS FL imaging system.

Live-cell microscopyTumor cells were plated in 35-mm cell culture dishes either

tissue culture treated or coated with 100 mg/mL rat tail collagen I(Sigma). Following attachment in complete medium, cells wereserum starved overnight, then stimulated with 10% FBS andimmediately imaged under phase contrast on an Axiovert 100M microscope fitted with a Zeiss Axiocam ICM1 camera andmaintained at 37�C and 5% CO2 using a Live Cell Pathologyincubator. Images were collected every minute for 4 hours. Theimages were analyzed using NIH ImageJ software (Version 1.50e)with theMTrackJ Plugin to determine distanced travelled by cells.The center of each cell nucleus was used as the point of tracking,and cells undergoing mitosis were excluded from analysis. Tracklength was measured for 10 individual cells per treatment andrepeated at least thrice.

Human thyroid cancer database analysisThe Oncomine platform (www.oncomine.org; ref. 29) was

used to compare the expression levels of COL1A1 and LOXmRNA between thyroid cancer subtypes (GSE27155; ref. 30),which included 4 normal thyroid samples, 15 follicular variantpapillary thyroid cancer samples (FVPTC), 10 tall cell PTCs (TCPTC), 10 follicular adenomas, 13 follicular thyroid cancers(FTC), 2 medullary thyroid cancer samples, 7 oncocytic ade-nomas, 8 oncocytic FTCs, 26 PTCs, and 4 undifferentiated/anaplastic thyroid cancer samples. The log2 median-centeredintensity values for COL1A1 and LOX were extracted for theanalysis using all samples. cBioPortal, the web-based openplatform for analyzing multidimensional cancer genomics data(31, 32), was used to obtain summary statistics on co-occur-rence of genomic alterations in BRAF, NRAS, HRAS, KRAS,COL1A1, and LOX in thyroid carcinomas in 397 thyroid cancercases. Odds ratios to indicate the likelihood of mutual exclu-sivity or co-occurrence of each pair of genes were calculated.P values were determined by the Fisher exact test.

MAPK and PI3K Activation Recruits a Fibrotic TME

www.aacrjournals.org Cancer Res; 76(7) April 1, 2016 1805




Statistical analysisAll data were analyzed using Prism 6 software (GraphPad).

Differences with P values of �0.05 were considered statisticallysignificant.

ResultsBrafV600E and PI3K signaling cooperate in the development ofPTCs that rapidly progress to PDTC in vivo

MAPK signaling plays a critical role in thyroid cancer initiation,as evidenced by our previous studies demonstrating thatendogenous expression of BrafV600E is sufficient to inducemurinePTCs that recapitulate human disease (25). BRAF mutations areassociated with more aggressive PTC and are often found inconjunction withmutations that result in constitutive PI3K/AKTsignaling, including PIK3CA and PTEN mutations, in PDTCs(33). This led to the hypothesis that simultaneous MAPKactivation via BrafV600E and PI3K activation via Pten loss wouldcooperate in thyroid cancer initiation and progression toadvanced disease. To determine whether BrafV600E and PI3Ksignaling could cooperate in thyroid cancer progression in vivo,LSL-BrafV600E/Ptenflox/floxmice were crossed with Ptenflox/flox/TPO-Cre mice to generate mice in which BrafV600E is conditionally

activated and Pten is homozygously inactivated through thy-roid-specific Cre recombinase activation (BrafV600E/Pten�/�/TPO-Cre). BrafV600E/Pten�/�/TPO-Cremice developed PTCs thatrapidly progressed to PDTCs with 100% penetrance and lethal-ity by weaning (Fig. 1A). In stark contrast to wild-type (WT)thyroid glands with normal follicular architecture (Fig. 1B),BrafV600E/Pten�/�/TPO-Cre tumors encompass the entire thy-roid gland and display many of the classical hallmarks of high-grade human PTC, including formation of papillae, fine chro-matin, and nuclear grooves (Fig. 1 C–E), as well as features ofPDTC, including central necrosis (Fig. 1 F and G) and invasioninto surrounding tissue (Fig. 1H). The very early lethality ofBrafV600E/Pten�/�/TPO-Cre mice precludes long-term studies todetermine factors involved in disease progression. However,the rapid tumor development and pathological features ofPDTC that are recapitulated in BrafV600E/Pten�/�/TPO-Cre miceprovide a model by which to investigate factors within the TMEthat may contribute to disease progression.

The TME of BrafV600E/Pten�/�/TPO-Cre tumors is enriched withtumor-associated fibroblasts

The TME is composed of many different cell types thatinfluence tumor progression, including CAFs. CAFs promote

Days

% S

urvi

val

0 5 10 15 20 250

50

100

Braf V600E Pten-/-

WT

A

20×

BrafV600E/Pten-/-WT

Tr

Th20×

B C

D E

F G H

BrafV600E/Pten-/- BrafV600E/Pten-/-

20×

20×

BrafV600E/Pten-/- BrafV600E/Pten-/- BrafV600E/Pten-/-

40×

40×

40×

Figure 1.Thyroid-specific activation of MAPK signaling via BrafV600E and Pten loss cooperate in the development of PTCs that progress to PDTCwith short latency. A, Kaplan–Meier survival curve of BrafV600E/Pten�/�/TPO-Cre mice in comparison with control WT mice. No BrafV600E/Pten�/�/TPO-Cre mice survived past weaning(P <0.0001, log-rank analysis). B andC, representativeH&E imageof thyroid tissue fromWT (left) andBrafV600E/Pten�/�/TPO-Cre (right)mice at 8days of age; scalebar, 200 mm. Arrows point to thyroid lobe (Th) and tracheal cartilage (Tr). D, BrafV600E/Pten�/�/TPO-Cre tumor displaying papillae formation; scale bar,200mm. E,BrafV600E/Pten�/�/TPO-Cre tumor displaying classical features of PTC, including nuclear grooves and fine chromatin; scale bar, 100mm.F, high-gradePTCwith central necrosis (arrows) in a BrafV600E/Pten�/�/TPO-Cre mouse at 8 days; scale bar, 200 mm. G, high magnification of high-grade PTC with centralnecrosis (arrows) in a BrafV600E/Pten�/�/TPO-Cre mouse at 8 days of age; scale bar, 100 mm. H, PTC with invasive focus pushing into surrounding adipose tissue(arrow) in a BrafV600E/Pten�/�/TPO-Cre tumor at 8 days scale bar, 100 mm.

Jolly et al.

Cancer Res; 76(7) April 1, 2016 Cancer Research1806




tumorigenesis in human cancers and in vivo model systems(11). Interestingly, fibroblast growth factors (FGF) and theirreceptors (FGFR) are overexpressed in thyroid cancer (34,35)and correlate with thyroid cancer progression (36). Hematox-ylin and eosin (H&E) staining of BrafV600E/Pten�/�/TPO-Cretumor sections revealed areas of fibrosis along the tumorperiphery and cells with fibroblast morphology (Fig. 2A andB, top, arrows). In contrast, no areas of fibrosis were observed inWT thyroid tissue. These cells were confirmed as fibroblasts viaimmunostaining with aSMA. In contrast to WT thyroid, inwhich no aSMA staining was observed, BrafV600E/Pten�/�/TPO-Cre tumors displayed robust peripheral and intratumoralaSMA staining, indicating fibroblast recruitment and infiltra-tion (Fig. 2A and B, bottom).

Tumor cells isolated from BrafV600E/Pten�/�/TPO-Cre micestimulate fibroblast proliferation and migration in vitro

Given the increased fibroblast infiltrate observed inBrafV600E/Pten�/�/TPO-Cre tumors, we asked whether Braf-driven thyroid tumor cells could stimulate the proliferationand/or migration of fibroblasts. We generated multiple stabletumor cell lines from BrafV600E/Pten�/�/TPO-Cre mice (Braf-1,B1180T, and B297T) and tested the ability of these cell linesto drive the proliferation and migration of two independentCAF lines in vitro (27). Conditioned media collected from Brafand B297T cells significantly increased mCAF and 4F fibro-blast proliferation in comparison with serum free controlsafter 48 hours of incubation (Fig. 3A and Supplementary Fig.S1A). To determine whether increased proliferation was spe-cific to factors secreted by Braf-driven thyroid tumor cells, theexperiments were repeated with the Hras-driven thyroidtumor cell line H340T. Conditioned medium isolated from

H340T cells had no effect on the proliferation of mCAF or 4Fafter 48 hours compared with the serum-free control (Fig. 3Aand Supplementary Fig. S1A). Additionally, conditionedmedia from Braf, B1180T, and B297T cells significantlyincreased the migration of mCAF and 4F fibroblasts in Trans-well assays compared with serum-free controls (Fig. 3B and C;Supplementary Fig. S1B and S1C), demonstrating that tumorcells from BrafV600E/Pten�/�/TPO-Cre tumors secrete factorsthat induce fibroblast migration and likely drives fibroblastrecruitment to BrafV600E/Pten�/�/TPO-Cre tumors in vivo. Con-sistent with the proliferation studies, conditioned media fromHrasG12V-driven murine thyroid tumor cell lines did notstimulate mCAF or 4F migration compared with serum-freecontrols (Fig. 3B and C; Supplementary Fig. S1B and S1C).Together, these results suggest that Braf, but not Hras, acti-vation results in secretion of tumor-derived factors thatinduce the proliferation and recruitment of fibroblasts inmurine thyroid cancer. TGFb is a key mediator of fibroblastactivation during wound healing and exerts promitogenic andchemotactic effects on fibroblasts (reviewed in 12). To deter-mine whether the induction of fibroblast migration andproliferation in response to conditioned medium fromBraf-driven tumor cells is TGFb dependent, proliferation andmigration experiments were repeated with TGFbRII knockoutfibroblasts (27). Treatment with conditioned medium fromBraf, B1180T, B297T, and H340T inhibited the proliferationof TGFbRII knockout fibroblasts compared with 0.5% FBS(Supplementary Fig. S2A). No migration through transwellswas observed in any treatment group, even after 24 hours ofexposure to 10% FBS (Supplementary Fig. S2B), indicatingthat intact TGFbRII signaling is required for the migration offibroblasts. To determine whether Braf-driven tumor cells

aSMA

H&E

WT

20×20×

BrafV600E/Pten-/- Braf V600E/Pten-/-

40×

40×

40×

40×

A B

aSMA

H&E

WT BrafV600E/Pten-/- Braf V600E/Pten-/-

20× 20×

Figure 2.BrafV600E/Pten�/�/TPO-Cre tumors are enrichedwith CAFs. A, top, lowmagnification of H&E-stained tissue fromaWT thyroid (left) andBrafV600E/Pten�/�/TPO-Crethyroid (right); scale bar, 200 mm. Black arrow, areas of fibrosis in BrafV600E/Pten�/�/TPO-Cre tumor. Bottom, low magnification of a-SMA–immunostainedtissue from a WT thyroid (left) and BrafV600E/Pten�/�/TPO-Cre thyroid (right); scale bar, 200 mm. Increased aSMA staining (red) was observed along theperiphery of BrafV600E/Pten�/�/TPO-Cre tumors, while no aSMA staining was evident in WT thyroid tissue. B, top, high magnification of H&E-stained thyroidtissue from a BrafV600E/Pten�/�/TPO-Cre tumor; scale bar, 100 mm. Arrows, areas of fibrosis and intratumoral fibroblasts. Bottom, immunostained thyroidtissue from a BrafV600E/Pten�/�/TPO-Cre tumor; scale bar, 100 mm. White arrows, extra- and intratumoral aSMAþ fibroblasts. Nuclei were stained with40 ,6-diamidino-2-phenylindole (DAPI).






could induce activation of TGFb signaling in fibroblasts, 4Ffibroblasts were treated with conditioned medium from Braf,B1180T, and B297T cells and Western blot analysis for phos-phorylated SMAD 2 and 3 performed. Treatment with con-ditioned medium from Braf-driven thyroid tumor cell linesdid not induce the phosphorylation of SMAD 2 or 3 infibroblasts (Supplementary Fig. S2C). Collectively, these datasuggest that while TGFb signaling is permissive for the induc-tion of fibroblast migration in response to Braf-driven tumorcell–derived signals, alternative pathways are likely beingactivated by tumor-derived factors to induce fibroblast pro-liferation and migration.

Increased total and fibrillar collagen deposition and Loxexpression in BrafV600E/Ptenhom/TPO-Cre tumors

Collagens, in particular collagen 1 (Col 1), are primarilyderived from fibroblasts and augment tumor cell invasion andmigration in vivo and in vitro (13–20). We hypothesized that therecruitment of fibroblasts to BrafV600E/Pten�/�/TPO-Cre tumorswould result in increased collagen deposition in the thyroid TME.Col1a1, which encodes the a1 chain of Col 1, expression levelswere consistently upregulated in BrafV600E/Pten�/�/TPO-Cretumors compared with WT thyroid (Fig. 4A). By contrast, Col1a1expression in tumor cell lines derived from BrafV600E/Pten�/�/TPO-Cre tumors (B1180T and B297T) was no different than WTcontrols (Fig. 4B), suggesting that Col1a1 expression in BrafV600E/Pten�/�/TPO-Cre tumors is not derived by tumor cells. Col1a1expression was significantly upregulated in the parent tumors(B1180 and B297) from which these cell lines were derived (Fig.4B), suggesting that the increased Col1a1 expression in wholeBrafV600E/Pten�/�/TPO-Cre tumors occurs primarily in stromalcells, likely fibroblasts, rather than tumor cells. Immunostainingrevealed increased Col 1 in BrafV600E/Pten�/�/TPO-Cre tumorscompared with WT thyroid (Fig. 4D). Additionally, Col 1 wasundetectable in ECMderived from BrafV600E-driven tumor cells invitro (data not shown). To test the hypothesis that fibroblasts arethe predominant source of Col 1 within the TME of BrafV600E/Pten�/�/TPO-Cre tumors, tumor sections were immunostainedwith Col 1 and aSMA to determine if they colocalized. Both Col 1and aSMA staining localized to the tumor–stromal interface

(Fig. 4D), supporting the hypothesis that fibroblasts are thepredominant source of Col 1 within the TME of BrafV600E/Pten�/�/TPO-Cre tumors.

The biomechanical properties and deposition of ECM proteinsare altered during tumorigenesis. Further, the activity of tumorand stromal derived matrix metalloproteinases (MMP) and col-lagen-cross-linking enzymes, which modulate the structural sta-bility of ECM proteins, is increased in different cancers (37). Lysyloxidase (Lox) is an ECM-modifying enzyme that catalyzes thecross-linking of collagen fibers, resulting in increased collagenfiber stability and ECM stiffness, which can enhance the invasivecapacity of tumor cells in vivo (38). Upregulation of LOX isobserved in a variety of solid tumors (39–41) and correlates withreduced metastasis-free survival in breast and head and neckcancers (39). LOX has recently been found to be upregulated inthyroid cancer and potentiates metastasis and invasion of ana-plastic thyroid cancer cell lines in vivo (40). Lox expression wassignificantly upregulated in BrafV600E/Pten�/�/TPO-Cre tumorscompared with WT thyroid controls (Fig. 4C). Lox expression wasalso increased in B1180T and B297T cell lines in comparison withWT controls (Supplementary Fig. S3). Picrosirius red staining oftumor sections revealed increased polarized intensity inBrafV600E/Pten�/�/TPO-Cre tumors, demonstrating higher content ofmatureand cross-linked collagen fibers (Fig. 4E). In WT thyroid, onlytracheal cartilage contained collagen. These results indicate thatBrafV600E/Pten�/�/TPO-Cre tumors promote increased collagensynthesis and cross-linking through upregulation of Col1a1 andLox, resulting in increased collagen deposition and stability in theTME of BrafV600E/Pten�/�/TPO-Cre tumors. To determine whetherCol 1 modulates tumor cell phenotype, cell motility was mea-sured on tissue culture plates and Col 1 coated plates. Live-cellmicroscopy and tracking analysis demonstrated that BrafV600E/Pten�/�/TPO-Cre tumor cell lines Braf and B297T displayed sig-nificantly increased motility when plated on Col 1 versus tissueculture plastic (Fig. 4F and G; Supplementary S4). No increase inmotilitywas observedwhenBrafV600E/Pten�/�/TPO-Cre tumor celllineB1180Twasplated onCol 1.However, B1180T cells platedonCol 1 exhibited an increase in the mitotic index; therefore, lesstotal cellswere included in thefinal analysis. Further, Col 1hadnoeffect on the motility HrasG12V/Pten�/�/TPO-Cre tumor cell lines

mCAFAR

LU (p

rolif

erat

ion)

+ 0.5%

FBS

+ 10%

FBS

+ Braf

CM

+ B29

7T C

M

+H34

0T C

M

+ 0.5%

FBS

+ 10%

FBS

+ Braf

CM

+ B297T

CM

+H34

0T C

M0

20,000

40,000

60,000

80,000 MCAF

4F

* **

*

CB 4F

RLU

(mig

ratio

n)

+ 0.5%

FBS

+ 10%

FBS

+Bra

f CM

+B11

80T CM

+ B297T

CM

+ H34

0T C

M

RLU

(mig

ratio

n)

+ 0.5%

FBS

+10%

FBS

+Braf

CM

+B11

80T C

M

+B29

7CM

+H34

0T C

M

+H24

5T C

M0

20,000

40,000

60,000 5,000

4,000

3,000

2,000

1,000

0

*

Figure 3.BrafV600E/Pten�/�/TPO-Cre thyroid tumor cells stimulate the proliferation and migration of CAFs. A, quantification of proliferation of mCAF and 4F fibroblasts inresponse to 48-hour treatment with conditioned media from three independent tumor cell lines isolated from BrafV600E/Pten�/�/TPO-Cre thyroid tumors.� , P<0.05. RLU, relative light units. B and C, quantification of mCAF and 4F fibroblast migration using Fluoroblok Transwell assays. Conditioned media fromBraf, B1180T, and B297T cells induced significant migration of mCAF and 4F fibroblasts in comparison with the serum-free controls after 8 hours of treatment.� , P < 0.05. Migration was quantified via fluorescent staining of migrated cells with Calcein-AM. Results are representative of two independent experiments.

Jolly et al.





H340T andHras1, suggesting that the increasedmotility responseto Col 1 is specific to Braf and notHras-driven thyroid tumor cells.

COL 1 and LOX are upregulated in human PTC and areassociated with aggressive histologic variants of PTC and PDTC

To determine whether these murine models recapitulatedhuman disease and reflected changes observed in patients, theOncomine database (29) was used to investigate COL1A1 andLOX expression in thyroid tumors from theGiordano cohort (30).COL1A1 and LOX expression were increased in PTC (Fig. 5A)compared with normal thyroid, FTC, and FVPTC, which is asso-ciated almost exclusively with RASmutations and displays manypathologic features similar to FTC (2). COL1A1 and LOX expres-sion levels were further increased in tall-cell variant PTC, a moreaggressive form of PTC, and highest in undifferentiated thyroidcancers (Fig. 5A). The cBioPortal was used to analyze thyroidcancer data in The Cancer Genome Atlas (TCGA) dataset in order

to correlate COL1A1 and LOX upregulation with mutationalstatus. COL1A1 and LOX upregulation occurred in 8% and10%, respectively, of all thyroid tumors analyzed (397 cases),and occur exclusively in thyroid tumors harboring BRAF, but notRAS, mutations (Fig. 5B). Strong tendencies in the rate of co-occurrence between BRAF mutations and COL1A1 upregulation,BRAF mutations and LOX upregulation, and COL1A1 and LOXupregulation in thyroid cancers were found (Table 1). Together,these results suggest that COL1A1 and LOX cooperate in thyroidcancer progression and that upregulation of COL1A1 and LOXoccurs in response to BRAF, but not RAS, activation in thyroidcancer.

Upregulation of COL1A1 and LOX is associated with decreasedoverall survival in thyroid cancer patients

Mutations in BRAF correlate with decreased overall survival inthyroid cancer patients (41). Given that upregulation of COL1A1

D

A

WT11

6612

8012

8112

8412

8612

9012

8311

80 297

0

100

200

300 *

Fold

cha

nge

vs

. WT

Fold

cha

nge

vs

. WT

Fold

cha

nge

vs

. WT

Col1a1

WT

B1180

T11

80B29

7T 297

0

20

40

60

80

100* *

B Col1a1

WT11

6612

8012

8112

8412

8612

9012

8311

80 297

0

100

200

300 *C Lox

Braf V600E/Pten−/− WT

αSMA

Col I

Overlay

20X

DAPI40X

E

WT BrafV600E/Pten−/−

10X

TrTr

fiber thickness

F

G

Braf

20×

20×

Plastic

Collagen

Plastic

Plastic

Plastic

Plastic

Plastic

Collagen

Collagen

Collagen

Collagen

Collagen

0

500

1,000

1,500

* *

Braf B1180T B297T Hras1 H340T

Dis

tanc

e in

μm

ol/L

Figure 4.Increased total and fibrillar collagen deposition and Lox expression inBrafV600E/Pten�/�/TPO-Cre tumors. A, qPCR analysis of Col1a1 expression fromRNA extractedfrom whole BrafV600E/Pten�/�/TPO-Cre tumors (n ¼ 9) and WT thyroids (n ¼ 2 independent pools of 5 thyroids). � , P < .05. B, qPCR analysis of Col1a1expression in B1180T and B297T tumor cell lines compared with Col 1 expression in parent BrafV600E/Pten�/�/TPO-Cre tumors (1,180 and 297). � , P < 0.05. C, qPCRanalysis of Lox expression from RNA extracts isolated from whole BrafV600E/Pten�/�/TPO-Cre tumors (n ¼ 9) and WT thyroids (n ¼ 2 independentpools of 5 thyroids). � ,P< .05. D, representative images (�20) of immunofluorescent staining for Col 1 (green) andaSMA (red) inWTandBrafV600E/Pten�/�/TPO-Crethyroids. Nuclei were stained with 40 ,6-diamidino-2-phenylindole (DAPI). E, representative images (�20) of Picrosirius Red staining of WT (left) andBrafV600E/Pten�/�/TPO-Cre thyroid (right) sections. Tr, trachea. Dashed line demarcates between the edge of tissue and background. F, distance traveled (mm)in 4 hours by Braf, B1180T, B297T, Hras1, and H340T tumor cells on tissue culture–coated (plastic) or collagen I–coated (collagen) dishes. Data analysis isrepresentative of three independent experiments per condition per cell line, and each data point represents the distance traveled by a single cell. � , P < 0.0001. G,representative (�20) images of tracks travelled by Braf tumor cells seeded on plastic and collagen. Each colored track represents the distance traveled percell over 4 hours.






and LOX occurs predominantly in thyroid tumors harboringBRAF mutations, we sought to determine whether COL1A1 andLOX overexpression was associated with reduction of overallsurvival in thyroid cancer. Co-upregulation of COL1A1 and LOXin thyroid cancer (N¼ 41 cases) results in a significant decrease inoverall survival in thyroid cancer patients (Fig. 5D) comparedwith patients with tumors without COL1A1 and LOX upregula-tion. These results suggest that overexpression of COL1A1 andLOX contributes to disease progression in thyroid cancer andmaycontribute to thyroid cancer–related mortality.

DiscussionEach of the components that make up the TME, including

tumor cells, non-malignant infiltrating stromal cells, and ECMproteins, work in concert to establish a permissive niche that is

essential for tumorigenesis (42). While many studies haveaddressed the involvement of a singular cell type, such as fibro-blasts or immune cells, or ECM component in tumor develop-ment, few studies have investigated the cross-talk between mul-tiple components within the TME and how these complex rela-tionships function together to promote tumor development. Inthis study, we dissected the cellular and non-cellular componentswithin the TME of thyroid cancer in order to understand howinteractions between these components contribute to thyroidcancer progression.

The BRAFV600E mutation is associated with a more aggressivetumor phenotype in thyroid cancer patients and has recently beenimplicated in the modulation of the TME through the regulationof ECM components (43). Genes associated with ECM remodel-ing, including integrins, TGFb1, and fibronectin, are upregulated

Figure 5.Upregulation ofCOL1A1 and LOX is associatedwith human thyroid cancer progression and increasedmortality. A, COL1A1 and LOXmRNA expression levels in humanthyroid cancers evaluated within the Oncomine database. COL1A1 and LOX are upregulated in PTC in comparison with normal thyroid (n ¼ 4), FTC (n ¼ 13),and FVPTC (n ¼ 15). COL1A1 and LOX expression levels are further upregulated in Tall Cell Variant PTC (n ¼ 10), a more aggressive form of PTC, and are highest inundifferentiated/anaplastic thyroid cancer (N ¼ 4). B, cBioPortal analysis of the frequency of genetic alterations in BRAF, NRAS, HRAS, KRAS, COL1A1,and LOX in human thyroid cancers (397 cases). C, Kaplan–Meier survival curve of thyroid cancer patients with (N ¼ 41) or without (N ¼ 273) both COL1A1 and LOXupregulation. Cases with increased COL1A1 and LOX expression have a significant decrease in overall survival (P ¼ 0.00674).

Table 1. BRAF mutations are associated with LOX and COL1A1 upregulation

Gene Gene P Log OR Association

BRAF COL1A1 <0.001 1.645 Tendency toward co-occurrenceBRAF LOX <0.001 >3 Tendency toward co-occurrenceCOL1A1 LOX <0.001 2.837 Tendency toward co-occurrence

NOTE: cBioPortal odds-ratio analysis indicates BRAFmutations and LOX or COL1A1 upregulation, and upregulation of both COL1A1 and LOX, have a strong tendencyto co-occur in thyroid carcinomas (P < 0.001).

Jolly et al.





in PTCs with BRAFV600E mutations when compared with PTCswithout the mutation (44), suggesting activation of BRAF iscritical for the development of a fibrotic tumor stroma. In agree-ment with these findings, our data demonstrate that activation ofBraf andPI3K signaling in thyrocytes results in the development ofa fibrotic and reactive tumor stroma in BrafV600E/Pten�/�/TPO-Cretumors, characterized by increased fibroblast recruitment andstromal deposition of Col 1 (Fig. 6). In this model, we proposethat fibroblasts are recruited to the thyroid TME by BrafV600E/Pten�/�/TPO-Cre tumor cells, which activate fibroblasts to pro-duce and deposit Col 1. In turn, tumor cells cross-link thefibroblast derived Col 1 fibers in the TME via upregulation ofLox, resulting in a stiffer Col 1 matrix that augments tumor cellmotility and promotes tumor progression.

BrafV600E/Pten�/�/TPO-Cre tumors contained higher levels oftotal and fibrillar collagen and increased expression of Lox. Col1 augmented the motility of BrafV600E/Pten�/�/TPO-Cre tumorcell lines in vitro. Interestingly, no fibroblast recruitment orcollagen deposition was observed in the TME in responseto Hras activation in HrasG12V /Pten�/�/TPO-Cre mice, aclosely related murine model of thyroid cancer in which micedevelop follicular carcinomas that progress to PDTC (A. Franco;manuscript in preparation). In addition, Col 1 had no effect onthe motility of Hras-driven tumor cell lines in vitro. These dataindicate that in the context of Pten loss, activation of Braf, butnot Hras, results in a fibrotic response in the TME of thyroidcancer that promotes tumor progression and potentially inva-sion. Further supporting fibroblast recruitment to the thyroidTME is BrafV600E specific, only conditioned media fromBrafV600E/Pten�/�/TPO-Cre cells was able to induce proliferationand migration of fibroblasts in vitro. These results suggest thatactivation of Braf, but not Hras, induces secretion of factors thatpromote fibroblast migration and likely leads to the increasedfibroblast recruitment observed in vivo. Interestingly, RAS acti-

vation is associated with increased inflammation and tumorimmune cell infiltration in murine models of lung and pan-creatic ductal adenocarcinoma (45, 46), and mutant BRAFV600E

induces fibroblast activation in melanoma cell lines (47).Future studies are needed to unravel the molecular mechanismsby which the activation of closely related MAPK effectors, likeRAS and RAF, lead to the development of distinct TMEs throughthe differential recruitment of various cell types or ECMremodeling.

The biomechanical properties of a tumor-associatedmatrix canhave a strong influence on cellular behavior (48). LOX is a knowndriver of ECM stiffness within the TME due to its ability to cross-link collagen fibers, and inhibition of LOX attenuates metastasisin mouse models of breast cancer and, more recently, thyroidcancer (39, 40, 49). Increased matrix stiffness also induces theactivation of integrin signaling and downstream ERK activationand promotes the stabilization of focal adhesion complexes thatcan drive malignancy (50). BrafV600E/Pten�/�/TPO-Cre tumorscontained abundant total and fibrillar collagen. Future studieswill investigate whether inhibition of Lox decreases matrix stiff-ness and can attenuate thyroid cancer progression in BrafV600E/Pten�/�/TPO-Cre tumors.

Advanced forms of thyroid cancer are associated with muta-tions in the MAPK pathway and additional mutations thatresult in constitutive PI3K activation (33). These data demon-strate that activation of BrafV600E and PI3K leads to the devel-opment of PTCs that rapidly progress to PDTCs. These tumorsare associated with a fibrotic TME characterized by increasedstromal collagen deposition and Lox upregulation. Thesemurine models faithfully recapitulate patient tumors by whichincreased COL1A1 and LOX expression is associated with PTCcompared with FTC and normal thyroid, and that COL1A1 andLOX are expressed at highest levels in PDTC and ATC. Together,these data support the critical role of these ECM components in

Figure 6.Proposed model of BrafV600E-driven remodeling of theTME that contributes to progression of thyroid cancer.Activation of BrafV600E and loss of Pten in thyroid cellsleads to rapid transformation of thyroid epithelial cells.Braf V600E/Ptenhom tumor cells stimulate the recruitment,proliferation, and activation of tumor-associatedfibroblasts in the TME. Activated fibroblasts synthesizeand deposit collagen fibrils in the TME. Braf V600E/Ptenhom tumor cell–derived LOX cross links fibroblast-derived collagen fibrils to form mature and cross-linkedcollagen fibers that augment tumor cell motility,promoting thyroid cancer progression.






promoting thyroid cancer progression. COL1A1 and LOXexpression in human PTCs is strongly correlated with BRAF,but not RAS mutations. RAS mutations are very common inFTC and FVPTC, while BRAF mutations are closely associatedwith classical PTC, suggesting that COL1A1 and LOX upregula-tion in thyroid cancer occurs in response to BRAF activation andmay drive PTC versus FTC development. Finally, COL1A1 andLOX upregulation is associated with decreased overall survivalin thyroid cancer, implicating COL1A1 and LOX as mediators ofcancer progression and may serve as a prognostic indicator ofdisease status in addition to the BRAFV600E mutation in thyroidcancer.

While it is now widely accepted that the TME is essential fortumorigenesis, most studies only address the contribution ofsingular TME component to cancer progression. Consideringthat the TME is composed of multiple components (bothcellular and non-cellular), studies that aim to investigate howthese components work together to establish a niche permissivefor tumorigenesis are needed to fully understand the mechan-isms of tumor development and therapeutic resistance. Thisstudy is the first to identify and describe the interaction betweentumor cells, fibroblasts, collagen, and Lox in the TME of thyroidtumors, providing a model by which this dynamic interactionmay drive thyroid tumor progression (Fig. 6). We hope thatthese results will lead to the development of more effectivetherapeutic strategies for thyroid cancer that account for thecomplexity of the TME in vivo.

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

Authors' ContributionsConception and design: L.A. Jolly, A.T. FrancoDevelopment of methodology: L.A. Jolly, P. Owens, W. Fang, A.T. FrancoAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): L.A. Jolly, P. Owens, A.T. FrancoAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): L.A. Jolly, P. Owens, A.T. FrancoWriting, review, and/or revision of the manuscript: L.A. Jolly, S. Novitskiy,P. Owens, N. Cheng, W. Fang, H.L. Moses, A.T. FrancoAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): S. Novitskiy, N. Cheng, W. Fang, A.T. FrancoStudy supervision: A.T. FrancoOther (pathology slide review): N. Massoll

AcknowledgmentsThe authors thank Dr. James Fagin for mouse strains and Drs. Julio Ricarte

Fihlo and Subhajyoti De for helpful advice and guidance with cBio Portal andOncomine.

Grant SupportThis work was supported by the University of Arkansas for Medical Sciences

CTSA grant NID UL1TR000039. The National Institute of General MedicalSciences supported thiswork through theCenter forMicrobial Pathogenesis andHost Inflammatory Responses at theUniversity of Arkansas forMedical SciencesCOBRE Grant 1P20GM103625-02, The American Thyroid Association/Thycaresearch grant (A. Franco), and UAMS Envoys Seeds of Science Award(A. Franco).

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 27, 2015; revised January 11, 2016; accepted January 14,2016; published OnlineFirst January 27, 2016.

References1. Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian

LM. Projecting cancer incidence anddeaths to 2030: theunexpected burdenof thyroid, liver, and pancreas cancers in the United States. Cancer Res2014;74:2913–21.

2. Nikiforov YE, NikiforovaMN. Molecular genetics and diagnosis of thyroidcancer. Nat Rev Endocrinol 2011;7:569–80.

3. Are C, Shaha AR. Anaplastic thyroid carcinoma: biology, pathogenesis, prog-nostic factors, and treatment approaches. Ann Surg Oncol 2006;13:453–64.

4. Ricarte-Filho JC, Ryder M, Chitale DA, Rivera M, Heguy A, Ladanyi M, et al.Mutational profile of advanced primary and metastatic radioactive iodine-refractory thyroid cancers reveals distinct pathogenetic roles for BRAF,PIK3CA, and AKT1. Cancer Res 2009;69:4885–93.

5. Garcia-Rostan G, Costa AM, Pereira-Castro I, Salvatore G, Hernandez R,Hermsem MJ, et al. Mutation of the PIK3CA gene in anaplastic thyroidcancer. Cancer Res 2005;65:10199–207.

6. Sherman SI. Targeted therapy of thyroid cancer. Biochem Pharmacol2010;80:592–601.

7. Krajewska J, Handkiewicz-JunakD, Jarzab B. Sorafenib for the treatment ofthyroid cancer: an updated review. Expert Opinion on Pharmacother2015;16:573–83.

8. Joyce JA. Therapeutic targeting of the tumor microenvironment. CancerCell 2005;7:513–20.

9. Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche incancer progression. J Cell Biol 2012;196:395–406.

10. Wels J, Kaplan RN, Rafii S, Lyden D. Migratory neighbors and distantinvaders: tumor-associated niche cells. Gen Devel 2008;22:559–74.

11. Madar S, Goldstein I, Rotter V. 'Cancer associated fibroblasts'–more thanmeets the eye. Trends Mol Med 2013;19:447–53.

12. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat Rev Cancer 2006;6:392–401.

13. Bhowmick NA, Neilson EG, Moses HL. Stromal fibroblasts in cancerinitiation and progression. Nature 2004;432:332–7.

14. YamaguchiH, YoshidaN, TakanashiM, ItoY, FukamiK, YanagiharaK, et al.Stromal fibroblasts mediate extracellular matrix remodeling andinvasion of scirrhous gastric carcinoma cells. PloS One 2014;9:e85485.

15. Karagiannis GS, Poutahidis T, Erdman SE, Kirsch R, Riddell RH,DiamandisEP. Cancer-associated fibroblasts drive the progression of metastasisthrough both paracrine and mechanical pressure on cancer tissue. MolCancer Res 2012;10:1403–18.

16. Loeffler M, Kruger JA, Niethammer AG, Reisfeld RA. Targeting tumor-associated fibroblasts improves cancer chemotherapy by increasing intra-tumoral drug uptake. J Clin Invest 2006;116:1955–62.

17. Armstrong T, PackhamG,Murphy LB, BatemanAC,Conti JA, FineDR, et al.Type I collagen promotes the malignant phenotype of pancreatic ductaladenocarcinoma. Clin Cancer Res 2004;10:7427–37.

18. Zou X, Feng B, Dong T, Yan G, Tan B, Shen H, et al. Up-regulation of type Icollagen during tumorigenesis of colorectal cancer revealed by quantitativeproteomic analysis. J Proteomics 2013;94:473–85.

19. Kauppila S, Stenback F, Risteli J, Jukkola A, Risteli L. Aberrant type I andtype III collagen gene expression in human breast cancer in vivo. J Pathol1998;186:262–8.

20. Ramaswamy S, Ross KN, Lander ES, Golub TR. A molecular signature ofmetastasis in primary solid tumors. Nat Gen 2003;33:49–54.

21. Netti PA, Berk DA, SwartzMA,Grodzinsky AJ, Jain RK. Role of extracellularmatrix assembly in interstitial transport in solid tumors. Cancer Res2000;60:2497–503.

22. Ryschich E, Khamidjanov A, Kerkadze V, Buchler MW, Zoller M, Schmidt J.Promotion of tumor cell migration by extracellular matrix proteins inhuman pancreatic cancer. Pancreas 2009;38:804–10.

23. Provenzano PP, InmanDR, Eliceiri KW, Knittel JG, Yan L, Rueden CT, et al.Collagen density promotes mammary tumor initiation and progression.BMC Med 2008;6:11.

24. Shintani Y, Maeda M, Chaika N, Johnson KR, Wheelock MJ. Collagen Ipromotes epithelial-to-mesenchymal transition in lung cancer cells via

Jolly et al.





transforming growth factor-beta signaling. Am J Resp Cell Mol Biol 2008;38:95–104.

25. Franco AT, Malaguarnera R, Refetoff S, Liao XH, Lundsmith E, Kimura S,et al. Thyrotrophin receptor signaling dependence of Braf-induced thyroidtumor initiation in mice. PNAS 2011;108:1615–20.

26. Miller KA, Yeager N, Baker K, Liao XH, Refetoff S, Di Cristofano A.Oncogenic Kras requires simultaneous PI3K signaling to induce ERKactivation and transform thyroid epithelial cells in vivo. Cancer Res 2009;69:3689–94.

27. Cheng N, Bhowmick NA, Chytil A, Gorksa AE, Brown KA, Muraoka R,et al. Loss of TGF-beta type II receptor in fibroblasts promotes mam-mary carcinoma growth and invasion through upregulation of TGF-alpha-, MSP- and HGF-mediated signaling networks. Oncogene 2005;24:5053–68.

28. Muller PY, Janovjak H, Miserez AR, Dobbie Z. Processing of gene expres-sion data generated by quantitative real-time RT-PCR. BioTechniques2002;32:1372–4, 76, 78–9.

29. Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, Ghosh D, et al.ONCOMINE: a cancer microarray database and integrated data-miningplatform. Neoplasia 2004;6:1–6.

30. Giordano TJ, Au AY, Kuick R, ThomasDG, Rhodes DR,WilhelmKGJr, et al.Delineation, functional validation, and bioinformatic evaluation of geneexpression in thyroid follicular carcinomas with the PAX8-PPARG trans-location. Clin Cancer Res 2006;12:1983–93.

31. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al.Integrative analysis of complex cancer genomics and clinical profiles usingthe cBioPortal. Sci Signal 2013;6:pl1.

32. Cerami E,Gao J,DogrusozU,Gross BE, Sumer SO, Aksoy BA, et al. The cBiocancer genomics portal: an open platform for exploring multidimensionalcancer genomics data. Cancer Discov 2012;2:401–4.

33. Liu Z, Hou P, Ji M, Guan H, Studeman K, Jensen K, et al. Highlyprevalent genetic alterations in receptor tyrosine kinases and phospha-tidylinositol 3-kinase/akt and mitogen-activated protein kinase path-ways in anaplastic and follicular thyroid cancers. J Clin EndocrinolMetab 2008;93:3106–16.

34. Shingu K, Fujimori M, Ito K, Hama Y, Kasuga Y, Kobayashi S, et al.Expression of fibroblast growth factor-2 and fibroblast growth factorreceptor-1 in thyroid diseases: difference between neoplasms and hyper-plastic lesions. Endocrine J 1998;45:35–43.

35. Pasieka Z, Stepien H, Komorowski J, Kolomecki K, Kuzdak K. Evalu-ation of the levels of bFGF, VEGF, sICAM-1, and sVCAM-1 in serum ofpatients with thyroid cancer. Recent Results in Cancer Research For-tschritte der Krebsforschung Progres dans les recherches sur le cancer2003;162:189–94.

36. St Bernard R, Zheng L, Liu W, Winer D, Asa SL, Ezzat S. Fibroblast growthfactor receptors as molecular targets in thyroid carcinoma. Endocrinology2005;146:1145–53.

37. Kessenbrock K, Plaks V, Werb Z. Matrix metalloproteinases: regulators ofthe tumor microenvironment. Cell 2010;141:52–67.

38. Levental KR, Yu H, Kass L, Lakins JN, Egeblad M, Erler JT, et al. Matrixcrosslinking forces tumor progression by enhancing integrin signaling. Cell2009;139:891–906.

39. Erler JT, Bennewith KL, NicolauM,Dornhofer N, KongC, LeQT, et al. Lysyloxidase is essential for hypoxia-induced metastasis. Nature 2006;440:1222–6.

40. Boufraqech M, Nilubol N, Zhang L, Gara SK, Sadowski SM, Mehta A, et al.miR30a inhibits LOX expression and anaplastic thyroid cancer progres-sion. Cancer Res 2015;75:367–77.

41. Yarchoan M, LiVolsi VA, Brose MS. BRAF mutation and thyroid cancerrecurrence. J Clin Oncol 2015;33:7–8.

42. Mbeunkui F, Johann DJJr. Cancer and the tumor microenvironment: areviewof an essential relationship. Cancer Chemother Pharmacol 2009;63:571–82.

43. Xing M. BRAF mutation in papillary thyroid cancer: pathogenic role,molecular bases, and clinical implications. Endocrine Rev 2007;28:742–62.

44. Nucera C, Porrello A, Antonello ZA,Mekel M, NehsMA, Giordano TJ, et al.B-Raf(V600E) and thrombospondin-1 promote thyroid cancer progres-sion. PNAS 2010;107:10649–54.

45. Bayne LJ, Beatty GL, Jhala N, Clark CE, Rhim AD, Stanger BZ, et al. Tumor-derived granulocyte-macrophage colony-stimulating factor regulates mye-loid inflammation and T cell immunity in pancreatic cancer. Cancer Cell2012;21:822–35.

46. Ji H, Houghton AM, Mariani TJ, Perera S, Kim CB, Padera R, et al. K-rasactivation generates an inflammatory response in lung tumors. Oncogene2006;25:2105–12.

47. Whipple CA, Brinckerhoff CE. BRAF(V600E) melanoma cells secretefactors that activate stromal fibroblasts and enhance tumourigenicity. BrJ Cancer 2014;111:1625–33.

48. NgMR,Brugge JS. A stiff blow from the stroma: collagen crosslinking drivestumor progression. Cancer Cell 2009;16:455–7.

49. Pickup MW, Laklai H, Acerbi I, Owens P, Gorska AE, Chytil A, et al.Stromally derived lysyl oxidase promotes metastasis of transforminggrowth factor-beta-deficient mouse mammary carcinomas. Cancer Res2013;73:5336–46.

50. Paszek MJ, Zahir N, Johnson KR, Lakins JN, Rozenberg GI, Gefen A, et al.Tensional homeostasis and the malignant phenotype. Cancer Cell 2005;8:241–54.






2016;76:1804-1813. Published OnlineFirst January 27, 2016.Cancer Res Lee Ann Jolly, Sergey Novitskiy, Phillip Owens, et al.

and Pten LossV600EDriven by BrafMicroenvironment Facilitates Progression of Thyroid Cancers Fibroblast-Mediated Collagen Remodeling Within the Tumor

Updated version

10.1158/0008-5472.CAN-15-2351doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2016/01/27/0008-5472.CAN-15-2351.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/76/7/1804.full#ref-list-1

This article cites 50 articles, 15 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/76/7/1804.full#related-urls

This article has been cited by 3 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/76/7/1804To 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-15-2351

http://cancerres.aacrjournals.org/content/suppl/2016/01/27/0008-5472.CAN-15-2351.DC1

http://cancerres.aacrjournals.org/content/76/7/1804.full#ref-list-1

http://cancerres.aacrjournals.org/content/76/7/1804.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/76/7/1804


Documents

Fibroblast-Mediated Collagen Remodeling Within the Tumor ... · Statistical analysis All data were analyzed using Prism 6 software (GraphPad). Differences with P values of 0.05 were