novelbispeciﬁcdomainantibodytolrp6inhibits wntandr ... · revealed two independent wnt-binding...

Signal Transduction

NovelBispecificDomainAntibodytoLRP6 InhibitsWnt andR-spondinLigand-InducedWntSignalingand Tumor GrowthHeather Jackson1, David Granger2, Gavin Jones2, Louisa Anderson2, Sarah Friel2,Daniel Rycroft2,William Fieles1, James Tunstead1, Michael Steward2, Trevor Wattam2,Adam Walker2, Jeremy Griggs2, Muhammad Al-Hajj1, and Christopher Shelton1

Abstract

Aberrant WNT signaling is associated with the formationand growth of numerous human cancer types. The low-densitylipoprotein receptor-related protein 6 (LRP6) is the leastredundant component of the WNT receptor complex with twoindependent WNT ligand-binding sites. Using domain anti-body (dAb) technology, a bispecific antibody (GSK3178022)to LRP6 was identified that is capable of blocking stimulationin the presence of a range of WNT and R-spondin (RSPO)ligands in vitro. GSK3178022 was also efficacious in reducing

WNT target gene expression in vivo, in both cancer cell lineand patient-derived xenograft models, and delays tumorgrowth in a patient-derived RSPO fusion model of colorectalcancer.

Implications: This article demonstrates the inhibition of a keyoncogenic receptor, intractable tomAb inhibition due tomultipleindependent ligand interaction sites, using an innovative dAbapproach. Mol Cancer Res; 14(9); 859–68. �2016 AACR.

IntroductionThe WNT pathway is a high-priority target for development

of anticancer therapies. Involvement of the WNT pathway intumorigenesis was discovered more than two decades ago, withidentification of WNT1 as an oncogenic driver in murine mam-mary carcinogenesis and identification of Adenomatous PolyposisColi protein (APC) as a tumor suppressor in colorectal cancer(1, 2). More recently, advances in sequencing methods haverevealed that up to one third of solid tumors contain loss-of-function mutations in APC or gain-of-function mutations inb-catenin (CTNNB1; ref. 3). Other genetic lesions have beenidentified, including R-spondin (RSPO) fusions in colorectalcancer and RNF43 mutations in pancreatic cancer (4, 5). Manytumors show evidence of epigenetic activation of the Wnt path-way through altered expression of negative regulators, such asWIF1 and SFRP1.

WNT ligands are a family of 19 palmitoylated secreted glyco-proteins that interact with receptor complexes containing mem-bers of two families of cell-surface receptors: 10 related seven-passtransmembrane (7TM) proteins, known as Frizzleds (FZD), andtwo type I single pass proteins, LRP5 and LRP6 (6, 7). Engagementof WNT ligands with the receptor complex leads to increased

cytoplasmic levels of b-catenin through inhibition of a "destruc-tion complex" composed of AXIN2, APC, andGSK3. Cytoplasmicb-catenin translocates to the nucleus and interacts with membersof the TCF family of transcription factors to activate expression oftarget genes that mediate proliferation.

RSPO proteins represent a recently identified second class ofligands capable of stimulating the Wnt pathway (8). RSPOproteins bind to a group of three related 7TMs, LGR4, LGR5, andLGR6, to modulate the activity of two related membrane-boundE3 ligases, RNF43, and ZNRF3. RNF43 and ZNRF3 appear tocontrol the turnover of FZD receptors, and the expression ofRNF43 and ZNRF3 is upregulated by WNT stimulation (9).Therefore, RSPO proteins increase WNT signaling by inhibitingthe action of a negative feedback loop in theWnt pathway. A totalof 10% of colorectal cancers contain fusions in the RSPO genesthat lead to overexpression and enhanced stimulation of the Wntpathway (4). These fusions are exclusive to b-catenin and APCmutations, providing strong evidence that RSPO alteration candrive tumorigenesis through the Wnt pathway.

OncogenicWnt pathway activation can occur throughmultiplereceptor-dependent and -independent mechanisms. Receptor-dependent mechanisms include increased expression of WNT/RSPO ligands, as well as epigenetic repression of other extracel-lular antagonists, such as DKKs, WIF1, and SFRPs. For example,downregulation of WIF1 or SFRP1 is found in a majority of non–small cell lung cancer tumors (10, 11). Receptor-independentmechanisms involve mutations downstream of the receptor com-plex, typically in components of the destruction complex. Incolorectal cancer, with the exception of RSPO fusions, activationof the pathway downstream of the receptor complex is common,typically through mutations in APC, b-catenin, or AXIN2.

LRP5 and LRP6 constitute the least redundant component ofthe receptor complex and are potential targets for blocking aber-rant receptor-dependent activation of the pathway using inhib-itory mAb approaches. Biophysical characterization of LRP6 has

1Oncology R&D, GlaxoSmithKline, Collegeville, Pennsylvania. 2Bio-pharmaceuticals R&D, GlaxoSmithKline, Medicines Research Centre,Cambridge, United Kingdom.

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

Corresponding Author: Christopher Shelton, GlaxoSmithKline, 1250 S. College-ville Rd, Collegeville, PA 19426. Phone: 610-917-5579; Fax: 610-917-4108; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-16-0088

�2016 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 859

on December 5, 2020. © 2016 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst July 11, 2016; DOI: 10.1158/1541-7786.MCR-16-0088

http://mcr.aacrjournals.org/

revealed two independent WNT-binding sites, and monospecificanti-LRP6 mAbs are unable to block both sites simultaneously,indicating that an alternative approach, such as bispecific, isrequired (12, 13). Using domain antibody (dAb) technology(14, 15), we identified a bispecific inhibitor of LRP6, termedGSK3178022. GSK3178022 is effective at inhibiting both WNTandRSPO stimulation in vitro, abrogatesWnt signaling in vivo, anddelays tumor growth in a patient-derived colorectal cancer tumormodel with an RSPO fusion.

Materials and MethodsCell lines

All cell lines were purchased from the ATCCwith the exceptionof EKVX andHEK293/6E, whichwere obtained from theNationalCancer Institute and Biotechnology Research Institute, NationalResearch Centre, respectively. PA-1 and HT-1080 cell lines weregrown in Eagle's Minimal Medium supplemented with 10% FBS,NTERA-2 cell line in DMEMwith 10% FBS, and HEK293/6E withFreeStyle 293 Medium (Gibco). All other lines were grown inRPMI1640with 10%FBS. All lines with the exception of EKVX arebanked internally, and authenticity has been confirmed using thePromega Cell ID System.

Reporter assaysHT-1080 or HEK293/SE cells were seeded in tissue culture–

treated microassay plates in Eagle's Minimum Medium supple-mented with 10% FBS and incubated overnight at 37�C, 5%CO2.Cellswere cotransfectedwith 75ngperwell TOPFlash or FopFlashplasmid (Millipore) and the appropriate WNT ligand expressionplasmid (generated in-house) using Lipofectamine 2000 (LifeTechnologies) in OptiMEMmedia (Life Technologies) with 0.5%FBS. Two hours after transfection, cells were treated with anti-LRP6 agents and RSPO ligands (R&D Systems) where appropriateand further incubated for 24 hours. Luciferase activity was mea-sured using Bright-Glo Luciferase Assay System (Promega).

Western blotsHT-1080 cells were seeded in tissue culture–treated dishes in

Eagle's Minimal Medium with 10% FBS and incubated overnightat 37�C, 5% CO2. Cells were treated with anti-LRP6 agent for1 hour and recombinant WNT3a ligand added for an additional1-hour incubation. Whole-cell lysates were quantitated for totalprotein content using BCA Protein Assay Kit (Thermo ScientificPierce). Immunoblots were performed with the followingantibodies: anti-phospho-LRP6, anti-LRP6, anti-b-catenin (CellSignaling Technology), anti-active-b-Catenin (Millipore), andanti-b-actin (Sigma-Aldrich).

Real-time PCR assaysCell lines were seeded in tissue culture–treated dishes in appro-

priate growth media with 10% FBS and incubated overnight at37�C, 5% CO2. Treatment mixes were made in Opti-Mem with0.5% FBS to contain 200 ng/mL of recombinant WNT3a ligandwith or without anti-LRP6 constructs. Frizzled8 CRD (R&D Sys-tems) was used at 10 mg/mL. Growth media were replaced bytreatment mixes, and plates were incubated for 24 hours at 37�C,5% CO2. cDNA was produced and RT-PCR reactions performedusing TaqMan Gene Expression Cells-to-CT Kit (Ambion) andvalidated TaqManGene Expression Assays (Life Technologies) forAXIN2 and SP5. Data were analyzed using the DDCt method bynormalizing data to b-actin (ACTB) Ct values.

dAb selection and affinity maturationdAbs were selected from a library using LRP6 ECD (R&D

Systems) as described previously (14, 16). Fusion proteins com-prising selected dAbs and the Fc-portion of human IgG1 wereexpressed in HEK293/6E cells, which were maintained in Free-Style 293 medium (Gibco) at 37�C, 5% CO2, and 125 rpm. dAbswere connected at the N-terminal and C-terminal, using a threeamino acid and eight amino acid linker, respectively.

Antibodies were affinity matured by scanning complemen-tarity determining region (CDR) positions with all codonsusing site-directed mutagenesis. Plasmid DNA encoding theantibody was used as a template in PCRs with PfuUltra (Agi-lent). Primers introducing NNB codons were used in reactionsand the resultant mutated antibody sequences tested for bind-ing and function in reporter assays.

Surface plasmon resonanceBinding analysis was performed on a BIACORE 4000 instru-

ment (Biacore) at 25�C in HBS-EP buffer (GE Healthcare). dAbmolecules fused to human IgG Fc domain were expressed at anunknown concentration and captured from clarified, 0.22-mmfiltered cell culture supernatants via IgG by injection at a flow rateof 10 mL/minute for 60 seconds. LRP6 ECD (5 mg/mL) in HBS-EPbuffer supplemented with 1%w/v BSA (Sigma) was then injectedfor 5 minutes at 30 mL/minute, followed by 10 minutes ofdissociation. Analysis was performed using the BIACORE 4000evaluation software (Biacore) using a global analysis 1:1 Lang-muir-binding model.

Pharmacokinetics of GSK3178022 in plasmaFor pharmacokinetics determination, GSK3178022 was dosed

into three groups (n ¼ 3) of male C57Bl6 mice. Each groupreceived an intravenous dose at either 0.25, 1, or 10mg/kg.Wholeblood was taken over a range of time points up to 72 hourspostdose for 0.25mg/kg dose group and up to 168 hours postdosefor the 1 and 10 mg/kg dose groups. Whole-blood samples weremixed 1:1 with water containing 5% EDTA to make blood–waterand stored at �80�C. Analysis of pharmacokinetic blood-watersamples was performed using the Meso Scale Discovery (MSD)platform. LRP6ECD(in-house)was coated onto 96-well standardbindMSDplates.Wellswere blockedwith assay buffer (3%BSA inPBS containing 0.1% Tween 20) and incubated for 1 hour withconstant shaking. Quality control samples and blood–waterpharmacokinetic samples were added at a range of dilutions. Acalibration curve was generated using GSK3178022 diluted incontrol mouse blood–water (SeraLab). Samples and standardswere incubated for 1 hour at room temperature with constantshaking. Bound GSK3178022 was detected with MSD sulfo-tagged mouse anti-human Vk mAb (in-house) or with MSDsulfo-tagged mouse anti-human Fc (in-house). MSD read bufferwas added and plates were read on a SECTOR 6000 MSD imager.Final assay results were fitted inWinNonLin by noncompartmen-tal analysis (NCA) according to standard methods. The meanpharmacokinetics profile of GSK3178022 in mice after 0.25, 1,and 10 mg/kg intravenous doses was plotted using GraphPadPrism version 6. Derived pharmacokinetics parameters wereobtained from the NCA fit.

In vivo pharmacodynamic and antitumor effect of GSK3178022NTERA-2 tumors were established by subcutaneous injection

of 8 � 106 NTERA-2 cells mixed with Matrigel (BD Biosciences)

Jackson et al.

Mol Cancer Res; 14(9) September 2016 Molecular Cancer Research860




1:1 with a total volume of 100 mL per injection in female NOD.CB17-Prkdcscid/NcrCrl mice (Charles River Laboratories). Tumorefficacy studies were performed at Crown Bioscience using femaleBALB/c nude mice. CR3150 and CR1560 patient-derived tumorswere established by implantation of tumor fragments fromBALB/c nude stockmice inoculated with selected primary human cancertissues. Each mouse was inoculated subcutaneously at the rightflank with the selected tumor fragment of 3� 3� 3mm size. Theimplantation day was denoted as D0. Treatments were startedwhen mean tumor size reached approximately 200 mm3. Forantitumor studies, mice were allocated randomly into four exper-imental groups of 10 mice each according to their tumor sizes. Anonbinding isotype control dual-dAb served as a negative control.Dual-dAbs and irinotecan were administered twice weekly byintraperitoneal injection at 40 mg/kg dose. To measure tumorgrowth, tumor sizes were determined twice weekly by calipermeasurement, and volume was determined, in mm3, using theformula: V ¼ 1/2(a) � (b)2, where a and b are the long andshort diameters of the tumor, respectively. To determine geneexpression, tumors were obtained by necropsy and homogenizedprior to isolating RNA using an RNeasy Kit (Qiagen). RNA wasconverted to cDNAusing theHighCapacity Reverse TranscriptionKit (Life Technologies). RT-PCR reactions were performed usingTaqMan Gene Expression Master Mix (Life Technologies) andvalidated TaqManGene Expression Assays (Life Technologies) forAXIN2 and SP5. Data were analyzed using the DDCt method bynormalizing data to ACTB.

RNAScopeSuperfrost Plus Slides (Fisher) were prepared with a 4-mm

section of paraffin-embedded tumor fromboth an isotype-treatedand a GSK3178022-treated animal. On the day of use, slides wereheated at 60�C for 45 minutes, cooled to room temperature,washed in xylene three times for 5 minutes, followed by 100%ethanol three times for 5minutes, and air dried. Slideswere boiledfor 10minutes at 98�C (LAUDA) inmRNAPretreatment Buffer #2(Ventana) diluted in diethyl pyrocarbonate (DEPC) water(Gibco) according to the kit instructions. Slides were then imme-diately placed in DEPC water at room temperature and storedthere until placement on the Ventana Ultra. On the Ultra, slideswere subjected to additional treatments of 8 minutes in Pretreat-ment A and Pretreatment B (Ventana), followed by incubation inthe specific probe (ACD proprietary sequences VS Probe - Hs SP5for NM_001003845.2), followed by signal amplification anddevelopment (Ventana cat #760-222 and #760-224). Slides weredehydrated in a series of ethanol washes, cleared in xylene,coverslipped, and imaged on a Hamamatsu NanoZoomer. Adja-cent sections were stained with hematoxylin and eosin.

ResultsLRP6 contains two independent binding sites for WNT ligand

engagement, and biopharmaceutical inhibition of signalingmediated by all WNT ligand classes, therefore, requires inhibitionat both sites (12, 13). dAbs are minimal binding regions derivedfrom the variable region of an antibody heavy or light chain(14, 15). To begin construction of a bispecific inhibitor of bothWNT ligand-binding sites,monospecific dAbswere first identifiedthat functionally blocked each site of LRP6. To identify dAbs,LRP6-binding dAbs were selected from a phage library using 3rounds of selection with decreasing antigen concentration to

apply selection pressure. Selected dAbs were expressed as bivalenthuman IgG1 Fc-domain fusion proteins, evaluated for specificityto LRP6 over LRP5 by ELISA and tested for ability to bind to LRP6ECD using surface plasmon resonance. dAb–Fc fusions were thentested in TOPFlash reporter assays using HEK293/6E cells toidentify molecules that functionally blocked different classes ofWNT ligands, known to bind each of the two binding sites. Thisidentified a dAb designated k020, which blocked functionalactivity of WNT3a, and dAbs designated H027, H035, andH037, which blocked functional activity of WNT1 (Fig. 1).

To determine the optimal combination of monospecific dAbsthat most effectively inhibited LRP6-mediated signaling by bothclasses of WNT ligands, "dual-dAb" constructs were generated,comprising dAbs fused to both the N- and C-termini of an IgG1Fc-domain. All pairings with k020 were generated: N-terminalK020 paired with C-terminal H027, H035, or H037 and vice versa.These bispecific, tetravalent proteinswere tested for their ability toinhibit WNT signaling in the human fibrosarcoma cell line, HT-1080, transfectedwith TOPFlash reporter plasmids and compared

0.005.00

10.0015.0020.0025.0030.0035.0040.0045.0050.00

TopF

lash

/Fop

Flas

h 10 µg/mL

1 µg/mL

0.1 µg/mL

0.005.00

10.0015.0020.0025.0030.0035.0040.0045.0050.00

TopF

lash

/Fop

Flas

h 10 µg/mL

1 µg/mL

0.1 µg/mL

A

B

Wnt1 S�mula�on

Wnt3a S�mula�on

Figure 1.

Screening of mono- or bispecific dAbs in the presence of WNT1 (A) or WNT3A(B). Inset figure represents the arrangement of dual-dAbs, with the dAbsconnected by short linkers, one dAb at the N-terminal of an Fc domainand a second dAb at the C-terminal of the Fc domain, as described in Materialsand Methods.

Wnt Pathway Inhibition by an LRP6 Bispecific dAb

www.aacrjournals.org Mol Cancer Res; 14(9) September 2016 861




with their monospecific counterparts (Fig. 1). The HT1080 cellline was used because it preferentially expresses LRP6, with nodetectable LRP5, by Western blot analysis (data not shown). Inthis assay, the bivalent, monospecific components of the dualdAbs tested (k020, H035, and H037) were each able to blocksignaling of one class of WNT ligands but elicited reciprocalpotentiation of signalingmediated by the alternative ligand class,as previously reported for monospecific mAbs (12, 13). Forexample, k020 (10 mg/mL) potentiated WNT signaling mediatedby WNT1 (Fig. 1A) but decreased signaling mediated by WNT3A(Fig. 1B). Conversely, and as expected, H035 and H037 eachinhibited WNT1-mediated signaling while potentiating signaling

byWNT3A. The functional results suggest that the different classesof dAbs interact with different domains of LRP6 that comprise thediscrete WNT1 class- and WNT3A class–binding sites. Whencombined into the dual dAb formats, with the exception ofk020-H027, all combinations tested were able to strongly inhibitsignaling in the presence of either WNT1 or WNT3A ligands or acombination of both. The reduced activity of the k020-H027 inblocking WNT1 stimulation may reflect a reduced ability of thisparticular bispecific construct to interact with both epitopes, orunfavorable biophysical characteristics of this construct, althoughthis was not explored further. H027-k020 demonstrated thestrongest activity and was selected for further characterization.

0

2

4

6

8

10

12

14

16

18

Wnt1 Wnt2 Wnt3 Wnt3a Wnt6 Wnt7b Wnt8a Wnt9a Wnt9b Wnt10a Wnt10b No WntNo Rspo

Top/

Fop

ra�o

No treatment

5 nmol/L GSK3178022

A

B

0

10

20

30

40

50

60

Rspo1 Rspo2 Rspo3 Rspo4 No Rspo Rspo1 Rspo2 Rspo3 Rspo4 No Rspo No Rspo

Top/

Fop

ra�o

No treatment

5 nmol/L GSK3178022

Wnt 1 Wnt 3a

“Ac�ve”β-Catenin

tβ-Catenin

tLRP6

pLRP6

Wnt3a – + – + –No treatment GSK3178022 Fzd8 CRD

β-Ac�n

+

C

Figure 2.

Inhibitory activity of GSK3178022 against a panel of WNT(A) and WNT and RSPO ligands (B). C, Western blotanalysis of HT-1080 cells treated first with anti-LRP6agent for 1 hour and recombinant WNT3A ligand for anadditional 1-hour incubation. Similar results were obtainedin three additional Western blot experiments.

Jackson et al.





Interestingly, each combination demonstrated superior activityover monospecific constructs, potentially due to the avidity effectachieved by the tetravalent, bispecific format. To further improvethe properties of H027-k020, the dAb moieties were subjected toaffinity maturation by scanning diversified residues within theCDRs by site-directedmutagenesis. Improvements in binding andbiological activity of novel variants were initially screened usingsurface plasmon resonance and single-point TOPFlash assays,which highlighted beneficial CDR mutations, which were theniteratively combined. In addition, it has been demonstrated thatresidues Leu234-Ser239 of the CH2 domain of human IgG Fc arenecessary for FcgR binding (17). The Fc effector functions ofH027-k020 were abrogated by mutating the equivalent of Leu-235 andGly-237 to alanine, abolishing Fc-g receptor binding andcomplement deposition. AnN-glycosylation site wasmutated outof the CDR2 of H027, and a C-terminal threonine residue wasincorporated into the molecule to minimize potential immuno-genicity. This yielded the lead molecule GSK3178022. Bindingexperiments demonstrated thatGSK3178022 is highly selective tohuman LRP6 over murine LRP6, with no significant cross-reac-tivity to human LRP5 (Supplementary Fig. S1).

To characterize the activity of GSK3178022 against a fullrange of WNT and RSPO ligands, the HT-1080 cell line andTOPFlash reporter assay were used with a broad panel of WNTligands. Expression vectors for all 19 WNT ligands were con-structed and tested for activity. In this format, 11 WNT ligands

were able to stimulate the assay, and the capacity ofGSK3178022 to inhibit signaling mediated by each was tested(Fig. 2A). GSK3178022 reduced stimulation by all active WNTligands except WNT7B, which only weakly stimulated in thisassay. It remains unclear whether the lack of inhibition is due toinability of GSK3178022 to block this ligand or whether itreflects a limitation of the assay.

RSPOs increase WNT signaling produced byWNT stimulation.To determine the ability of GSK3178022 to block signaling in thepresence of RSPO agonists, all four RSPO family members wereprofiled in the reporter assay in the presence of representativeWNT ligands to each of the binding sites on LRP6: WNT1 andWNT3A (Fig. 2B). All four RSPOs potentiated stimulation byWNT1 and WNT3A. The strongest stimulation was seen whenRSPO2 and WNT3A were combined, which approached 50-foldstimulation above baseline. GSK3178022 (5 nmol/L) completelyabrogated this signal, confirming that it has broad and potentinhibitory activity against a range ofWNT ligands, includingwhenamplified by RSPO ligands.

RSPOs are thought to stimulate WNT signaling by affectingthe turnover of FZD receptors (9, 18). WNT ligands lead tophosphorylation of LRP6 and stabilization of b-catenin in thecytosol (19). GSK3178022 may block WNT and RSPO stimu-lation by affecting turnover of LRP6 or by blocking signalingdownstream of the receptor complex. To determine the bio-chemical mechanism by which GSK3178022 inhibits WNT

A

B

050

100150200250300350400450

A549EKVX

HCC1143NCI-H

2122PA-1

BT-549

MDA-MB-468

NTERA-2

A549EKVX

HCC1143

NCI-H2030

NCI-H2030

NCI-H2122PA-1

BT-549

MDA-MB-468

NTERA-2

1.00

0.33

0.11

0.04

0.01

1.00

0.33

0.11

0.04

0.01

1.00

0.33

0.11

0.04

0.01

Fzd8

CRD

No

trea

tmen

t

No

Wnt

3a

Nor

mal

ized

SP5

mRN

A ex

pres

sion

GSK3178022 Monotopic k020 Monotopic H027µg/mL

SP5 Expression Axin2 Expression110 12

Wnt3a

Wnt3a + 10 nmol/LGSK3178022

10

8

6

4

2

0

90

16141210

864Re

la�v

e SP

5 ex

pres

sion

Rela

�ve

Axin

2 ex

pres

sion

20

Figure 3.

Inhibitory activity of GSK3178022 onWNT target gene expression in a panel ofcancer cell lines.A, expressionof theWnttarget genes AXIN2 and SP5 in thepresence of recombinantWNT3A ligand,with and without GSK3178022treatment, in a panel of cancer cell lines.B, ability of monospecific dAbs andGSK3178022 to decreasemRNA levels ofSP5 in NTERA-2 cells treated withrecombinant Wnt3a ligand.






signaling, the levels of total and phosphorylated LRP6 and"active" or unphosphorylated b-catenin levels were measuredin HT-1080 cells treated with WNT3A ligand and GSK3178022(Fig. 2C). WNT3A ligand treatment caused a marked increase inboth of these biochemical markers of Wnt pathway activationwithout changing the levels of either total LRP6 or b-catenin.GSK3178022 blocked phosphorylation of LRP6 as well as theunphosphorylated state of b-Catenin without altering levels oftotal LRP6. Therefore, GSK3178022 inhibits the receptor com-plex by blocking the biochemical events resulting from WNTligand engagement.

To establish the potential effectiveness of GSK3178022 acrossmultiple tumor types and on endogenous target gene expression,inhibitory activity was assessed in a panel of cancer cell lines usingRT-PCR against the known Wnt pathway target genes SP5 andAXIN2 (20, 21). Cell lines were stimulated by WNT3A with andwithout treatment with GSK3178022 and the expression of Wntpathway target genes SP5 and AXIN2 assessed (Fig. 3A). In all celllines examined, GSK3178022 inhibited stimulation of SP5 andAXIN2 expression, in most cases returning expression levels tobaseline. In five cell lines, expression was reduced below baseline,

suggesting inhibition of autocrine WNT signaling in these celllines.

The human testicular embryonal carcinoma cell line, NTERA-2,was themost responsive to stimulation of the cell lines tested andwas therefore used to further characterize the endogenousresponse of GSK3178022 treatment. Both monospecific dAb–Fcconstructs and GSK3178022 were tested. Neither the H027 dAbnor the k020 dAb alone were capable of blocking stimulation ofSP5 expression by WNT3A. However, GSK3178022 was highlyeffective at reducing levels of SP5 below baseline, with an EC50 ofaround 380 pmol/L (40 ng/mL; Fig. 3B). As a control, FZD8-CRDtested at 10 mg/mL was effective at reducing the stimulation, butnot to the levels achieved with GSK3178022. Therefore,GSK3178022 blocks expression of WNT target genes in a rangeof cancer cell lines with picomolar potency.

To permit detailed analysis of the in vivo activity ofGSK3178022, pharmacokinetic studies were performed in mice.GSK3178022 was dosed into three groups of mice, each receivinga single intravenous dose at either 0.25, 1, or 10 mg/kg. Nonlin-earity was observed across the dose escalation, with most para-meters not scaling as expected with dose, for example, from NCA

A

BMolecule Assay Dose T½ Tmax Cmax AUC 0-Inf_Pred MRT Vss Cl_Pred

(mg/kg) h h mg/mL h*mg/mL h mL/kg mL/h/kgAn�-Vk 4.6 0.1 2.9 14.6 6.2 106.6 17.4An�-Fc 4.8 0.1 2.8 14.5 6.6 114.3 17.4An�-Vk 8.9 0.2 9.8 92.2 11.8 128.7 10.9An�-Fc 8.8 0.2 10.7 97.3 11.8 122.1 10.3An�-Vk 19.2 0.2 94.7 440.8 19.5 438.5 23.2An�-Fc 18.6 0.2 93.9 480.7 20.4 420.3 21.2

GSK3178022

1,0000.25 mg/kg (Vk)1 mg/kg (Vk)10 mg/kg (Vk) 10 mg/kg (Fc)

1 mg/kg (Fc)0.25 mg/kg (Fc)

100

10

1

0.1

0.01

Log1

0 G

SK31

7802

2 pl

asm

a eq

uiva

lent

con

cent

ra�o

n (m

g/m

L)

0 10 20 30 40 50 60 70 80 90 100Time (h)

110 120 130 140 150 160 170 180

10

1

0.25

GSK3178022

GSK3178022

Figure 4.

Murine pharmacokinetics of GSK3178022. Mouse pharmacokinetics of GSK3178022 as determined by MSD using two assay formats (A and B), anti-Vk detection andanti-Fc detection, to inform on the integrity of the molecule in vivo.

Jackson et al.





analysis the half-lives and MRT increased with dose whileAUC (0–inf) increased with higher doses as expected but not ina dose-proportional manner. The volume of distribution alsoincreased at increasing dose; however, clearance was slowest forthe 1 mg/kg dose group (Fig. 4) The nonlinearity in pharmaco-kinetic parameters obtained suggests possible target-mediateddisposition through interaction with LRP6. The maximum half-life obtained in mice was 19.2 hours, with a Cmax of 94 mg/mL atthe 10 mg/kg dose.

Pharmacodynamic studies were undertaken to determine theeffect of in vivo treatment on endogenous WNT target genes.NTERA-2 tumor-bearing mice were treated with GSK3178022,and gene expression in tumors was determined by RT-PCR andin situ hybridization. Following a single-dose treatment withGSK3178022, NTERA-2 tumors were harvested at 5, 24, and 72hours posttreatment. RT-PCR was used to assess the expression ofSP5 as comparedwith tumors harvested fromuntreatedmice (Fig.5A). We observed a significant decrease in SP5 message levels inNTERA-2 tumors at both 5 and 24 hours posttreatment withGSK3178022. At 72 hours, SP5 message is reduced, but not at astatistically significant level. The pharmacodynamic effect of

GSK3178022 is consistent with the relatively short half-life ofthe molecule (<24 hours).

The tumor penetration and spatial modulation of wnt genemodulation achieved by GSK3178022 was analyzed by in situhybridization analysis of SP5 expression in tumor sections (Fig.5B). Sections of tumor from mice treated with an isotype controldual-dAb, which were incubated with the SP5-specific probe,showed a strong signal overmost regions of tumor cells indicatingsignificant levels of SP5 mRNA. Areas of connective tissue andbloodwithin the tumorwere devoidof signal. In contrast, sectionsof tumor frommice treated with GSK3178022 incubatedwith theSP5probe showed a significant reduction in signal inmost areas ofthe tumor, while others were devoid of signal, indicating reducedlevels of SP5 mRNA throughout the tumor. These data indicatexenograft penetrance and inhibition of WNT target gene expres-sion by GSK3178022 following intravenous administration.

Having determined that acute dosing of GSK3178022 inhib-ited WNT signaling in xenograft tumors, we tested whetherGSK3178022 treatment could affect growth in patient-relevanttumor models. RSPO fusions in colorectal cancer tumors havebeen found at a frequency of 10% (4). These fusions are exclusiveto the more common APC and b-catenin mutations in colorectalcancer, providing strong evidence that the fusions play a role intumorigenesis. In addition, efficacy has been demonstrated bytargeting of RSPOs in patient-derived xenograft (PDX) modelsselected on the basis of high RSPO expression (22). BecauseGSK3178022 effectively inhibits both WNT and RSPO stimula-tion in vitro, PDX models with genetically defined RSPO fusionswere identified and used for in vivo testing of the efficacy ofGSK3178022 to inhibit Wnt signaling and tumor growth.

RSPO fusion colorectal cancer PDX models were identified byexamining microarray expression data from a large panel ofcolorectal cancer models. Models with clearly elevated RSPOexpression were tested by PCR as described previously (4). Twomodels, CR1560 and CR3150, were identified with Crown Bio-science as having RSPO3 translocations by PCR and sequencing.The b-catenin and APC genes in CR1560 and CR3150 weresequenced and found to be wild type. Additional oncogenicmutations were present in these PDX models (SupplementaryTable S1). Mutations in KRAS, TP53, and EGFR were present inboth models. In addition, a mutation in PI3Ka (PIK3CA) waspresent in CR3150. To determine the pharmacodynamics andefficacy effect of GSK3178022 treatment, cohorts of mice wereestablished with subcutaneous CR1560 and CR3150 tumors.Tumors were isolated 24 hours after the final treatment dosingin the efficacy groups (below), and the expression of WNT targetgenes was examined by RT-PCR in tumor homogenates. Strong(>60%) suppression of theWNT target gene AXIN2was seen onlyin the CR1560 tumor model, with no effect on AXIN2 expressioninCR3150 (Fig. 6A). Serum levels ofGSK3178022 inbothmodelswere determined and found to be similar: 12.4 mg/mL� 1.9 (SE)for CR3150 and 11.0mg/mL� 0.1 (SE) for CR1560. Therefore, thelack of gene response in CR3150 suggests that this model isinsensitive to LRP6 inhibition by GSK3178022.

To establish the effect of GSK3178022 on tumor growth,groups of tumor-bearingmicewere treatedwith an isotype controldual-dAb, GSK3178022, irinotecan, or a combination ofGSK3178022 and irinotecan. In both models, treatment withirinotecan alone led to regression in tumor volume, and this effectwas not increased in combinationwith GSK3178022 (Fig. 6B andC). Single-agent treatment with GSK3178022 led to an initial

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6Re

la�v

e SP

5 ex

pres

sion

5 Hour

24 Hour

72 Hour

90 mg/kgIsotype

90 mg/kgGSK3178022

*

* P < 0.05

*

Untreated

A

Isotype GSK3178022-treatedSP5 RNAscope H&E SP5 RNAscope H&E

B

A B C D

Figure 5.

In vivo inhibitory activity of GSK3178022 on Wnt target gene expression inNTERA-2 tumor xenografts. A, relative expression of SP5 mRNA levels at 5 and24 hours postdosing in mice bearing NTERA-2 tumor xenografts and leftuntreated or dosed once with either GSK3178022 or an isotype control.B,in situ hybridization analysis of SP5 expression. Tumors were harvested at 24hours, sectioned, and probed with an SP5-specific probe or stained withhematoxylin and eosin (H&E).






delay in tumor growth in CR1560 (Fig. 6C) but had no effect inCR3150 (Fig. 6B). No in vitro effect of GSK3178022 treatmentwasseen in cell lines derived from either model, under 2D or 3Dgrowth conditions (data not shown).

DiscussionUsing dAb technology, we constructed an anti-LRP6 bispecific

molecule, GSK3178022, which potently inhibits signaling by a

range of WNT and RSPO ligands in cancer cell lines in vitro. Inaddition, GSK3178022 reduced WNT-mediated gene upregula-tion in vivo, both in the cancer cell line Ntera2 and in one of twoPDX models selected for activated RSPO status. GSK3178022delayed tumor growth in the xenograft model that respondedmolecularly but had no effect on the nonresponding model. Thismolecule demonstrates the ability to construct bispecific mole-cules, of reduced molecular weight compared with a mAb, thatcan functionally inhibit independent ligand-binding sites on thesame receptor. In addition, with respect to LRP6, GSK3178022meets the key challenge of simultaneously inhibiting ligand-mediated signaling through either binding site without theadverse reciprocal increase signaling previously described withconventional anti-LRP6 antibodies.

Both small-molecule and biological attempts have previouslybeen made to develop targeted therapeutics against Wnt stimu-lation. For example, LGK-974 is a small-molecule inhibitor ofPorcupine (PORCN), an enzyme required for lipidation of WNTligands (23). Because lipidation is a general property of WNTproteins, and required for secretion, LGK-974 is capable ofcompletely blocking WNT ligand stimulation of the pathway. Inaddition, OMP-18R5 is an antibody in phase I trials that recog-nizes a subset of FZD coreceptors and blocks WNT ligand engage-ment (24). However, targeting the action of WNT ligands alonemay be insufficient in tumors that have alterations in other Wntpathway modulators, such as in the RSPO/LGR pathway (6, 8).GSK3178022 is capable of blocking activation by both WNT andRSPO ligand classes, and therefore may be effective in a broaderrange of WNT-activated tumors.

Wnt pathway activity is also required in normal adult tissues,such as the intestine, bone, and liver, and can be modulated byeither LRP5 or LRP6 in a tissue-specific manner. For example,WNT signaling in bone regulates osteoblast differentiation and isrequired to maintain bone density (25). Human genetic dataindicate that LRP5 is an essentialWnt signaling component in thistissue andboth gain- and loss-of-functionmutations in LRP5havebeen described that lead to osteopetrosis and osteoporosis,respectively. A role for LRP6 in bone in humans has also beendescribed. Individuals with a loss-of-function mutation in LRP6develop early onset coronary artery disease, features of metabolicsyndrome, and osteoporosis (26). Mice with compound muta-tions in LRP5 and LRP6 globally or specifically in the embryonicmesenchyme show deficits in bone and limb formation in a dose-dependent manner (27, 28). In the intestine, conditional mousestudies have demonstrated a functional redundancy betweenLRP5 and LRP6 (29). Intestinal morphology was normal in micewith single conditional knockouts of LRP5 or LRP6, but disruptedinmicewith double conditional knockout of both. LRP6 has beendescribed as a more potent biochemical mediator of WNT stim-ulation than LRP5 (30). Consistent with this, we show thatGSK3178022 is broadly inhibitory of WNT3A stimulation incancer cell lines expressing both LRP5 and LRP6, indicating thatLRP5 cannot functionally substitute for LRP6 in these cell lines.However, GSK3178022 may affect normal tissue homeostasiswhere LRP6 is essential and nonredundant with LRP5. In addi-tion, GSK3178022 may fail to block Wnt pathway activation intumors that rely on LRP5 for Wnt activation.

Because GSK3178022 is an effective inhibitor of WNT andRSPO stimulation through LRP6, tumors with a genetic RSPOfusion may represent a tumor type responsive to GSK3178022treatment. Two patient-derived RSPO3 fusion positive colorectal

1.6

1.4

1.2

1

0.8

0.6

0.4

0.2

Nor

mal

ized

Axin

2 m

RNA

expr

essi

on

CR31

50 T

umor

size

(mm

3 )CR

1560

Tum

or si

ze (m

m3 )

0CR3150 CR1560

Isotype

A

B

C

GSK3178022

1,4001,3001,2001,1001,000

900800700600500400300200100

0

1,4001,3001,2001,1001,000

900800700600500400300200100

033 35 37 39 41 43 45 4947 51 53 55 57 59 61 63

30 32 34

Isotype

GSK3178022

GSK3178022 +Irinotecan

Irinotecan

Isotype

GSK3178022

GSK3178022 + Irinotecan

Irinotecan

36 38 40

Days a�er tumor inocula�on

Days a�er tumor inocula�on

42 44 46 48 50 52 54 56 58 60 62 64

Figure 6.

Effect of GSK3178022 in colorectal PDXmodels harboring RSPO3 fusions.A, twocolorectal PDX tumor models were treated with either an isotype control orGSK3178022; tumors were harvested at necropsy at the end of the study andAxin2 mRNA levels assessed. B, tumor volumes of xenograft model CR3150treatedwith isotype control andGSK3178022 either aloneor in combinationwithirinotecan. C, tumor volumes of xenograft model CR1560 treated with isotypecontrol and GSK3178022 either alone or in combination with irinotecan.Treatment groups isotype and GSK3178022 were compared using theWilcoxon–Mann–Whitney two-sample rank-sum test; asterisks (�) indicateresults that were significant at P � 0.05.

Jackson et al.





cancer xenograft models were tested for sensitivity toGSK3178022. Of the two identified models, one model did notrespond at themolecular level, indicating that RSPO fusion statusalone is insufficient to predict response. Pathway inhibition inmutationally activated colorectal cancer cell lines has also shownvariable dependence on the Wnt pathway (31–33). It is alsopossible that RSPO ligands can signal through parallel pathways,in addition to effects on the Wnt pathway. However, the secondidentified RSPO3 fusion positivemodel demonstrated a responseto GSK3178022, both at the level of WNT target gene expressionand in tumor growth delay, supporting the hypothesis that RSPOfusions activate Wnt signaling through LRP6.

We demonstrated GSK3178022 is a potent inhibitor ofsignaling mediated by a range of WNT ligands in multiple celllines, including in the presence of RSPO ligands. We sought toextend these studies into in vivo tumor models, to begin tocharacterize the pharmacokinetic and pharmacodynamic prop-erties of GSK3178022 in mice and to further explore its ther-apeutic potential.

GSK3178022 has a half-life of less than 20 hours inmice whichmay be driven, in part, by target-mediated clearance in normaltissues, as GSK3178022 is cross-reactive for mouse LRP6. None-theless, administration of a single 90mg/kg dose was sufficient toachieve a significant reduction in expression of a key WNTresponse gene, SP5, in NTERA-2 tumors, which was detectable5 hours after administration and was sustained for at least 24hours. These data and in situ hybridization analysis were consis-tent with GSK3178022 achieving tumor penetration and inhib-itory activity against WNT signaling in vivo. Similarly, AXIN2expression was significantly decreased in CR1560 tumors sub-jected to twice weekly GSK3178022 administration.. Despiteachieving such functional inhibition of WNT signaling, admin-istration of GSK3178022 under the conditions tested failed toprovide a therapeutic benefit. The reason for this observationremains to be determined, but may reflect inadequate potency ofGSK3178022 against multiple WNT ligands in an in vivo contextor insufficient exposure of GSK3178022 following the treatmentregimen used. Further investigation will be required to (i) bettercharacterize the relationship between GSK3178022 pharmacoki-netics and pharmacodynamics; (ii) define optimal pharmacody-namic marker(s) for GSK3178022-mediated wnt inhibition,which predict efficacious tumor response; and (iii) determine themagnitude ofWNT response gene inhibition, which is required toaffect tumor growth arrest or regression in vivo. Conversely, thesedata may reflect redundancy between LRP5 and LRP6 in this

context, which would limit the expected benefit of anti-LRP6therapy.

Using dAb technology, we demonstrate effective simultaneousinhibition of the two independent ligand-binding domains of theLRP6 receptor. We demonstrate that LRP6 is necessary for WNTstimulation in a broad range of cancer cell lines and is alsonecessary for RSPO stimulation of the Wnt pathway. In addition,we found that combining two domain antibodies into a bispecificmolecule increased potency compared with admixtures of theirmonospecific counterparts. Therefore, engineered bispecificmolecules using domain antibodies are a promising approachto generating potent inhibitors against complex receptors thatmediate multiple ligand interactions.

Disclosure of Potential Conflicts of InterestFunding for theworkwas provided byGlaxoSmithKline (GSK), the employer

of H. Jackson, D. Granger, G. Jones, L. Anderson, S. Friel, D. Rycroft, W. Fieles,J. Tunstead, M. Steward, T. Wattam, J. Griggs, M. Al-Hajj, and C. Shelton. Allstudies were conducted in accordance with the GSK Policy on the Care, Welfareand Treatment of Laboratory Animals and were reviewed at the InstitutionalAnimal Care and Use Committee either by GSK or by the ethical review processat the institution where the work was performed.

Authors' ContributionsConception and design:H. Jackson, D. Granger, G. Jones, D. Rycroft, W. Fieles,J. Tunstead, M. Steward, A. Walker, J. Griggs, M. Al-Hajj, C. SheltonDevelopment of methodology:H. Jackson, D. Granger, L. Anderson, W. Fieles,J. Tunstead, T. Wattam, J. Griggs, C. SheltonAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): H. Jackson, D. Granger, G. Jones, L. Anderson,S. Friel, D. Rycroft, W. Fieles, T. Wattam, C. SheltonAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis):H. Jackson, D. Granger, S. Friel, D. Rycroft, W. Fieles,J. Tunstead, M. Steward, T. Wattam, C. SheltonWriting, review, and/or revision of the manuscript: H. Jackson, D. Granger,S. Friel, D. Rycroft, W. Fieles, J. Tunstead, M. Steward, A. Walker, J. Griggs,M. Al-Hajj, C. SheltonAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): C. SheltonStudy supervision: D. Rycroft, C. Shelton

AcknowledgmentsThe authors acknowledge Crown Bioscience for performing in vivo studies.

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

Received March 16, 2016; revised June 3, 2016; accepted June 22, 2016;published OnlineFirst July 11, 2016.

References1. Nusse R, vanOA,CoxD, Fung YK, VarmusH.Modeof proviral activation of

a putative mammary oncogene (int-1) on mouse chromosome 15. Nature1984;307:131–6.

2. Powell SM, Zilz N, Beazer-Barclay Y, Bryan TM, Hamilton SR, ThibodeauSN, et al. APC mutations occur early during colorectal tumorigenesis.Nature 1992;359:235–7.

3. Forbes SA, Beare D, Gunasekaran P, Leung K, Bindal N, Boutselakis H, et al.COSMIC: exploring theworld's knowledge of somaticmutations in humancancer. Nucleic Acids Res 2015;43:D805–D811.

4. Seshagiri S, Stawiski EW, Durinck S, Modrusan Z, Storm EE, Conboy CB,et al. Recurrent R-spondin fusions in colon cancer.Nature 2012;488:660–4.

5. Jiang X, Hao HX, Growney JD, Woolfenden S, Bottiglio C, Ng N,et al. Inactivating mutations of RNF43 confer Wnt dependency in

pancreatic ductal adenocarcinoma. Proc Natl Acad Sci U S A 2013;110:12649–54.

6. Clevers H, Nusse R. Wnt/beta-catenin signaling and disease. Cell2012;149:1192–205.

7. Niehrs C. The complex world ofWNT receptor signalling. Nat RevMol CellBiol 2012;13:767–79.

8. de LW, Peng WC, Gros P, Clevers H. The R-spondin/Lgr5/Rnf43 module:regulator of Wnt signal strength. Genes Dev 2014;28:305–16.

9. Hao HX, Xie Y, Zhang Y, Charlat O, Oster E, Avello M, et al. ZNRF3promotesWnt receptor turnover in an R-spondin-sensitivemanner. Nature2012;485:195–200.

10. Fukui T, Kondo M, Ito G, Maeda O, Sato N, Yoshioka H, et al. Transcrip-tional silencing of secreted frizzled related protein 1 (SFRP 1) by promoter






hypermethylation in non-small-cell lung cancer. Oncogene 2005;24:6323–7.

11. Mazieres J, He B, You L, Xu Z, Lee AY,Mikami I, et al. Wnt inhibitory factor-1 is silenced by promoter hypermethylation in human lung cancer. CancerRes 2004;64:4717–20.

12. Ettenberg SA, Charlat O, Daley MP, Liu S, Vincent KJ, Stuart DD, et al.Inhibition of tumorigenesis driven by different Wnt proteins requiresblockade of distinct ligand-binding regions by LRP6 antibodies. Proc NatlAcad Sci U S A 2010;107:15473–8.

13. Gong Y, Bourhis E, Chiu C, Stawicki S, DeAlmeida VI, Liu BY, et al. Wntisoform-specific interactions with coreceptor specify inhibition orpotentiation of signaling by LRP6 antibodies. PLoS One 2010;5:e12682.

14. Holt LJ, Herring C, Jespers LS, Woolven BP, Tomlinson IM. Domainantibodies: proteins for therapy. Trends Biotechnol 2003;21:484–90.

15. Walker A, Dunlevy G, Rycroft D, Topley P, Holt LJ, Herbert T, et al. Anti-serum albumin domain antibodies in the development of highlypotent, efficacious and long-acting interferon. Protein Eng Des Sel2010;23:271–8.

16. Ignatovich O, Jespers L, Tomlinson IM, de Wildt RM. Creation of the largeand highly functional synthetic repertoire of human VH and Vkappadomain antibodies. Methods Mol Biol 2012;911:39–63.

17. Morgan A, Jones ND, Nesbitt AM, Chaplin L, Bodmer MW, Emtage JS. TheN-terminal end of the CH2 domain of chimeric human IgG1 anti-HLA-DRis necessary for C1q, FcyRI and FcyRIII binding. Immunology 1995;86:319–324.

18. Binnerts ME, Kim KA, Bright JM, Patel SM, Tran K, Zhou M, et al. R-Spondin1 regulates Wnt signaling by inhibiting internalization of LRP6.Proc Natl Acad Sci U S A 2007;104:14700–5.

19. Bilic J, Huang YL, Davidson G, Zimmermann T, Cruciat CM, BienzM, et al.Wnt induces LRP6 signalosomes and promotes dishevelled-dependentLRP6 phosphorylation. Science 2007;316:1619–22.

20. Jho EH, Zhang T, Domon C, Joo CK, Freund JN, Costantini F. Wnt/beta-catenin/Tcf signaling induces the transcription of Axin2, a negative regu-lator of the signaling pathway. Mol Cell Biol 2002;22:1172–83.

21. Takahashi M, Nakamura Y, Obama K, Furukawa Y. Identification of SP5 asa downstream gene of the beta-catenin/Tcf pathway and its enhancedexpression in human colon cancer. Int J Oncol 2005;27:1483–7.

22. Chartier C, Raval J, Axelrod F, Bond C, Cain J, Dee-Hoskins C, et al.Therapeutic targeting of tumor-derived r-spondin attenuates beta-cateninsignaling and tumorigenesis in multiple cancer types. Cancer Res 2016;76:713–23.

23. Liu J, Pan S, Hsieh MH, Ng N, Sun F, Wang T, et al. Targeting Wnt-drivencancer through the inhibition of Porcupine by LGK974. Proc Natl Acad SciU S A 2013;110:20224–9.

24. Gurney A, Axelrod F, Bond CJ, Cain J, Chartier C, Donigan L, et al. Wntpathway inhibition via the targeting of Frizzled receptors results indecreased growth and tumorigenicity of human tumors. Proc Natl AcadSci U S A 2012;109:11717–22.

25. Burgers TA, Williams BO. Regulation of Wnt/beta-catenin signaling withinand from osteocytes. Bone 2013;54:244–9.

26. Mani A, Radhakrishnan J, Wang H, Mani A, Mani MA, Nelson-Williams C,et al. LRP6mutation in a family with early coronary disease andmetabolicrisk factors. Science 2007;315:1278–82.

27. Holmen SL, Giambernardi TA, Zylstra CR, Buckner-Berghuis BD, Resau JH,Hess JF, et al. Decreased BMD and limb deformities in mice carryingmutations in both Lrp5 and Lrp6. J Bone Miner Res 2004;19:2033–40.

28. Joeng KS, Schumacher CA, Zylstra-Diegel CR, Long F, Williams BO. Lrp5and Lrp6 redundantly control skeletal development in the mouse embryo.Dev Biol 2011;359:222–9.

29. Zhong Z, Baker JJ, Zylstra-Diegel CR, Williams BO. Lrp5 and Lrp6 playcompensatory roles in mouse intestinal development. J Cell Biochem2012;113:31–8.

30. MacDonald BT, Semenov MV, Huang H, He X. Dissecting moleculardifferences between Wnt coreceptors LRP5 and LRP6. PLoS One 2011;6:e23537.

31. Verma UN, Surabhi RM, Schmaltieg A, Becerra C, Gaynor RB. Smallinterfering RNAs directed against beta-catenin inhibit the in vitro andin vivo growth of colon cancer cells. Clin Cancer Res 2003;9:1291–300.

32. Roh H, Green DW, Boswell CB, Pippin JA, Drebin JA. Suppression of beta-catenin inhibits the neoplastic growth of APC-mutant colon cancer cells.Cancer Res 2001;61:6563–8.

33. Zhang HH, Walker F, Kiflemariam S, Whitehead RH, Williams D,Phillips WA, et al. Selective inhibition of proliferation in colorectalcarcinoma cell lines expressing mutant APC or activated B-Raf. Int JCancer 2009;125:297–307.


Jackson et al.




2016;14:859-868. Published OnlineFirst July 11, 2016.Mol Cancer Res Heather Jackson, David Granger, Gavin Jones, et al. R-spondin Ligand-Induced Wnt Signaling and Tumor GrowthNovel Bispecific Domain Antibody to LRP6 Inhibits Wnt and

Updated version

10.1158/1541-7786.MCR-16-0088doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2016/07/09/1541-7786.MCR-16-0088.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/14/9/859.full#ref-list-1

This article cites 33 articles, 13 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/14/9/859.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://mcr.aacrjournals.org/content/14/9/859To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-16-0088

http://mcr.aacrjournals.org/content/suppl/2016/07/09/1541-7786.MCR-16-0088.DC1

http://mcr.aacrjournals.org/content/14/9/859.full#ref-list-1

http://mcr.aacrjournals.org/content/14/9/859.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/14/9/859


novelbispeciﬁcdomainantibodytolrp6inhibits wntandr ... · revealed two independent wnt-binding...

Documents