Journal of Controlled Release 186 (2014) 1–10
`
`Contents lists available at ScienceDirect
`
`Journal of Controlled Release
`
`j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j c o n r e l
`
`In vivo near-infrared fluorescence imaging of FAP-expressing tumors
`with activatable FAP-targeted, single-chain Fv-immunoliposomes
`
`Ronny Rüger a,⁎,1, Felista L. Tansi b,⁎⁎,1, Markus Rabenhold a, Frank Steiniger c,
`Roland E. Kontermann d, Alfred Fahr a,⁎, Ingrid Hilger b,⁎⁎
`a Department of Pharmaceutical Technology, Friedrich-Schiller-University Jena, Lessingstrasse 8, 07743 Jena, Germany
`b Dept. of Experimental Radiology, Institute of Diagnostic and Interventional Radiology I, Jena University Hospital-Friedrich Schiller University Jena, Erlanger Allee 101, 07747 Jena, Germany
`c Center for Electron Microscopy, Jena University Hospital-Friedrich Schiller University Jena, Ziegelmuehlenweg 1, 07743 Jena, Germany
`d Institute of Cell Biology and Immunology, University Stuttgart, Allmandring 31, 70569 Stuttgart, Germany
`
`a r t i c l e
`
`i n f o
`
`a b s t r a c t
`
`Article history:
`Received 23 December 2013
`Accepted 24 April 2014
`Available online 5 May 2014
`
`Keywords:
`FAP
`Immunoliposomes
`Fluorescence-quenching
`NIRF
`Optical imaging
`
`Molecular and cellular changes that precede the invasive growth of solid tumors include the release of proteolytic
`enzymes and peptides in the tumor stroma, the recruitment of phagocytic and lymphoid infiltrates and alteration
`of the extracellular matrix. The reactive tumor stroma consists of a large number of myofibroblasts, characterized
`by high expression of fibroblast activation protein alpha (FAP). FAP, a type-II transmembrane sialoglycoprotein is
`an attractive target in diagnosis and therapy of several pathologic disorders especially cancer. In the underlying
`work, a fluorescence-activatable liposome (fluorescence-quenched during circulation and fluorescence activa-
`tion upon cellular uptake), bearing specific single-chain Fv fragments directed against FAP (scFv′FAP) was de-
`veloped, and its potential for use in fluorescence diagnostic imaging of FAP-expressing tumor cells was evaluated
`by whole body fluorescence imaging. The liposomes termed anti-FAP-IL were prepared via post-insertion of
`ligand–phospholipid-conjugates into preformed DY-676-COOH-containing liposomes. The anti-FAP-IL revealed
`a homogeneous size distribution and showed specific interaction and binding with FAP-expressing cells
`in vitro. The high level of fluorescence quenching of the near-infrared fluorescent dye sequestered in the aqueous
`interior of the liposomes enables fluorescence imaging exclusively upon uptake and degradation by cells, which
`results in fluorescence activation. Only FAP-expressing cells were able to take up and activate fluorescence of
`anti-FAP-IL in vitro. Furthermore, anti-FAP-IL accumulated selectively in FAP-expressing xenograft models
`in vivo, as demonstrated by blocking experiments using free scFv′FAP. The local tumor fluorescence intensities
`were in agreement with the intrinsic degree of FAP-expression in different xenograft models. Thus, anti-FAP-IL
`can serve as a suitable in vivo diagnostic tool for pathological disorders accompanied by high FAP-expression.
`© 2014 Elsevier B.V. All rights reserved.
`
`1. Introduction
`
`A crucial step towards defining a correct personalized anticancer
`therapy is the identification of the genes and pathways altered in the
`tumor of the patient. In this respect a significant number of studies
`have revealed fibroblast activation protein alpha (FAP), as a promising
`tumor target for designing novel pharmaceuticals for both cancer diag-
`nosis and therapy [1]. FAP is a type-II transmembrane sialoglycoprotein
`and a member of the conserved family of serine proteases which are
`capable of cleaving the post-prolyl bond of peptides and proteins [2].
`FAP is selectively expressed by tumor-associated fibroblasts in over
`90% of malignant breast, colorectal, skin, and pancreatic tumors, but
`
`⁎ Corresponding authors. Tel.: +49 3641 949905.
`⁎⁎ Corresponding authors.
`E-mail addresses: ronny.rueger@uni-jena.de (R. Rüger), felista.tansi@med.uni-jena.de
`(F.L. Tansi), alfred.fahr@uni-jena.de (A. Fahr), ingrid.hilger@med.uni-jena.de (I. Hilger).
`1 Contributed equally to this work.
`
`not in normal adult tissues [3–5]. Opposed to its protease family mem-
`bers, FAP possesses an additional gelatinase activity [6,7]. Due to these
`properties, FAP serves as a biomarker of cancer-associated fibroblasts
`(CAFs) in the tumor stroma, where it modulates proliferation, invasion
`and metastasis [8–10]. Thus, FAP has been considered as a target in sev-
`eral preclinical approaches and clinical trials. For example a humanized
`anti-FAP monoclonal antibody (sibrotuzumab) revealed safety and dose
`tolerance after repeated administrations in phase II clinical trials [11].
`Considering that, FAP plays a central role in the invasiveness of solid
`tumors and given that 90% of the stromal fibroblasts of many cancer
`types express the protein, the development of diagnostic approaches
`by targeting FAP would be of great importance. In vivo optical imaging
`using near infrared (NIR) fluorescent probes is a non-invasive approach
`that exploits the fact that biological tissues show very low absorption
`and auto-fluorescence in the NIR spectrum window [12]. The potential
`of this approach in diagnostic and intraoperative imaging and in the
`monitoring of treatment response has been repeatedly demonstrated
`[13–15]. Thus, molecular imaging probes bearing peptide ligands or
`
`http://dx.doi.org/10.1016/j.jconrel.2014.04.050
`0168-3659/© 2014 Elsevier B.V. All rights reserved.
`
`

`

`2
`
`R. Rüger et al. / Journal of Controlled Release 186 (2014) 1–10
`
`antibodies directed towards FAP would enable classification of cancer
`patients for FAP-targeted therapy. So far, FAP-targeted probes that
`serve as non-invasive imaging tools for early detection of cancer have
`mostly been reported in in vitro studies [16–18]. In an in vivo approach
`to target FAP in optical imaging the covalent coupling of the NIRF dye
`Cy5.5 to a restriction peptide sequence for FAP was used [19].
`However, this approach only detects enzymatically active FAP
`whereas modified forms of the protein expressed on tumor cells would
`not be addressed. Furthermore, there are some drawbacks using protein
`and peptide-based molecular imaging probes. The amount of fluorescent
`molecules per ligand is limited to a defined number per ligand [20]. In-
`creasing the number of imaging molecules per ligand without reaching
`a quenching concentration increases the background signal during circu-
`lation [21]. Non-site directed conjugation processes used for labeling
`of e.g. antibodies can reduce the binding activity of the ligand to the tar-
`get due to the induced conformational changes or physically blocking
`the antigen binding sites [22]. Furthermore, the smaller the ligands
`the lower is their blood circulation half-life. Due to their small size and
`the lack of recycling processes mediated by the neonatal Fc receptor
`(FcRn) for whole antibodies, ligands such as peptides, scFv, diabodies
`and single-chain diabodies are quickly cleared by renal filtration [23,24].
`To circumvent these drawbacks, we developed an immunoliposome-
`based NIRF-FAP-imaging probe for in vivo application. The use of lipo-
`somes is advantageous in terms of high cargo payload and biocompati-
`bility [25,26]. Although, the intravenous administration of nanoparticle
`formulations such as the clinical doxorubicin-containing, PEGylated
`liposome, Doxil can induce immunological reactions such as liposo-
`mal complement activation-related pseudoallergy (CARPA) [27–30],
`premedication with Doxil-like empty liposomes (Doxebo) and a slow in-
`fusion showed improved tolerance to Doxil [31] supporting their poten-
`tial as suitable drug delivery vesicles. Recently, we demonstrated that
`non-targeted, PEGylated liposomes can be selectively taken up by phago-
`cytosis and activated in phagolysosomes thereby enhancing the imaging
`of inflammatory areas in vivo [32]. The use of liposomes circumvents the
`rapid clearance seen with peptide and antibody fragment-based probes
`alone and enables the combination of imaging with therapy.
`Here we sought to address the FAP in tumors by utilization of target-
`affine fluorescence activatable liposomes. The use of a scFv as binding
`moiety ensures specific binding to the target and at the same time re-
`duces possible adverse immune responses that may result from applying
`whole antibodies [33]. We evaluated whether covalent linkage of
`scFv′FAP at the surface of fluorescence-activatable liposomes will enable
`FAP-targeted delivery of fluorescent dyes in FAP-expressing tumors
`in vivo, thereby enhancing molecular imaging of FAP-expressing tumor
`models in mice. We demonstrate that specific uptake and activation of
`anti-FAP-IL take place exclusively in FAP-expressing cells in vitro. Further-
`more, the uptake and activation of anti-FAP-IL by FAP-expressing tumor
`cells enhanced whole body in vivo near-infrared fluorescence imaging of
`xenograft models in mice. Hence, anti-FAP-IL can be used to predict prog-
`nosis, guide intraoperative treatment and monitor treatment success.
`
`2. Abbreviations
`
`FAP
`NIR
`p.i
`ROI
`scFv
`
`fibroblast activation protein-alpha;
`near-infrared;
`post injection;
`regions of interest
`single-chain Fv fragment
`
`3. Materials and methods
`
`3.1. Materials
`
`Egg phosphatidylcholine (EPC) was purchased from Lipoid
`(Ludwigshafen, Germany); cholesterol, Tris(hydroxymethyl)-
`aminomethane (Tris), 2-(4-(2-hydroxyethyl)-1-piperazinyl)-
`
`ethansulfonsäure (HEPES) and 4-(1,1,3,3-tetramethylbutyl)phenyl-poly-
`ethylene glycol (Triton-X100) were purchased from Sigma (Taufkirchen,
`Germany); and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-
`[methoxy (polyethylene glycol)-2000] (ammonium salt) (mPEG2000-
`DSPE), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[maleimide
`(polyethylene glycol)-2000] (ammonium salt) (MalPEG2000-DSPE)
`and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-
`1,3-benzoxadiazol-4-yl) (ammonium salt) (NBD-DOPE) were pur-
`chased from Avanti Polar Lipids (Alabaster, USA). The free acid of the
`near infrared fluorescent dye DY-676-COOH and DY-676-maleimide
`was purchased from Dyomics GmbH (Jena, Germany).
`
`3.2. Preparation of quenched DY-676-COOH containing liposomes (Lip-Q)
`
`All liposomes were prepared by the film hydration and extrusion
`method [34]. Vesicles were composed of EPC:Chol:mPEG2000-DSPE at
`a molar ratio of 6.5:3:0.5. Anti-FAP-IL used for plasma stability experi-
`ments were stained with 0.3 mol% of the lipophilic, red fluorescent
`1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate
`(Sigma-Aldrich, Taufkirchen, Germany). Double stained liposomes
`were prepared whereby the hydrophilic DY-676-COOH was encapsu-
`lated within the liposomes and the lipophilic liposome-membrane
`marker NBD-DOPE (0.3 mol.%) was incorporated into the liposomal
`phospholipid bilayer. 100 μmol (100 mM) lipid was hydrated by
`using DY-676-COOH (6.18 mM) dissolved in 10 mM Tris buffer, pH
`7.4. Subsequently, the dispersions were frozen in liquid nitrogen and
`thawn at room temperature seven times [32]. The resulting vesicle
`dispersions were homogenized via hand extrusion, 21 times through
`a 100 nm polycarbonate membrane, using a LiposoFast-Basic extruder
`(Avestin, Ottawa, Canada). A subsequent gel filtration via a self-
`prepared Sephadex G25 column (length 26 cm, diameter 0.8 cm)
`was accomplished to purify DY-676-COOH containing liposomes from
`non-encapsulated, free DY-676-COOH. Due to dilution during gel filtra-
`tion the DY-676-COOH containing liposomes were centrifuged (2 h,
`100,000 g, 8 °C) using a Beckman XL80 equipped with rotor SW55-Ti
`(Beckman Coulter GmbH, Krefeld, Germany) to concentrate the sam-
`ples. This procedure results in 100 μmol plain liposomes filled with
`DY-676-COOH (Lip-Q), assuming no lipid loss during preparation.
`
`3.3. Preparation and purification of single-chain Fv fragments directed
`against FAP
`
`The scFv′FAP were produced in Escherichia coli (E. coli) by periplas-
`mic preparation and purified via immobilized metal affinity chromatog-
`raphy (IMAC) [34]. The protein concentration was determined by
`measuring the absorbance at 280 nm. The scFv fragments were analyzed
`via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-
`PAGE according to Laemmli) 12% under reducing and non-reducing
`conditions and stained with Coomassie brilliant blue G250 (Carl
`Roth GmbH, Karlsruhe, Germany).
`
`3.4. Preparation of anti-FAP-immunoliposomes and scFv′FAP-DY-676
`conjugates
`
`Plain DY-676-COOH containing liposomes (100 μmol, Lip-Q) were
`divided into two halves. The first half (50 μmol lipid) was used for the
`post-insertion with 0.1 mol.% cysteine-conjugated MalPEG2000-DSPE
`micelles as control (meaning 50 nmol MalPEG2000-DSPE anchor lipid)
`and the second half was used for the post-insertion with 0.1 mol.%
`MalPEG2000-DSPE micelles conjugated to scFv′FAP. The ligand–micelle
`conjugates were inserted into preformed PEGylated liposomes as
`described elsewhere [17]. Mal-PEG2000-DSPE micelles were prepared
`and conjugated with reduced ligands (scFv′FAP or cysteine) according
`to literature [35] with slight differences.
`Coupling reaction between scFv′FAP and thioreactive micelles was
`performed at room temperature for 60 min. The coupling efficiency
`
`

`

`R. Rüger et al. / Journal of Controlled Release 186 (2014) 1–10
`
`3
`
`was analyzed by SDS-PAGE and evaluated using the freeware ImageJ
`version 1.45 s. The scFv′FAP-coupled MalPEG2000-DSPE micelles were
`inserted into preformed PEGylated liposomes by incubation at 50 °C in
`a water bath for 30 min. Unbound scFv molecules were removed by
`gel-filtration using a Sepharose CL4B column (Amersham, Braunschweig,
`Germany). The liposomes were centrifuged (2 h, 100,000 g, 8 °C) using a
`Beckman XL80 equipped with a SW55Ti rotor and re-suspended in 1 ml
`sterile Tris buffer (10 mM, pH 7.4). To directly label scFv′FAP fragments
`with DY-676 molecules, the scFv′FAP was reduced and dialysed as
`above. The reduced scFv′FAP was incubated with a 10 fold molar excess
`of the thioreactive DY-676-maleimide derivative for 60 min at room
`temperature. Afterwards, the resulting scFv′FAP-DY-676 was washed
`up to 30 times using a Microcon® centrifugal filter device with a 3000
`Da molecular weight cut off.
`
`3.5. Characterization of liposomes
`
`vein endothelial cells (HUVECs) were isolated from human umbili-
`cal vein and cultured in Endothelial cell Growth Medium (ECGM)
`(Promocell, Germany). The cells were grown at 37 °C in a 5% CO2 and
`95% humidified atmosphere.
`
`3.10. Flow cytometry
`
`To analyze the interaction of anti-FAP-IL with FAP-expressing cells,
`0.5 × 106 HT1080-mFAP in 100 μl RPMI (5% FCS) were incubated with
`liposomes (496 μM, 1 h, on ice, in the dark). For competition experi-
`ments the cells were simultaneously incubated with anti-FAP-IL and
`10 μg free scFv′FAP under the same conditions as described above.
`This reflects a nearly 40 fold excess of free scFv′FAP over liposomal
`bound scFv′FAP. After incubation, cells were washed three times with
`3 ml PBA (PBS, 2% FCS, 0.02% sodium azide), re-suspended in 450 μl
`PBA and analyzed using an Epics XL-MCL (Beckmann Coulter, Krefeld,
`Germany). Data were evaluated using the freeware WinMDI 2.9.
`
`Sizes of liposomes were measured via dynamic light scattering
`and the zeta potential was determined by using a Zetasizer Nano ZS
`(Malvern, Herrenberg, Germany).
`
`3.11. Plasma stability experiments
`
`3.6. Cryo-transmission electron microscopy
`
`Approximately 20 μl of the liposome dispersion was applied to a cop-
`per grid covered by holey carbon film Quantifoil R3.5/1 (Micro Tools
`GmbH, Jena, Germany). Excess of liquid was blotted automatically for
`3 s between two strips of filter paper. Subsequently, the samples were
`rapidly plunged into liquid ethane (cooled to about − 180 °C) in a
`cryobox (Carl Zeiss NTS GmbH, Oberkochen, Germany). Excess ethane
`was removed with a piece of filter paper. The samples were transferred
`with a cryo-transfer unit (Gatan 626-DH) into the pre-cooled cryo-
`transmission electron microscope Philips CM 120 (Philips Research,
`Eindhoven, The Netherlands) operated at 120 kV and viewed under
`low dose conditions. The images were recorded with a 1 K CCD Camera
`FastScan F114 (TVIPS, Gauting, Germany).
`
`3.7. Quantitation of liposomal encapsulated DY-676-COOH
`
`To quantify the amount of DY-676-COOH encapsulated, a calibration
`curve was prepared by dissolving DY-676-COOH (0, 82, 124, 247, 494,
`741, 988 nM) in 10 mM Tris buffer, pH 7.4 containing 0.1% Triton
`X100. The liposome samples were incubated 5 min at RT in presence
`of 1% Triton X100 and afterwards diluted with 10 mM Tris buffer,
`pH 7.4 to a final Triton-X100 concentration of 0.1%. All samples were
`measured in triplicates on a Fluostar Optima microplate reader (BMG
`Labtech, GmbH, Ortenberg, Germany) equipped with adequate filters
`for excitation at λ = 645 nm and measuring the emitted fluorescence
`light at λ = 700 nm. The quantitation of DY-676-COOH was performed
`with the final liposomal formulations (Cys-Lip/anti-FAP-Lip).
`
`3.8. Liposomal quenching assay
`
`The fluorescence-quenching and activation effect of anti-FAP-IL and
`Lip-Q was analyzed in vitro, by measuring the fluorescence intensity
`before and after liposomal freeze-damaging as described earlier [32].
`
`Anti-FAP-IL was incubated with equal volumes of PBS or human
`plasma (Institute of Transfusion Medicine, University Hospital, Jena,
`Germany) for 24 h at 37 °C. Afterwards, the indicated amount of anti-
`FAP-IL was incubated with HT1080-mFAP cells for 1 h at 4 °C. The
`cells were then analyzed by flow cytometry as afore mentioned. For
`long term investigation (Supporting Information S5) liposomes were
`pre-treated with human plasma and PBS at 37 °C for up to 72 h.
`
`3.12. Uptake of liposomal probes and imaging
`
`For confocal microscopy, 30,000 cells (HT-1080, HT1080-hFAP,
`MDA-MB435S) were grown on poly-L-lysine-coated 8-well culture
`slides (BD Biosciences) for 16 h. Afterwards, 200 nmol (final lipid) of
`the liposomes was added and the cells were further cultured for 6–8 h
`at 37 °C or 4 °C. For time-course comparison of anti-FAP-IL and free
`scFv′FAP-DY-676 uptake, the HT-1080 and HT1080-hFAP cells were
`prepared in small culture flasks (2.0 × 106 cells) or on coated chamber
`slides (30,000 cells) then treated with anti-FAP-IL (100 nmol final lipid)
`or 2 μg scFv′FAP-DY-676 for 4 h, 8 h, 12 h and 24 h. The cells were
`washed 3 times with HBSS and fixed with 3.7% (v/v) formaldehyde
`(Roth, Karlsruhe, Germany) in PBS for 30 min at RT. Cells in culture flasks
`were scraped in 500 μl HBSS and pelleted by centrifugation and imaged
`in the MaestroTM in vivo fluorescence imaging system and signal inten-
`sities quantified (as described under “in vivo near-infrared fluorescence
`imaging” below). The cells were washed twice with PBS, mounted with
`PermaFluor (Thermo Fisher Scientific GmbH, Schwerte, Germany) con-
`taining the DNA stain Hoechst-33258 (Applichem GmbH, Darmstadt,
`Germany) then covered with glass cover-slips and subsequently imaged
`on the LSM510-META confocal microscope (Zeiss, Jena, Germany). The
`nuclei were visualized with a 405 nm laser diode and a 420–480 nm
`bandpass filter. A 543 nm HeNe laser and a 550–615 nm band pass filter
`were applied, whereby fluorescence of DY-676-COOH was excited with
`a 633 nm argon laser and emission was detected with a 650 nm longpass
`filter. Fluorescence emission of NBD-DOPE was detected with the GFP
`filter at 530 nm after excitation at 488 nm.
`
`3.9. Cell lines
`
`3.13. Animals and implantation
`
`The human melanoma cell line, MDA-MB435S was purchased from
`the Cell Lines Service, (CLS, Heidelberg, Germany). The fibrosarcoma
`cell line, HT1080 and its FAP-transfected counterparts HT1080-hFAP
`(human FAP) and HT1080-mFAP (murine FAP), SKBR3, B16F10,
`HEK293 were cultured in a RPMI medium supplemented with 5% (v/v)
`fetal calf serum (both Life Technologies GmbH, Darmstadt, Germany)
`whereas the MDA-MB435S cells were grown in Dulbecco's modified
`Eagle's medium containing 10% FCS and 1% HEPES. Human umbilical
`
`All procedures were approved by the regional animal committee
`and were in accordance with international guidelines on the ethical
`use of animals. 10–14 week-old female athymic nude mice (n = 55;
`Hsd:Athymic Nude-Foxn1nu nu/nu; Harlan Laboratories) weighing ap-
`proximately 20 g were housed under standard conditions with food and
`water ad libitum. 14–21 days prior to in vivo imaging, xenografts were
`induced by implanting 1.0 × 106 cells (HT1080-wt), 1.5 × 106 cells
`(HT1080-hFAP) or 2.0 × 106 cells (MDA-MB435S) subcutaneously.
`
`

`

`4
`
`R. Rüger et al. / Journal of Controlled Release 186 (2014) 1–10
`
`Briefly, the cells were pelleted and dispensed in 100 μl cold Matrigel™
`(BD Biosciences, Heidelberg, Germany) then injected subcutaneously
`into the lower back of the mice. Seven days prior to probe injection and
`imaging, all mice received a low pheophorbide diet C1039 (Altromin
`Spezialfutter GmbH & Co. KG, Lange, Germany) in order to reduce tissue
`auto-fluorescence. Throughout all procedures, animals were anesthetized
`with 2% isoflurane.
`
`3.14. In vivo near-infrared fluorescence imaging
`
`Mice bearing subcutaneous xenografts with approximate diameters
`of 5–9 mm (with average tumor volume of 138 mm3 for HT1080-wt,
`66 mm3 for HT1080-hFAP and 32 mm3 for MDA-MB435S) were anaes-
`thetized with 2% isoflurane and the respective liposomal solutions
`(30 μmol (lipid)/kg weight) applied intravenously through the tail
`vein. To control for signal specificity and contribution of EPR effect, a
`group of the animals received liposomal probe together with 10 μg
`free scFv′FAP (which is comparable to 8.15 μg liposomal scFv′FAP in
`anti-FAP-IL) or only the free scFv′FAP-DY-676, which was covalently
`coupled to DY-676-maleimide (Supporting Information S6). The ani-
`mals were immediately imaged with the MaestroTM in vivo fluorescence
`imaging system (MaestroTM, CRi Woburn, UK) by exciting with filters
`for the excitation range 615–665 nm and acquiring emission with a
`cut-in filter (N 700 nm) and recordings made as time point, t = 0 h
`post injection (p.i.). The animals were further imaged every 2 h for 10 h
`and again at t = 24 h–32 h and at 48 h p.i. The evaluation and extraction
`of background auto-fluorescence of native animals were performed
`with the MaestroTM-software according to the manufacturer's in-
`structions. Semi-quantitative analysis of fluorescence intensities of
`tumors was acquired by assigning a region of interest (ROI) to each
`tumor (target) and another on the hind leg region of the animals
`(muscle/skin = background). Fluorescence intensities were deduced
`as average signal (scaled counts/s). This represents count levels after
`scaling for exposure time, camera gain, binning and bit depth, and
`makes the measurements comparable among each other.
`
`3.16. Immunohistochemistry
`
`After imaging, mice were euthanized and tumors were collected,
`fixed in methanol-stabilized 5% formaldehyde solution (Otto-Fischar
`GmbH, Saarbrücken, Germany) for 24 h at 4 °C, then embedded in par-
`affin. Using a microtome system (Microm HM 340E) (Thermo-Fisher
`Scientific GmbH, Schwerte, Germany), three-micron thick serial sec-
`tions of the samples were cut, mounted on (poly-L-lysin)-coated glass
`slides and air-dried. The paraffin-embedded sections were dewaxed,
`and antigen retrieval was performed in 10 mM sodium citrate buffer,
`pH 7.4, for 20 min at 95 °C. Slides were washed three times in TBS-T
`(containing 1% Tween 20) at RT and hydrated in PBS for 10 min.
`Blocking for unspecific receptor binding was performed with 1% BSA
`and 0.1% Tween-20 in PBS pH 7.4, for 1 h at RT. Thereafter, the
`slides were washed 2 times with PBS for 5 min each. Tissue sections
`were incubated overnight at 4 °C with anti-alpha-smooth muscle
`actin primary antibody (anti-αSMA mAb, A2547, Sigma Aldrich,
`Taufkirchen, Germany) in order to detect stromal fibroblasts. This was
`followed by a 3 times washing step and incubation with Alexa-Fluor-
`488-coupled goat anti-mouse IgG, (aza11029, Mobitec GmbH,
`Goettingen, Germany) for 30 min at RT. The sections were washed 3
`times, mounted with Permafluor (Applichem GmbH, Darmstadt,
`Germany) containing a 1:50 dilution of the DNA stain Hoechst-33258
`and then cover slipped. Fluorescence images were acquired using the
`Zeiss LSM510-META confocal laser scanning microscope as indicated
`for in vitro uptake experiments.
`
`3.17. Statistical data
`
`Except otherwise indicated, Student's t-test was used to deduce the
`level of significance, if normality and equal variance tests were passed.
`Else Mann–Whitney-Rank sum test was applied. All experiments were
`done at least in duplicates. In vivo animal trials were performed with
`4 or more animals/group. Differences resulting in P b 0.05 were consid-
`ered significant.
`
`4. Results
`
`3.15. Ex vivo determination of the biodistribution of anti-FAP-IL
`
`4.1. Preparation of FAP-specific scFv and fluorescence-activatable
`scFv-immunoliposomes
`
`48 h post contrast agent injection and imaging, the animals were
`sacrificed and the organs and tumors were imaged. Semi-quantitative
`determination of the respective fluorescence intensities was performed
`with the MaestroTM imaging system as described above.
`
`Fig. 1. Analysis of scFv′FAP by 12% SDS-PAGE: A.) 2 μg of purified scFv′FAP under reducing
`(1) and non-reducing conditions (2). B.) The scFv′FAP analyzed under non-reducing con-
`ditions before (1) and after (2) conjugation to MalPEG2000-DSPE micelles. C.) Schematic
`overview of the preparation of anti-FAP-IL.
`
`The FAP-specific single-chain Fv fragments (scFv′FAP) purified from
`E. coli contained a distal sulfhydryl group for a site-directed conjugation
`to thioreactive groups. The yield was 0.35 mg per liter bacteria suspen-
`sion. The scFv′ showed a characteristic monomer band under reducing
`conditions (Fig. 1A, lane 1) and formed approximately 27.5% disulfide-
`linked dimers under non-reducing conditions (Fig. 1A, lane 2). The re-
`duced scFv′FAP was conjugated covalently to micelles composed of
`Mal-PEG2000-DSPE via thioether linkage. SDS-PAGE revealed a coupling
`efficiency of approximately 88% (Fig. 1B). The anti-FAP-IL was then
`prepared via post-insertion of 0.1 mol.% scFv′FAP-MalPEG2000-DSPE
`conjugates into activatable fluorescent liposomes (Lip-Q) [32] which
`were prepared with high quenching concentrations of the NIRF dye,
`DY-676-COOH in the aqueous interior and the green fluorescent phos-
`pholipid NBD-DOPE on the lipid interface (Fig. 1C). Cys-Lip were pre-
`pared as control liposomes by coupling of MalPEG2000-DSPE micelles
`to L-cysteine. L-cystein-MalPEG2000-DSPE micelles were post-inserted
`into plain liposomes (Lip-Q).
`
`4.2. Characterization of activatable scFv-targeted immunoliposomes
`
`Different liposomal formulations were prepared and compared. The
`Z-averages of all the liposomal formulations were below 160 nm with
`polydispersity indices (PDI) below 0.1 (Table 1). These data were fur-
`ther confirmed by cryo-electron microscopy (Fig. 2A). Most of the vesi-
`cles in Lip-Q (before post-insertion) as well as Cys-Lip and anti-FAP-IL
`
`

`

`R. Rüger et al. / Journal of Controlled Release 186 (2014) 1–10
`
`5
`
`Table 1
`Characterization of liposomes by dynamic light scattering.
`
`Liposome
`
`Size
`[nm]
`
`Polydispersity
`index (PDI)
`
`Zeta potential
`[mV]
`
`CDy676-COOH
`[μg/ml]
`
`CLipid
`[mM]
`
`Cys-Lip
`Anti-FAP-IL
`Lip-Q
`
`126 ± 5.1
`139.1 ± 25
`118.5 ± 0.7
`
`0.069 ± 0.03 −13.9 ± 6.7
`0.079 ± 0.02 −15.4 ± 4.9
`0.04 ± 0.02
`−9 ± 2
`
`135.2 ± 21.5
`148.6 ± 5
`68.76 ± 5.9
`
`50
`50
`21
`
`(after post-insertion) were unilamellar with less than 42% bilamellar
`structures and exhibited a vesicular morphology.
`The fluorescence-activatable liposomes retained their fluorescence
`quenching effect even after insertion of the scFv′FAP-MalPEG2000-
`DSPE conjugates. This could be detected in a double absorption peak
`of the anti-FAP-IL (Fig. 2B, anti-FAP-IL—bold line) which was compara-
`ble to that seen with the liposomes prior to insertion (Fig. 2B, LipQ—bold
`line). Following rupturing of the anti-FAP-IL by harsh freezing at
`−80 °C and thawing at 30 °C, an increase in the absorption maximum
`and the disappearance of the blue-shifted peak was observed (Fig. 2B,
`broken lines). This revealed fluorescence quenching of the encapsulated
`DY-676-COOH, which was not altered by the scFv′ insertion performed
`above the liposomal phase transition temperature.
`
`4.3. Specificity of liposomal uptake by FAP-expressing cells
`
`To verify the binding specificity of double-labeled anti-FAP-IL,
`the stable transfected human fibrosarcoma cell line HT1080-mFAP
`was incubated with different liposomal formulations and analyzed
`by flow cytometry with the aid of the green fluorescent phospholipid
`NBD-DOPE localized in the liposomal membrane. Whereas specific
`binding of anti-FAP-IL to the FAP-expressing cells was detected,
`(Fig. 3A, anti-FAP-IL) only very weak interaction was observed with
`control liposomes (Fig. 3A, Cys-Lip) and with the anti-FAP-IL when
`competing with 10 μg of free FAP-specific scFv′ (Fig. 3A, anti-FAP-
`IL + free scFv′FAP). Furthermore, the anti-FAP-IL did not bind to various
`eukaryotic FAP-negative cell lines, for example the human breast
`adenocarcinoma cell
`line SKBR3, the mouse melanoma cell line
`B16F10, and the human embryonic kidney cell line HEK293, whereas
`
`Fig. 2. A.) Representative cryo-transmission electron micrographs of liposomes prior to
`(Lip-Q) and post insertion of Ligand-MalPEG2000-DSPE micelles (cysteine as control, Cys-
`Lip and scFv′FAP (anti-FAP-IL)) and the structure of encapsulated DY-676-COOH. B.) Ab-
`sorption and emission spectra of activatable liposomes before and after rupturing by
`freeze and thaw treatment which reveals the release and activation of the dye and an in-
`crease in fluorescence signal. For more information see [32].
`
`only weak interaction could be observed with human umbilical vein
`endothelial cells, HUVEC (Fig. 3C).
`To analyze the stability of anti-FAP-IL under physiological conditions
`the liposomes were incubated for 24 h with 50% human plasma or PBS
`at 37 °C. Here, anti-FAP-IL was labeled with the red fluorescent mem-
`brane probe DiI. Only a slightly decreased binding was observed
`for anti-FAP-IL pre-incubated with human plasma compared to anti-
`FAP-IL pre-incubated with PBS (Fig. 3B). Anti-FAP-IL pre-incubated
`with 50% human plasma or PBS for up to 72 h (Supporting Information
`S5) revealed distinctive binding to FAP-expressing cells.
`To confirm binding and uptake of the anti-FAP-ILs, monolayer cul-
`tures of different FAP-expressing and FAP-negative cell lines were
`grown on cover slips and analyzed by confocal laser scanning microsco-
`py after incubation with the liposomes (Fig. 3D). Anti-FAP-IL could be
`taken up exclusively by FAP-expressing cells which was evident in a
`high uptake by the stably transfected human fibrosarcoma HT1080-
`hFAP cell line and the melanoma cell line MDA-MB-435S with low en-
`dogenous FAP expression (Fig. 3D, 37 °C). Opposed to this, the wild-
`type human fibrosarcoma cell line HT1080 incubated with the liposomes
`for 8 h revealed no uptake of the liposomes (Fig. 3D, HT1080-wt).
`In order to verify whether the anti-FAP-IL is taken up actively by
`receptor-mediated internalization, HT1080-hFAP cells were incubated
`for 6 h at 4 °C immediately after addition of liposomes. Confocal micros-
`copy of the cells revealed specific binding of the anti-FAP-IL to the cell
`membrane (Fig. 3D, 4 °C). A diffuse intracellular NIR-fluorescence of
`DY-676-COOH (red fluorescence) was observed, which differs from the
`intracellular fluorescence caused by internalized liposomes at 37 °C.
`The difference is a vesicular localization (Fig. 3D, 37 °C) versus diffuse
`NIR fluorescence due to membrane fusion and release of dye into the
`cytosol (Fig. 3D, 4 °C). The non-targeted Cys-Lip revealed no uptake by
`wild-type or FAP-expressing cell lines (Fig. 3D, Cys-Lip), which indicates
`that uptake of the anti-FAP-IL is specifically mediated by the FAP protein
`expressed on the cell surface.
`
`4.4. Whole body in vivo optical imaging of human fibrosarcoma and
`melanoma in mice xenograft models
`
`A gradual and persistent increase in the fluorescence of tumors
`could be detected 2 to 24 h post-intravenous injection of the anti-
`FAP-IL (Fig. 4). After 24 h there was a gradual decrease in the fluores-
`cence intensity with time. However, there was still a strong fluorescence
`signal in the tumors 48 h post injection. The targeted anti-FAP-IL re-
`vealed distinct uptake by the different tumor models with especially
`high accumulations in both the HT1080-wt and HT1080-hFAP-derived
`xenograft models. Consistent with the specific receptor-mediated accu-
`mulation of anti-FAP-IL in high FAP-expressing tumors, competitive
`blocking by co-injection of the anti-FAP-IL together with 10 μg