Pericytes Elicit Resistance to Vemurafenib and Sorafenib ... · Pericytes Elicit Resistance to Vemurafenib and Sorafenib Therapy in Thyroid Carcinoma via the TSP-1/TGFb1 Axis Alessandro

Translational Cancer Mechanisms and Therapy

Pericytes Elicit Resistance to Vemurafenib andSorafenib Therapy in Thyroid Carcinoma via theTSP-1/TGFb1 AxisAlessandro Prete1,2, Agnes S. Lo3, Peter M. Sadow4, Swati S. Bhasin5, Zeus A. Antonello1,2,DanicaM.Vodopivec1,2, SoumyaUllas6, JenniferN.Sims1,2, JohnClohessy7,AnnM.Dvorak2,Tracey Sciuto2, Manoj Bhasin5, Joanne E. Murphy-Ullrich8, Jack Lawler2,S. Ananth Karumanchi3, and Carmelo Nucera1,2,9

Abstract

Purpose: The BRAFV600E oncogene modulates the papillarythyroid carcinoma (PTC) microenvironment, in which peri-cytes are critical regulators of tyrosine-kinase (TK)-dependentsignaling pathways. Although BRAFV600E and TK inhibitors areavailable, their efficacy as bimodal therapeutic agents inBRAFV600E-PTC is still unknown.

Experimental Design: We assessed the effects of vemurafe-nib (BRAFV600E inhibitor) and sorafenib (TKI) as single agentsor in combination in BRAFWT/V600E-PTC and BRAFWT/WT cellsusing cell-autonomous, pericyte coculture, and an ortho-topic mouse model. We also used BRAFWT/V600E-PTC andBRAFWT/WT-PTC clinical samples to identify differentiallyexpressed genes fundamental to tumor microenvironment.

Results: Combined therapy blocks tumor cell prolifer-ation, increases cell death, and decreases motility viaBRAFV600E inhibition in thyroid tumor cells in vitro.Vemurafenib produces cytostatic effects in orthotopictumors, whereas combined therapy (likely reflectingsorafenib activity) generates biological fluctuations with

tumor inhibition alternating with tumor growth. We dem-onstrate that pericytes secrete TSP-1 and TGFb1, andinduce the rebound of pERK1/2, pAKT and pSMAD3 levelsto overcome the inhibitory effects of the targeted therapyin PTC cells. This leads to increased BRAFV600E-PTC cellsurvival and cell death refractoriness. We find thatBRAFWT/V600E-PTC clinical samples are enriched in peri-cytes, and TSP1 and TGFb1 expression evoke gene-regulatory networks and pathways (TGFb signaling, metas-tasis, tumor growth, tumor microenvironment/ECM remo-deling functions, inflammation, VEGF ligand–VEGF recep-tor interactions, immune modulation, etc.) in the micro-environment essential for BRAFWT/V600E-PTC cell survival.Critically, antagonism of the TSP-1/TGFb1 axis reducestumor cell growth and overcomes drug resistance.

Conclusions: Pericytes shield BRAFV600E-PTC cells from tar-geted therapy via TSP-1 and TGFb1, suggesting this axis as anew therapeutic target for overcoming resistance to BRAFV600E

and TK inhibitors. Clin Cancer Res; 24(23); 6078–97. �2018 AACR.

IntroductionThe Surveillance, Epidemiology, andEndResults (SEER) cancer

registry reveals that advanced-stage papillary thyroid cancer (PTC)is less tractable to therapy than localized PTC, as implied byincreasing mortality rates (1). BRAFV600E is a potent regulator ofthe MAPK (e.g., ERK1/2) pathway which is highly mutated inhuman cancers, and has been identified as a crucial and commononcogene in PTC. This mutation is associated with loss of radio-

iodine avidity, higher rates of recurrence and metastases, poorerprognosis, and lower survival rates (2–4). Thus effective treatmentof advanced PTC through BRAFV600E would meet an urgentclinical need.

Vemurafenib is the first FDA-approved BRAFV600E inhibitor forthe treatment of BRAFV600E-positive metastatic melanoma (5).Vemurafenib binds to the ATP binding cassette of mutatedBRAFV600E and inhibits its pathway through ERK1/2. It has beenused against metastatic BRAFV600E-PTC that is refractory to

1Laboratory of Human Thyroid Cancers Preclinical and Translational Research,Division of Experimental Pathology, Cancer Research Institute (CRI), CancerCenter, Department of Pathology, Beth Israel Deaconess Medical Center,Harvard Medical School, Boston, Massachusetts. 2Department of Pathology,Center for Vascular Biology Research (CVBR), Beth Israel Deaconess MedicalCenter, Harvard Medical School, Boston, Massachusetts. 3Department ofMedicine, Center for Vascular Biology Research (CVBR), Beth Israel DeaconessMedical Center, HarvardMedical School, Boston,Massachusetts. 4Department ofPathology, Massachusetts General Hospital, Harvard Medical School, Boston,Massachusetts. 5Bioinformatic and Systems Biology Unit, Department ofMedicine, Beth Israel Deaconess Medical Center, Harvard Medical School,Boston, Massachusetts. 6Longwood Small Animal Imaging Facility (LSAIF),Department of Radiology, Beth Israel Deaconess Medical Center, HarvardMedical School, Boston, Massachusetts. 7Division of Cancer Genetics, Depart-ment of Medicine, Beth Israel Deaconess Medical School, Harvard Medical

School, Boston, Massachusetts. 8Departments of Pathology, Cell Developmentaland Integrative Biology, and Ophthalmology, University of Alabama at Birming-ham, Birmingham, Alabama. 9Broad Institute of MIT and Harvard, Cambridge,Massachusetts.

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

A. Prete and A.S. Lo contributed equally to this article.

Corresponding Author: Carmelo Nucera, Beth Israel Deaconess Medical Center,Harvard Medical School, Broad Institute of MIT and Harvard, Cambridge, MA02215. Phone: 617-667-5964; Fax: 617-667-3591; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-18-0693

�2018 American Association for Cancer Research.

ClinicalCancerResearch

Clin Cancer Res; 24(23) December 1, 20186078

on July 17, 2020. © 2018 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 3, 2018; DOI: 10.1158/1078-0432.CCR-18-0693

http://crossmark.crossref.org/dialog/?doi=10.1158/1078-0432.CCR-18-0693&domain=pdf&date_stamp=2018-11-6

http://clincancerres.aacrjournals.org/

radioiodine, but response has been variable and relapse is com-mon (6). Numerousmechanisms can promote resistance throughbypass of pharmacologic BRAFV600E inhibition, which enablesrebound of ERK1/2 (7, 8). Another mechanism of resistanceinvolves upregulation of proangiogenic molecules such as thetyrosine kinases (TK) in the microenvironment (9). BRAFV600E-PTC growth is influenced by the tumormicroenvironment, whichis in turn altered by the tumor itself, leading to abnormal extra-cellular matrix (ECM) deposition and activation of angiogenicpathways (10, 11). Many of the processes involved in thyroidtumor growth andmetastasis aremediatedby signalingmoleculesdownstream of activated TKs.

Another FDA-approved drug, sorafenib, inhibits both TKs andBRAF intracellular signaling. Sorafenib targets RET (includingRET/PTC) to inhibit proangiogenic pathways such as VEGFR2and PDGFRB (12), and is currently used to treat hepatocellularcarcinoma, advanced renal carcinoma, and metastatic thyroidcarcinoma (13). However, resistance to sorafenib can developin the first-line setting in patients with thyroid carcinoma (12, 14,15), and BRAFV600E may impact treatment duration and promoteresistance mechanisms to TK inhibitors.

Angiogenic factors confer survival advantages and are over-expressed in both tumor cells and blood vessels in PTC (16). Theangiogenic microenvironment includes pericytes, which are het-erogeneous stromal cell populations that are fundamental tovessel stabilization and maturation and express a range of angio-genic factors (e.g., PDGFRB, VEGF, etc.; ref. 17). Pericytes regulateparacrine communications between tumor cells and microvascu-lar endothelial cells in the thyroid gland (18) and human tumors(17). In the tumor vasculature, pericytes protect endothelial cellsfrom antiangiogenic therapies, andmay be players in resistance tovascular and microenvironment targeting drugs. They likewiseconfer survival advantages to endothelial cells through the secre-tion of proangiogenic factors (19). However, pericyte depletionusing genetic methods has led to more metastasis and enhancedepithelial-to-mesenchymal transition (EMT) in breast cancermodels (20).

Thrombospondin-1 (TSP-1; THBS1 gene) is produced bytumor and stromal cell types and plays a fundamental role inregulating the angiogenic microenvironment as well as cell pro-

liferation, adhesion, migration, and invasion, and angiogenesis(21). Its functional domains are crucial for cell–cell or cell–ECMinteractions. It is enriched in the thyroid carcinoma microenvi-ronment and plays an important role in tumor aggressiveness(22). TSP-1 is also a key regulator of latent TGFb activation, theconversion of latent TGFb to its biologically active form in certaindiseases (23). TGFb is an early tumor suppressor that is also aplayer in the metastatic switch of tumors, and promotes EMT andmetastasis (24). TGFb-induced SMAD phosphorylation andEMT induction required MAPK pathway activation in murinethyrocytes derived from BRAFV600E mice, indicating thattumor initiation by BRAFV600E predisposes murine thyroid cellsto TGFb-induced EMT, through a MAPK-dependent process (25).

The inefficacy of single-agent BRAFV600E or TKs inhibitors, andthe eventual resistance to these agents, even in combination,highlight the need for a better understanding of the tumormicroenvironment, including the cross talk between pericytesand tumor cells. Therefore, in this study, we have analyzed theeffects of combined vemurafenib plus sorafenib therapy inBRAFV600E PTC patient-derived cells using cell cultures andin vivo models. Our results demonstrate that pericyte-derivedsecretomes increase pERK1/2, pAKT, and pSMAD3 levels inthyroid tumor cells to overcome the inhibitory effects of vemur-afenib and sorafenib either alone or in combination. Pericyte-derived factors also increased survival of BRAFV600E tumor cellsand refractoriness to tumor cell death. We demonstrate thatpericytes are a source of both TSP-1 and TGFb, and that antag-onism of TSP1-dependent activation of latent TGFb1 overcomesresistance to BRAFV600E inhibitors or TKI. Together, these dataprovide evidence. Thus, pericytes elicit resistance to vemurafeniband sorafenib therapy via the TSP-1/TGFb1 axis, suggesting thisaxis as a promising new target in overcoming therapy resistance.

Materials and MethodsCell cultures

We used authenticated (short tandem repeat and DNAsequencing for KTC1; DNA sequencing and RT-PCR for TPC1)KTC1 (BRAFWT/V600E) and TPC1 (BRAFWT/WT) human thyroidcarcinoma cell lines, and human pericytes (BRAFWT/WT) wereobtained from Promo Cell (18). The use of these cell lines wasapproved from the committee on microbiological safety [COMS,Beth Israel Deaconess Medical Center (BIDMC), Boston, MA].KTC1 is a spontaneously immortalized human thyroid carcinomacell line that harbors BRAFWT/V600E mutation. It was establishedfrom the metastatic pleural effusion from recurrent and radio-iodine (RAI)-refractory PTC in a 60-year-old male patient (26) byDr. J. Kurebayashi (Department of Breast and Thyroid SurgeryKawasaki Medical School Kurashiki, Japan) and provided byDr. Rebecca E. Schweppe (University of Colorado, Aurora, CO).

Drug treatmentsFor our in vitro assays, we used 10 mmol/L vemurafenib

(PLX4032, RG7204, catalog no. S1267) (Selleckchem) dissolvedin 100% dimethyl sulfoxide (DMSO, vehicle). Sorafenib tosylate(catalog no.S1040, Selleckchem), a multikinase inhibitor, wasdissolved in 100% DMSO (Sigma) according to the manufac-turer's instructions to produce 10 mmol/L stock solution. Inter-mediate doses of vemurafenib or sorafenib were prepared in100% DMSO and diluted in 0.2% fetal bovine serum (FBS)DMEM to achieve desired final concentrations, maintaining a

Translational Relevance

Metastatic, radioiodine refractory heterozygousBRAFWT/V600E-PTC is resistant to targeted therapies andrepresents an unanswered clinical challenge. We showthat combined therapy with vemurafenib and sorafenibin vivo does not provide therapeutic efficacy as comparedto single-agent treatments, likely due to tumor angiogenicactivity in the microenvironment. Furthermore, wepropose a novel model of drug resistance mediated byPTC microenvironment-associated pericytes via the TSP-1/TGFb1 axis. Antagonizing the TSP-1/TGFb1 axis mayrepresent a novel therapeutic approach with translationalapplications for BRAFWT/V600E-PTC resistant to targetedtherapies. Finally, TSP-1 is a potential biomarker forassessing therapeutic response to BRAFV600E and TK inhi-bitors in patients with invasive BRAFWT/V600E-PTC.

TSP-1, TGFb1, and Drug Resistance in Thyroid Cancer

www.aacrjournals.org Clin Cancer Res; 24(23) December 1, 2018 6079




constant final concentration at 2% DMSO for optimal solubility(see Supplementary Methods). Synergy, subadditive, or additiveactivity for the combined treatments of vemurafenib plus sora-fenibwere estimated usingGeoGebra Classic and applying Loewetestmethod according to Tallarida (27) to assess drug synergy andantagonism. Cells were treated for 48 hours in the presence of0.2% FBS DMEM at final 2% DMSOwith: 1, 2.5, 5, or 10 mmol/Lof either vemurafenib or sorafenib; or combined therapy withvemurafenib plus sorafenib combining all above doses. Vehiclewas used as untreated control (2% DMSO diluted in 0.2% FBSDMEM). Before adding treatments, cells were washed with PBSfrom 10% FBS DMEM. Quantitative analysis was performed bycrystal violet assays (see Supplementary Methods) of adherentcells (magnification: 10�). Vehicle (control) was 2% DMSOdiluted in 0.2% FBS DMEM.

SRI31277 peptidePeptide SRI31277 (24) was synthesized by BioMatik and

purity confirmed at Southern Research. We reconstituted thepeptide in 0.2% FBS DMEM) to achieve the stock concentrationof 2.6 mmol/L. SRI31277 was diluted in 0.2% FBS DMEMto achieve final concentration of 1 mmol/L, 2.5 mmol/L,5 mmol/L, 10 mmol/L, 25 mmol/L, 50 mmol/L, or 100 mmol/L.

Model of pericyte secretomePericytes were seeded at about 90% confluence in 6-well dishes

in DMEM growth medium supplemented with 10% FBS. Forty-eight hours following cell seeding, pericytes were treated for5 hours with 10 mmol/L vemurafenib, 2.5 mmol/Lsorafenib, combined therapy with 10 mmol/L vemurafenib plus2.5 mmol/L sorafenib, or vehicle (2% DMSO) in the presence of0.2%FBSDMEMgrowthmedium. Following treatment, the 0.2%FBS DMEM cell growthmedium enriched by cell-derived secretedprotein factors was defined as secretome and was normalized tothe same cell growthmedium to subtract background; then, it wascollected and separated from dead cell debris by short spin. Wecollected an aliquot of secretome volume for ELISA analysis. Inaddition, the remaining volumeof all four secretomeswas used totreat BRAFWT/V600E-KTC1 andBRAFWT/WT-TPC1 for 5 hours. At thesame time, another condition included BRAFWT/V600E-KTC1 andBRAFWT/WT-TPC1 cells (both cell lines were seeded at 90% con-fluence in the presence of 10% FBS DMEM growth medium theday prior to treatments) directly treated (without pericyte secre-tome) for 5 hours with 10 mmol/L vemurafenib, 2.5 mmol/Lsorafenib, combined therapy with 10 mmol/L vemurafenib plus2.5 mmol/L sorafenib, or vehicle (2% DMSO) in the presenceof 0.2% FBS DMEM growth medium. Also, after secretomecollection, adherent pericytes were lysed for protein extractionto perform Western blotting assays. After treatment of theBRAFWT/V600E-KTC1 and BRAFWT/WT-TPC1 thyroid tumor cells,we collected an aliquot of secretome volume for ELISA analysis(Supplementary Materials and Methods). Then we lysed theadherent thyroid tumor cells for protein extraction in order toperform Western blotting assays.

Model of cell coculturemCherry-KTC1 and pericytes were seeded at 1.5 � 105 cells

(1:1) per well in 6-well dishes or 3� 104 cells per well in 24-welldishes. Forty-eight hours after cell seeding, cells were treated for48 hours with 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib,combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L

sorafenib, or vehicle (2% DMSO) in the presence of 0.2% FBSDMEM growthmedium. After 48 hours, the 0.2% FBSDMEM cellgrowth medium enriched by cell-derived secreted protein factorswas defined as secretome and was normalized to the same cellgrowth medium to subtract background, then it was collectedfor ELISA analysis (Supplementary Materials and Methods) andseparated from dead cell debris by short spin. Adherent cellswere fixed with 10% formalin for 20 minutes at room tempera-ture. Cells were then washed with PBS. Cells were stainedwith 5 mmol/L Hoechst 33342 (Nexcelom) diluted in PBSfor 15 minutes. Total number of cells (cell growth)/well wasanalyzed by Celigo image cytometer (Nexcelom). Data wereplotted as matrix of cell count using both GraphPad Prism 6 andexcel software.

Western blottingCells were grown in 10-cm dishes in 10% FBS DMEM. They

were treated with vehicle (DMSO), vemurafenib, sorafenib, orcombined vemurafenib plus sorafenib therapy in the presence of0.2% FBS DMEM when reached about 90%–100% confluence.Western blotting assays were performed according to standardprocedure (11). The intensity of each protein band was normal-ized to housekeeping protein band (tubulin or actin) and quan-tified by densitometry analysis (ImageJ software).

Orthotopic mouse modelAll animal work was approved and done in accordance with

federal, local, and institutional guidelines (IACUC) at the BIDMC(Boston, MA). Human metastatic KTC1 tumor-derived cells har-boring the heterozygous BRAFV600E mutation and engineered toexpress luciferase were cultured in 10-cm dishes and grown inDMEM supplemented with 10% FBS, penicillin, streptomycin,and amphotericin at 37�C with 5% CO2 atmosphere. Prior toimplantation, cells were trypsinized, gently centrifuged, and sus-pended in serum-free DMEM growth medium to achieve a cellsuspension concentration ranged between 3.5� 106 and 5� 106

cells/10 mL. The cells were kept on ice until implantation.KTC1 cells were orthotopically injected in the right thyroid of9-week-old male NSG mice (strain name: NOD.Cg-Prkdcscid

Il2rgtm1Wjl/SzJ; stock number: 005557; n¼ 5per group) accordingto our previous experimental procedures (22). Mice wererandomly divided into four groups of 5 for the purpose ofestablishing a timeline of tumor cell growth and response totherapy with vemurafenib, sorafenib, combined therapy (vemur-afenib plus sorafenib), or vehicle. Treatments were started 6weeksafter KTC1 tumor cells implantation and performed for 5 weeks.

Vemurafenib and sorafenib preparation for mouse treatmentFor in vivo studies, drug suspensions were prepared for vemur-

afenib (10 mg/mL in 2% hydroxypropylcellulose) and sorafenib(3.75 mg/mL according to Fendrich and colleagues; ref. 28).Freshly prepared drug suspensions were stored at 4�C and usedwithin 48 hours. Mice were dosed once daily with vehicle alone(control), vemurafenib (100 mg/kg), sorafenib (30 mg/kg), orcombination of vemurafenib and sorafenib as indicated by oralgavage using a 22G needle.

Differential gene expression, regulatory networks, pathwaysanalysis, and pericytes abundance score in PTC clinical samples

To determine the association of a select set of 23 genes linked toextracellular matrix functions, pericyte functions, angiogenesis,

Prete et al.

Clin Cancer Res; 24(23) December 1, 2018 Clinical Cancer Research6080




cell growth, adhesion/migration/invasion, and metastasis path-ways with BRAF mutational status, we performed analysis onthe genes PTC TCGA data. We downloaded RNA-seq data ofPTC from TCGA to analyze 23 genes differentially expressedusing 211 BRAFWT/V600E-PTC, 23 PTC harboring BRAFWT/V600E

and hTERT mutations, and 256 BRAFWT/WT-PTC samples. Afterperforming analysis on the 23 selected genes, we consideredonly those with raw P values < 0.05 and fold-change (FC) �1.2or ��1.2 as significantly associated with BRAF mutationalstatus. Network analysis was also performed. More details arereported in the Supplementary Methods. Pericyte abundanceanalysis was assessed by the ssGSEA algorithm using RNAseqexpression data from NT and PTC TCGA samples. SSGSEA calcu-lates separate enrichment scores for each pairing of a sampleand gene set; each enrichment score represents the degree towhich the genes in a particular set are coordinately upregulatedor downregulated. We specified the positive signature for peri-cytes based on the expression of NG2, PDGFRB, aSMA, andCD90 genes. To identify samples enriched with pericytes, weused the genes PECAM1, LYVE1, and CD34 as negative signature.On the basis of score differences, we ranked samples from thehighest enrichment of pericyte signature to the least enrichment.

Statistical analysisStatistical analysis was carried out using GraphPad Prism 6

software, Microsoft Excel, and GeoGebra Classic statistical tools.c2test, Student t test, Mann–Whitney test, one-way ANOVA formultiple comparisons tests, and Pearson correlation analysis wereused. Data are reported as the averaged value, and error barsrepresent the SD of the average for each group. Results withP values below 0.05 were considered statistically significant.

We also used virus transduction assays (for gene overexpressionor knockdown); gene-regulatory networks/pathway analyses; andTEM (for more details, see the Supplementary Methods).

ResultsAnti-BRAFV600E (vemurafenib) and anti-TK (sorafenib)combined therapy blocks cell proliferation, increases celldeath, and decreases motility in PTC cells.

As a first step toward suppressing PTC cell survival, we assessedanti-BRAFV600E (vemurafenib) and anti-multi-TK (sorafenib)combined therapy in heterozygous BRAFWT/V600E (KTC1) orBRAFWT (TPC1) tumor cells derived from invasive PTC(Fig. 1A). We have previously characterized the PTC microenvi-ronment, which has pericytes (11), crucial components of thevasculature known to express TKs (e.g., PDGFRB), as well as othermarkers such asNG2 andaSMA (17).Our results confirmed thesefindings (Fig. 1B), and human pericytes were also negative formesenchymal and endothelial cell markers (Fig. 1B). Further-more, transmission electron microscopy (TEM) revealed thathuman well-differentiated thyroid carcinoma tissue were charac-terized by endothelial cells and pericytes with a large nucleus andlittle cytoplasm as compared with the normal thyroid (NT) tissue(Fig. 1C). Pericytes may impact resistance in cancer by regulatingTK-dependent angiogenic signaling pathways. We therefore ana-lyzed the expression of twomajor TKs (PDGFRB and VEGFR2) inPTC-derived cells and human pericytes (Fig. 1D). Pericytesshowed 3.7-fold and 5.1-fold changes in phospho(p)-PDGFRB-Y751 levels upon PDGFB treatment, and 1.1-fold and 1.38-foldchanges in pVEGFR2-Y1059 levels upon VEGFA treatment when

compared with KTC1 and TPC1 cells, respectively (Fig. 1D).PDGFB and VEGFA stimulated phosphorylation of PDGFRB andVEGFR2 in pericytes and BRAFWT/V600E-PTC, but not in BRAFWT/

WT-PTC, which showed low levels of these receptors (Fig. 1D).PDGFB stimulated VEGFR2 phosphorylation more than VEGFAdid, possibly due to potential PDGFRB/VEGFR2 heterodimers;however, further studies are needed to define this phenomenon.To determine whether simultaneous inhibition of BRAFV600E andTKs was effective against tumor cells, we combined vemurafeniband sorafenib. Previous studies had identified the dose–responsecurve for vemurafenib (IC50, 50% maximal inhibitory concen-tration) in thyroid cancer cells (11); to assess the most effectivedoses of combined vemurafenib plus sorafenib, we treated tumorcells for 48 hours with seven different drug dose combinations(Supplementary Fig. S1). We used isobolographic analysis(Fig. 1E and F) to assess synergy or additivity. Our results showedthat 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib hadthe highest therapeutic efficacy (synergistic effect) againstBRAFWT/V600E-PTC cells (Fig. 1E), but was subadditive inBRAFWT/WT-PTC cells (Fig. 1F) and pericytes (SupplementaryFig. S2A). Specifically, the targeted therapy achieved a significantlylower viability (31%, 46%, and 49% by vemurafenib, sorafenib,and combined therapy, respectively) in BRAFWT/V600E-PTC cellsthan in pericytes (Supplementary Fig. S2B). Similarly, combinedtherapy significantly decreased BRAFWT/V600E-KTC1 cell viabilitycompared with vehicle (60% decrease), vemurafenib (31%decrease), or sorafenib (20% decrease; Supplementary Fig. S1).In BRAFWT/WT-TPC1 cells, combination therapy yielded a 23%decrease in cell viability compared with vehicle, 18% decreasecompared with vemurafenib, and a 52% decrease compared withsorafenib (Supplementary Fig. S3). We also found significantinduction of cell death in PTC-derived cells when we used com-bined therapy, as shown by cell death analysis (Fig. 1G). Com-bined therapy was significantly more effective than vehicle(21-fold increase), vemurafenib (7.3-fold increase), or sorafenib(10.8-fold increase) in inducingdeath inBRAFWT/V600E-KTC1 cells(Fig. 1G). Also, vemurafenib (2.2-fold change increase) orcombined therapy (2.1-fold change increase) induced significant-ly higher rate of cell death in BRAFWT/V600E-KTC1 than BRAFWT/

WT-TPC1 cells, indicating the higher specificity of vemurafenib intargeting BRAFWT/V600E PTC cells versus BRAFWT/WT PTC cells. Incontrast, sorafenib was more effective on BRAFWT/WT-TPC1 cellscompared with BRAFWT/V600E-KTC1 (0.51-fold change).

Furthermore, analysis of cell proliferation by BrdU (5-bromo-2-deoxyuridine) assay showed that targeting BRAFV600E by vemur-afenib and sorafenib significantly inhibitedDNA synthesis of PTCcells compared with vehicle or single agents. This effect was likelydriven by combined therapy in BRAFWT/V600E-KTC1 cells, but bysorafenib alone in BRAFWT/WT-TPC1 cells (Fig. 1H and I). Signif-icantly, high doses of vemurafenib were also required to inhibitmelanoma cell viability (29). Since BRAFV600E is a strong regulatorof tumor cell migration (30), we applied a monolayer wound-healing (cell motility) assay to study the effects of cell migrationduring treatment (Fig. 1J and K). We measured the initial woundat baseline (time zero), treated the cellswith vehicle, vemurafenib,sorafenib, or a combination of the two for 7 hours (a time pointprior to any observed effects on cell proliferation), and thenquantified healing area at 7 hours versus baseline in each condi-tion (Fig. 1J and K). Importantly, BRAFWT/V600E-KTC1 cells dis-played a significant decrease in cell motility upon combinationtreatment as compared with vehicle at 7 hours (Fig. 1J and K). In






Figure 1.

Combined therapy with vemurafenib plus sorafenib induces stronger cell death compared to single-agent treatment in PTC cells. A, Diagram of targetedtherapy with BRAFV600E inhibitor and tyrosine kinase (TK) inhibitors against human invasive thyroid carcinoma cells harboring the heterozygous BRAFV600E

mutation. B, Immunocytochemistry and hematoxylin and eosin (H&E)-stained sections of formalin-fixed, paraffin-embedded (FFPE) cell blocks ofrepresentative human pericytes in vitro. Immunocytochemistry staining shows cytoplasmic to membranous staining with antibodies against PDGFRB, NG2,and aSMA. Desmin and CD31 immunostain was negative. C, Representative transmission electron microscopy (TEM). Above, venule from normal humanthyroid shows a typical endothelial cell (asterisk) with vascular lumen (pound sign) filled with plasma and enveloped by a pericyte (black arrow). A small pericyte(arrow) surrounds the abluminal surface of the endothelium. Below, in contrast, venule from a well-differentiated human thyroid carcinoma (n ¼ 3) shows anendothelial cell (asterisk) with large, activated nucleus and thinned cytoplasm (arrowhead) enveloped by an activated pericyte (black arrow). Vascular lumen isfilled with densely packed red blood cells (white pound sign) and with little plasma present, which is a characteristic sign (55) of vascular permeability to plasma.Scale bar size, 500 nm.D,Western blotting analysis (protein loading: 70 mg/lane) of protein expression levels in human BRAFWT-pericytes, BRAFWT/V600E-KTC1 andBRAFWT/WT-TPC1 cells at 20 minutes post stimulation with VEGF (20 ng/mL) or PDGFB (20 ng/mL). (Continued on the following page.)

Prete et al.





addition, vemurafenib treatment caused a decrease in cellmotilitycompared with vehicle, while sorafenib did not (Fig. 1J and K). Insummary, combined therapy blocked tumor cell proliferation,increased cell death, and decreased motility in BRAFWT/V600E-PTCcells, likely via BRAFV600E inhibition.

Effects in vivo on tumor growth by targeted therapy in anorthotopic mouse model of human BRAFWT/V600E-PTC

We have developed the first interventional preclinical mousetrial of vemurafenib therapy in BRAFWT/V600E-PTCpatient-derivedcells (Fig. 2A). Immunocompromised mice were orthotopicallyimplanted with human KTC1 cells derived from recurrentBRAFV600E-positive PTC, and engineered to express luciferase(Fig. 2B). Orthotopic tumors developed in all mice and wereanalyzed 6 weeks after injection as baseline (Fig. 2B). Mice werethen randomized for treatment with vehicle, vemurafenib, sor-afenib, or combined therapywith vemurafenib plus sorafenib. Alltumors in the vehicle-treated mice exhibited a 3-fold increase ingrowth over baseline at week 5 (Fig. 2C). The vemurafenib dosagewas similar to that of other studies (31); here we found that onlyvemurafenib resulted in consistent reduction in tumor growth,likely due to cytostatic effects, with 61% significant reduction intumor growth atweek 4 and48%reduction atweek 5 as comparedwith vehicle-treatedmice (Fig. 2B andC). Therapeutic response tosorafenib was fluctuant and resulted in a smaller reduction (23%)in tumor growth at week 4 and increased reduction (59%) atweek5 as compared with vehicle (Fig. 2B and C). Combined therapyfollowed the pattern of sorafenib activity and yielded a 5.1-foldincrease (36%) in tumor growth at week 4 versus vehicle, while areduction (43%) in tumor growth occurred at week 5 versusvehicle (Fig. 2B and C). We observed no apparent toxic sideeffects upon either single-agent treatment or combined therapy.Because inhibition of BRAFV600E may redifferentiate thyroidtumor cells (32), we analyzed this phenomenon by micro-SPECT/CT imaging. Vemurafenib-treated mice at 5 weeks post-treatment showed >2-fold increase in 99mTc uptake, suggestingthat targeting BRAFV600E could block not only tumor growth butalso induce thyroid tumor redifferentiation more substantiallythan sorafenib (Fig. 2D and E).

Pericyte secretome via the TSP-1/TGFb1 axis evokes resistanceto targeted therapy in BRAFWT/V600E-PTC cells

Wehave hypothesized that pericytes, which are fundamental tovessel maturation (33), are also fundamental to thyroid tumorcell viability, and limit the efficacy of BRAFV600E inhibitors and

TKI. To test this hypothesis, we developed an experimental modelusing secretome derived fromhuman pericytes treated for 5 hourswith either vehicle, vemurafenib, sorafenib, or combined therapy.Pericytes were grown in medium with low concentration of FBS(i.e., 0.2%) during treatment. The 0.2% FBS cell growth mediumenriched by pericyte-derived secreted factors within 5 hours ofdrug or vehicle treatment was defined as the secretome and wasnormalized to the same cell growth medium to subtract back-ground. A multiplex ELISA assay (Supplementary Materials andMethods) including themost important cytokines and angiogenicfactors showedno changes in secretion levels upondrug treatment(Supplementary Fig. S4). Intracellular TSP-1 protein expression inpericytes was upregulated within 5 hours of drug treatments(2.9-fold change with vemurafenib, 1.6-fold change with sorafe-nib, and 1.7-fold change with combined therapy vs.vehicle; Fig. 3A). In contrast, levels of secreted TSP-1 either fellor remained unchanged after treatment (12.1% decrease withvemurafenib, unchanged with sorafenib, and 15.3% decreasewith combined therapy vs. vehicle; ELISA assay, SupplementaryMaterials andMethods; Fig. 3B). Levels of secreted TGFb1 likewisefell or remained unchanged after treatment (17.8% decrease withvemurafenib, unchanged with sorafenib, and 27.3% decreasewith combined therapy vs. vehicle; ELISA assay, SupplementaryMaterials and Methods; Fig. 3B). TSP-1 is a mediator of TGFb1activation, which regulates many cell functions through SMAD,ERK1/2, AKT proteins (23, 24). Moreover, we found that TKslevels were affected by the targeted therapy, that is, PDGFRBphosphorylation levels were reduced upon drug treatment(43.6% decrease upon vemurafenib, 70.1% decrease upon sor-afenib, and 58.8% decrease upon combined therapy) versusvehicle (Fig. 3A). pVEGFR2 protein decreased upon treatmentwith vemurafenib (20.0%) or sorafenib (22.4%), however, nosubstantial changes were observed upon combined therapy(Fig. 3A). Protein expression of prosurvival factors likewise fellin pericytes after drug treatment versus vehicle treatment: (i) pAKT(94.5% decrease upon vemurafenib, 80.2% decrease uponsorafenib and 93.9% decrease upon combined therapy), (ii)pERK1/2 (15.7% decrease upon vemurafenib, 74.8% decreaseupon sorafenib and 82.7% decrease upon combined therapy),and (iii) pSMAD3 (7.7% increase upon vemurafenib, 15.6%increase upon sorafenib, and 21.1% decrease upon combinedtherapy; Fig. 3A).

It is known that BRAFV600E inhibitors such as vemurafenibselectively inhibit MAPK signaling (e.g., ERK1/2) in BRAFWT/V600E

thyroid tumor cells (32, 34, 35). To investigate the effects of

(Continued.) Both tubulin blots are different and derive from two different membranes. These results were validated by two independent experiments.E and F, Visualization of drug combinations: dose-effect analysis of combined therapy with vemurafenib plus sorafenib versus vehicle on BRAFWT/V600E-KTC1and BRAFWT/WT-TPC1 cell viability (cells were grown in 0.2% FBS DMEM growth medium during treatment). Each point represents the mean of threereplicates from two independent measurements. This method uses the dose-effect data of the individual drugs and drugs combined doses. The area ofsynergyor subadditive effects is distinguished by the line (isobole curve),which indicates additive effects. The red dot highlights the best dose-effect using combinedtreatment with vemurafenib plus sorafenib. G, Quantification of cell death by annexin V and propidium iodide (PI) dual staining assay in BRAFWT/V600E-KTC1 andBRAFWT/WT-TPC1 cells at 48 hours treatment with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, and combined therapy with 10 mmol/Lvemurafenib plus 2.5 mmol/L sorafenib; cells were grown in 0.2% FBS DMEM growth medium during treatment. These data represent the average � SD(error bars) of two independent replicate measurements (n ¼ 3 for each condition, � , P < 0.05; �� , P < 0.01; ��, P < 0.001). H and I, Quantification of cellproliferation by combined BrdU (5-bromo-2-deoxyuridine) pulse/PI (propidium iodide) by flow cytometry analysis of BRAFWT/V600E-KTC1 and BRAFWT/WT-TPC1cells at 48 hours treatment with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, and combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/Lsorafenib; cells were grown in 0.2% FBS DMEM growth medium during treatment. These data represent the average � SD (error bars) of two independentreplicate measurements (� , P < 0.05; �� , P < 0.01). J and K,Quantification of BRAFWT/V600E-KTC1 and BRAFWT/WT-TPC1 cell motility was analyzed by wound-healingassay. Cells were grown in 0.2% FBS DMEM growth medium during treatment. Images were captured at 0 and 7 hours after culture scratch. These data arerepresentative of two independent replicate measurements calculating percentage of healing area at 7 hours compared with 0 hours (basal area) for each condition.Statistical analysis was performed comparing drug treatments versus vehicle (�� , P < 0.001).






Figure 2.

Effects of targeted therapy with vemurafenib and sorafenib in a late intervention model of an orthotopic mouse using PTC-derived KTC1 cells harboring theheterozygous BRAFV600E mutation. A, Experimental design of an in vivo late intervention orthotopic preclinical model using BRAFWT/V600E-KTC1 cells derivedfrom a patient with invasive PTC harboring the heterozygous BRAFWT/V600E mutation. Human BRAFWT/V600E-KTC1 cells, engineered to express luciferase,were implanted in 20 nine-week-old male NSG mice, which were then treated with vemurafenib, sorafenib, combined therapy, or vehicle (n ¼ 5 for eachexperimental condition). Their orthotopic tumors were evaluated by histology and bioluminescence imaging. Either vehicle, vemurafenib (100 mg/kg, onetime daily), sorafenib (30 mg/kg, one time daily), or combined therapy vemurafenib (100 mg/kg, one time daily) plus sorafenib (30 mg/kg, one time daily)treatmentswere begun at 6weeks posttumor implantation, and the response to drugswas evaluatedweekly for 5weeks.B,BLI (bioluminescence imaging, emissionof photons/second) analysis for tumor growth assessment (by luciferase signal) in mice treated daily with vehicle, vemurafenib, sorafenib, or combined therapy forfive weeks. C, Total flux analysis of BLI (emission of photons/second) and distribution of data showed that five weeks of vemurafenib treatment resulted insignificantly lower orthotopic tumor growth than in controls (vehicle treatment; � , P < 0.05, Mann–Whitney test). Sorafenib therapy was fluctuant and resulted in asmaller reduction in tumor growth. Combined therapy showed a rising trend in tumor growth at week 4 followed by a decrease atweek 5 versus vehicle treatment. B,baseline. D, MicroSPECT/CT representative images of results plotted in E. T, mouse thyroids; S, mouse salivary glands. E, Quantification of technetium-99 (99mTc)uptake in BRAFWT/V600E-positive KTC1 orthotopic tumors based on the standardized uptake value (SUV) calculated by tissue radioactivity concentration/injectedactivity/body weight in grams and obtained by combining CT (computed tomography) and microSPECT (single photon emission computed tomography)imaging analysis performed in mice at four weeks posttreatment with vehicle (n ¼ 2), vemurafenib (100 mg/kg, one time daily; n ¼ 2), sorafenib (30 mg/kg,one time daily; n ¼ 2), or combined therapy with vemurafenib (100 mg/kg, one time daily) plus sorafenib (30 mg/kg, one time daily; n ¼ 2). These data representthe fold change (FC) � SD (error bars) compared with the vehicle treatment.

Prete et al.





pericytes on PTC-derived KTC1 or TPC1 cells, we treated thesethyroid tumor cells for 5 hours with vemurafenib, sorafenib, orcombined therapy in the presence or absence of the pericyte-derived conditioned medium (secretome) containing secretedTSP-1 and TGFb1 (Fig. 3C). Importantly, pericytes andBRAFWT/V600E-KTC1 cells were substantially responsive to treat-ment with exogenous recombinant human latent TGFb1 protein(while response in BRAFWT/WT-TPC1 cells was less robust), whichupregulated pSMAD3 protein levels (Fig. 3D), suggesting thepresence of endogenous regulators of latent TGFb1. When weused a TSP-1 antagonist (i.e., SRI31277) derived from the LSKLsequence of latent TGFb1 that blocks TSP1-mediated TGFb1activation (24) plus latent TGFb1, we found pSMAD3 proteinexpression decreased 19% in pericytes versus latent TGFb1 treat-ment alone (Fig. 3D).

Because TSP-1 is also a key-player in aggressive anaplasticthyroid carcinoma (ATC) harboring BRAFV600E (22), we analyzedTSP-1 protein levels in aggressive PTC-derived cells. Interestingly,BRAFWT/WT-TPC1 cells showed low intracellular TSP-1 proteinlevels (Supplementary Fig. S5); also, secreted TSP-1 levels (as wellas TKsproangiogenic factors, i.e., VEGFR2or PDGFRB levels)weresubstantially lower (20–50 folds) in BRAFWT/WT-TPC1 cellscompared with BRAFWT/V600E-KTC1 cells across all treatments(Supplementary Fig. S6A), suggesting expression of these factorsmight depend on the BRAFV600E pathway. Indeed, TSP-1 proteinexpressionwas downregulated by direct drug treatments (withoutthe presence of pericyte secretome) compared with vehicle inBRAFWT/V600E-KTC1 cells, i.e., 15.5% by vemurafenib, 32.4% bysorafenib and 51.5%by combined therapy (Fig. 3E). Importantly,direct treatment by combined therapy more effectively down-regulated both TSP-1 (51.5% vs. vehicle), and the intracellularsignaling effectors pERK1/2 (88% vs. vehicle), pAKT (58.8% vs.vehicle), and pSMAD3 (33.8% vs. vehicle), as well as TKspVEGFR2 (38.2% vs. vehicle) and pPDGFRB (9.5% vs. vehicle)than single agents in BRAFWT/V600E-KTC1 cells (Fig. 3E).Vemurafenib treatment upregulated pERK1/2, as expected, inBRAFWT/WT-TPC1 cells (Supplementary Fig. S5), likely due toparadoxical effects (36).

Because ourmouse data suggested that kinase and angiogenesisinhibitors such as vemurafenib or sorafenib elicited cytostaticeffects with differing levels of pharmacologic action (Fig. 2B andC), probably due to angiogenic microenvironment-mediatedeffects, we focused our attention on pericytes, denizens of thetumormicroenvironment, which are critical to vessel stabilizationand angiogenic endothelial functions (33). To understand thefunctional role of pericytes in paracrine communication with PTCcells, we assessed the ability of conditioned media (secretome)derived frompericytes to influence thyroid tumor cell intracellularsignaling upon vehicle, vemurafenib, sorafenib, or combinedtreatment. As all pericyte secretome was collected within 5 hoursof treatment, it was unlikely to have been produced duringderegulation of pathways related to cell death. Across all treat-ments, the presence of pericyte secretome as compared to nopericyte secretome (Fig. 3E) consistently and substantiallyincreased levels of: (i) pERK1/2 (2.9-fold change with vehicle,6.9-fold change with vemurafenib, 4.5-fold change with sorafe-nib, and 4.0-fold change with combined therapy); (ii) pAKT(1.7-fold change with vehicle, 11-fold change with vemurafenib,3.5-fold change with sorafenib and 8.6-fold change with com-bined therapy); (iii) pSMAD3 (0.9-fold change with vehicle,1.2-fold changewith vemurafenib, 1.4-fold changewith sorafenib

and 2-fold change with combined therapy); (iv) TSP-1 (1.6-foldchange with vehicle, 1.4-fold change with vemurafenib, 1.9-foldchange with sorafenib and 1.6-fold change with combinedtherapy); and (v) pVEGFR2 (1.02-fold change with vehicle,1.9-fold change with vemurafenib, 1.6-fold change withsorafenib and 1.1-fold change with combined therapy) inBRAFWT/V600E-KTC1 cells (Fig. 3E). As a result, pericytes providedsignificant growth advantages to BRAFV600E-KTC1 cells,even when treated with vehicle (2.1-fold change compared withBRAFWT/V600E-KTC1 tumor cells not cocultured withpericytes; Fig. 3F andG). These resultsmay be linked to the abilityof BRAFV600E-KTC1 cells when stimulated by pericyte secretome(5hours) to significantly increase secretionof TSP-1 (150.2%withvehicle, 112.5% with vemurafenib, 104.9% with sorafenib, and81.1% with combined therapy) and TGFb1 (20.7% with vehicle,83.9% with vemurafenib, 38.8% with sorafenib, and 47.2% withcombined therapy) compared with the BRAFWT/V600E-KTC1 cellswithout pericyte secretome (Fig. 3H). Pericyte secretome promot-ed a moderate rebound of pERK1/2 (but not when treated withvemurafenib), pAKT, and pSMAD3 in BRAFWT/WT-TPC1 cellsupon treatment with vehicle or drugs (Supplementary Fig. S5).

Furthermore, we used one of the most efficient short hairpinRNA (shRNA; ref. 22) to knockdown TSP-1 in pericytes (Fig. 3I),downregulating TSP-1 protein levels by more than 50%, andreducing secreted TSP-1 levels even more robustly upon drugtreatment (Supplementary Fig. S6B). Importantly, knockdown ofTSP-1 (by shTSP-1) in pericytes substantially reduced the capa-bility of the shTSP-1 pericyte secretome (compared with shGFPpericyte secretome, control) to trigger rebound of pERK1/2 (butnot with combined therapy), pAKT, and pSMAD3 levels inBRAFWT/V600E-KTC1 tumor cells (Fig. 3J). Also, secretion ofTSP-1 decreased by 12.8% in BRAFWT/V600E-KTC1 cells in thepresence of the combined therapy–treated shTSP-1 pericyte secre-tome compared with shGFP secretome alone (SupplementaryFig. S6C). ShTSP-1 pericyte secretome did not have substantialadditive effects in combination with drug treatments in suppres-sing pERK1/2 (except with sorafenib, 29%) and pAKT levels, andexertedmoderate effects indownregulating pSMAD3(22.5%withvemurafenib) in BRAFWT/WT-TPC1 cells (Supplementary Fig. S7),with no changes observed in the secreted levels of TSP-1 (only amoderate decrease with sorafenib, Supplementary Fig. S6D),suggesting that BRAFWT/WT-PTC cells may have a different TSP-1–regulatory pathway than BRAFWT/V600E-PTC cells.

Overall, these results demonstrate that pericyte-derived secre-tome (e.g., TSP-1, TGFb1) induces the rebound of prosurvivaland proangiogenic factors and overcomes the inhibitory effectsof targeted therapy in BRAFWT/V600E-PTC cells, and ultimatelycontributes to an increase in BRAFWT/V600E-thyroid tumorcell survival.

Pharmacologic antagonism of TSP-1 by SRI31277 impairsTGFb1-dependent signaling, reduces growth of BRAFWT/V600E-PTC cells, and overcomes resistance to targeted therapy

The inhibition of TGFb1 activation is a therapeutic strategyagainst cancer (24).Weused anantagonist (i.e., SRI31277; ref. 24)that blocks TSP1-mediated TGFb1 activation in the extracellularenvironment to determine the role of the TSP-1/TGFb1 pathwayin BRAFWT/V600E-KTC1 cells using pSMAD3 protein expression.We assessed dose–response (IC50) for the SRI31277 peptidein BRAFWT/V600E-KTC1 or BRAFWT/WT-TPC1 cells, and inpericytes treated with a matrix of different doses (Supplementary






Figure 3.

Model of coculture with PTC patient-derived cells harboring the heterozygous BRAFV600E mutation and pericytes reveals resistance to BRAFV600E andtyrosine kinase inhibitors via the TSP-1/TGFb1 axis. A, Western blot analysis of proteins expression levels in pericytes at 5-hour treatment with DMSO (vehicle),10 mmol/L vemurafenib (Vemu.), 2.5 mmol/L sorafenib (Sora.), and combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib. These resultswere validated by three independent replicate measurements. (Continued on the following page.)

Prete et al.





Fig. S8A–S8C). Our results showed that compared with vehicle,10 mmol/L SRI31277 provided a significant therapeutic effectagainst PTC cells and pericytes, reducing cell viability by 3.1-foldinBRAFWT/V600E-KTC1, 4.1-fold inBRAFWT/WT-TPC1, and2.6-foldin pericytes (Supplementary Fig. S8A–S8C). SRI31277 upregu-lated pSMAD3 protein levels in pericytes within 5 hours upontreatment with vehicle (37%) or vemurafenib (4.5%), and down-regulated its levels in the presence of sorafenib (4.7%) orcombined therapy (27%; Fig. 4A). Also, SRI31277 treatmentsubstantially reduced both secreted TSP-1 (11%) and TGFb1(25.5%) in pericytes treated with combined therapy (Fig. 4B).Importantly, in BRAFWT/V600E-KTC1 cells treated with pericytesecretome the pharmacologic antagonism of TSP-1 by SRI31277treatment substantially downregulated protein levels of pERK1/2(23.4% with vehicle, 17.9% with vemurafenib, 37.7% withsorafenib, and 8.2% with combined therapy) and pSMAD3(30.1% with vehicle, 27.4% with vemurafenib, 14.8% withsorafenib and 24.4% with combined therapy; Fig. 4C, right).Intriguingly, we found that SRI31277 upregulated both pERK1/2 (100% with vehicle, 101% with vemurafenib, 24% with sor-afenib, and 205%with combined therapy) and pSMAD3 (34.8%with vehicle, 21.1%with vemurafenib, 40.3%with sorafenib, and25%with combined therapy) in the absence of pericyte secretomecompared with vehicle in BRAFWT/V600E-KTC1 cells (Fig. 4C, left).SRI31277 less robustly affected pERK1/2 and pSMAD3 levels inBRAFWT/WT-TPC1 cells treated with pericyte secretome (Supple-mentary Fig. S9). Secreted levels of both TSP-1 (5% with vehicle,58.6% with vemurafenib, 26.7% with sorafenib, and 52.3% withcombined therapy), and TGFb1 (17.1% with vehicle, 45.6%with vemurafenib, 27.9% with sorafenib, and 32% with com-bined therapy conditions) increased in BRAFWT/V600E-KTC1 cells

culturedwithpericytes for 48hours comparedwithBRAFWT/V600E-KTC1 cells in single culture (Fig. 4D). Importantly, the pharma-cologic antagonism of TSP-1 by SRI31277 combined withBRAFV600E inhibition (vemurafenib) significantly reduced secre-tion of TSP-1 and TGFb1 in BRAFWT/V600E-KTC1 cells coculturedwith pericytes compared to the KTC1 cell coculture withoutSRI31277 treatment (Fig. 4D).More importantly, the antagonismof TSP-1 by SRI31277 significantly overcame therapeutic resis-tance of BRAFWT/V600E-KTC1 cell tumor growth inhibition toeither vemurafenib (67.1% reduction vs. no SRI31277), sorafenib(63.1% reduction vs. no SRI31277), and combined therapy(66.5% reduction vs. no SRI31277; Fig. 4E). Wemeasured within5 hours of treatment, the concentration of endogenous activeTGFb1 by ELISA in pericytes and KTC1 cells (SupplementaryMaterials and Methods). Active TGFb1 represented a small frac-tion of the total TGFb1 (reported in Fig. 3B and H). Specifically itwas 21.8 pg/mL and 14.5 pg/mLwith vehicle; 14.4 pg/mL and 6.6pg/mL with vemurafenib, 22.6 pg/mL and 10.6 pg/mL withsorafenib; and 13.5 pg/mL and 6.8 pg/mLwith combined therapyin BRAFWT/V600E-KTC1 cells and pericytes, respectively. Finally,to overcome SRI31277 antagonism upon TGFb1 activation byTSP-1, we treated BRAFWT/V600E-KTC1 cells in pericytes cocultureusing the recombinant human active TGFb1 (Fig. 4F). Important-ly, cell stimulation by recombinant human active TGFb1 proteinsignificantly rescued BRAFWT/V600E-KTC1 tumor cell growth inthe coculture with pericytes within 48 hours of treatmentwith SRI31277 plus sorafenib (1.34-fold, P ¼ 0.0075). Aftertreatment with SRI31277 plus combined therapy, TGFb1produced an even more robust rescue effect (8.7-fold, P ¼0.0044; Fig. 4F). Stimulation with recombinant human activeTGFb1 was also effective in significantly enabling BRAFWT/V600E-

(Continued.) Cells were grown in 0.2% FBS DMEM growth medium during treatment. B, Measurements of secreted TSP-1 and TGFb1 total protein levels inpericytes treated for 5 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/Lsorafenib, in the presence of 0.2% FBS DMEM growth medium. The secretome (0.2% FBS DMEM cell growth medium enriched by cell-derived secretedprotein factors) was collected and protein levels (ng/mL or pg/mL) were determined by ELISA. Secreted protein levels were normalized to cell growth medium(DMEM supplemented with 0.2% FBS) which was measured to determine subtracted background. These data represent the average � SD (error bars) of twoindependent replicate measurements (� , P < 0.05). C, Diagram of secreted factors (i.e., TSP-1, TGFb1) derived from human pericytes treated for 5 hours withvehicle, 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapywith 10mmol/L vemurafenib plus 2.5mmol/L sorafenib grown in the presence of 0.2% FBSDMEM growth medium as shown in A; secreted factors represent pericyte conditioned medium to treat PTC cells. D, Western blot analysis of pSMAD3proteins expression levels in human pericytes, BRAFWT/V600E-KTC1 and BRAFWT/WT-TPC1 cells treated for 5 hours with vehicle, recombinant human latent TGFb1protein (2 ng/mL), 10 mmol/L SRI31277, or recombinant human latent TGFb1 protein (2 ng/mL) plus 10 mmol/L SRI31277 in the presence of 0.2% FBS DMEMgrowthmedium. E,Western blot analysis of proteins expression levels in BRAFWT/V600E-KTC1 cells at 5 hrs direct treatment (without pericytes conditionedmedium;cellswere grown in0.2%FBSDMEMgrowthmediumduring treatment)withDMSO (vehicle), 10mmol/L vemurafenib, 2.5mmol/L sorafenib, or combined therapywith10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib; or 5 hours treatments with pericyte-derived conditioned medium (defined secretome) treated with DMSO(vehicle), 10 mmol/L vemurafenib, 2.5mmol/L sorafenib or combined therapywith 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib as shown inA. These resultswerevalidated by three independent replicates. F, Fluorescence imaging for fixed BRAFWT/V600E-KTC1mCherry cells alone (highlighted by mCherry and Hoechststaining, with red and white signal, respectively) or cocultured with pericytes [highlighted by mCherry (red signal) specific to label KTC1 cells and Hoechst (whitesignal) staining to label both KTC1 cells and pericytes]. KTC1 cells or cocultures were treated for 48 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/Lsorafenib, or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib. Cells were grown in 0.2% FBS DMEM growth medium during treatment.G, Quantification of only BRAFWT/V600E-KTC1mCherry cells reported in F. These data represent the average � SD (error bars) of three independent replicates(� ,P<0.05, �� ,P<0.01).H,Measurements of secreted TSP-1 and TGFb1 total protein levels in: BRAFWT/V600E-KTC1 cells directly treated (without pericyte conditionedmedium, blue bars) for 5 hourswith DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapywith 10 mmol/L vemurafenib plus 2.5 mmol/Lsorafenib in the presence of 0.2% FBS DMEM growth medium; and in BRAFWT/V600E-KTC1 cells treated for 5 hours with pericyte-derived conditioned medium(secretome, red bars) treated with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/Lsorafenib, in the presence of 0.2% FBS DMEM growth medium. The secretome (pericyte-derived conditioned medium) was collected and protein levels (ng/mLor pg/mL) were determined by ELISA. Secreted protein levels were normalized to cell growth medium (DMEM supplemented with 0.2% FBS) that was measuredto determine subtracted background. These data represent the average � SD (error bars) of two independent replicates (� , P < 0.05; �� , P < 0.01; ��, P < 0.001).I, Western blot analysis of protein expression levels in sh-GFP (control) or sh-TSP-1 pericytes at 5-hour treatment with DMSO (vehicle), 10 mmol/L vemurafenib,2.5 mmol/L sorafenib and combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib in the presence of 0.2% FBS DMEM growth medium. Theseresults were validated by two independent replicates. J, Western blot analysis of protein expression levels in BRAFWT/V600E-KTC1 cells at 5-hour direct treatment(without pericyte-derived conditioned medium) with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy with 10 mmol/Lvemurafenib plus 2.5 mmol/L sorafenib in the presence of 0.2% FBS DMEM growthmedium; or 5-hour treatments with sh-GFP (control) or sh-TSP-1 pericyte-derivedconditioned medium (secretome) treated with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy with 10 mmol/L vemurafenibplus 2.5 mmol/L sorafenib as shown in I. These results were validated by two independent replicates.






Figure 4.

The SRI31277 peptide overcomes resistance to vemurafenib plus sorafenib therapy in PTC-derived cells harboring the heterozygous BRAFV600E mutation.A, Western blot analysis of proteins expression levels in pericytes at 5-hour treatment with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafeniband combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib in the presence or absence of the SRI31277 peptide that blocks TSP-1/TGFb1activation (cells were grown in 0.2% FBS DMEM growth medium during treatment). (Continued on the following page.)

Prete et al.





KTC1 tumor cell growth in the absence of SRI31277 treatment(1.4-, 3.1-, and 2.6-fold change upon vehicle, vemurafenib andcombined therapy, respectively).

Taken together, these results indicate BRAFWT/V600E-KTC1 cellselicit paracrine signals in synergy with pericytes, which contributeto tumor survival (Fig. 4E and F) via the TSP-1/TGFb1 axis, andthat pericytes trigger resistance to the targeted therapy.

Effects of vemurafenib and sorafenib on the TSP-1/TGFb1 axisin an orthotopic mouse model of human BRAFWT/V600E-PTC

Our in vivo mouse results (Fig. 2B and C) did not confirm theunique and synergistic effect of combined therapy with vemur-afenib and sorafenib observed in our in vitromodels (Fig. 1E–G),indicating the likely elicitation of drug resistance by the tumormicroenvironment. Therefore, to assess the potentialmechanismsof the apparent resistance to this targeted therapy, we haveperformed IHC on the orthotopic thyroid tumors. We foundluciferase expression substantially decreased in orthotopicmousetumors (Fig. 2B and C) after 5 weeks of treatment witheither vemurafenib, sorafenib, or combined therapy (Fig. 5A).To corroborate these results, we analyzed the expression level ofPAX8, a thyroid-specific marker, to identify BRAFWT/V600E humanthyroid tumor cells (11), which matched the downregulation ofthe luciferase marker in drug-treated mice (Fig. 5A). We also usedterminal deoxynucleotidyl transferase dUTP nick end labeling(TUNEL) to quantify cell death. Importantly, we found thatvemurafenib moderately increased cell death in a subpopulationof tumor cells (Fig. 5A), whereas sorafenib treatment was lesseffective, as confirmed by in vitro results using the Annexin V/PIassay (Fig. 1G). Combined therapy with vemurafenib and sor-afenib was found to induce cell deathmore effectively than singleagents in a subpopulation of tumor cells (Fig. 5A), indicating that

either BRAFV600E inhibitor (i.e., vemurafenib) or TKI (i.e., sora-fenib) abrogates BRAFWT/V600E-PTC through complex mechan-isms of action, ultimately resulting in cytostatic effects likely dueto the presence of different cloneswith intrinsic primary resistance(34). BRAFWT/V600E stimulates PTC cell proliferation (Fig. 1H),and affects the expression of TSP-1 and TGFb1-dependentpSMAD3 and pERK1/2 (Figs. 3E and 4C), suggesting thatthis oncogene promotes thyroid tumor aggressiveness via theTSP-1/TGFb1 axis. TSP-1 can inhibit VEGF-stimulated VEGFR2phosphorylation in microvascular endothelial cells and blockangiogenesis (37). Because our in vitro data suggested a role forTSP-1 in the paracrine communication between BRAFWT/V600E-PTC cells and pericytes, we assessed the expression of the TSP-1/TGFb1 axis and proangiogenic factors in vivo in the angiogenicmicroenvironment of orthotopic BRAFWT/V600E-PTC treated withvemurafenib, sorafenib, combined therapy, or vehicle (Fig. 5B).Combined therapy reflected the action of both vemurafenib andsorafenib in downregulating cytosolic TSP-1 and TGFb1 in tumorcellsmore than 50% comparedwith vehicle or single agents. Also,combined therapy reduced protein expression levels of prosurvi-val factors such as pAKT and tumor growth-relatedmolecules suchas pERK1/2. Critically, no inhibition (specifically by vemurafeniband combined therapy) of the TSP-1/TGFb1 axis was observed inthe vascular compartment, including endothelial cells and peri-cytes (Fig. 5B), suggesting that stromal vascular cells in themicro-environment can elicit drug resistance and provide advantagesto PTC growth. With reduction of TSP-1 and TGFb1 in tumorcells, we also found a substantial decrease (2.5-fold change) inpSMAD3expression involved in the TGFb1pathway inorthotopicthyroid tumor cells after treatment with vemurafenib or com-bined therapy, with no associated changes in the vascular/endothelial compartment (Fig. 5B). In addition, vemurafenib

(Continued.) These results were validated by two measurements. B, Measurements of secreted TSP-1 and TGFb1 total protein levels in pericytes treated for48 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib(cells were grown in 0.2% FBS DMEM growth medium during treatment) in the presence or absence of the SRI31277 peptide (10 mmol/L). The secretome(0.2% FBS DMEM growthmedium enriched by cell-derived secreted protein factors) was collected and protein levels (ng/mL or pg/mL) were determined by ELISA.Secreted protein levels were normalized to cell growth medium (DMEM supplemented with 0.2% FBS) which was measured to determine subtractedbackground. These data represent the average � SD (error bars) of two independent replicate measurements (�, P < 0.05; �� , P < 0.01). C, Western blotanalysis of proteins expression levels in BRAFWT/V600E-KTC1 cells at 5-hour direct treatment with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib,or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib in the presence or absence of the SRI31277 peptide; or 5-hour treatments with pericyte-derived secretome treated with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/Lsorafenib as shown in A in the presence or absence of the SRI31277 peptide. Cells were grown in 0.2% FBS DMEM growth medium during treatment. Theseresults were validated by two independent replicate measurements.D,Measurements of secreted TSP-1 and TGFb1 total protein levels in both BRAFV600E-KTC1 cellsand coculture with BRAFWT/V600E-KTC1mCherry cells and pericytes treated for 48 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, orcombined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib in the presence or absence of the SRI31277 peptide (10 mmol/L); cells were grown in 0.2%FBS DMEM growth medium during treatment. The secretome (0.2% FBS DMEM cell growth medium enriched by cell-derived secreted protein factors) wascollected and protein levels (ng/mL or pg/mL) were determined by ELISA. Secreted protein levels were normalized to cell growth medium (DMEM supplementedwith 0.2% FBS) that was measured to determine subtracted background. These data represent the average � SD (error bars) of two independent replicatemeasurements (� , P < 0.05; �� , P < 0.01; �� , P < 0.001). E, Fluorescence imaging for fixed BRAFWT/V600E-KTC1mCherry cells (highlighted by mCherry and Hoechststaining, with red and white signal, respectively) treated for 48 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combinedtherapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib in the presence or absence of the SRI31277 peptide (10 mmol/L); cells were grown in 0.2%FBS DMEM growth medium during treatment. Quantification of only BRAFWT/V600E-KTC1mCherry cells number is reported in the histogram. Fluorescence imaging forfixed coculture with BRAFWT/V600E-KTC1mCherry cells (highlighted by mCherry and Hoechst staining, with red and white signal, respectively) and pericytes(highlighted by Hoechst staining, white signal) treated for 48 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy,with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenib in the presence or absence of the SRI31277 peptide (10 mmol/L); cells were grown in 0.2% FBS DMEMgrowthmediumduring treatment.Quantification of onlyBRAFWT/V600E-KTC1mCherry cells number is reported in thehistogram. Thesedata represent the average�SD(error bars) of three independent replicate measurements (� , P < 0.05; �� , P < 0.01). F, Fluorescence imaging for fixed coculture with BRAFWT/V600E-KTC1mCherry cells(highlighted by mCherry and Hoechst staining, with red and white signal, respectively) and pericytes (highlighted by Hoechst staining, white signal) treatedfor 48 hours with DMSO (vehicle), 10 mmol/L vemurafenib, 2.5 mmol/L sorafenib, or combined therapy with 10 mmol/L vemurafenib plus 2.5 mmol/L sorafenibin the presence or absence of the SRI31277 peptide (10 mmol/L); cells were grown in 0.2% FBS DMEM growth medium during treatment. Cell coculture wasstimulated every 8 hours for 48 hours with recombinant human active TGFb1 (7.75 ng/500 mL in 24-well dishes) and compared with cell coculture withoutrecombinant human active TGFb1 as a control. Quantification of only BRAFWT/V600E-KTC1mCherry cells numbers are reported in the histogram. These datarepresent the average � SD (error bars) of three replicate measurements (�� , P < 0.01).







Prete et al.




substantially reduced pERK1/2 and pAKT levels in tumor cells butnot in the vascular/endothelial compartment (Fig. 5B). In con-trast, sorafenib treatment alone was ineffective in suppressing theexpression of these intracellular signaling targets in vivo in bothtumor and vascular cells (Fig. 5B). Vemurafenib (but not sorafe-nib)was ineffective at downregulatingmarkers of vascular density(e.g., CD31; Fig. 5C), suggesting that vemurafenib specificallytargets BRAFV600E-thyroid tumor cells but not BRAFWT-stromalcells, and also that possibly subpopulations of stromal vascularcells elicit resistance to vemurafenib. Therefore, the tumor-asso-ciated vascular milieu may contribute to paracrine signaling andsustained BRAFV6000E-thyroid tumor cell survival; indeed, wefound no cell death effects from vemurafenib or combinedtherapy treatment in either CD31þ (vascular endothelial cells),NG2þ, or PDGFRBþ (markers of pericytes; ref. 17) cells in thevascular compartment, including endothelial cells and pericytes(Fig. 5C). Interestingly, sorafenib produced a substantial reduc-tion (2-fold change) of viability in the aSMAþ cell population(microvessels; Fig. 5C), and all drug treatments substantiallydecreased levels of proangiogenic factors VEGF and VEGFR2 inthe tumor cells; in addition, vemurafenib reduced PDGFRB levelsin the tumor cells. Only sorafenib and combined therapy (likelyreflecting sorafenib activity) downregulated PDGFB levels in cellsof the vascular/endothelial compartment (Fig. 5C). BecauseBRAFWT/V600E-PTC expresses high levels of adhesion molecules,which play important roles in the ECM of the tumor microenvi-ronment (10, 22) we also performed trichrome staining ofBRAFWT/V600E-orthotopic PTC that revealed robust collagen depo-sition (Fig. 5D). Vemurafenib was associated with decreasedcollagen deposition (Fig. 5D), while sorafenib proved an ineffec-tive mediator. When combined with vemurafenib, sorafeniblimited its suppressive efficacy as well.

Collectively, these results indicate that combined therapy sup-presses proangiogenic molecules and TSP-1/TGFb1 expression inthyroid tumor cells but not robustly in the vascular compartment,including pericytes and endothelial cells. This heterogeneous

therapeutic response may be linked to the presence of differentsubpopulations (lineage) of pericytes (NG2þ, PDGFRBþ, aSMAþ

cells) in the tumormicroenvironment thatmight trigger resistanceto targeted therapy (i.e., BRAFV600E and TK inhibitors; Fig. 3E). Allinformation about antibodies used in this study is reported inSupplementary Table S1.

The TSP-1/TGFb1 axis regulates pathways fundamental forECM angiogenic microenvironment and tumor growth inBRAFWT/V600E-PTC compared with BRAFWT/WT-PTC clinicalsamples

TSP-1 mediates the interaction of tumor cells with the ECM(38), and plays a key role in progression when the BRAFV600E

mutation is present (22). TSP-1 also profoundly influencestumor cell proliferation, adhesion, andmigration (22). We there-fore validated our TSP-1/TGFb1 in vitro and mouse findings byassessing the functional interactions of TSP1-dependent regula-tory gene networks in clinical samples of BRAFWT/V600E orBRAFWT/WTPTC. We used TCGA (The Cancer Genome Atlas sam-ples (39) and applied the Linearmodel for RNA-seqdata (Limma)for moderate T-statistics to identify genes fundamental to tumorgrowth and microenvironment functions that were differentiallyexpressed in these two tumor groups.We analyzed 23 TCGAgenesthat are known to regulate ECM or pericyte function, angiogen-esis, inflammation, immune response, cell viability and growth,cytoskeleton organization, adhesion/migration/invasion, andmetastasis. Nineteen of 23 genes (82.6%) were significantly dif-ferentially expressed in BRAFWT/V600E compared with BRAFWT/WT

PTC; 10 of 19 (52.7%) were upregulated, and 9 of 19 (47.3%)were downregulated (Supplementary Table S2). The set of upre-gulated genes included TSP-1 (THBS1; Fig. 6A), which signifi-cantly increased (1.56-fold change) in BRAFWT/V600E-PTC (Sup-plementary Table S2). Furthermore, FN1, COL1A1, ITGA3, TGFb,and THBS2, etc., which are crucial microenvironment-associatedECM components, were also upregulated with significant fold-change ranging from 1.24 to 11.3. We found similar results when

Figure 5.IHC analysis of angiogenesis factors and TSP-1/TGFb1 expression in the orthotopic human PTC harboring the heterozygous BRAFV600E mutation treated withtargeted therapy. A, BRAFWT/V600E-positive human KTC1 orthotopic tumor growth in vivo in the late intervention mouse model described in Fig. 2A–C.Control mice (H&E, luciferase, and PAX8 stain) showed tumor growth within a circumscribed area, with mixed spindled features and atypical cells.Magnifications, 400� or 1,000�. Scale bars are shown in the bottomor top panel for the 400� and 1,000� photomicrographs, respectively: 100 mm (400� images)or 50 mm (1,000� images). Arrowhead (shown in the left panel for the 400� photomicrographs) highlights orthotopic KTC1 tumor area, asterisk highlightsmouse normal thyroid follicles area. Right panel for the 1,000� photomicrographs: arrows mark tumor cells and arrowheads mark VE ¼ vascular/endothelialcompartment. Abbreviations: H&E, hematoxylin and eosin. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) IHC for detecting DNAfragmentation/dead cells. Luciferase, PAX8 and TUNEL stains were assessed semiquantitatively using the following scoring method: 0 (negative), 1 (<50% positivecells), and 2 (�50% positive cells). B, IHC protein expression of TSP-1, TGFb1, phospho(p)-SMAD3, pERK1/2, and pAKT in BRAFWT/V600E-KTC1 orthotopictumors mice at week 5 posttreatment with vehicle, vemurafenib (100 mg/kg, one time daily), sorafenib (30 mg/kg, one time daily), or combined therapy withvemurafenib (100 mg/kg, one time daily) plus sorafenib (30 mg/kg, one time daily). Arrowhead (shown in the left for the 400� photomicrographs) highlightsorthotopic KTC1 tumor area, asterisk highlights mouse normal thyroid follicles area. Right panel for the 1,000� photomicrographs: arrows mark tumor cellsand arrowheads mark VE ¼ vascular/endothelial compartment. IHC stains were assessed semiquantitatively using the following scoring method: 0 (negative),1 (<50% positive cells), and 2 (�50% positive cells). VE ¼ vascular/endothelial compartment. Magnifications: 400� or 1,000�. Scale bars are shown in thebottomor top panel for the 400� and 1,000�photomicrographs, respectively: 100mm(400� images) or 50mm(1,000� images).C, IHC protein expression of CD31,aSMA, PDGFRB, PDGFB, VEGFR2, VEGF, andNG2 in BRAFWT/V600E-KTC1 orthotopic tumorsmice atweek 5 posttreamentwith vehicle, vemurafenib (100mg/kg, onetime daily), sorafenib (30 mg/kg, one time daily), or combined therapy with vemurafenib (100 mg/kg, one time daily) plus sorafenib (30 mg/kg, one time daily).Arrowhead (shown in the left panel for the 400� photomicrographs) highlights orthotopic KTC1 tumor area, asterisk highlights mouse normal thyroid follicles area.Right panel for the aSMA and CD31 1,000� photomicrographs: arrowheads mark VE ¼ vascular/endothelial compartment (i.e., microvessels). IHC stainswere assessed semiquantitatively using the following scoring method: 0 (negative), 1 (<50% positive cells), and 2 (�50% positive cells). VE¼ vascular/endothelialcompartment. Magnifications: 400� or 1,000�. Scale bars are shown in the bottom or top panel for the 400� and 1,000� photomicrographs, respectively: 100 mm(400� images) or 50 mm (1,000� images). D, Trichrome staining highlights the intratumoral abundant amount of collagen deposition (blue staining) inBRAFWT/V600E-KTC1 orthotopic tumors mice at week 5 posttreatment with vehicle, vemurafenib (100 mg/kg, one time daily), sorafenib (30 mg/kg, one time daily),or combined therapy with vemurafenib (100 mg/kg, one time daily) plus sorafenib (30 mg/kg, one time daily). Magnification: 400�. Scale bar is shown onthe bottom photomicrograph: 100 mm (400� images).






Figure 6.

TSP-1 (THBS1) gene-regulatory networks and pericytes abundance in BRAFWT/V600E-PTC versus BRAFWT/WT-PTC clinical samples. A, In the heatmap,rows depict differentially expressed genes and columns depict BRAFWT/V600E-PTC versus BRAFWT/WT-PTC (PTC TCGA database) samples. The relativeexpression level of genes is shown using a color scale. (Continued on the following page.)

Prete et al.





comparing PTC harboring both BRAFWT/V600E and hTERT muta-tions (which has been reported in TCGA to cooccur in a very smallnumber of PTC samples; ref. 39) versus either BRAFWT/V600E-PTC(Supplementary Table 3; Supplementary Fig. S10) or BRAFWT/WT-PTC (Supplementary Tables 4; Supplementary Fig. S11), indicat-ing that BRAFV600E is important in the transcriptional regulationof TSP-1 and other genes with functions linked to the tumormicroenvironment, including the vascular compartment andassociated endothelial cells and pericytes. The significantly dif-ferentially expressed genes were used to build a TSP-1 generegulatory network enrichment using the Cytoscape Genemaniaalgorithm that also included gene coexpression results fromTCGA. Each node identified genes and each edge representedfunctional interactions between genes. We found significantinteractions between the upregulated genes in BRAFWT/V600E-PTCversus BRAFWT/WT-PTC samples. TSP-1 significantly interactedwith all upregulated genes, and more importantly was coex-pressed with the TGFb1 gene (Fig. 6B), suggesting the importanceof TSP-1 in the direct regulation of TGFb1 activation and path-ways. We next carried out a TSP-1 pathways enrichment analysis(Fig. 6C), which identified significant pathways crucial for TGFbsignaling, metastasis, inflammation, immune modulation,tumor microenvironment–associated ECM remodeling func-tions, tumor growth, and VEGF ligand–VEGF receptor interac-tions, etc. Importantly, many genes involved in these pathwaysare known to play roles in endothelial cell and pericyte functionsin the vascular compartment. We found similar results when wecompared PTC harboring both BRAFWT/V600E and hTERT muta-tions versus BRAFWT/WT-PTC (Supplementary Fig. S12). Further-more, we quantified pericyte abundance (Fig. 6D) using canon-ical markers such as aSMA, PDGFRB, NG2 (17), and CD90(THY1) and the single sample Gene Set Enrichment Analysis(ssGSEA) algorithm applied to PTC TCGA data. From a set of 538samples (59 NT and 479 PTC), we identified the 5% mostpericyte-enriched (n ¼ 27, all PTC) and 5% least pericyte-enriched (n ¼ 27, 21 NT and 6 PTC) according to the PTC TCGAdata (Fig. 6D). The remaining 90%of samples that included 38 of59 NT or 446 of 479 PTC (185 BRAFWT/V600E-PTC, 20 PTC withboth BRAFWT/V600E and hTERT mutations, and 241 BRAFWT/WT-PTC) ranked in themiddle (mediocre) range (defined as "averagesamples" with neither high nor low pericyte enrichment) ofpericyte abundance scores (Fig. 6D). Specifically, NT sampleswere significantly overrepresented (3.5 folds, P < 0.001) in thelow pericytes–enriched group (21 of 27, 77.7%) as comparedwith PTC samples (6/27, 22.2%). In contrast, 27 of 33 PTC

samples (81.8%) showed substantial enrichment (4.5-foldincrease) in pericytes while only 6 of 33 samples (18.1%) showedlower enrichment. Of the samples exhibiting a high abundanceof pericytes (n ¼ 27), 16 (59.2%) were BRAFWT/V600E-PTC(P < 0.001, compared with NT), 3 (11.1%) were PTC with bothBRAFWT/V600E and hTERT mutations, and 8 (29.6%) wereBRAFWT/WT-PTC, and none were NT. Of the samples with theleast enrichment (n ¼ 27), 5 (18.5%) were BRAFWT/V600E-PTCand 1 (3.7%) were BRAFWT/WT-PTC, whereas the vast majority(n ¼ 21) were NT (P < 0.001, compared with BRAFWT/V600E-PTCwith high abundance of pericytes). Therefore high pericyteenrichment aligned with BRAFWT/V600E-PTC, and was 2-fold or5.3-fold higher than levels observed in BRAFWT/WT-PTC or PTCwith both BRAFWT/V600E and hTERT samples, respectively (Fig.6D). Importantly, intermediate risk of recurrence as assessed bythe PTC TCGA clinical database was associated with the 33 PTCsamples as follows: 15/33 (45.4%) were BRAFWT/V600E-PTC, 4/33(12.1%) were BRAFWT/WT-PTC, and 1/33 (3%) were PTC withBRAFWT/V600E and hTERT mutations. Importantly, among theBRAFWT/V600E-PTC samples with intermediate risk of recurrence,13 of 15 (86.6%) exhibited a high abundance of pericytes,whereas 2 of 15 (13.3%) showed low pericyte enrichment. Threeof 4 (75%) of BRAFWT/WT-PTC samples with intermediate risk ofrecurrence showed high pericyte enrichment, and 1 of 4 (25%)showed low pericyte enrichment. No associations were foundin PTC samples between BRAF mutational status, the low- orhigh-risk category, and pericytes abundance score. Overall, ourdata analysis indicated that pericytes population increased inBRAFWT/V600EPTC than PTC with other genetic alterations, oreven more robust than in NT samples (Fig. 6D). Collectively, ourdata indicate different activity by targeted therapy with vemur-afenib or sorafenib on the regulation of angiogenesis and ECMmolecules expression in the BRAFWT/V600E-KTC1 orthotopictumor cells and vascular/endothelial compartment (Fig. 6E).BRAFWT/V600E-PTC cells evoke paracrine regulatory networks torecruit pericytes, which ultimately sustain tumor cell survival.Overall, our findings reveal a new model of resistance to vemur-afenib and sorafenib therapy in BRAFWT/V600E-PTC via the TSP-1/TGFb1 axis triggered by pericytes (Fig. 6F and G).

DiscussionAlthough thyroid cancer mortality rates are lower relative to

incidence rates, thyroid cancer mortality has neverthelessincreased significantly since the late 1980s (1). Patients diagnosed

(Continued.) Colors indicate standardized values (green represents gene downregulation and red represents gene upregulation). B, THBS1 (TSP-1)regulatory networks with genes differentially expressed in BRAFV600E-PTC versus BRAFWT-PTC human samples (PTC TCGA data base) built up by Cytoscape.Each node represents a gene and each edge represents the interaction between the genes. We observed that there is a significant interaction between thedifferentially expressed genes that are identified from BRAFWT/V600E-PTC versus BRAFWT/WT-PTC human samples analysis. The coexpression analysis forthese genes was performed on the basis of PTC TCGA data (Pinteraction < 0.05 from correlation test were considered significant). Image shows genes that depictsignificant interaction with TSP-1 or TGFb1. Circles indicate genes deregulated in BRAFWT/WT-PTC versus BRAFWT/V600E-PTC samples; the color intensitycorresponds to the extent of upregulation/downregulation based on the log2FC values. C, Pathways enrichment analysis for significantly differentiallyexpressed genes shown in A in BRAFWT/V600E-PTC versus BRAFWT/WT-PTC. The significance of effect/enrichment on pathways is shown along x axis as �log10 Pvalue. For pathways analysis:�log10 P; 1.3¼ P¼ 0.05;�log10 P, 2¼ P¼ 0.01;�log10 P; 3¼ P¼ 0.001;�log10 P; 4¼ P¼ 0.0001. D, Plot of the pericytes abundancecalculated by the pericytes abundance score in normal thyroid (NT) samples, BRAFWT/V600E-PTC, PTC harboring both BRAFV600E and hTERT mutations, andBRAFWT/WT-PTC samples (PTC TCGA database) arranged in the order of their scores from highest to lowest. Only the 5% most pericytes enriched samplesand the 5% least pericytes enriched samples are highlighted in color with the rest of the average samples shown in gray. The color code is green for NT samples,blue for BRAFWT/V600E-PTC; black for BRAFWT/WT-PTC, and orange for PTC with BRAFWT/V600E and hTERT mutations. E, Flow chart that summarizeseffects of vemurafenib, sorafenib, or combined therapy on the angiogenic molecules expression, collagen deposition, and TSP-1/TGFb1/SMAD3 expression intheBRAFWT/V600E-KTC1 orthotopic tumor cells and vessels compartment (see Fig. 5). Blue line, downregulation/inhibitory effects; red line, no inhibitory effects; ECM,extracellular matrix. F and G, Overview of treatment with vemurafenib, sorafenib, or combined therapy. Combined therapy determines cell death in thyroid tumorcells; role of pericytes which might trigger resistance to these agents via the TSP1/TGFb1 axis.






with thyroid cancer can be treated with radioactive iodine, but asubset of patients fails to respond to this treatment and suffer lowsurvival rates (40), due to the BRAFV600E mutation. To date, mostclinical trials for metastatic thyroid cancer have focused on singleagents with low response rates; more effective treatment optionsfor this disease are urgently needed. The prevalence and criticalrole of genetic mutations in cancer cells have led to targetedmolecular therapy. Current therapies target BRAFV600E, MEK,PI3K, TKs, etc. Because BRAFV600E is the most frequent oncogeneimplicated in PTC initiation (32) and aggressiveness (22), target-ing this mutation holds great promise for future therapies. Small-molecule kinase inhibitors allow for fewer side effects thantraditional chemotherapy; however, most cancers are heteroge-neous and have the capacity to develop resistance to targetedtherapies (41). Sorafenib was the first targeted therapy approvedfor patients with advanced differentiated thyroid carcinoma(DTC; ref. 42). It is an oral multikinase inhibitor that is used asfirst-line treatment for metastatic DTCs. When tumors grow, theygenerate new blood vessels by angiogenesis to supply adequatenutrients (43). Researchers have therefore focused on antiangio-genic therapy, including disruption of new blood vessels, tosuppress tumor growth. Sorafenib has been considered a possibleangiogenesis disruptor (44). Here we have shown, however, thatthyroid tumor cells harboring the BRAFV600E mutation elicitresistance to sorafenib and limit its therapeutic efficacy.

Other systemic therapies have been used against DTC in theclinic; we chose to investigate vemurafenib, the first FDA-approved selective oral inhibitor of BRAFV600E (45). Vemurafenibhas been used in patients with metastatic and radioiodine-refractory BRAFV600E-PTC and continuously administered twicea day in cycles of 28 days (6). This therapy showed antitumoractivity with partial response in 10 of 26 patients (38.5%, bestoverall response). Fourpatients (15%)died after amedian follow-up of 18.8 months. It is unclear whether the similarities of drugtargets in DTC could lead to complete cross-resistance andwheth-er sequential treatment would be efficacious. Previous studiesrevealed that treatment of BRAFV600E-positive thyroid cancer cellswith vemurafenib initially produced a therapeutic response, how-ever, mechanisms of resistance were quickly elicited which led tocell death refractoriness (8, 11, 34, 35). Also, studies onBRAFV600E

humanmelanoma cells showed resistance to completely suppressERK1/2 activation even with high doses of vemurafenib (46),similar to our results using the 10 mmol/L dosage. Our study is thefirst to show data on combined inhibition of both BRAFV600E andtyrosine kinase as a potential targeted therapeutic option. Com-bined therapywith vemurafenib and sorafenib showed synergisticeffects in BRAFWT/V600E-PTC cells, whereas it was subadditive inBRAFWT/WT-PTC cells and pericytes. As a result, this combinedinhibition was significantly more effective than vehicle or singleagents in inducing higher rates of death in BRAFWT/V600E-PTC thanBRAFWT/WT-PTC cells. However, our mouse data showed thatvemurafenib produced cytostatic effects in orthotopic tumors,whereas combined therapy (likely reflecting sorafenib activity)generated biological fluctuations with tumor inhibition alternat-ing with tumor growth. To understand this phenomenon, wesubsequently analyzed the microenvironments of orthotopicBRAFWT/V600E-PTC, and found these tumors were enriched withheterogeneous populations of pericytes, in particular PDGFRBþ,aSMAþ, and NG2þ. Many genetic alterations, including theBRAFV600E oncogene, confer a cell-autonomous fitness advantageto tumors by providing independence from growth factors or

suppressing cell death responses. In addition, tumor progressionis affected bymicroenvironment-associated factors that cannot beovercome by adopting targeted therapies against only tumor cellsper se. The combined vemurafenib plus sorafenib therapyshowed synergy effects against BRAFWT/V600E-PTC cells but notBRAFWT/WT-PTC cells or pericytes, suggesting biological cooper-ation between BRAFV600E and TK pathways. This therapeuticsynergy inhibited activation of ERK1/2, AKT, and TKs (e.g.,VEGFR2, PDGFRB) to suppress tumor survival and motility.

The depletion of pericytes in transgenic mice under control ofNG2 promoter inhibited breast tumor growth and defectivetumor vasculature, but in one study was also found to increasemetastasis (20). In the context of BRAFWT/V600E-PTC, pericytescould elicit resistance driven by higher intracellular levels of TSP-1. Intracellular TSP-1 was found to behave differently in pericytesthan in BRAFWT/V600E-PTC cells upon the administration of tar-geted therapy; that is, pericytes treated with the targeted therapyupregulated intracellular TSP-1, whereas BRAFWT/V600E-PTC cellsdownregulated it. This might be due to the fact that thrombos-pondins, including TSP-1, are ER-resident effectors of adaptive ERstress (47), and can function inside the cell during stress, tissuedamage, or active remodeling to augment ER function (47). Here,we found that pericytes secreted TSP-1 and TGFb1, and inducedthe rebound of pERK1/2, pAKT and pSMAD3 levels to overcomethe inhibitory effects of the targeted therapy in BRAFWT/V600E-PTCcells. This led to increased BRAFV600E-PTC cell survival and celldeath refractoriness, indicating that accumulated intracellularTSP-1might drive this effect. These intracellular signaling effectorswere suppressed by direct treatment with vemurafenib or sorafe-nib, and more strongly by combined therapy, suggesting thatmicroenvironment-associated pericytes represent one of the con-straints for these targeted therapies. It has been reported thatpericytes can communicate with tumor cells and endothelial cellsby paracrine communication to ultimately elicit tumor cell sur-vival (19). Combined therapy against BRAFWT/V600E-PTC cellsalone strongly induced cell death and inhibited proliferation,but when these tumor cells were cocultured with pericytes, theyshowed increased cell survival. More importantly, our in vivofindings showed that combined therapy effectively suppressedthe expression of both TSP-1/TGFb1 and TKs pathways in tumorcells, but not in the vascular compartment. We can argue, there-fore, that vascular microenvironment-derived TSP-1 is an impor-tant factor indrug resistance inPTC,which functions in a signalingnetwork with intracellular TSP-1, TGFb1/SMAD, and TKs (i.e.,VEGFR2, PDGFRB) for the regulation of tumor growth, as wasconfirmed in our PTC clinical sample data analysis. Importantly,specific sequences in the thrombospondin type1 repeats (TSR) arecritical for activating latent TGFb1 through disruption of the LSKLsequence in the latency-associated peptide region of latent TGFb1within the mature domain, which reconforms the latent complextomake it accessible to TGFb signaling receptors (48). Peptides ofthe LSKL sequence or related analogues such as SRI31277, antag-onize TGFb activation by preventing TSP-1 binding to the latentcomplex (24). Like TSP-1, TGFb1 plays an important role intumorigenesis and angiogenesis. TGFb signaling can be protu-morigenic or tumor suppressive (49, 50). TGFb acts throughmembrane receptor kinases to activate SMADtranscription factorsfor many biological functions (51). To assess the role of thispathway inPTC cells andpericytes, weused the SRI31277peptide,which antagonizes TSP-1 activity on TGFb1 activation. This pep-tide showed significant antitumor effects against multiple

Prete et al.





myeloma (24). In our study, we analyzed the specificity of theTSP1-driven drug resistancemechanism promoted by pericytes inPTC cells through both an SRI31277-based pharmacologicapproach and an shRNA strategy; together these showed thatantagonism or knockdown of TSP-1, respectively, downregulatedTGFb1 pathways in PTC cells when these cells were stimulatedwith pericyte secretome. Ultimately, this disruption of the TSP-1/TGFb1 axis recovered the ability of the targeted therapy to inhibitBRAFWT/V600E-PTC intracellular cell signaling, leading to signifi-cant decrease in tumor cell growth. Our data also indicated thatpericyte populations expressing PDGFRB and NG2 wereunchanged by any treatment in our orthotopic mouse model ofhuman BRAFWT/V600E-PTC, whereas pericytes expressing aSMAwere substantially reduced upon administration of sorafenibalone, despite this treatment yielding no significant tumor growthreduction, which suggests that the aSMAþ cell subpopulationmay be susceptible to induction of cell death by TK inhibitors (e.g., sorafenib). Recruitment of aSMAþ pericytes around teratocar-cinoma vessels indicated vessel differentiation and higher tumorcell proliferation (52). Our results therefore suggest that diversepericyte subpopulations play multiple roles in the PTC microen-vironment, as shown in other tumor models (17), and maydestabilize vessels and promote/sustain BRAFWT/V600E-PTC cell-to-ECM adhesion for survival, motility and ultimately vascularinvasion. Further studieswill beneeded, however, to explore thesepossibilities, characterize pericyte subpopulations, anddeterminethe level of activation of pericytes in thyroid carcinoma. Inglioblastoma, a contact-dependent interaction with tumor cellsswitched on the tumor-promoter character of pericytes, inducingtheir participation in tumor progression (53). Importantly,BRAFWT/V600E-PTC samples showed higher pericyte enrichmentthan BRAFWT/WT-PTC or NT samples, as well as a significantabundance of ECM molecules, which may be linked to theactivation of regulatory gene networks and pathways fundamen-tal for tumor growth, angio-invasion, and metastasis. Pericytesexpress and secrete ECM molecules that are important for cellsurvival, adhesion, and migration (18), and support processesimportant to the efficient alignment of endothelial cells requiredfor maturation of capillary structures and delivery of nutrients.Ultimately this processmight contribute toBRAFWT/V600E-PTC cellsurvival, extra-thyroidal extension, angio-invasion, and metasta-sis. Because humans and mice have different oral drug bioavail-ability, physiology, and metabolism, vemurafenib and sorafenibdoses used in mice are not comparable with those appropriate tohuman patients. Vemurafenib was substantially more effectivethan sorafenib at inhibiting tumor growth in vivo, and caused bothcytotoxic and cytostatic effects, as well as downregulation of TSP-1, TGFb1, SMAD3, ERK1/2 and AKT protein levels in tumor cells,but not in the vascular compartment, including endothelial cellsand pericytes. Overall, our results indicate that the TSP-1/TGFb1axis is expressed and functional in thePTCmicroenvironment andmight contribute to targeted therapy resistance. Our findings alsorevealed an abundance of collagen deposition, another key playerinmigration and angio-invasion, in the orthotopic tumors treatedwith either sorafenib or combined therapy, but not vemurafenib.Sorafenib primarily acted as an antiangiogenic agent, suppressingVEGFR2 expression in the orthotopic tumor cells, and VEGF andPDGFB protein expression levels in the vascular compartment/endothelial; however, it was ineffective in blocking the likelyexpression of ECMmolecules (e.g., collagen) in the thyroid tumormicroenvironment, and maintained high protein levels of

pERK1/2, pSMAD3 and pAKT. Ultimately, this effect substantiallyextended tumor survival compared with vemurafenib treatment,and negatively impacted the therapeutic efficacy of the combinedtreatment as well. Critically, it has been suggested that TSP-1 mayregulate the expression of collagen in response to stress and tissueremodeling, independent of TGFb1 activation (54). One possi-bility is that ER stress response may lead to a lack of TSP-1suppression and cause hyper-secretion of ECM molecules suchas collagen and fibronectin. Furthermore, we found these ECMmolecules transcriptionally upregulated in BRAFWT/V600E-PTCcompared to BRAFWT/WT-PTC clinical samples, indicating thatthe BRAFV600E oncogene deregulates ECM composition, and notonly provides growth advantages to PTC cells but also promotescell adhesion/migration, leading to intravascular invasion andmetastasis (e.g., to the neck lymph node compartments), which isfrequent in BRAFWT/V600E-PTC.

Collectively, our results indicate that the action of the TSP-1/TGFb1 axis enables drug resistance toBRAFV600E andTK inhibitorsin BRAFWT/V600E-PTC cells. Our study is limited by a lack ofhuman samples from patients with metastatic BRAFV600E-PTCtreated with BRAFV600E inhibitors and TKI; however, our resultsare substantiated by our integrated cell cultures, cell cocultures,and mouse model, as well as the use of TCGA clinical samples ofPTC. We cannot exclude the possibility that this resistance mech-anism is specific only to sorafenib. Further studies will be neededto assess BRAFV600E inhibitors in combination with other TKinhibitors such as lenvatinib. Also, as TGFb1 is known to evokeactive immune suppression, future research could assess theeffects of pharmacologic antagonism of the TSP-1/TGFb1 axis onthe immune microenvironment of thyroid carcinomas.

In summary, our findings indicate that pericytes may be adouble-edged sword in BRAFWT/V600E-PTC therapy, and secretesoluble factors, such as TSP-1 and TGFb1, that trigger resistance toBRAFV600E inhibitors and TKI in BRAFWT/V600E-PTC. Pharmaco-logic antagonism of TSP-1 to block TGFb1 activation may becrucial for the inhibition of tumorigenic effects by TGFb1pathwayin BRAFWT/V600E-PTC cells. This might represent a novel thera-peutic strategy with translational applications in clinical trialsagainst BRAFWT/V600E-PTC resistant to targeted therapies. Finally,TSP-1 is a potential biomarker for assessing therapeuticresponse to BRAFV600E and TK inhibitors in patients with invasiveBRAFWT/V600E-PTC.

Disclosure of Potential Conflicts of InterestJ.E. Murphy-Ullrich has ownership interest (including stock, patents, etc.) in

EMDMilipore license for anti-TSP antibody (not used in paper) and aUS patenton SRI31277. C. Nucera received expert testimony for NIH as an ad hoc peer-reviewer. No potential conflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: J. Lawler, C. NuceraDevelopment of methodology: A. Prete, A.S. Lo, Z.A. Antonello,D.M. Vodopivec, J. Clohessy, J. Lawler, C. Nucera, A.M. DvorakAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A. Prete, A.S. Lo, P.M. Sadow, Z.A. Antonello,D.M. Vodopivec, S. Ullas, J.N. Sims, J. Clohessy, T. Sciuto, J. Lawler, C. NuceraAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): A. Prete, A.S. Lo, P.M. Sadow, S.S. Bhasin,Z.A. Antonello, D.M. Vodopivec, S. Ullas, J.N. Sims, T. Sciuto, M. Bhasin,J.E. Murphy-Ullrich, S.A. Karumanchi, C. NuceraWriting, review, and/or revision of the manuscript: A. Prete, A.S. Lo,P.M. Sadow, S.S. Bhasin, D.M. Vodopivec, S. Ullas, J.N. Sims, M. Bhasin,J.E. Murphy-Ullrich, J. Lawler, S.A. Karumanchi, C. Nucera, A.M. Dvorak






Administrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D.M. Vodopivec, C. Nucera, A.M. DvorakStudy supervision: J. Clohessy, C. NuceraOther (figure preparation): A. Prete, C. NuceraOther (provided SRI312777): J.E. Murphy-Ullrich

AcknowledgmentsC. Nucera (Principal Investigator, Human Thyroid Cancers Preclinical and

Translational Research at the Beth Israel Deaconess Medical Center (BIDMC)/Harvard Medical School) was awarded grants by the National Cancer Institute/NIH (1R21CA165039-01A1 and 1R01CA181183-01A1), the American ThyroidAssociation (ATA), and ThyCa:Thyroid Cancer Survivors Association Inc. forThyroid Cancer Research. C. Nucera was also a recipient of the Guido Berlucchi"Young Investigator" research award 2013 (Brescia, Italy) and BIDMC/CAOGrants. SRI31277 production and J.E. Murphy-Ullrich's contributions weresupported by NIH grant 1R01CA175012. We are grateful to Prof. Harold F.

Dvorak and Miss Elizabeth McGonagle (BIDMC/HMS) for critical reading ofour manuscript. We are grateful to Dr. Andrew L. Kung (Memorial SloanKettering Cancer Center, NYC, USA) for kindly providing the FUW-Luc-mCherry-puro (luciferase/cherry) plasmid, and to Prof. Judy Lieberman andDr.MinhT.N. Le (Children'sHospital,HarvardMedical School, Boston,MA) forkindly providing packaging plasmids (psPAX2, pMD2-G, and RTR2). We alsothank Miss Nicole Pandell for the technical support for the oral gavages in themice.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received February 28, 2018; revised June 27, 2018; accepted July 30, 2018;published first August 3, 2018.

References1. LimH, Devesa SS, Sosa JA, Check D, Kitahara CM. Trends in thyroid cancer

incidence and mortality in the United States, 1974–2013. JAMA 2017;317:1338–48.

2. Xing M, Alzahrani AS, Carson KA, Viola D, Elisei R, Bendlova B, et al.Association between BRAF V600Emutation and mortality in patients withpapillary thyroid cancer. JAMA 2013;309:1493–501.

3. Xing M, Alzahrani AS, Carson KA, Shong YK, Kim TY, Viola D, et al.Association between BRAF V600E mutation and recurrence of papillarythyroid cancer. J Clin Oncol 2015;33:42–50.

4. Liu R, Bishop J, Zhu G, Zhang T, Ladenson PW, Xing M. Mortality riskstratification by combining BRAF V600E and TERT promoter mutations inpapillary thyroid cancer: genetic duet of BRAF and TERT promoter muta-tions in thyroid cancer mortality. JAMA Oncol 2016 Sep 1[Epub ahead ofprint].

5. Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J, et al.Improved survival with vemurafenib in melanoma with BRAF V600Emutation. N Engl J Med 2011;364:2507–16.

6. Brose MS, Cabanillas ME, Cohen EE, Wirth LJ, Riehl T, Yue H, et al.Vemurafenib in patients with BRAF(V600E)-positive metastatic or unre-sectable papillary thyroid cancer refractory to radioactive iodine: a non-randomised, multicentre, open-label, phase 2 trial. Lancet Oncol 2016;17:1272–82.

7. Nucera C. Evolution of resistance to thyroid cancer therapy. Aging2016;8:1576–7.

8. Montero-Conde C, Ruiz-Llorente S, Dominguez JM, Knauf JA, Viale A,Sherman EJ, et al. Relief of feedback inhibition of HER3 transcription byRAF andMEK inhibitors attenuates their antitumor effects in BRAF-mutantthyroid carcinomas. Cancer Discov 2013;3:520–33.

9. Nazarian R, Shi H, Wang Q, Kong X, Koya RC, Lee H, et al. Melanomasacquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS upregula-tion. Nature 2010;468:973–7.

10. Nucera C, Lawler J, Parangi S. BRAF(V600E) and microenvironment inthyroid cancer: a functional link to drive cancer progression. Cancer Res2011;71:2417–22.

11. Duquette M, Sadow PM, Husain A, Sims JN, Antonello ZA, Fischer AH,et al. Metastasis-associated MCL1 and P16 copy number alterationsdictate resistance to vemurafenib in a BRAFV600E patient-derivedpapillary thyroid carcinoma preclinical model. Oncotarget 2015;6:42445–67.

12. BroseMS,NuttingCM, Jarzab B, Elisei R, Siena S, Bastholt L, et al. Sorafenibin radioactive iodine-refractory, locally advanced or metastatic differenti-ated thyroid cancer: a randomised, double-blind, phase 3 trial. Lancet2014;384:319–28.

13. Pitoia F, Jerkovich F. Selective use of sorafenib in the treatment of thyroidcancer. Drug Des Devel Ther 2016;10:1119–31.

14. Kloos RT, RingelMD, KnoppMV,Hall NC, KingM, Stevens R, et al. Phase IItrial of sorafenib in metastatic thyroid cancer. J Clin Oncol 2009;27:1675–84.

15. Sherman EJ, Dunn LA,HoAL, Baxi SS, Ghossein RA, FuryMG, et al. Phase 2study evaluating the combination of sorafenib and temsirolimus in the

treatment of radioactive iodine-refractory thyroid cancer. Cancer2017;123:4114–21.

16. BunoneG, Vigneri P,Mariani L, But�o S, Collini P, Pilotti S, et al. Expressionof angiogenesis stimulators and inhibitors in human thyroid tumors andcorrelation with clinical pathological features. Am J Pathol 1999;155:1967–76.

17. Song S, Ewald AJ, StallcupW,Werb Z, Bergers G. PDGFRbetaþ perivascularprogenitor cells in tumours regulate pericyte differentiation and vascularsurvival. Nat Cell Biol 2005;7:870–9.

18. Sadow PM, Priolo C, Nanni S, Karreth FA, Duquette M, Martinelli R, et al.Role of BRAFV600E in the first preclinical model of multifocal infiltratingmyopericytoma development and microenvironment. J Natl Cancer Inst2014;106:pii:dju182.

19. Franco M, Roswall P, Cortez E, Hanahan D, Pietras K. Pericytes promoteendothelial cell survival through induction of autocrine VEGF-A signalingand Bcl-w expression. Blood 2011;118:2906–17.

20. CookeVG, LeBleuVS,KeskinD,KhanZ,O'Connell JT, TengY, et al. Pericytedepletion results in hypoxia-associated epithelial-to-mesenchymal transi-tion and metastasis mediated by met signaling pathway. Cancer Cell2012;21:66–81.

21. Yee KO, StreitM,Hawighorst T, DetmarM, Lawler J. Expression of the type-1 repeats of thrombospondin-1 inhibits tumor growth through activationof transforming growth factor-beta. Am J Pathol 2004;165:541–52.

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

23. Crawford SE, Stellmach V, Murphy-Ullrich JE, Ribeiro SM, Lawler J, HynesRO, et al. Thrombospondin-1 is amajor activator of TGF-beta1 in vivo. Cell1998;93:1159–70.

24. Lu A, Pallero MA, Lei W, Hong H, Yang Y, Suto MJ, et al. Inhibition oftransforming growth factor-beta activation diminishes tumor progressionand osteolytic bone disease in mouse models of multiple myeloma. Am JPathol 2016;186:678–90.

25. Knauf JA, Sartor MA, Medvedovic M, Lundsmith E, Ryder M, Salzano M,et al. Progression of BRAF-induced thyroid cancer is associated withepithelial-mesenchymal transition requiring concomitant MAP kinase andTGFbeta signaling. Oncogene 2011;30:3153–62.

26. Kurebayashi J, Tanaka K, Otsuki T, Moriya T, Kunisue H, Uno M, et al. All-trans-retinoic acid modulates expression levels of thyroglobulin and cyto-kines in a new human poorly differentiated papillary thyroid carcinomacell line, KTC-1. J Clin Endocrinol Metab 2000;85:2889–96.

27. Tallarida RJ. Quantitative methods for assessing drug synergism. GenesCancer 2011;2:1003–8.

28. Fendrich V, Maschuw K, Rehm J, Buchholz M, Holler JP, Slater EP, et al.Sorafenib inhibits tumor growth and improves survival in a transgenicmouse model of pancreatic islet cell tumors. ScientificWorldJournal 2012;2012:529151.

29. Azimi A, Tuominen R, Costa Svedman F, Caramuta S, Pernemalm M,Frostvik Stolt M, et al. Silencing FLI or targeting CD13/ANPEP lead todephosphorylation of EPHA2, amediator of BRAF inhibitor resistance, and

Prete et al.





induce growth arrest or apoptosis in melanoma cells. Cell Death Dis2017;8:e3029.

30. Nucera C, Nehs MA, Nagarkatti SS, Sadow PM, Mekel M, Fischer AH, et al.Targeting BRAFV600E with PLX4720 displays potent antimigratory andanti-invasive activity in preclinical models of human thyroid cancer.Oncologist 2011;16:296–309.

31. Hong DS, Morris VK, El Osta B, Sorokin AV, Janku F, Fu S, et al. Phase IBstudy of vemurafenib in combination with irinotecan and cetuximab inpatients with metastatic colorectal cancer with BRAFV600E mutation.Cancer Discov 2016;6:1352–65.

32. Chakravarty D, Santos E, RyderM, Knauf JA, Liao XH,West BL, et al. Small-molecule MAPK inhibitors restore radioiodine incorporation in mousethyroid cancers with conditional BRAF activation. J Clin Invest 2011;121:4700–11.

33. Jain RK, Booth MF. What brings pericytes to tumor vessels? J Clin Invest2003;112:1134–6.

34. Antonello ZA, Hsu N, Bhasin M, Roti G, Joshi M, Van Hummelen P, et al.Vemurafenib resistance via de novoRBMgenesmutations and chromosome5 aberrations is overcome by combined therapywith palbociclib in thyroidcarcinoma with BRAF(V600E). Oncotarget 2017;8:84743–60.

35. Danysh BP, Rieger EY, Sinha DK, Evers CV, Cote GJ, Cabanillas ME, et al.Long-term vemurafenib treatment drives inhibitor resistance through aspontaneous KRAS G12D mutation in a BRAF V600E papillary thyroidcarcinoma model. Oncotarget 2016;7:30907–23.

36. Poulikakos PI, Zhang C, Bollag G, Shokat KM, Rosen N. RAF inhibitorstransactivate RAF dimers and ERK signalling in cells with wild-type BRAF.Nature 2010;464:427–30.

37. Zhang X,Kazerounian S,DuquetteM, Perruzzi C,Nagy JA,DvorakHF, et al.Thrombospondin-1 modulates vascular endothelial growth factor activityat the receptor level. FASEB J 2009;23:3368–76.

38. Kazerounian S, YeeKO, Lawler J. Thrombospondins in cancer. CellMol LifeSci 2008;65:700–12.

39. Cancer Genome Atlas Research Network. Integrated genomic characteri-zation of papillary thyroid carcinoma. Cell 2014;159:676–90.

40. Cabanillas ME, Dadu R, Hu MI, Lu C, Gunn GB, Grubbs EG, et al.Thyroid gland malignancies. Hematol Oncol Clin North Am 2015;29:1123–43.

41. Shibue T, Weinberg RA. EMT, CSCs, and drug resistance: the mechanisticlink and clinical implications. Nat Rev Clin Oncol 2017;14:611–29.

42. Dadu R, Devine C, Hernandez M, Waguespack SG, Busaidy NL, Hu MI,et al. Role of salvage targeted therapy in differentiated thyroid cancerpatients who failed first-line sorafenib. J Clin Endocrinol Metab 2014;99:2086–94.

43. Dvorak HF. Tumor stroma, tumor blood vessels, and antiangiogenesistherapy. Cancer J 2015;21:237–43.

44. Zhu AX, Duda DG, Sahani DV, Jain RK. HCC and angiogenesis: possibletargets and future directions. Nat Rev Clin Oncol 2011;8:292–301.

45. Bollag G, Tsai J, Zhang J, Zhang C, Ibrahim P, Nolop K, et al. Vemurafenib:the first drug approved for BRAF-mutant cancer. Nat Rev Drug Discov2012;11:873–86.

46. Su F, Bradley WD, Wang Q, Yang H, Xu L, Higgins B, et al. Resistance toselective BRAF inhibition can be mediated by modest upstream pathwayactivation. Cancer Res 2012;72:969–78.

47. Lynch JM,MailletM, VanhoutteD, SchloemerA, SargentMA, BlairNS, et al.A thrombospondin-dependent pathway for a protective ER stress response.Cell 2012;149:1257–68.

48. Young GD, Murphy-Ullrich JE. Molecular interactions that confer latencyto transforming growth factor-beta. J Biol Chem 2004;279:38032–9.

49. Calon A, Espinet E, Palomo-Ponce S, Tauriello DV, Iglesias M, C�espedesMV, et al. Dependency of colorectal cancer on a TGF-beta-driven programin stromal cells for metastasis initiation. Cancer Cell 2012;22:571–84.

50. David CJ, Huang YH, Chen M, Su J, Zou Y, Bardeesy N, et al. TGF-betatumor suppression through a lethal EMT. Cell 2016;164:1015–30.

51. Massague J. TGFbeta signalling in context. Nat Rev Mol Cell Biol 2012;13:616–30.

52. Taverna D, Hynes RO. Reduced blood vessel formation and tumor growthin alpha5-integrin-negative teratocarcinomas and embryoid bodies.Cancer Res 2001;61:5255–61.

53. Caspani EM, Crossley PH, Redondo-Garcia C, Martinez S. Glioblastoma: apathogenic crosstalk between tumor cells and pericytes. PLoS One 2014;9:e101402.

54. SweetwyneMT, PalleroMA, Lu A, VanDuynGraham L,Murphy-Ullrich JE.The calreticulin-binding sequence of thrombospondin 1 regulates collagenexpression and organization during tissue remodeling. Am J Pathol 2010;177:1710–24.

55. Nagy JA, Feng D, Vasile E, Wong WH, Shih SC, Dvorak AM, et al. Perme-ability properties of tumor surrogate blood vessels induced by VEGF-A. LabInvest 2006;86:767–80.






2018;24:6078-6097. Published OnlineFirst August 3, 2018.Clin Cancer Res Alessandro Prete, Agnes S. Lo, Peter M. Sadow, et al.

1 Axisβin Thyroid Carcinoma via the TSP-1/TGFPericytes Elicit Resistance to Vemurafenib and Sorafenib Therapy

Updated version

10.1158/1078-0432.CCR-18-0693doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2018/08/03/1078-0432.CCR-18-0693.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/24/23/6078.full#ref-list-1

This article cites 54 articles, 11 of which you can access for free 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://clincancerres.aacrjournals.org/content/24/23/6078To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-18-0693

http://clincancerres.aacrjournals.org/content/suppl/2018/08/03/1078-0432.CCR-18-0693.DC1

http://clincancerres.aacrjournals.org/content/24/23/6078.full#ref-list-1

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

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/24/23/6078


Documents

Pericytes Elicit Resistance to Vemurafenib and Sorafenib ... · Pericytes Elicit Resistance to Vemurafenib and Sorafenib Therapy in Thyroid Carcinoma via the TSP-1/TGFb1 Axis Alessandro