HAL Id: tel-03033429
https://tel.archives-ouvertes.fr/tel-03033429
Submitted on 1 Dec 2020
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Development of theranostic molecules targeted against
prostate cancer
Jungyoon Choi
To cite this version:
Jungyoon Choi. Development of theranostic molecules targeted against prostate cancer. Human health and pathology. Université Grenoble Alpes, 2017. English. �NNT : 2017GREAV025�. �tel-03033429�
THÈSE
Pour obtenir le grade de
DOCTEUR DE LA COMMUNAUTÉ UNIVERSITÉ
GRENOBLE ALPES
Spécialité : Biotechnologie
Arrêté ministériel : 7 août 2006
Présentée par
« Jungyoon CHOI »
Thèse dirigée par « Jean-Luc COLL »
préparée au sein du Centre de recherche INSERM-UGA U1209 pour l'Avancée des Biosciences dans l'École Doctorale Chimie et Sciences du vivant
Développement de molécules
théranostiques ciblantes contre
le cancer de la prostate
Thèse soutenue publiquement le « 6 juillet 2017 », devant le jury composé de :
M. Alain, DUPERRAY
Directeur de recherche, INSERM, Grenoble, Président
Mme. Catherine, LADAVIERE
Directeur de recherche, CNRS, Lyon, Rapporteur
M. Cédric, GAGGIOLI
Charge de recherche, INSERM, Nice, Rapporteur
M. Didier, BOTURYN
Directeur de recherche, CNRS, Grenoble, Examinateur
Mme. Elisabeth, GARANGER
Directeur de recherche, CNRS, Bordeaux, Examinateur
Mme. Isabelle, TEXIER-NOGUES
Senior scientist-directeur de recherche, CEA, Grenoble, Examinateur
M. Jean-Luc, COLL
Acknowledgements
Publications and Communications
• Coll, J.L., Choi, J. Challenges of Applying Targeted Nanostructures with
Multifunctional Properties in Cancer Treatments. Nanoscience and
Nanotechnology for Human Health, Voorde, B. M. a. M. V. d., Ed. John Wiley & Sons: 2017; pp 127-156.
• Choi, J., Rustique, E., Henry, M., Guidetti, M., Josserand, V., Sancey, L., Boutet, J., and Coll, J. L. (2017). Targeting tumors with cyclic RGD-conjugated lipid
nanoparticles loaded with an IR780 NIR dye: In vitro and in vivo evaluation. Int
J Pharm. DOI: 10.1016/j.ijpharm.2017.03.007.
• Annual presentations in Ph.D. committee in the institute, Institute for Advanced Biosciences (IAB), La Tronche, 2014, 2015, 2016
• ‘Theranostic nanoparticles for prostate cancer’, Bio Korea in Europe (BiKiE) symposium, KIST Europe, Saarbrucken, Germany, 2016
• ‘Targeted nanoparticles for prostate cancer imaging’, Journée Sciences et Métiers, UJF, Grenoble, 2014
• ‘Cancer-targeted theranostic nanovectors that can help early diagnosis and advanced therapy of prostate cancer’, Annual meeting of EDCSV, Grenoble, 2015
• [Prize] ‘Cancer-targeted theranostic nanovectors that can help early diagnosis and advanced therapy of prostate cancer’, Euro-Korean Conference on science and technology, Strasbourg, 2015
• ‘Theranostic nanoparticles to help early diagnosis and targeted therapy for prostate cancer’, SFNano European Nanomedicine Meeting, Grenoble, 2015 • ‘Theranostic nanoparticles for prostate cancer’, European Molecular Imaging
Meeting, Utrecht, Netherland, 2016
• ‘cRGD-Lipid nanoparticle loaded with IR780 iodide for cancer
NIR/photoacoustic imaging’, SFNano European Nanomedicine Meeting, Paris, 2016
Table of contents
αvβ3
αvβ αvβ3
β α
List of Figures
α β
β
β β
β α
α β
Résumé
Développement de molécules théranostiques dirigées contre le cancer de la prostate
α β β β α
α β α β
Mots clés: integrin alpha v beta 3, Neuropilin-1, lipid nanoparticles, silica nanoparticles, prostate cancer, nanomedicine, cancer imaging, dual targeting
Abstract
Development of theranostic molecules targeted to prostate cancer
α β
β β α
α β α β
Key words: integrin alpha v beta 3, Neuropilin-1, lipid nanoparticles, silica nanoparticles, prostate cancer, nanomedicine, cancer imaging, dual targeting
Introduction
Prostate cancer and its diagnosis
Incidence and mortality of prostate cancer
Figure 1. Worldwide map of estimated prostate cancer incidence in 2012.
Different colors stand for the indicated range of the estimated age-standardized rates (world) per 100,000. The figure was adapted from GLOBOCAN 2012.
Figure 2. Trends in incidence and mortality of prostate cancer in selected countries.
Figure 3. Percentage of cases and 5-year prevalence by prostate cancer stage.
Figure 4. Prostate zones and TRUS-guided prostate biopsy.
(A) Image was adapted from reference (De Marzo et al., 2007). (B) Illustration of TRUS prostate biopsy
(available from the National Cancer Institute website, www.cancer.gov). (C) An example of cancer in echography which is easily distinguishable due to its big size. Adapted from reference (Patel, 2004).
Prostate photoacoustic imaging Figure 5. Photoacoustic imaging
(A) A photoacoustic imaging device, VisualSonics VEVO LAZR, installed in our in vivo imaging facility
(OPTIMAL). (B) A close view of the transducer head in VEVO LAZR. The image was adapted from a company’s brochure, available at https://www.visualsonics.com/. (C) Echography (left) and photoacoustic image superposed on echography (right) in tumor area after the injection of EGFR-targeted gold nanoparticles. Nanoparticles in yellow, oxygenated blood in red, and deoxygenated blood in blue. Reused from a publication of Homan et al. (Homan et al., 2012).
α β
Figure 6. Absorption spectra of endogenous chromophores
and exogenous contrast agents
Images were reused from a review article of Mallidi et al (Mallidi et al., 2011)
In vivo optical imaging technologies
Figure 7. NIR imaging system of Fluoptics for the intraoperative use
Images were reused from Fluoptics’ webpage (https://www.fluoptics.com/).
Figure 8. Examples of intraoperative fluorescence imaging systems.
(a) The Novadaq SPY™ system, (b) ArtemisTM, (c) Hamamatsu’s Photodynamic Eye (PDE™), (d) Fluoptics’
Fluobeam®. Functional intraoperative FMI systems: (e) FLARE™ imaging system, (f) Multispectral FMI system from Technische Universität München & Helmholtz Zentrum, (g) Surgical navigation system GXMI Navigator from the Institute of Automation, Chinese Academy of Sciences. This figure was reused from a review article of Chi et al. (Chi et al., 2014).
Tumor-targeted contrast agents for NIR imaging
αvβ
αvβ αvβ
Table 1. Targeted NIR fluorescence probes including small molecules and nanoparticles
Reused from a review article of Yi et al (Yi et al., 2014). Reference number in this table corresponds to the review.
Figure 9. Schemes of cRGD-based integrin αv3β-targeting molecules developed in our group with
collaborators
(A) RAFT-cRGD4 (Boturyn et al., 2004), (B) lipid nanoparticles (Goutayer et al., 2010), (C)
cRGD-conjugated gadolinium-based Small Rigid Platforms (SRPs) (Morlieras et al., 2013). Images were adapted from the original articles.
Table 2. Clinical trials using targeted fluorophores (small molecules) for intraoperative cancer detection
Figure 10. Lipid nanoparticles loaded with NIR dyes
(A) Structure of lipid nanoparticle. (B) NIR dyes ICG (indocyanine green) and IR780-oleyl dye that were
embedded in the lipid nanoparticles. (C) EPR (Enhanced permeability and retention) effect in subcutaneous tumor-bearing mice after injections of IR780-oleyl dye-loaded lipid nanoparticles (50 nm in size); in vivo imaging was taken during 22 days post-injection. Images were adapted from the original articles of Gravier et al and Jacquart et al (Gravier et al., 2011; Jacquart et al., 2013).
Table 3. Clinical-stage nanomedicines for cancer therapy
Table 3 (continued). Clinical-stage nanomedicines for cancer therapy
Tumor-targeting via Integrin αvβ3
Integrins in the tumor environment
α-
β-α-
β-Figure 11. Integrin heterodimers found in humans and their ligands
α β αvβ αvβ αvβ αvβ αvβ α β αvβ α β α β α β α β α β α β α β αvβ αvβ α β αvβ αvβ αvβ αvβ αvβ α β α β αvβ α β α β αvβ α β αvβ
αv αvβ αvβ α β αvβ αvβ α β αvβ αvβ
Integrins and prostate cancer α β3 αvβ5 αIIbβ3 αvβ3 αvβ5 αvβ3 αvβ5 αvβ3 αvβ5 αvβ3 αvβ5
Table 4. Expressions of RGD-binding integrins observed in clinical tissue
samples of prostate adenocarcinoma
RGD peptide as an inhibitor of αvβ3 integrin
α β α β
RGD peptides for cancer imaging Table 5. Integrin antagonists tested in clinical trials for cancer therapy
Adapted from a review of Avraamides et al from 2008 (Avraamides et al., 2008). Cilengitide has completed a phase III study with unsuccessful outcomes in 2014 (Stupp et al., 2014).
αvβ αvβ
Dual-targeting of angiogenesis via Integrin αvβ3 and
Neuropilin-1 (NRP1)
Multiple regulators are controlling angiogenesis
β β β
α β α β
Crosstalk between integrins and NRP1 in angiogenesis
Figure 12. Integrins-VEGFR2 signaling
α β
α β α β
α β
Figure 13. iRGD’s targeting mechanism and its accumulation in tumors
(A) A schematic representation of the targeting and internalization mechanisms of iRGD. (B) Accumulation of
iRGD peptide, and iRGD-decorated phage (65 nm) and micelles (10‒25 nm) in orthotopic 22Rv1 human prostate cancer xenograft mice. CRGDC peptide: integrin αvβ3-binding peptide without penetration activity. Scale bar: 50 µm. Adapted from a publication of Sugahara et al, 2009 (Sugahara et al., 2009).
Examples of RGD/ATWLPPR dual-targeting
Objectives
Presentation of Project 2: cRGD/ATWLPPR-Silica
nanoparticles (SiNP)
Results
Characteristics of cell lines
β β
β β
Expression profile of integrin αvβ3 and neuropilin-1
αvβ αvβ
Table 6. List of cell lines
Name of cell line Description
PC3 Human prostate adenocarcinoma, derived from a bone metastasis PC3-luc A sub-cell line of PC3, transfected with the luciferase reporter vector
DU145 Human prostate adenocarcinoma, derived from a brain metastasis RWPE1 Human prostate normal epithelial cell line
U87MG Human glioblastoma cell line, derived from a malignant glioma M21 Human melanoma cell line, known to highly express the αvβ3 integrin M21L A variant cell line of M21 which lacks the expression of αv integrin
HEK293 A human embryonic kidney cell line (hypotriploid), known to express the αv integrin and lack the β3 integrin
HEK293-β3 A sub-cell line of HEK293, transfected with the β3 integrin cDNA HEK293-β1 A sub-cell line of HEK293, transfected with the β1 integrin cDNA
HEK293-β3-αvRFP A sub-cell line of HEK293-β3, transfected with a red fluorescent protein (RFP) fused in the extracellular domain (N-terminus part) of the αv integrin elsewhere the RGD binding domain
HUVEC Primary human umbilical vein endothelial cells
EA.hy926 A human umbilical vein cell line, established by fusing primary human umbilical vein cells with a thioguanine-resistant clone of A549 by the exposure to PEG
αvβ β β αvβ
Figure 14. Expression profiles of αvβ3 and NRP-1 in human prostate cancer cell lines and other reference cell
lines
54
Project 1, cRGD-LNP
We initially tested two different synthesis methods to produce the targeted LNPs: (1) using labeled peptides on the LNPs or (2) starting with maleimide-labeled LNPs and the peptides. All synthesis were performed by Emilie Rustique in CEA-LETI and I verified the LNPs’ functionality at first.
Characteristics of LNPs
Lipid nanoparticle (LNP), initially called LipImage™ 815, was synthesized following the previously described method (Jacquart et al., 2013). LNPs are composed of an oily core mixed with oleyl-IR780 near-infrared (NIR) dyes and a PEGylated phospholipid monolayer which encapsulates the core (Figure 15). To conjugate the peptides on the surface of LNPs, PEG surfactants and the peptides were linked by the maleimide-thiol coupling reaction: R-maleimide + R’-SH R-(thiol ether bond)-R’. The conjugation was performed in two directions in order to select the better one. In the first method (Fuctionalization1, F1), maleimide-labeled peptides were reacted with SH-labeled LNPs, while in the second method (Functionalization2, F2) maleimide-LNPs were reacted with SH-peptides. cRGD-LNP (αvβ3-targeting), cRAD-LNP (no affinity for αvβ3 due to a nonsense peptide sequence with alanine instead of glycine), and standard/OH LNP (negative control without peptide) were generated (Figure 15C).
The F1-LNPs were slightly larger than the F2-LNPs; the hydro-dynamic diameter of F1-LNP was 50-55 nm, whereas the F2-LNP was approximately 47 nm, measured by DLS. Zeta potentials were similar amongst all the LNPs being in a range of -4 to -9 mV. Based on the number of dye molecules and of particles, we calculated that each LNP was loaded with 104 to 112 molecules of dye/particle (Table 7). Unfortunately, exact numbers of ligands that are present on the LNPs could
Figure 15. Structure of LNPs
(A) Appearance of final product LNPs in white light. (B). Full structure of LNP and components (drawing
credits to CEA-Leti). (C) Structural formulas of peptide ligands on the LNP surface: the functionalization 1 LNPs as representative examples. The functionalization 2 LNPs have the maleimide linker in the inversed orientation between the peptide and the PEG residue (adapted drawing).
Table 7. Physico-chemical characteristics of each LNP lots
LOT #1-F1 (functionalization1) LOT #1-F2 (functionalization2) LOT #2-F1 (functionalization1) cRGD cRAD OH cRGD cRAD OH cRGD cRAD Std Final lipid concentration 10%, 100 g/L 10%, 100 g/L 8.5%, 85 g/L 9%, 90 g/L 9%, 90 g/L 7.5%, 75 g/L 9%, 90 g/L 9%, 90 g/L 10%, 100 g/L Maleimide quantity (initial)
748 nmol 748 nmol 374 nmol 748 nmol 748 nmol 374 nmol 748 nmol 748 nmol -
Peptide quantity (theory)
750 nmol 750 nmol - 750 nmol 750 nmol - 750 nmol 750 nmol -
Particle concentration (/mL)
1.5E+15 1.5E+15 1.25E+15 1.35E+15 1.35E+15 1.1E+15 1.31E+15 1.31E+15 1.5E+15
Fluorophore concentration 260 µM 260 µM 230 µM 250 µM 250 µM 200 µM 240 µM 240 µM 270 µM Number of dyes/particle (approx.) 104 104 111 112 112 109 108 108 110 Size (nm) 54.28 53.67 53.51 47.46 47.25 47.73 55.35 55.52 49.49 Polydispersity index 0.187 0.125 0.142 0.116 0.098 0.122 0.126 0.130 0.126 Zeta potential (mV) -9.43 -8.60 -7.75 -7.25 -6.94 -6.84 -4.53 -4.87 -4.16
Figure 16. Normalized absorbance and emission spectra of F1-LNPs
Measured in 1X PBS. The emission spectrum was measured with excitation at 790nm.
Figure 17. Fluorescence intensity curve of three F1-LNPs measured under the NIR in
Specific binding of cRGD-LNPs
αvβ αvβ
β
Figure 18. Cell attachment assay using HEK293-β3 cells in the presence of LNPs
The cells were incubated with different LNPs for 15 min at 37°C. Mean values are connected in lines. (n=6, repeated twice in triplicate)
Targeting
tumors
with
cyclic
RGD-conjugated
lipid
nanoparticles
loaded
with
an
IR780
NIR
dye:
In
vitro
and
in
vivo
evaluation
JungyoonChoia,EmilieRustiqueb,MaximeHenrya,MélanieGuidettia, VéroniqueJosseranda,LucieSanceya,Jérôme Boutetb,Jean-Luc Colla,*
aInstituteforAdvancedBiosciences,INSERMU1209,CNRSUMR5309,UniversitéGrenobleAlpes,38000Grenoble,France b
CEA,LETI,MINATECCampus,TechnologiesforBiologyandHealthDivision,17AvenuedesMartyrs,38054Grenoble,France
ARTICLE INFO Articlehistory:
Received26January2017
Receivedinrevisedform27February2017 Accepted6March2017 Availableonlinexxx Keywords: Cancerimaging Theranostic Integrins RGD Nanoemulsion NIRimaging ABSTRACT
Likeseveral50nm-largenanocarriers,lipidnanoparticles(LNPs)canpassivelyaccumulateintumors throughtheEnhancedPermeabilityandRetention(EPR)effect.Inthisstudy,wedevelopedPEGylated LNPsloadedwithIR780iodideasacontrastagentforNIRfluorescenceimagingandmodifiedthemwith cyclicRGDpeptidesinordertotargetintegrinavb3.Wedemonstrateaspecifictargetingofthereceptor withcRGD-LNPsbutnotwithcRAD-LNPandstandardLNPusingHEK293(b3),HEK293(b3)-avRFP,DU145 andPC3celllines.WealsodemonstratethatcRGD-LNPsbindtoavb3,interferewithcelladhesionto vitronectinandco-internalizewithavb3withinonehour.Wetheninvestigatedtheirbiodistributionand tumortargetinginmicebearingDU145orM21tumors.Weobservednosignificantdifferencesbetween cRGD-LNPandthenon-targetedonesregardingtheirbiodistributionandaccumulation/retentionin tumors.ThissuggestedthatdespiteanefficientformulationofthecRGD-LNPs,thecRGD-mediated targetingwasnotincreasingthetotalamountofLNPthatcanalreadyaccumulatepassivelyinthe subcutaneoustumorsviatheEPReffect.
©2017ElsevierB.V.Allrightsreserved.
1.Introduction
Therehasbeenalotofefforttogeneratenanovectorssuitable forthedeliveryofcontrastagentsanddrugs.Severaldrugdelivery systems are approved by the FDA, but only a few targeted formulations are currentlytested in humans (Shi et al., 2016). Amongthedifferentformulations,LipidNanoparticles(LNPs)are promising in particularbecause they areshowing: 1) excellent biocompatibility, 2) long circulation time in blood, 3) passive tumoraccumulationviatheenhancedpermeabilityandretention (EPR)effect,4)capacityofloadinglargeamountofpoorlysoluble drugs as wellas imaging probes, 5) possible chemical surface modificationsthatmayhelpforactivespecificreceptortargeting (Goutayeretal.,2010;Gravieretal.,2011;Hirsjarvietal.,2013; Jacquart et al., 2013; Merian et al., 2015). LNPs are made of a
lipophilic core encapsulated by a PEGylated mono-layer of phospholipids.
LNPsloadedwithalipophilic-derivativeofIR780dye(addition of a C18chain toIR780iodide dye)arestable for 6months in phosphatebuffersalineat4Cinthedark.Theypresenta long-lastingEPR-mediatedretentioninsubcutaneoustumors(Jacquart etal.,2013)duetothepresenceofaPEG-coating.Unfortunately, PEGylationisalsoloweringtheirinteractionwithcellmembranes andthisisfinallyending-upwithalowefficiencyofintracellular deliveryofthecargo(BozzutoandMolinari,2015).Weexpectthat thepresenceofatargetingligandshouldimprovethisfinalstep. The cyclic Arg-Gly-Asp (RGD) peptide known to target preferentially the avb3 integrin (Dechantsreiter et al., 1999;
Haubneret al.,1996)hasbeenlargelyused forthis purposeas recentlyreviewedinArosioetal.(ArosioandCasagrande,2016). RGDwasusedinparticularforapplicationssuchascancerimaging, radiotherapy, phototherapy and nanomedicine(Abdollahi et al., 2005;Chenetal.,2012;Kunjachanetal.,2015;LiuandWang,2013; Rayetal.,2011;Shuhendleretal.,2012;Wangetal.,2014;Yanetal., 2016;Yuanetal.,2015).Ourgroupalsodescribed cyclic-RGDfK-decoratedLNPsloadedwithDiD,adyeadaptedtoNear-infraredin * Correspondingauthorat:CentredeRechercheUGA-INSERMU1209,CNRSUMR
5309,InstituteForAdvancedBiosciences,SiteSanté,AlléedesAlpes,LaTronche 38700,France.
E-mailaddress:Jean-luc.coll@univ-grenoble-alpes.fr(J.-L.Coll).
http://dx.doi.org/10.1016/j.ijpharm.2017.03.007 0378-5173/©2017ElsevierB.V.Allrightsreserved.
InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
ContentslistsavailableatScienceDirect
International
Journal
of
Pharmaceutics
vivoimaging(Goutayeretal.,2010).ThiscRGD-LNPshowedagood tumor accumulation and longer retention in tumors. We then generatedIndocyanineGreen(ICG)-containingLNPbecauseICGis approvedfor humaninjection, butICG was leakingout rapidly fromtheLNPinvivo(Gravieretal.,2011;Merianetal.,2015).We thusfinallygeneratedstableand brightLNPscontaining IR780-Oleylmodifieddye, which turnedouttobesuitable forinvivo imaging(Jacquartetal.,2013).
In the present study, we developed the IR780-Oleyl-LNPs targeted by cyclic RGD. We demonstrate that cRGD-LNP are functionalbecausetheybindtointegrinavb3andareinternalized. However,despitethesesatisfyingresultsinvitro,thepresenceof cRGD targeting does not augment the global accumulation of targetedversusnon-targetedNP.
2.Materialsandmethods 2.1.Materials
SuppocireNBTM(semi-syntheticsC12-C18saturated triglycer-ide)waspurchasedfromGattefossé(Saint-Priest,France);Lipoid S75(soybeanlecithin at>75%phosphatidylcholine)fromLipoid (Ludwigshafen, Germany); MyrjTM S40 (polyethylene glycol 40 stearate); Super refined soybean oil from Croda Uniqema (Chocques, France); Cyclic RGD and cyclic RAD from Pepscan Presto(Lelystad,Netherland).IR780iodidedye ((2-[2-[2-Chloro-3- [(1,3-dihydro-3,3-dimethyl-1-propyl-2H-indol-2-ylidene)ethyli- dene]-1-cyclohexen-1-yl]ethenyl]-3,3-dimethyl-1-propylindo-liumiodide) and otherchemicalsfor theLNPpreparation were purchasedfrom Sigma-Aldrich(Saint-Quentin-Fallavier, France). CyclicRGDstandsforc[DfK(Mal)RG]andcRADforc[DfK(Mal)RA] withf=DphenylalanineandMal=3MaleimidePropionicAcid.
CellculturemediaandbufferswerepurchasedfromGibco(Life technologies, Paisley, UK); Bovine serum albumin (BSA) from Interchim(Montluçon,France);Puromycindihydrochloridefrom Sigma-Aldrich(USA);Recombinanthumanvitronectin(rhVTN-N, truncated)and Geneticin(G418 sulfate) fromLifeTechnologies. Unless otherwise specified, phosphate buffer saline (PBS) and Hanks'BalancedSalt Solution(HBSS) arealways supplemented with1mMCaCl2and1mMMgCl2.
2.2.Targetedandnon-targetedLNPpreparation
The formulation and composition of LNPs(previously called Lipimage815) have been already described in details (Jacquart etal.,2013).Briefly,theIR780-Oleyldyewassynthesizedstarting fromthecommercialIR780iodidebyaddingaC18lipidchain.A mixtureofoil,SuppocireNBTMandlecithinwasprepared,then,the dyewasintroducedinit.AnaqueoussolutionwithamixofMyrjTM S40(PEG-40Stearate)andSACOHN-PEG100-SPDP(3%ofthetotality ofPEG)in154mMNaClbufferwasaddedtotheoilphase.Fifty nm-large lipid particles were formed through the emulsification process during a 5 min-sonication using a VCX750 Ultrasonic processor(poweroutput190W,3mmprobediameter,Sonics).A solutionofDTTwasaddedtotheLNPsandagitatedfor2htoobtain LNPs-SH. The LNP dispersions were dialyzed overnight at RT against1000timestheirvolumein154mMNaClbuffer(12–14KDa MWcut-offmembranes,ZelluTrans,CarlRoth,France).
Following the first step of LNP-SH synthesis, the peptide conjugation was carried out by the thiol-maleimide coupling reaction.2mLofLNP-SH(15%lipids)solutionweremixedwith 75mLofeachpeptidelinkedwithmaleimide(cyclicRGDandcyclic RAD 750nmolinDMSO).After4hofacouplingreactionatRTin the dark, the remaining SH groups were neutralized using a solutionof 1-(2-hydroxyethyl)-1H-pyrrole-2,5-dione for 30min. Anotherdialysiswascompletedovernighttopurifythe
peptide-conjugatedLNPs.ThefinalLNPdispersionswerefilteredthrougha 0.22mmMilliporemembraneforsterilizationandstoredat4Cin thedark.
2.3.CharacterizationofLNPs
LNPs’ hydrodynamic diameters, polydispersity indexes (PDI) andzetapotentialsweredeterminedinsterile0.1XPBS(noCaCl2/ MgCl2)usinga dynamic lightscattering(DLS) instrument(Zeta Sizer NanoZS, Malvern,UK), performed intriplicate.The absor-banceandfluorescencespectraweremeasuredin1XPBSusinga spectrophotometer Thermo Scientific Evolution 201 and a fluorescence spectrometer Perkin ElmerLS55, respectively. The emissionspectrumwasmeasuredwithexcitationat790nm.Each measurementwas performedintriplicate.Concentrations men-tionedforLNPsrepresenttheconcentrationsofparticles. 2.4.Celllinesandcultureconditions
HEK293(b3)cells,ahumanembryonickidneycelllineHEK293 that was stably transfected with the human b3 integrin gene (kindlyprovidedbyJ-F.Gourvest,Aventis,France),wereculturedin DMEM-GlutaMAXTMsupplementedwith10%FBSand700mg/mL Geneticin.HEK293(b3)-avRFPcells,avariantcelllineofHEK293 (b3)transfectedwiththehumanavintegringenetaggedbyRFPat itsextracellulardomain(N-terminuspart),wereculturedasthe HEK293(b3)cellsbutinthepresenceof1mg/mLPuromycin(See Supplementary information). DU145 cells, a human prostate carcinoma cell line (purchased from ATCC), and M21 cells, a humanmelanomacellline,wereculturedinDMEM-GlutaMAXTM with10%FBS.PC3cells,ahumanprostateadenocarcinomacellline (ATCC),wereculturedinDMEM-F12-GlutaMAXTMwith10%FBS. Cellsweremaintainedat37C,5%CO2inahumidifiedincubator. 2.5.Evaluationofavb3expressionbyflowcytometry
Cellsweregrownandharvestedat60%confluency.Thecells werebrieflywashedwith1XPBSandresuspendedwithtrypsin. After being centrifuged at 1200rpm for 5min, the cells were suspended in1XPBS. 5105 cells per tubeweresuspendedin 100mLcold1XPBSandmixedwith20mLofanti-avb3monoclonal antibody (PE-conjugated mouse anti-human CD51/CD61, clone 23C6,BDPharmingen)during1honiceinthedark.Afterrinsing twice they were finally suspended in 200mL cold PBS. Their fluorescenceintensitywasmeasuredusingaflowcytometer(BD AccuriC6;BDBiosciences).
2.6.Cellattachmentassaywithvitronectin
A96-wellplate(NuncMaxisorpELISAplates)wascoatedwith 5mg/mLvitronectin(50mLperwell)overnightat4C.Then,the wellsweredrainedandincubatedwith3%(w/v)BSA(50mLper well)foranhourat37C.Thenegativecontrolwellswerecoated with3%BSAonly.TheplatewasrinsedoncewithHBSS.50mLof25 000cellsand50mLoftheLNPperwellwereaddedinthewellsand incubatedat37Cfor15min(HEK293(b3)andDU145)or30min (PC3).Whenneeded,thecellsandLNPsweredilutedinHBSS.The platesweredrainedandgentlywashedtwicewith200mLof1X HBSSperwell.Thesurface-boundcellswerefixedwithethanolfor 15minatRTanddriedatRTor37C. Thefixed cellswerethen stainedwith1%methyleneblue(50mLperwell)for15minatRT, rinsedinrunningwateranddried.Thebluedyewasfinallyeluted in 0.1N HCl (50mL per well)and theplateswerescanned ata 620nmabsorbancewavelengthusingaBeckmanCoulterAD340 platereader.Allthemeasurementswereperformedintriplicate (orquadruplicate)andrepeatedtwice.
2 J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
2.7.Microscopy
2.7.1.Brightfieldmicroscopy
HEK293(b3)cells(5104cells/well)wereseededinan8-well chambered coverglass (Nunc Lab-Tek, Thermo FisherScientific) and grown overnight in DMEM with 10% FBS and 700mg/mL Geneticinat37C.Thecellswerewashedoncewithpre-warmed 1XPBSandthechamberedcoverglasswasmovedtoamicroscope (Zeiss AxioObserver Z1) equipped with a humidified culture chamber.Afterbeing placed onthe microscope,the cells were treatedwiththeLNPsdilutedto45or4.5nMinHBSS.Then,3%(v/ v)FBSwereshortlyaddedtoeachwell.Theimagesweretaken every5minstartingfromthetreatmenttimepoint(T0). 2.7.2.Confocallaserscanningmicroscopy
HEK29(b3)-avRFPcells wereseeded(at5104 cells/500mL/ well)in a 4-wellchambered coverglass(Nunc Lab-Tek, Thermo Fisher Scientific) and kept overnight in DMEM with 10% FBS, 700mg/mLGeneticinand1mg/mLPuromycin at37C.Thecells werewashedoncewithpre-warmedPBSandtreatedwith500mL of different LNPs diluted at 45 or 4.5nM in HBSS. After an incubationtimeof5,10,30or60min,thecellsweregentlywashed once with HBSSand fixed with 4% PFA containing 1mg/mL of Hoechst33342for15minatRTinthedark.Thenthecellswere washed and stored in 1X PBS and observed using a confocal microscopeLSM510NLOMETAFLIM(ZeissAxiovert200M).A63 oil immersion objective was used. The Hoechst-stained nuclei
wereexcitedwitha720nmbiphotonlaseranddetectedwitha 390–465nmfilter.TheRFPswereexcitedwitha543nmlaserand signalswerecollectedwitha565–615nmfilter.
Fortheinternalizationstudy,theHEK293(b3)-avRFPcellswere seededinthe4-wellchambercoverglassandincubatedovernight. Theincubation,fixation,andHoechst-stainingwereasdescribed above.Thecellswerethenobservedusingaconfocalmicroscope LSM710 NLO(ZeissAxioObserverZ1)witha 63xoil immersion objective.TheLNPsignalwasexcitedusinga633nmlaser,setupat 50%ofitsmaximumintensityanda636–758nmdetectionfilter was used. The Hoechst was excited with a 405nm laser and detectedwitha415–502nmfilter.TheRFPwasexcitedwitha560 nm laser and detected with a 563–631nm filter. The scanned imageswerepseudo-coloredinbluefornuclei,ingreenforavRFP andinredfortheLNP.
2.8.Humantumorxenograftmice
Allanimalproceduresareincompliancewiththeguidelinesof theEuropeanUnion(regulationn86/609),takenintheFrenchlaw (decree87/848)regulatinganimalexperimentation.Alleffortsare madetominimizeanimalsufferingandtoreducethenumberof animalsused.Allanimalmanipulationsareperformedwithsterile techniquesandareapprovedbytheethicalcommitteeofGrenoble fortheuseofanimalresearch(France).Subcutaneoustumormice bearingDU145orM21tumorwereusedforaninvivoevaluationof theLNPs.EightweeksoldmaleNMRInudemicewerepurchased
Fig.1.LNPcomposition.(i)SchematicrepresentationoftheLNPs.(ii–iv)Chemicalstructureofthetargetedornon-targetedsurfactantsusedinthedifferentLNPs.Thecyclic RGDfK(iii,targeting)andcRADfK(iv,control)peptideswereconjugatedonthePEGbythemaleimide-thiolcouplingreaction.Theillustrationwasadaptedfrom(Merianetal., 2015).
J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx 3
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
fromJanvier,LeGenest-Saint-Isle,France.Micewereanesthetized (isoflurane/oxygen3.5%–4%forinductionand1.5%–2%thereafter, CSP,Cournon,France)andweresubcutaneouslyinjectedwith107 ofDU145cellsor4106ofM21cellspermouseintherighttrunk. Whenthetumorsreachedavolumebetween200and500mm3, themicewereusedforLNPinjections.Tumorgrowthstook70days fortheDU145tumorsand43daysfortheM21tumorsinaverage. 2.9.NIRinvivoimagingafterinjectionsofLNPs
Isoflurane-anesthetizedmicewereinjectedinthetailveinwith 200mLof 450nM LNPs(equal to10nmol offluorophores).NIR fluorescenceimageswereacquired1,3,5and24hafterinjection withaFluobeam1800device fromFluoptics(Grenoble,France) equipped with a 780nm excitation laser and an LP820nm detectionfilter.Themiceweresacrificedafterthelastacquisition andfluorescencepresentintheextractedorgansandserumwere also acquired. Major organs were cryo-conserved for ex vivo studies.ImageanalysiswasperformedusingtheWasabisoftware (Hamamatsu,Massy,France).
2.10.Statisticalanalysis
Statisticalanalysisofthedatawasperformedbyusingtwo-way ANOVAwithBonferronipost-tests.Adifferenceofpvalue<0.05 wasconsideredstatisticallysignificant.Allthedataareshownas meanSD.
3.Results
3.1.CharacteristicsofLNPs
Theouterlayercompositionandchemicalstructureof peptide-coated LNPs are illustrated in Fig. 1. LNPs are made of a phospholipidmonolayerthatdelineatestheiroilycorecontaining thefluorophores, and an outershell of PEG surfactant withor withoutthepeptideligandfortumortargeting.ThecyclicRGDfK (avb3 targetingpeptide)orcyclicRADfK(anegativecontrolof cRGDinwhichtheglycinewasreplacedbyanalanine,Fig.1-iv) peptideswerelinkedonthePEGsbyamaleimide-thiolcoupling reaction(Fig.1-iii).
The two peptide-conjugated LNPs showed a slightly larger diameterof55nmascomparedtothe50nmofthestandardLNP withoutpeptide.Thepolydispersityindexes(PDI)ofthethreeLNPs werelowerthan0.13,indicatingthattheyaremonodispersedand donotformaggregates.Thezetapotentialswerekeptclosedto neutrality around 4mV for each LNPs without significant differencedue to the presence of the peptides (Table 1). Such zeta-potentialsareavoidingnon-specificelectrostaticinteractions. DuetotheinclusionofIR780-Oleyl,themaximumabsorbance wasobservedfor wavelengthscenteredat790–792nmandthe maximumemissionbandwasdetected at810–812nm (Supple-mentaryFig.1).BasedonthemeasuredconcentrationsofIR780 and of LNPs, we calculated that each nanoparticle contained approximately108moleculesofdye(Table1).
3.2.CRGD-LNPspreventcellbindingonvitronectin
We investigated ifthe presence of cRGD could prevent the binding ofthreedifferentcelllinesonvitronectin,a“natural”proteinligand Table1
Physico-chemicalcharacteristicsoftheLNPs(triplicatevaluesSEM). cRGD-LNP cRAD-LNP LNP
Size(nm) 551 551 491
Polydispersityindex,PDI 0.1260.027 0.1300.008 0.1260.021 Zetapotential(mV) 4.530.38 4.870.39 4.130.81 [Approximately108fluorophores(1.810 22
mol)pernanoparticle.]
Fig.2. Integrin-specificbinding ofcRGD-LNPs.(A) Thepresenceof the avb3
integrinwasdetectedbyflowcytometryonHEK293(b3),DU145andPC3cells.Grey:
antibody-labeledcells;black:controlcells(noantibody).(B)Thecapacityofcellsto sticktothevitronectincoatingwasmeasuredforeachcellline(HEK293(b3),DU145
andPC3)inthepresenceofdifferentLNPsat4.5,9,45and90nM.Thenumberof adherentcellswasmeasuredbymethylenebluestaining.MeanSD,n=6-8.The valueswerenormalizedwiththoseofthestandardLNP.Asterisks:**=p<0.01, ***=p<0.001,comparedtothestdLNP.(Forinterpretationofthereferencesto colourinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle.) 4 J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
recognizedbytheavb3integrin.Asestablishedbyflowcytometry, HEK293(b3)andDU145expresstheintegrinwhilePC3cellsdonot (Fig.2A).Cellsweremixedinsolutionwith4.5,9.0,45.0and90.0nM ofcRGD-,cRAD-orStandard-LNPsandthenaddedon vitronectin-coatedmultiwellplates(Fig.2B).Thetreatmentofthecellswith cRGD-LNPonlydecreasedtheadherenceofHEK293(b3)andDU145 cells. HEK293(b3) cell adhesion was strongly decreased in the presenceof45nMcRGD-LNP(p<0.001)butwasalready statisti-callydecreasedwith9nMcRGD-LNPs(p<0.01).BecauseDU145 cells are expressing lower amounts of integrins they are more sensitivetothepresenceofcRGDand4.5nMcRGD-LNP(p<0.001) weresufficienttoimpactstronglyontheiradhesion.Incontrast,the adherenceofavb3-negativePC3cellswasnotparticularlyaffected bycRGD-LNPs(Fig.2B).
3.3.cRGD-mediateddetachmentofavb3-positivecells
Having confirmed that cRGD-LNPs canbind ontothe freely-accessibleavb3receptorwhenthecellsareinsuspension,wenext exploredtheircapacitytoreachandinterferewithoccupiedavb3 -integrinsoncells grownina monolayer.ConfluentHEK293(b3) cellsweretreatedwith4.5nMor45nMcRGD,cRAD,orstandard LNPinasolutioncontaining3%FBS.Thecellswereobservedfor morethan3honamicroscopeequippedwithaculturechamber (Fig. 3). In the presence of 45nM cRGD-LNP the cells started retractingduringthefirst5minandgraduallylostadhesionover time.Whenusedat4.5nM,cRGD-LNPstillaffectedcelladhesion, althoughlessseverely,andthiseffectwasonlytransientandthe cell–cellcontactswererestoredafter3h.ThecRADandstandard LNPsdidnotaffectcellshapewhentestedat45and4.5nMLNP (Fig.3andSupplementaryFig.2).
Fig.3. ConfluentHEK293(b3)cellstreatedwithLNPsat45or4.5nM.Thecellswereobservedforaperiodof3hinrealtimeunderthemicroscopeequippedwithaculture
systemforlivecellimaging.Scalebar=100mm.
J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx 5
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
WethenusedHEK293(b3)-avRFPcellsexpressingintegrinav -labeledwithred fluorescentprotein(RFP) (butpseudo colorin greeninthefigures).Inthiscellline,theavb3integrinwaslocated mainlyontheplasmamembraneandenrichedinareaofcell–cell contacts.Cellretractionwasclearlyvisibleonehourafteraddition of45nMcRGD-LNP(Fig.4Arightpanel).Ascanbeseenonthe bottompanel,cellshrinkage,lossoffocalandofcell–celljunctions occurredvery rapidly and were visible as soon as 5min after additionofthecRGD-LNPs(Fig.4B).
ThisindicatedthatcRGD-LNPswerefunctionalandcouldbind toavb3evenwhentheintegrinwasalreadyengaged.
3.4.InternalizationofcRGD-LNP
UsingtheHEK293(b3)-avRFPcells,wefurtherinvestigatedthe internalizationofcRGD-LNPs.Pleasenotethatthe“red”signalof theavRFPproteinispresentedingreen.IncontrastwiththeLNPs andcRAD-LNPswhichdidnotprovideasignificantcelllabeling, anintense labeling of the cell membranes was observedwith cRGD-LNPsat45nM(Fig.5).Similarresultswerealsoobtained whenthe different LNPs were used at 4.5nM (Supplementary Fig.3). Importantly,intracellular vesicles(probably endosome) werealsopositively stainedby boththe cRGD-LNPand bythe integrin-fusion protein after 60min of incubation (Fig. 5, first row), but not after 10 or 30min (Supplementary Fig. 4), suggestingthat the internalizationwas mediatedby the avb3 integrin via a slow process that necessitated between 30 and 60min.Thevesiclesco-stainedbycRGD-LNPs/integrinavb3were alsofoundwhenthecellswerewashed15minafteradditionofthe cRGD-LNPs(Fig.5,secondrow).Intheseconditions,theplasma membranes were not strongly labeled: the integrin/cRGD-LNP complexesinitiallylocatedtothemembranewereinternalized, whilethecell mediumcontaining thecRGD-LNPwas removed, preventinganybindingaftertherinsing.
3.5.Invivoevaluation
InordertodemonstratethecRGD-specifictargetingoftumors invivo,wecomparedthebiodistributionofthethreeLNPsinmice
bearingDU145orM21tumors.M21isahumanmelanomacellline that stronglyexpresses theavb3 integrin((Felding-Habermann
etal.,1992),SupplementaryFig.5).Afteri.v.injectionsof200mLof 450nMLNPs,wefollowedthedistributionoftheNIR-labeledLNPs in real time using a NIR camera. Unexpectedly, no major differencesinthetumoraccumulationwerefoundbetweenthe cRGD-,cRAD-andstandardLNPsinbothDU145orM21tumors (Fig.6andSupplementaryFig.6).Thiswasconfirmedex-vivoafter sacrifice of the mice and excision of the tumors (Fig. 7 and SupplementaryFig.7).Aswell,nosignificantdifferencesof EPR-mediated accumulation were detected when the mice bearing M21tumorswerefollowedfor4–5daysaftertheinjections.
Thus,the3typesLNPsaccumulatepassivelyviatheEPReffectin the2tumortypes.However,thepresenceofcRGDdoesnotprovide a significant augmentation of the number of LNPs that can be detectedmacroscopicallyinthetumors.
4.Discussion
Themainobjectiveofthisstudywastodemonstrateaspecific bindingandinternalizationofthecRGD-LNPinordertogenerate tumor-targetednanoparticlesfortheranosticpurpose.
WefoundthatcRGD-LNPcanbindandinterferefunctionally with integrin avb3 in vitro. This indicated that the chemical couplingofcRGDonthePEGylatedlipidnanocapsuleswascorrect andresultedinanefficientrecognitionoftheintegrinreceptorson thecellsurface.cRGD-LNPswereabletobindto“free”integrins presentonthesurfaceofthecellsinsuspensionaswellasonthose alreadyboundtovitronectin.Thislatereffectcausedanefficient shrinkingofadherentcellsduetotheinhibitionofcell–celland cell-matrix contacts. This effect was strong enough to detach integrinpositive cells when cRGD-LNPs were used at a 45nM concentration.Atalowerconcentration(4.5nM),thedetachment was transient and cell adhesionand spreading wererecovered after3h.AsexpectedthiseffectwasmorepronouncedonDU145 cellsexpressinglowlevelsofintegrinsthanonHEK293(b3)cells whicharestronglyavb3positive.
Ofnote,theHEK293(b3)celllineisintegrinb5negativewhile DU145andPC3areknowntoexpresslowtomoderatelevelsof Fig.4.ConfocallasermicroscopyofHEK293(b3)-avRFPcellsinthepresenceof45nMLNPsat37C.Blue:nuclei,green:avRFP.Scalebar=20mm.(A)Thecellsweretreated
withdifferentLNPsandvisualizedafter60minofincubation.NoLNP:treatedwithHBSS.(B)ThecellswereincubatedwithcRGD-LNPsfor5,10,30or60min. 6 J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
avb5anda5b1 andnegligiblelevelsofaIIbb3(Sutherlandetal.,
2012).AllthesedifferentintegrinscanrecognizetheRGDmotif, buttheimportantspecificityofcyclicRGDfKtowardavb3isthus largelypredominant(Mas-Morunoetal.,2010).
Integrin internalization can occur via macro-pinocytosis, clathrin-dependentorclathrin-independentendocytosis( Bridge-wateretal.,2012).Wepreviouslydemonstratedthat internaliza-tion of a tetravalent cRGD presenting scaffold is mediated via clathrin-mediated endocytosis (Sancey et al., 2009). Other integrin-targeted NPs such as trimethyl chitosan-based NPs labeled with FQSIYPpIK (an avb3 targeting peptide) are using clathrin-mediated endocytosis (Xu et al., 2016). In the present study, we also observed that only the cRGD-LNPs are actively endocytosedandcanbeseentraffickingwithinav-RFPand IR780-Oleylco-labeledvesicles.Inaddition,aweakeranddiffuseIR780
fluorescentsignalwasalsobuildingupovertimeinthecytoplasm and thisdiffusesignalwasalsodetected inthepresenceofthe negativecontrolscRAD-andnon-targeted-LNPs.Thismayindicate that while only cRGD-LNPs can enter specifically via a cRGD-integrin-dependent pathway(s), all 3 types LNPscan fuse non-specifically withthelipidsofthecellmembranesand canthen releasetheircontentdirectlytothecytoplasm.Thiscouldhavean importantimpactforLNP-mediateddrugdelivery.Inparticular,we arecurrentlyinvestigatingthepossiblecytotoxiceffectgenerated underan800nmlight-activationoftheIR780dyesincethisdyeis known to generate toxic reactive oxygen species usable for photodynamictherapy.
Invivo,nospecificaccumulationwasnoticedin theintegrin positive DU145 orM21 subcutaneoustumors after intravenous injection of the cRGD-LNPs versus cRAD- or Standard-LNPs. Fig.5.Co-localizationofcRGD-LNPwithavb3inHEK293(b3)-avRFPcells.Thecellswereincubatedwith45nMLNPsforanhourat37C.Blue:nuclei,red:LNPs,green:
avRFP.Scalebar=10mm.(Firstrow)ThecellswereincubatedwithcRGD-LNPsfor60mincontinuously.(Fromthesecondtothefourthrow)ThecellsweretreatedwithLNPs
and,15minlater,theLNPsolutionwasreplacedbyfreshHBSS.Theincubationat37Cwascontinuedforanother45min.(Bottomrow)ThecellswereincubatedwithHBSS
bufferonly.(Forinterpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle.)
J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx 7
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
Similar fluorescence intensities as well as similar kinetics of accumulationinthetumorsweremeasuredwiththe3typesof LNPs.This suggested that the functionalizationof theLNPs by cRGDmotifsdidnotgenerateasignificantlyaugmentedcaptureor retentionofthetargetedversusnot-targetedLNPsinthetumor mass.Thisisinagreementwithourformerworkinwhichweused cRGD-targetedDID-loadednanoemulsions.Inthispreviouswork, wehadaverymodestaugmentationofcRGD-dependenttargeting invivowhenusingHEK293(b3)tumorsbutnotTSa/pc’s(Goutayer
et al., 2010). Actually, HEK293(b3) cells are very strongly expressing the integrin while TSa/pc are weakly expressing it. On the other hand, we classified HEK293(b3) subcutaneous tumors as “EPR-deficient” because an intravenous injection of lipid nanoemulsions generated a Tumor/Skin (T/S) ratio of 1.5, while TSa/Pc were found EPR-positive with a T/S ratio of 3.0
(Karageorgisetal.,2016).Accordingtothisscheme,DU145and M21tumorscanthusbeclassifiedasEPR-positivebasedontheirT/ Sratioof3,aswellasintegrinpositive.Thissuggeststhatactive, cRGD-mediated,targetingisnotaugmentingtheaccumulationof the nanoemulsions when the EPR effect is present. However, several groups including ours (Dufort et al., 2011; Karageorgis et al., 2017) have shown that RGD-vectorization is efficiently inducingatherapeuticeffect(forveryrecentexamples,see:(Fang etal.,2017;Zhuetal.,2016),andarecentreviewbyZhongetal. (Zhongetal.,2014).ThissuggeststhatcRGD-activetargetingmay beimportantattheverylaststep,i.e.todeliverthecargomore efficientlyintothecytoplasmofintegrinpositivecells.Wearenow activelytryingtovalidatethishypothesisthatmaybeimportant for therapeutic applications inparticular when thetoxic agent cannotcrosstheplasmamembraneonitsown.
Fig.6. NIRinvivoimagingofDU145subcutaneoustumorinmiceafterinjectionsoftheLNPs.NIRfluorescenceimagesareshownforeachmouseingreylevelswithMin-Max valuesof:0-46912.Theimagesweretaken24hafterinjectionsofLNPs.L:liver,T:tumor.GraphsbelowtheimagesshowNIRfluorescencelevelsontumorsandskin.X-axis: timeafterinjection(hours),LeftY-axis:RLU/pixel/10ms,RightY-axis:tumor/skinratio.
8 J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
Acknowledgements
JungyoonChoiwassupportedbytheLigueNationalecontrele Cancer(GB/MA/SC-12742),theFondationARCpourlaRecherche surle Cancer(DOC20150602724)and byCampusFrance (refer-ence: 796786A). This work was supported by a public grant overseenbytheFrenchNationalResearchAgency(ANR)aspartof the “Investissements d’Avenir” program (reference: ANR-10-NANO-01),but alsobytheInstitutNationaldelaSantéetdela Recherche Médicale (INSERM U1209) and by the France Live Imaging program. The authors acknowledge the team of the “OPTIMALsmallanimalopticalimagingfacility”inourinstitute. AppendixA.Supplementarydata
Supplementarydataassociatedwiththisarticlecanbefound,in the online version, at http://dx.doi.org/10.1016/j. ijpharm.2017.03.007.
References
Abdollahi,A.,Griggs,D.W.,Zieher,H.,Roth,A.,Lipson,K.E.,Saffrich,R.,Grone,H.J., Hallahan,D.E.,Reisfeld,R.A.,Debus,J.,Niethammer,A.G.,Huber,P.E.,2005. Inhibitionofalpha(v)beta3integrinsurvivalsignalingenhancesantiangiogenic andantitumoreffectsofradiotherapy.Clin.CancerRes.11,6270–6279. Arosio,D.,Casagrande,C.,2016.Advancementinintegrinfacilitateddrugdelivery.
Adv.DrugDeliv.Rev.97,111–143.
Bozzuto,G.,Molinari,A.,2015.Liposomesasnanomedicaldevices.Int.J.Nanomed. 10,975–999.
Bridgewater,R.E.,Norman,J.C.,Caswell,P.T.,2012.Integrintraffickingataglance.J. CellSci.125,3695–3701.
Chen,Z.,Deng,J.,Zhao,Y.,Tao,T.,2012.CyclicRGDpeptide-modifiedliposomaldrug deliverysystem:enhancedcellularuptakeinvitroandimproved
pharmacokineticsinrats.Int.J.Nanomed.7,3803–3811.
Dechantsreiter,M.A.,Planker,E.,Mathä,B.,Lohof,E.,Hölzemann,G.,Jonczyk,A., Goodman,S.L.,Kessler,H.,1999.N-MethylatedcyclicRGDpeptidesashighly activeandselectiveaVb3integrinantagonists.J.Med.Chem.42,3033–3040.
Dufort,S.,Sancey,L.,Hurbin,A.,Foillard,S.,Boturyn,D.,Dumy,P.,Coll,J.L.,2011. Targeteddeliveryofaproapoptoticpeptidetotumorsinvivo.J.DrugTarget.19, 582–588.
Fang,Y.,Jiang,Y.,Zou,Y.,Meng,F.,Zhang,J.,Deng,C.,Sun,H.,Zhong,Z.,2017. TargetedgliomachemotherapybycyclicRGDpeptide-functionalizedreversibly core-crosslinkedmultifunctionalpoly(ethylene glycol)-b-poly(epsilon-caprolactone)micelles.ActaBiomater.50,396–406.
Felding-Habermann,B.,Mueller,B.M.,Romerdahl,C.A.,Cheresh,D.A.,1992. InvolvementofintegrinalphaVgeneexpressioninhumanmelanoma tumorigenicity.J.Clin.Invest.89,2018–2022.
Goutayer,M.,Dufort,S.,Josserand,V.,Royere,A.,Heinrich,E.,Vinet,F.,Bibette,J., Coll,J.L.,Texier,I.,2010.Tumortargetingoffunctionalizedlipidnanoparticles: assessmentbyinvivofluorescenceimaging.Eur.J.Pharm.Biopharm.75,137– 147.
Gravier,J.,Navarro,F.P.,Delmas,T.,Mittler,F.,Couffin,A.C.,Vinet,F.,Texier,I.,2011. Lipidots:competitiveorganicalternativetoquantumdotsforinvivo fluorescenceimaging.J.Biomed.Opt.16,096013.
Haubner,R.,Gratias,R.,Diefenbach,B.,Goodman,S.L.,Jonczyk,A.,Kessler,H.,1996. StructuralandfunctionalaspectsofRGD-Containingcyclicpentapeptidesas highlypotentandselectiveintegrinaVb3antagonists.J.Am.Chem.Soc.118, 7461–7472.
Hirsjarvi,S.,Dufort,S.,Gravier,J.,Texier,I.,Yan,Q.,Bibette,J.,Sancey,L.,Josserand,V., Passirani,C.,Benoit,J.P.,Coll,J.L.,2013.Influenceofsize,surfacecoatingandfine chemicalcompositionontheinvitroreactivityandinvivobiodistributionof lipidnanocapsulesversuslipidnanoemulsionsincancermodels.Nanomed.: Nanotechnol.Biol.Med.9,375–387.
Jacquart,A.,Keramidas,M.,Vollaire,J.,Boisgard,R.,Pottier,G.,Rustique,E.,Mittler, F.,Navarro,F.P.,Boutet,J.,Coll,J.L.,Texier,I.,2013.LipImage815:novel dye-loadedlipidnanoparticlesforlong-termandsensitiveinvivonear-infrared fluorescenceimaging.J.Biomed.Opt.18,101311.
Karageorgis,A.,Dufort,S.,Sancey,L.,Henry,M.,Hirsjarvi,S.,Passirani,C.,Benoit,J.P., Gravier,J.,Texier,I.,Montigon,O.,Benmerad,M.,Siroux,V.,Barbier,E.L.,Coll,J.L., 2016.AnMRI-basedclassificationschemetopredictpassiveaccessof5to 50-nmlargenanoparticlestotumors.Sci.Rep.6,21417.
Karageorgis,A.,Claron,M.,Juge,R.,Aspord,C.,Thoreau,F.,Leloup,C.,Kucharczak,J., Plumas,J.,Henry,M.,Hurbin,A.,Verdie,P.,Martinez,J.,Subra,G.,Dumy,P., Boturyn,D.,Aouacheria,A.,Coll,J.L.,2017.Systemicdeliveryoftumor-Targeted bax-derivedmembrane-activepeptidesforthetreatmentofmelanomatumors inahumanizedSCIDmousemodel.Mol.Ther.25,534–546.
Kunjachan,S.,Detappe,A.,Kumar,R.,Ireland,T.,Cameron,L.,Biancur,D.E., Motto-Ros,V.,Sancey,L.,Sridhar,S.,Makrigiorgos,G.M.,Berbeco,R.I.,2015. Nanoparticlemediatedtumorvasculardisruption:anovelstrategyinradiation therapy.NanoLett.15,7488–7496.
Liu,Z.,Wang,F.,2013.DevelopmentofRGD-basedradiotracersfortumorimaging andtherapy:translatingfrombenchtobedside.Curr.Mol.Med.13,1487–1505. Mas-Moruno,C.,Rechenmacher,F.,Kessler,H.,2010.Cilengitide:thefirst
anti-angiogenicsmallmoleculedrugcandidatedesign,synthesisandclinical evaluation.AnticancerAgentsMed.Chem.10,753–768.
Merian,J.,Boisgard,R.,Bayle,P.A.,Bardet,M.,Tavitian,B.,Texier,I.,2015. Comparativebiodistributioninmiceofcyaninedyesloadedinlipid nanoparticles.Eur.J.Pharm.Biopharm.93,1–10.
Ray,A.,Larson,N.,Pike,D.B.,Gruner,M.,Naik,S.,Bauer,H.,Malugin,A.,Greish,K., Ghandehari,H.,2011.Comparisonofactiveandpassivetargetingofdocetaxel forprostatecancertherapybyHPMAcopolymer-RGDfKconjugates.Mol. Pharm.8,1090–1099.
Sancey,L.,Garanger,E.,Foillard,S.,Schoehn,G.,Hurbin,A.,Albiges-Rizo,C.,Boturyn, D.,Souchier,C.,Grichine,A.,Dumy,P.,Coll,J.L.,2009.Clusteringand internalizationofintegrinalphavbeta3withatetramericRGD-synthetic peptide.Mol.Ther.17,837–843.
Shi,J.,Kantoff,P.W.,Wooster,R.,Farokhzad,O.C.,2016.Cancernanomedicine: progress,challengesandopportunities.Nat.Rev.Cancer17,20–37.
Shuhendler,A.J.,Prasad,P.,Leung,M.,Rauth,A.M.,Dacosta,R.S.,Wu,X.Y.,2012.A novelsolidlipidnanoparticleformulationforactivetargetingtotumoralpha(v) beta(3)integrinreceptorsrevealscyclicRGDasadouble-edgedsword.Adv. HealthcareMater.1,600–608.
Sutherland,M.,Gordon,A.,Shnyder,S.D.,Patterson,L.H.,Sheldrake,H.M.,2012. RGD-bindingintegrinsinprostatecancer:expressionpatternsandtherapeutic prospectsagainstbonemetastasis.Cancers4,1106–1145.
Wang,F.,Chen,L.,Zhang,R.,Chen,Z.,Zhu,L.,2014.RGDpeptideconjugated liposomaldrugdeliverysystemforenhancetherapeuticefficacyintreating bonemetastasisfromprostatecancer.J.Controll.Release196,222–233. Xu,Y.,Xu,J.,Shan,W.,Liu,M.,Cui,Y.,Li,L.,Liu,C.,Huang,Y.,2016.Thetransport
mechanismofintegrinalphavbeta3receptortargetingnanoparticlesinCaco-2 cells.Int.J.Pharm.500,42–53.
Yan,F.,Wu,H.,Liu,H.,Deng,Z.,Liu,H.,Duan,W.,Liu,X.,Zheng,H.,2016.Molecular imaging-guidedphotothermal/photodynamictherapyagainsttumorby iRGD-modifiedindocyaninegreennanoparticles.J.Controll.Release224,217–228. Yuan,A.,Yang,B.,Wu,J.,Hu,Y.,Ming,X.,2015.Dendriticnanoconjugatesof
photosensitizerfortargetedphotodynamictherapy.ActaBiomater.21,63–73. Zhong,Y.,Meng,F.,Deng,C.,Zhong,Z.,2014.Ligand-directedactivetumor-targeting
polymericnanoparticlesforcancerchemotherapy.Biomacromolecules15, 1955–1969.
Zhu,Y.,Zhang,J.,Meng,F.,Deng,C.,Cheng,R.,Feijen,J.,Zhong,Z.,2016. cRGD-functionalizedreduction-sensitiveshell-sheddablebiodegradablemicelles mediateenhanceddoxorubicindeliverytohumangliomaxenograftsinvivo.J. Controll.Release233,29–38.
Fig.7. FluorescencelevelsinthedissectedorgansandplasmafromDU145 tumor-bearingmice24hafterinjectionsoftheLNPs.Meanfluorescenceintensities/pixel/ 10msofeachLNPgroupareshownwithstandarddeviation.
J.Choietal./InternationalJournalofPharmaceuticsxxx(2016)xxx–xxx 9
GModel
IJP16479No.ofPages9
Pleasecitethisarticleinpressas:J.Choi,etal.,TargetingtumorswithcyclicRGD-conjugatedlipidnanoparticlesloadedwithanIR780NIRdye: Invitroandinvivoevaluation,IntJPharmaceut(2017),http://dx.doi.org/10.1016/j.ijpharm.2017.03.007
Cell labeling by LNPs
β
Figure 19. Fluorescence signals from adherent HEK293-β3 cells after incubation with F1-LNPs, observed
by NIR camera Fluobeam®800
The cells were incubated for 15 min on ice or at 37°C with each LNP at 45 or 90 nM LNP concentrations. Mean±SD (n=6, two repeats in triplicate).
Figure 20. Fluorescence signals from HEK293-β3 cell suspensions after incubation with F1-LNPs, observed
by NIR camera Fluobeam®800
The cells were incubated for 15 min on ice or for 30 to 60 min at 37°C with each LNP at 90 nM LNP concentrations. (A) The same volume of each cell suspensions was transferred in a multiwell plate and pictured. Cell numbers were counted after the fluorescence acquisitions. (B) MFI values normalized by cell numbers are represented in a graph. Mean±SD (n=6, two repeats in triplicate).
αvβ αvβ
Internalization of cRGD-LNPs
αvβ Chemical formulas of ICG, IR780-oleyl, and DiD dyes
Figure 21. Confocal images of HEK293-β3-αvRFP cells at different times of incubation with LNPs
The cells were incubated with 45 nM LNPs at 37°C for 10, 30 and 60 min. Blue: nuclei, red: LNP, green: αvRFP; the colocalized LNP with the αvβ3 are seen in yellow. Scale bar = 20 µm.
Figure 23. Confocal images of HEK293-β3-αvRFP cells at different times following 15 min of incubation with
LNPs
The cells were incubated with 45 nM LNPs at 37°C for 15 min and the LNPs-containing solution was removed (left column). The incubation was continued for another 15 min (mid column) or 45 min (right column). Blue: nuclei, red: LNP, green: αvRFP; the colocalized LNP with αvβ3 are seen in yellow. Scale bar=20 µm.
HUVEC tube formation
αvβ
Figure 25. HUVEC Tube formation on matrigel in the presence of different LNPs
Tumor accumulation and biodistribution of LNPs in different tumors
Figure 26. Curves of mean tube lengths during 10 hours of HUVEC tube formation
Cells were incubated on matrigel in the absence or presence of different LNPs during 10 hours. Data are shown with mean±SD of triplets.
25 30 35 40 45 50 55 60 65 70 0h 2h 4h 6h 8h 10h Me an tube le ngt h (pi xe l)
Time after incubation
No LNP
cRGD-LNP, 45 nM cRAD-LNP, 45 nM Std-LNP, 45 nM cRGD-LNP, 4.5 nM
Figure 27. NIR in vivo imaging of DU145 s.c. tumor mice after i.v. injections with F1-LNPs
NIR images, taken 24 hours after injections, are shown for each mouse in gray levels with Min-Max value of 0-27744 (integration time 20 ms). L: liver, T: tumor. Graphs below the images show NIR fluorescence levels on tumors and skin. X-axis: hours after injection, left Y-axis: RLU/pixel/10ms, right Y-axis: tumor/skin signal ratio.
Figure 28. Images of DU145 tumor mice at different time points after LNP injections (supplementary to
Figure 27)
Fluorescence images are shown in gray levels with Min-Max: 0-32093 (integration time 20 ms). B/W = black and white images under white light.
Figure 29. Fluorescence analysis of dissected organs from DU145 tumor mice
Organs were dissected 24 hours after LNP injections and pictured using the NIR camera.
0 10000 20000 30000 40000 blood heart lung muscle kidney brain skin adrenal gland bladder intestine spleen stomach testis liver pancreas fat tumor RLU/px/10ms cRGD (n=3) cRAD (n=3) OH (n=3)
Figure 30. NIR fluorescence images of dissected DU145 tumors
Fluorescence images are shown in gray levels with Min-Max: 0-56352 (integration time 50 ms). Numbers represent mouse identification numbers. (a, b) planar view of whole tumor from the top; (c, d) section view of tumor cut into two; (a,c) white light images; (b, d) NIR fluorescence images.
Figure 31. Ex vivo confocal images of 100 µm-thick sections from DU145 tumors, injected with cRGD- or
cRAD-LNPs
3D reconstruction of stack images with a saturated volume rendering mode. Objective 64× was used. Blue: DAPI, red: LNP.
Figure 32. Ex vivo confocal images of 100 µm-thick sections from DU145 tumors, injected with OH-LNP or vehicle
solution (HBSS buffer)
3D reconstruction of stack images with a saturated volume rendering mode. Objective 64× was used. Blue: DAPI, red: LNP.
Figure 33. NIR in vivo imaging of U87MG s.c. tumor mice after i.v. injections with F1-LNPs
NIR images, taken 24 hours after injections, are shown for each mouse in gray levels with Min-Max values of 0-39424 with an integration time 20 ms. Graphs below the images show NIR fluorescence levels on tumors and skin. X-axis: hours after injection, left Y-axis: RLU/pixel/10ms, right Y-axis: tumor/skin signal ratio.
Figure 34. Images of U87MG tumor mice at different time points after LNP injections (supplementary to
Figure 33)
Fluorescence images are shown in gray levels with Min-Max: 0-32416 for the cRGD, cRAD, Std LNPs (integration time 20 ms) and 0-39424 for the OH-LNP. B/W = black and white images under white light.
Figure 35. Fluorescence analysis of dissected organs from U87MG tumor mice
Organs were dissected 24 hours after LNP injections and pictured using the NIR camera.
0 10000 20000 30000 40000 50000 blood heart lung muscle kidney brain skin adrenal gland bladder intestine spleen stomach ovary uterus liver pancreas fat tumor RLU/px/10ms cRGD (n=3) cRAD (n=3) Std (n=3) OH (n=3)
Figure 36. Comparison of tumor accumulation and biodistribution between different tumor models
Organs were dissected and pictured under a NIR camera 24 hours after injections of 200 µl of 450 nM OH-LNP.
0 20000 40000 60000 80000 blood heart lung muscle kidney brain skin adrenal gland bladder intestine spleen stomach ovary uterus testis liver pancreas fat tumor RLU/px/10ms U87MG (n=3, females) HEK293-β3 (n=2, females) M21L (n=3, females) M21 (n=3, females) DU145 (n=3, males) PC3 orthotopic (n=3, males)
β
β
β
Figure 37. Photoacoustic signals from normal and degraded LNPs
Measures were done in 1X PBS using pulsed laser excitation. LNPs were scanned in the 680‒960 nm window. The curves represent the intensity of PA signals. Acetonitrile and methanol were used to degrade LNPs in order to measure the PA signal of the free dye.
Conclusion
Project 2, cRGD/ATWLPPR-SiNP
Structure of SiNPs
β
Figure 38. A scheme of synthesis process and structures of SiNPs
Drawing was adapted from an article of Ciccione et al. (Ciccione et al., 2016; Sancey et al., 2009). Fluorescein-5-Isothiocyanate (FITC) was used to label these NPs.
Figure 39. Flow cytometry analysis of PC3-luc cells, after incubation with non-PEGylated SiNPs in a cell
suspension, comparing PBS vs. serum-containing medium
SiNPs at 2 nM were incubated with 1 million cells in suspension (volume 200 µL) for 30 min at RT. Histogram: representative result of experiments performed in triplicate. X-axis (FL1-A) = FITC intensity. Bar graphs: mean ±SD of triplicate experiments.
Figure 40. Influence of a 30 min incubation with cRGD-siNP in comparison with other NPs on the cell viability
(supplementary to Figure 39)
Figure 41. Flow cytometry analysis of PC3-luc cells, after incubation with non-PEGylated SiNPs on adherent
cells in a serum-containing medium
PC3-luc cells, placed in 24-well plate (5x104/well), were incubated with 2 nM SiNPs diluted in 300 µL of 10%
FBS-containing medium during 30 min at 37°C. Histogram: representative replicate. X-axis (FL1-A) = FITC intensity. Bar graph: mean±SD of duplicate experiments.
Figure 42. Flow cytometry analysis of PC3-luc cells, after incubation with PEGylated SiNPs on adherent cells,
comparing PBS vs. serum-containing medium
PC3-luc cells, in a 12-well plate (2x104/well), were incubated with 1 nM PEGylated SiNPs (volume 500 µL) for
30 min at 37°C. Histogram: representative replicate. X-axis (FL1-A) = FITC intensity. Bar graphs: mean±SD of duplicate experiments.
Figure 43. Confocal microscopy images of PC3-luc cells after incubation with non-PEGylated and PEGylated SiNPs
PC3-luc cells were incubated with 2 nM NPs diluted in a 10% FBS-containing medium during 30 min at 37°C and washed twice. Cells were fixed with PFA and nuclei were stained by Hoechst 33342. Blue: nuclei, green: FITC (SiNP).
αvβ3
αvβ5 αvβ3
αvβ5 αvβ
αv β β
Impact of cRGD/ATWLPPR-PEG-SiNPs on HUVEC tube formation Figure 44. Expression level of αvβ5 integrin in PC3-luc cells
HUVEC cells were used as a positive control of αvβ5-expressing cells. PC3 luc cells (1x106) were incubated
with a FITC-conjugated mouse mAb anti-human αvβ5 (clone P1F6) during 30 min at 4°C, in PBS containing 1% BSA and 1mM MgCl2/CaCl2. Control cells were treated with an antibody isotype IgG1.
Figure 45. Influences of different PEG-SiNPs on HUVEC tube formation
Cells were diluted in a serum-containing medium, placed in a 24-well plate coated with normal matrigel (at 50,000 cells/250 µL/well), and incubated during a half day in the presence of different PEG-SiNPs. (A) Images taken after 10 hours of incubation. A 5x objective was used. (B) Characteristics of tube formations at T10h
Cytotoxicity assay with cRGD/ATWLPPR-PEG-SiNPs
Figure 46. Cytotoxicity test using EA.hy926 cells by MTS assay
Cells were incubated in a serum-free (SF) medium in the presence of different PEG-SiNPs for 3 days. (A) Cell viability was measured by MTS assay. Data are shown with mean±SD from triplicate experiments. (B) Microscopy images at D3.
Figure 47. Fluorescence microscopy images of PC3-luc subcutaneous tumors collected 3 hours after
injections of PEG-SiNPs
(A) Tumors were sectioned at 5 µm and mounted with a DAPI-containing solution. Slides were scanned
using a mosaic function with a 63x/oil objective. The presented images were montages of mosaic images.
Conclusion
αvβ
Figure 48. Bindings of PEG-SiNPs in various cell lines adapted from Dr T. Jia’s work
Two columns on the left show expression levels of αvβ3 and NRP1. Incubation was carried out for 30 min at 37°C.
