Trastuzumab deliver encapsulated cargo into the nuclei of tumor cells and the tumor microenvironment simultaneously

Accepted Manuscript Activatable bispecific liposomes bearing fibroblast activation protein directed single chain fragment / Trastuzumab deliver encapsulated cargo into the nuclei of tumor cells and the tumor microenvironment simultaneously Felista L. Tansi, Ronny Rüger, Claudia Böhm, Frank Steiniger, Roland E. Kontermann, Ulf K. Teichgraeber, Alfred Fahr, Ingrid Hilger PII: DOI: Reference:

S1742-7061(17)30203-9 http://dx.doi.org/10.1016/j.actbio.2017.03.033 ACTBIO 4801

To appear in:

Acta Biomaterialia

Received Date: Revised Date: Accepted Date:

5 November 2016 17 March 2017 22 March 2017

Please cite this article as: Tansi, F.L., Rüger, R., Böhm, C., Steiniger, F., Kontermann, R.E., Teichgraeber, U.K., Fahr, A., Hilger, I., Activatable bispecific liposomes bearing fibroblast activation protein directed single chain fragment / Trastuzumab deliver encapsulated cargo into the nuclei of tumor cells and the tumor microenvironment simultaneously, Acta Biomaterialia (2017), doi: http://dx.doi.org/10.1016/j.actbio.2017.03.033

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Activatable bispecific liposomes bearing fibroblast activation protein directed single chain fragment / Trastuzumab deliver encapsulated cargo into the nuclei of tumor cells and the tumor microenvironment simultaneously Felista L. Tansia,*, Ronny Rügerb,*, Claudia Böhma, Frank Steinigerc, Roland E. Kontermannd, Ulf K. Teichgraebera, Alfred Fahrb & Ingrid Hilgera* a

Institute of Diagnostic and Interventional Radiology, Experimental Radiology, Jena University Hospital

Am klinikum 1, 07747 Jena, Germany. b

Department of Pharmaceutical Technology, Friedrich-Schiller-University Jena, Lessingstrasse 8, 07743

Jena, Germany c

Center for Electron Microscopy, Jena University Hospital, Ziegelmuehlenweg 1, 07743 Jena, Germany

d

Institute of Cell Biology and Immunology, University Stuttgart, Allmandring 31, 70569 Stuttgart

* Authors to whom correspondence should be addressed: Felista L. Tansi: [email protected]; Tel. +4936419324993; Fax. +4936419325922 Ronny Rüger: [email protected]; Tel. +493641949905; Fax. +493641949902 Ingrid Hilger: [email protected]; Tel. +4936419325921; Fax. +4936419325922

1

Abstract Molecular targeting plays a significant role in cancer diagnosis and therapy. However, the heterogeneity of tumors is a limiting obstacle for molecular targeting. Consequently, clinically approved drug delivery systems such as liposomes still rely on passive targeting to tumors, which does not address tumor heterogeneity. In this work, we therefore designed and elucidated the potentials of activatable bispecific targeted liposomes for simultaneous detection of fibroblast activation protein (FAP) and the human epidermal growth factor receptor 2 (HER2). The bispecific liposomes were encapsulated with fluorescence-quenched concentrations of the nearinfrared fluorescent dye, DY-676-COOH, making them detectable solely post processing within target cells. The liposomes were endowed with a combination of single chain antibody fragments specific for FAP and HER2 respectively, or with the FAP single chain antibody fragment in combination with Trastuzumab, which is specific for HER2. The Trastuzumab based bispecific formulation, termed Bi-FAP/Tras-IL revealed delivery of the encapsulated dye into the nuclei of HER2 expressing cancer cells and caused cell death at significantly higher rates than the free Trastuzumab. Furthermore, fluorescence imaging and live microscopy of tumor models in mice substantiated the delivery of the encapsulated cargo into the nuclei of target tumor cells and tumor stromal fibroblasts. Hence, they convey potentials to address tumor plasticity, to improve targeted cancer therapy and reduce Trastuzumab resistance in the future. Key words: Molecular targeting, fluorescence quenching, optical imaging, liposomes, tumor heterogeneity.

2

1. Introduction Nanoformulations, which are sophisticated tools of several nanometer sizes and compositions protect instable and toxic drugs from rapid elimination and enhance their effective delivery to diseased sites [1]. Combining nanomedicine and optical molecular imaging is essential for studying disease pathogenesis [2], drug development, image-guided surgery and therapy monitoring. Though nanomaterials such as liposomal chemotherapeutics [3] are approved for the treatment of cancer, their application still strongly relies on passive delivery to tumor cells and the tumor microenvironment. The benefit of furnishing them with multiple ligands for simultaneous delivery into distinct tumor cells could be enormous. For example, Laginha and colleagues [4] showed that the modification of doxorubin-bearing liposomes with ligands directed towards the tumor markers, CD19 and CD20 improved the selectivity of delivery to lymphoma models, thereby improving therapeutic response. Considering the heterogeneity of cancers and tumor markers, we therefore hypothesized that endowing liposome formulations with ligands directed towards more universal tumor markers that allow selective use in a diverse spectrum of candidate tumor types could be beneficial for many patients. Unique target markers to tackle the heterogeneity of tumors include fibroblast activation protein (FAP) and the human epidermal growth factor receptor 2 (HER2). FAP, is a transmembrane glycoprotein which is overexpressed exclusively on activated fibroblasts in healing wounds [5], rheumatoid arthritis [6] and on 90% of tumor associated fibroblasts (TAFs) of a broad spectrum of epithelial tumors [7], but not on healthy tissues. The overexpression of FAP by TAFs aggravates tumor growth and metastasis formation [8], making it a spotlighted target for both diagnosis and therapy of many diseases [9]. In mice and humans, FAP can be targeted with the same targeting antibodies, thanks 3

to an 89% homology between human and mouse amino acid sequence [10]. Opposed to FAP, HER2 is overexpressed on 20-30% of breast cancers [11, 12] and on other cancer types such as gastric [13], ovarian [14], salivary gland [15] and prostate [16] cancers, and is associated with poor therapeutic response, increased cancer relapse with multiple metastasis [17] and patient deaths. Treatment of cancers with high HER2 levels was shown to be more accurate with a combination of chemotherapy and the humanized monoclonal antibody, Trastuzumab [18]. Thus simultaneous targeting of FAP and HER2 conveys potential for addressing the tumor heterogeneity with a unique probe. To elucidate the potentials of such a probe, we therefore prepared activatable, bispecific liposomes with three main peculiarities. First, the liposomes were encapsulated with a high concentration of a near infrared fluorescent (NIRF) dye, which under the encapsulated condition is fluorescence-quenched [19] and only allows NIR-fluorescence imaging after active uptake and degradation by cells. This leads to subsequent dye release and fluorescence activation within the cells, enhancing their fluorescence detection. Fluorescence quenching is exploited in in vivo fluorescence imaging of tumors to distinguish between accumulation of delivered cargo within the extracellular environment of the tumors due to enhanced permeability and retention (EPR) effect [20, 21] and active uptake by the tumor and tumor microenvironment cells. The second peculiarity of the bispecific liposomes is a non-quenched green fluorescent phospholipid embedded within the lipid bilayer. This enables tracking of the intact (quenched) liposomes within target cells. Finally, the third and unique feature of the bispecific liposomes is the presence of single chain variable antibody fragments (scFv) directed to FAP, together with either variable antibody fragments (scFv) or Trastuzumab, directed to HER2. The bispecific liposomes were termed Bi-FAP/HER2-IL and Bi-FAP/Tras-IL which contained the HER2-binding moiety, HER2’scFv and Trastuzumab, respectively. 4

The Bi-FAP/Tras-IL caused a significantly higher therapeutic depletion of HER2 expressing tumor cells than the free Trastuzumab or monospecific Tras-IL. Furthermore, the activatable bispecific FAP and HER2-targeting liposomes showed FAP and HER2-binding properties and enhanced delivery of cargoes into tumor cells and TAFs in vivo, which enabled their fluorescence detection. This delivery was especially interesting, since the cargo was detected in the nuclei of target cells in cell culture and also in vivo in animal studies. Thus, the bispecific Bi-FAP/Tras-IL constitutes a suitable tool for addressing the tumor heterogeneity. Furthermore it could potentially be implemented to manage Trastuzumab resistance which results in part due to the inability of Trastuzumab to enter the nucleus [22] and target nuclear localized HER2 [23].

2. Materials and methods 2.1. Phospholipids used for liposome preparation The following phospholipids: egg phosphatidylcholine (EPC) from Lipoid (Ludwigshafen, Germany),

cholesterol,

Tris(hydroxymethyl)-aminomethane

Tetramethylbutyl)phenyl-polyethylene (Taufkirchen,

Germany),

glycol

(Triton-X100)

(Tris) purchased

and

4-(1,1,3,3-

from

Sigma

1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy

(polyethylene glycol)-2000] (ammonium salt) (mPEG2000-DSPE), 1,2-distearoyl-sn-glycero-3phosphoethanolamine-N-[maleimide(polyethylene (MalPEG2000-DSPE)

and

glycol)-2000]

(ammonium

salt)

1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-1,3-

benzoxadiazol-4-yl) (ammonium salt) (NBD-DOPE) acquired from Avanti Polar Lipids (Alabaster, USA) were used. The near infrared fluorescent dye DY-676-COOH was got from DYOMICs GmbH (Jena, Germany).

5

2.2. Preparation of quenched DY-676-COOH containing liposomes Quenched liposomes were prepared by the film hydration and extrusion method with the lipid composition EPC:Chol:mPEG2000-DSPE at a molar ratio of 6.5:3:0.5 and 0.3 mol% of the lipophilic marker NBD-DOPE. The lipid film was subsequently hydrated with 6.18 mM of the hydrophilic near infrared fluorescent dye, DY-676-COOH in 10 mM Tris buffer pH 7.4. The detailed procedure of dispersion and extrusion of the resulting quenched liposomes (termed LipQ) is reported elsewhere [19, 24]. 2.3. Post insertion of ligand-micelles into preformed LipQ The FAP’scFv and HER2’scFv were overexpressed and purified from periplasmic preparations of Escherichia coli and characterized as described earlier [25]. Trastuzumab was commercially acquired from Hoffmann La Roche (Germany) as a lyophilized powder and reconstituted to a final 2 mg/ml in PBS buffer. The FAP’scFv and HER2’scFv were covalently coupled to MalPEG2000-DSPE micelles by maleimide coupling of reduced ligands (60 min at RT), whereas Trastuzumab was conjugated to PEG2000-NHS-DSPE micelles by amine coupling of protonated ligands (2 h, RT). Aliquots of the target-specific micelles were subjected to SDS-PAGE and evaluated for coupling efficiency using ImageJ version 1.45s. Subsequently, the respective micelles were inserted individually to the preformed LipQ to acquire the respective monospecific liposomes (FAP-IL, HER2-IL and TrasIL) or inserted simultaneously to get the respective bispecific liposomes (Bi-FAP/HER2-IL and Bi-FAP/Tras-IL). Allen and colleagues demonstrated that the post-insertion of micelles into preformed liposomes is time and temperature dependent and has an upper limit of micellar concentration that can be inserted [26, 27]. Based on this, titration experiments were conducted to determine suitable micelle concentrations for the post-insertion procedure (Supplementary data

6

S2). Based on the titration results, 0.1 mol % (with respect to the liposomal lipid) of the Trastuzumab and FAP ligand-micelles were used, whereas 0.3 mol % of the HER2-specific micelles were post inserted. To adjust the size of LipQ used as control in subsequent experiments, bare micelles (reduced MalPEG2000-DSPE) were used for post insertion. The post insertion procedure was conducted by incubation of the respective micelles and preformed LipQ* at 50°C for 30 min in a water bath. The temperature was selected with respect to the respective melting temperatures of the targeting ligands as deduced from dynamic light scattering studies (Supplementary data S1) and titration results (Supplementary data S2). After insertion, residual unbound ligand molecules were removed by gel-filtration on a Sepharose CL4B column (Amersham, Braunschweig, Germany). The resulting liposomes were centrifuged (2 h, 100,000 g, 8°C) using a Beckman XL80 equipped with a SW55Ti rotor and re-dispersed in 1 ml sterile Tris buffer (10 mM, pH 7.4). 2.4. Estimation of the number of ligands conjugated per liposomal vesicle With respect to literature [26, 27] and ligand / lipid titration experiments (Supplementary data S1 / S2), the amount of final phospholipid and protein per liposomal vesicle was estimated under the assumption that no or only marginal lipid and protein loss occurred during liposome preparation. First, the sizes, d of the liposomes were determined by photon correlation spectroscopy. Considering the thickness of the phospholipid bilayer (between 3-5 nm) and the area of the phospholipid head-group of phosphatidylcholine (0.71 nm2) the total number of phospholipids per liposome (Ntot) could be calculated using the following formula:


    = 17.69 ∗  + − 3  2 2

For the post-insertion process, defined amounts (pi) of thioreactive anchor lipid micelles were used. The number of anchor lipids per liposomes (NAL) was calculated as follows: 7


 =


 ∗  100

A 5-fold molar excess of the micelle forming anchor lipids, AL were used for the conjugation to the various antibody ligands (ratioAL to ligand is 5:1). With respect to literature, all the micelles are expected to be inserted into preformed liposomes during the post-insertion procedure. Analysis of the coupling efficiency (ce) between ligands and the anchor lipids via SDS-PAGE (Supplementary data S1) was then used to determine the final amount of proteins coupled per liposome vesicle (NLig) applying the following formula:
 =


 ∗  100 ∗  !"  #

%$2.5. Physicochemical characterization of liposomes All the liposomal formulations were subjected to dynamic light scattering /photon correlation spectroscopy on a Zetasizer Nano ZS (Malvern, Herrenberg, Germany) to deduce their sizes, zeta potentials and polydispersity indices. 2.6. Cryo-transmission electron microscopy Approximately 20 µl of the respective liposomal formulations were subjected to cryotransmission electron microscopy as reported in detail in our previous studies [2]. 2.7. Quantitation of liposomal encapsulated DY-676-COOH The amount of DY-676-COOH encapsulated was quantified as reported in detail earlier [19]. 2.8. Liposomal quenching assay and dye release in serum The fluorescence-quenching and activation potential of the control and target-specific liposomes (200 nmol final lipid concentration) was analyzed in vitro, by measuring the absorption and 8

fluorescence intensity before and after liposomal freeze-damaging as described earlier [19]. The release of the liposomal encapsulated DY-676-COOH in serum was also deduced by measuring the fluorescence emission of liposomal probes before and after incubation in serum for 24 h at 37°C. For this purpose, 200 nmol (final lipids) of the liposomes were diluted in same volumes of fetal calf serum to achieve a total 50% serum. The samples were incubated for 24 h at 37°C, and then subsequently diluted to 100 µl final volume with 10 mM Tris pH 7.4 prior to the measurements. To substantiate the stability and concentration of DY-676-COOH in the liposomes, equivalent amounts were subjected to the freeze-damaging before incubation with serum. 2.9. Preparation of dye-coupled Trastuzumab (Tras-DY) The humanized anti-HER2 monoclonal antibody, Trastuzumab (from Hoffmann La-Roche AG, Germany) was covalently conjugated to the amine reactive N-hydroxysuccinimidyl (NHS) ester of the near-infrared fluorescent dye, DY-652 (DYOMICS GmbH, Jena, Germany) by the amine coupling method as described earlier [28]. 2.10.

Cell culture

The human melanoma cell line, MDA-MB435S and the human breast carcinoma cell line SKBR3 were purchased from the Cell Lines Service, (CLS, Germany) and grown in Dulbecco’s modified Eagle’s medium containing 10% (v/v) fetal calf serum (FCS). Likewise, the human breast carcinoma MCF-7 was got from CLS and grown in DMEM/F12 supplemented with 10% FCS. The human FAP-transfected fibrosarcoma cell line, HT1080-hFAP was cultured in RPMI medium supplemented with 5% (v/v) fetal calf serum (both Life Technologies GmbH, Darmstadt, Germany). All the cells were grown at 37°C in a 5% CO2 and 95% humidified atmosphere.

9

2.11.

Flow cytometry

The binding selectivity of the respective liposomes to FAP- (HT1080-hFAP) and HER2(SKBR3) expressing cells was compared to target negative HEK293 and B16F10 murine melanoma cells. For this, the liposomes (490 nmol final lipid concentration) were added to 2.5 x 105 cells in 1 ml culture media (RPMI containing 5% FCS) and incubated on ice for 2 h in the dark. For competitive blocking experiments, the respective free ligands (FAP’scFv, HER2’scFv or Trastuzumab at 10 µg) were simultaneously added to the cells together with the liposomes. Thereafter, the cells were centrifuged for 5 min at 200 xg, washed three times with 3 ml PBA (PBS, 1% serum albumin) then dispersed in 450 µl PBA and subjected to flow cytometry on an Epics XL-MCL (Beckmann Coulter, Krefeld, Germany). Detection of the binding was based on the green fluorescent NBD-DOPE incorporated in the lipid layer of the liposomes. 2.12.

Effect of Trastuzumab bearing liposomes on cell viability

The effect of free or liposomal conjugated Trastuzumab on cell viability was deduced by cell count with an automated cell counter (Casy cell counter, Innovatis Germany). Briefly, the high HER2-expressing SKBR3 at 1.106 cells were seeded on small tissue culture flasks and grown for 16 h, then treated with 200 nmol (final lipids) of the respective liposomes or 1 µg free Trastuzumab-DY676 for 24 h or 48 h at 37°C. The cells were washed with HBSS and detached with biotase then counted. The relative effects of the probes on cell vitality were presented as relative levels of viable cells detected compared to untreated control which were considered 100% viable. All experiments were performed in duplicates and repeated at least 2 times. 2.13.

Uptake of liposomal probes and confocal microscopic imaging

30,000 cells (HT1080-hFAP, MDA-MB435S, and MCF-7) or 60,000 cells (SKBR3) were grown on poly-L-lysine-coated 8-well culture slides (BD Biosciences) for 16 h. Thereafter, 200 nmol 10

(final lipid) of the respective liposomes or corresponding concentrations of the dye conjugated Trastuzumab (Tras-DY) were added and the cells further cultured for different durations at 37°C or 4°C. All further steps applied for preparation of the slides and for microscopy were as reported in detail earlier [2]. 2.14.

Animals and implantation

All procedures involving animals 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 (Hsd:Athymic Nude-Foxn1nu nu/nu; Harlan Laboratories) weighing approximately 20 g were housed under standard conditions with ad libitum access to mouse chow and water. 2-6 weeks prior to in vivo imaging, xenografts were induced by subcutaneously coimplanting 1.5 x 106 cells (MCF-7), and 3.0 x 106 cells (SKBR3) adjacent to each other on the same mouse. Briefly, the respective cell numbers were pelleted and dispensed in 120 µl cold Matrigel™ (BD Biosciences, Heidelberg, Germany) then injected subcutaneously into the lower back of the mice. To reduce tissue auto-fluorescence during subsequent imaging, the mice were given a low pheophorbide diet C1039 (Altromin Spezialfutter GmbH & Co. KG, Lange, Germany) seven days prior to imaging. All the imaging procedures were performed with anesthetized animals (2% isoflurane). 2.15.

Determination of tumor volumes

The length, width and heights of tumors were measured with a digital caliper and the corresponding volumes calculated in two dimensions according to [29] using the formula: V = π/6 x (length x width x height).

11

2.16.

In vivo whole body near-infrared fluorescence imaging

Mice bearing double subcutaneous xenografts with approximate diameters of 4-10 mm and average tumor volume of 30 mm³ (SKBR3), 73 mm³ (MCF-7) were anesthetized with 2% isoflurane and the respective liposomal solutions in HBSS (20 µmol (lipid) /kg weight) applied intravenously through the tail vein. The animals were immediately imaged with the MaestroTM in vivo fluorescence imaging system (Maestro, CRi Woburn, UK) with the red filter set composed of the excitation range 615-665 nm and a cut-in emission filter (>700 nm). Images were subsequently made every 2 h for 10 h post injection and recorded as time point, t=0 h-10 h post injection (p.i.) and again at t= 24 h – 32 h and at 48 h p.i. Image analysis and all evaluations and extraction of background auto-fluorescence of native animals were performed with the Maestrosoftware according to the user manual. Semi-quantitative levels of the fluorescence intensities of tumors were got by assigning regions of interest (ROI) to each tumor (target) and another on the hind leg region of the animals (muscle/skin = background). Fluorescence intensities were given as average signal (scaled counts/s) and thereby represent the count levels after software-based scaling for exposure time, camera gain, binning and bit depth. As such, the measurements can be directly compared with each other. 2.17.

In vivo confocal microscopic imaging of tumors

Mice bearing double subcutaneous xenografts SKBR3 and MCF-7 were anesthetized with 2 % isoflurane and the skin surrounding the SKBR3 was carefully removed with a sterile scalpel. Thereafter, the respective liposomes (20 µmol (lipid) / kg weight) diluted in HBSS to a 150 µ l volume were applied intravenously through the tail vein. The animals were immediately placed in a temperature controlled cavity of an inverted confocal laser scanning microscope (LSM780, Zeiss GmbH, Jena), with the exposed tumor surface lying on the coverslip of the microscope.

12

Images were acquired with an air corrected 20X plan apochromatic objective. All excitation and acquisition settings were similar to those used for confocal microscopy of cells as described above. In this setup, the tissue autofluorescence is imaged with the blue filter (blue), whereas the liposomal NBD-DOPE (green) and NIRF DY-676-COOH (red) could be acquired without interference by the tissue autofluorescence. 2.18.

Euthanasia

After whole body fluorescence imaging mice were first anesthetized with 2 % isoflurane till they no longer reacted to touch, then euthanized with carbon dioxide and controlled till they completely stopped breathing. 2.19.

Confocal microscopy of freshly isolated tumors

60-70 min post-injection of liposomes, the mice were sacrificed and the tumors were excised and quickly rinsed in sterile PBS. A small piece of the freshly excised tumor was placed with a smooth surface lying on a glass coverslip of the LSM780 confocal microscope, (Zeiss GmbH, Germany) and imaged with similar settings used for the live tumor microscopy as stated above. 2.20.

Ex vivo determination of the biodistribution of liposomes

After imaging, the mice were sacrificed 48 h post injection and the organs and tumors were excised and imaged. Semi-quantitative determination of the respective fluorescence intensities was performed with the MaestroTM imaging system and software as described above. 2.21.

Statistical data

If not mentioned otherwise, student’s t-test or one way annova were used to deduce the level of significance, if normality and equal variance tests were passed. Else, Mann-Whitney-Rank sum

13

test was applied. Every experiment was repeated at least twice. In in vivo animal trials 5 or more animals / group were used. Differences leading to P < 0.05 were considered significant.

3. Results 3.1. Preparation of activatable bispecific liposomes Activatable bispecific FAP / HER2- targeting and control liposomes were prepared by first encapsulating quenched concentrations of the NIR fluorescent dye, DY676-COOH in the aqueous interior by the film hydration method [19]. The preformed quenched liposomes (LipQ*) were subjected to a post insertion process, whereby the targeting ligands were covalently conjugated to micelles (Supplementary data S1) then fused with the quenched liposomes (Figure 1A). The resulting liposomes were termed FAP-IL (with FAP’scFv), HER2-IL (with HER2’scFv), Tras-IL (with Trastuzumab), Bi-FAP/HER2-IL (with both FAP’scFv and HER2’scFv) or Bi-FAP/Tras-IL, which contained both FAP’scFv and Trastuzumab (Figure 1A). The liposomal formulations retained comparable properties to the non-targeted DY-676-COOH loaded liposomes. This was evident in their zeta potentials, polydispersity indices (PDI) and size as determined by dynamic light scattering (Table 1/ Supplementary data S2). Only minimal increases in the size were seen in ligand-targeted liposomes compared to the non-targeted LipQ. This is based on the size contribution by the post-inserted ligand-micelles (Supplementary data S2). Table 1: Properties of the activatable liposomes Parameter

Size [nm]

PDI

Zeta Potential [mV]

LipQ

129.6 ± 2

0.050 ± 0.30

- 18.6 ± 0.40

FAP-IL

139.1 ± 2

0.079 ± 0.02

- 15.4 ± 4.90

HER2-IL

138.7 ± 2.5

0.064 ± 0.01

- 14.6 ± 0.97

Tras-IL

136.0 ± 1.8

0.089 ± 0.02

- 15.8 ± 0.10 14

Bi-FAP/HER2-IL

147.4 ± 4.8

0.096 ± 0.09

- 15.1 ± 1.25

Bi-FAP/Tras-IL

147.0 ± 2.5

0.118 ± 0.02

- 14.1 ± 0.40

Transmission electron microscopy substantiated the size and morphology, showing a predominant unilamellary vesicle pattern, irrespective of the number of ligands conjugated (Figure 1B).

Figure 1: Activatable fluorescent liposomes for image-guided delivery. A) Scheme of the preparation of liposomes, depicting film hydration with DY-676-COOH (abs. / em. 674 nm /699 nm), the respective ligands used and the dedicated terms assigned for the resulting formulations. LipQ* were additionally fused with empty micelles to get LipQ that was used in the work as control probe (not depicted). B) Electron micrographs of the respective liposome formulations reveal predominant unilamellary vesicles.

15

3.2. Fluorescence properties and serum stability of the quenched liposomes Intact DY-676-COOH-containing liposomes revealed double absorption maxima and low fluorescence emission (Figure 2A, blue lines), which is indicative of fluorescence quenching. Upon freeze-damage of the liposomal membrane (slow freezing at -80°C and thawing at 30°C) release and activation of the DY-676-COOH in the surrounding buffer resulted in an increase in the absorption maximum and the disappearance of the blue-shifted peak, giving a single absorption maximum at 660 nm and a 3 fold increase in fluorescence emission. This was observed in all the liposomal formulations, irrespective of the number of ligands conjugated on the lipid bilayer (Figure 2A, red lines). Contrarily, the free DY-676-COOH used at equivalent concentrations of the LipQ dye content, showed only one absorption maximum, and high fluorescence emission, irrespective of freezing at -80°C (Figure 2B, DY-676-COOH). Since the interaction of DY-676-COOH with serum proteins influences its spectroscopic properties, causing almost a 10-fold increase in fluorescence emission (Figure 2B, DY-676COOH, purple lines), the stability of the liposomes in serum was studied based on the release of the liposomal DY-676-COOH into serum and the resulting increase in fluorescence emission. Only minimal release of the DY-676-COOH was detected in all the liposome formulations, following their incubation in serum (Figure 2B, purple lines). The serum caused approximately 35 arbitrary units of fluorescence of the liposomal formulations, except the HER2-IL which revealed a slightly higher release of the dye in serum, leading to about 50 arbitrary units of fluorescence after 24 h incubation. Freeze damaging the liposomes prior to incubation in serum caused approximately 4-5 fold increase in fluorescence emission which was comparable with the emission level detected with equivalent amounts of the free dye used as control (Supplementary data S3). This substantiates the high concentration of the DY-676-COOH encapsulated, and

16

suggests that the HER2-IL is relatively instable compared to the other liposomal formulations. The instability, which is probably based on the HER2’scFv used, was partially compromised in the Bi-FAP/HER2-IL probe, exposing one potential benefit of bispecific targeting.

Figure 2: The liposomes reveal fluorescence activatability and serum stability in vitro. A.) Liposomal formulations (200 nmol final lipids) or the free DY-676-COOH (at a concentration equivalent to encapsulated amount in 200 nmol LipQ used) were recovered from storage at 4°C and diluted to 100 µl in 10 mM Tris buffer pH 7.4 then their absorption and emission maxima were measured (blue lines). Likewise, equivalent amounts of the probes were freeze-damaged at -80°C then thawed at 30°C before dilution in Tris buffer. The freeze-damaged vesicles show a single absorption maximum and a 3-fold increase in fluorescence emission of activated DY-676COOH (red lines). B.) Stability and dye release from the liposomes in serum. Similar concentrations of the liposomes and free DY-676-COOH were incubated in same volumes of fetal calf serum (to a final 50% FCS) for 24 h at 37°C then diluted with Tris buffer to a final 100 µl volume prior to emission measurements (purple lines).

3.3. The bispecific liposomes are selective for FAP and HER2- expressing cells Flow cytometric detection based on the liposomal green fluorescent phospholipid (NBD-DOPE) showed strong binding of the liposomes to target cells. The bispecific liposomes (Bi-FAP/HER2IL and Bi-FAP/Tras-IL) revealed comparable binding efficiency to HER2- and FAP expressing 17

cells (Figure 3A, SKBR3 and HT1080-hFAP), but not the target negative cells (Figure 3A, HEK293 and B16F10), whereby the Trastuzumab-based formulations showed superior binding to HER2 than the HER2’scFv based liposomes. Pre-incubation of the cells with the respective free antibodies or scFv fragments inhibited the binding. Thus, FAP-based binding was successfully blocked with FAP’scFv (Figure 3B, Bi-I +FAP* and Bi-II +FAP*) and HER2-based binding could be blocked with free HER2’scFv or Trastuzumab (Figure 3B, Bi-I +HER2* and Bi-II +Tras*). Likewise, the monospecific formulations FAP-IL and Tras-IL bound the respective target cells, whereas the HER2-IL showed only low binding. Interestingly, the Bi-FAP/HER2-IL revealed a 4-fold higher binding to HER2-expressing cells than the HER2-IL (Figure 3A, SKBR3, compare HER2-IL and Bi-I). Confocal microscopy qualitatively confirmed the selectivity of the respective liposomes for target cells (Figure 3C). In accordance, the FAP-based formulations (FAP-IL, Bi-FAP/HER2-IL, BiFAP/Tras-IL) were strongly taken up and activated by the HT1080-hFAP cells or mildly by the low FAP-expressing melanoma cell line, causing a strong red (DY-676-COOH) and green (NBDDOPE) fluorescence of the cells (Figure 3C, HT1080-hFAP and MDA-MB435S). Likewise, the HER2-specific formulations, Tras-IL and Bi-FAP/Tras-IL were strongly taken up and activated by the high HER2-positive cell line SKBR3, whereas HER2-IL and Bi-FAP/HER2-IL (both bearing the less stable HER2’scFv) were taken up to a lesser degree (Figure 3C, SKBR3). The HER2’scFv based (HER2-IL and Bi-FAP/HER2-IL) binding and processing was relatively rapid, causing predominantly red fluorescence of the HER2-expressing cells in particular, and only minimal residual green fluorescence. No detectable uptake was seen in the low-HER2 expressing cell line (Figure 3C, MCF-7). Also, the non-targeted LipQ and the free DY-676-COOH showed no uptake in any of the cell lines used (Figure 3C, LipQ and DY-676-COOH).

18

Figure 3: The targeted activatable liposomes are selective for respective target cells. Flow cytometric analysis of the binding of respective liposomes (490 nmol final lipid concentration; Bi-I = Bi-FAP/HER2-IL and Bi-II = BiFAP/Tras-IL) to 2.5 x105 cells expressing either FAP (HT1080-hFAP), HER2 (SKBR3) or no targets (HEK293 and B16F10) after incubation for 2 h at 4°C in the absence (A), or after pre-incubation in the presence of free antibodies (B). FAP*, HER2* and Tras* depict 10 µg free FAP’scFv, HER2’scFv or Trastuzumab respectively, n=3 ± SD. C.) Confocal microscopic images of the stable FAP-transfected human fibrosarcoma HT1080-hFAP, the high HER2expressing SKBR3, and the low-HER2 expressing MCF-7 after incubation with 200 nmol of the indicated liposomes, control LipQ or free DY-676-COOH (at a concentration equivalent to the DY676-COOH content in 200 nmol FAPIL) for 8 h at 37 °C.

19

3.4. Trastuzumab conjugated to the liposomal surface retain therapeutic ability to deplete cultured cancer cells To verify whether Trastuzumab retains its therapeutic ability to deplete HER2 expressing cancer cells even after conjugation to the surface of liposomes, time-course experiments and subsequent cell counts and microscopy were performed. A time-dependent and selective increase in fluorescence of the HER2-expressing cells (Figure 4A) as well as FAP-expressing cells (Figure 4B) were detected. Only the less stable HER2-IL and Bi-FAP/HER2-IL revealed poor uptake in the HER2-expressing cells, with no peculiar increase in fluorescence after 4 h incubation (Figure 4, Bi-FAP/HER2-IL and HER2-IL). Free Trastuzumab conjugated to a NIR fluorescent dye (TrasDY) could only be internalized by the HER2-expressing cells (Figure 4C). This substantiates the selectivity of Trastuzumab for HER2 expressing cells and underlines the feasibility of the bispecific liposomes to detect both FAP and HER2 expressing cells based on the presence of both ligands. Interestingly, microscopic investigations revealed that the Tras-IL and Bi-FAP/Tras-IL partly deliver the encapsulated DY-676-COOH into the nuclei of a subset of the HER2 expressing cells (Figure 4A and D, white arrows), whereas the free Tras-DY was predominantly located on the cell membrane and vesicles even after 48 h incubation (Figure 4D). This suggests that conjugation to lipids or liposomes greatly influence the intracellular traffic and nuclei localization of Trastuzumab. We therefore verified if the liposomal-Tras induced internalization to the nuclei impacts tumor cell death. We determined a significantly higher (P < 0.05) level of cell death of the high HER2-expressing cell line, SKBR3 after incubation with the liposomal probes than with the Tras-DY conjugate (Figure 4E). Furthermore, the depletion was distinct for the individual probes, and time dependent as compared to the untreated control. For example the Tras-IL induced a significant level of cell death only after 48 h incubation (P = 0.001 compared 20

to control), whereas the Bi-FAP/Tras-IL caused a higher level of cell death after 24 h (P = 0.02) than after 48 h incubation (P = 0.05). The FAP-IL and the free Tras-DY revealed no significant induction of cell death under the investigation conditions, as compared to untreated control. Furthermore, the Tras-IL revealed a significantly higher (P < 0.001 at 48 h) effect on cell viability than the Tras-DY after 48 h treatment. Although the Bi-FAP/Tras-IL effect on cell viability was not significantly higher than the free Tras-DY, it showed a slight tendency (P = 0.087 / 0.081 at 24 and 48 h respectively) to influence cell viability than the Tras-DY. These results reveal a potential benefit of the conjugation of Trastuzumab to a lipidic backbone for efficient internalization to the nuclei of cells and their subsequent depletion.

Figure 4: Confocal microscopic images of (A) the endogenous HER2-expressing human breast carcinoma SKBR3, and (B) the stable FAP-transfected human fibrosarcoma HT1080-hFAP, after incubation with 200 nmol of the

21

indicated liposomes for the indicated durations at 37°C. The characteristic liposomal green and red fluorescence in target cells substantiates the selectivity and time dependent uptake /activation of the liposomes. (C) Uptake of Trastuzumab-DY (Tras-DY, 12 pmol) in SKBR3 versus HER2-negative HT1080-hFAP cells over time. (D) Opposed to the free Tras-DY uptake, liposomal conjugated Trastuzumab drives delivery of the encapsulated DY-676-COOH into the nuclei of the HER2-positive cells (SKBR3, white arrows). (E) Relative levels of SKBR3 cell depletion by free and liposomal conjugated Trastuzumab. Each bar represents the mean percentage of viable cells (n≥3 ± SD) relative to untreated controls (ctrl) at the given time points post probe (200 nmol Liposomes, 24 pmol Tras-DY) application. The Bi-FAP/Tras-IL showed a significantly higher effect on cell depletion after 24 h (* P=0.02 by One way Annova /Holm-Sidak method) compared to untreated control, whereas both the Tras-IL (*** P=0.001) and BiFAP/Tras-IL (** P=0.05) also showed cell depleting ability after 48 h treatment.

3.5. Liposome-based in vivo imaging of HER2-expressing tumor models in mice Mice co-implanted with the high HER2-expressing SKBR3 (left flank) and the low HER2expressing MCF-7 (right flank) were used for in vivo validation. Following intravenous injection of liposomes, a distinct fluorescence distribution in tumors over time was observed. This was characteristic of the individual binding selectivity of the probes, in relation to the heterogenic nature of the tumors (Figure 5A/B and Supplementary data S4 of whole animals). The bispecific Bi-FAP/Tras-IL, Tras-IL and FAP-IL accumulated in both tumor models with no significant difference (P > 0.05) between the fluorescence intensities of the different tumor models. However, the Bi-FAP/Tras-IL revealed a slightly higher tendency (P = 0.059) to accumulate in the high HER2 expressing tumor model at 2-8 h post injection (Figure 5A/B, Bi-FAP/Tras-IL; compare SKBR3 with MCF-7 / Supplementary data S5). Admittedly, the non-targeted LipQ accumulated in both tumor models and also strongly on the skin (Figure 5A, LipQ), revealing a significantly higher (P < 0.05 by Mann Whitney rank sum test) fluorescence intensity of the low HER2-expressing (MCF-7) than the high HER2 positive 22

tumor model (SKBR3) at the time points 8-24 h post injection (Figure 5, LipQ). Furthermore, the accumulation of LipQ increased with increase in tumor volumes, opposed to the targeted liposomes. This exposes a bias of small tumors by the non-targeted LipQ which is activatable by the phagocytic tumor macrophages. The slight instability of the HER2’scFv-based probes (HER2-IL and Bi-FAP/HER2-IL), and rapid elimination of their NBD-DOPE as observed in vitro, demanded a more detailed analysis to convincingly validate their potentials in vivo. Hence, the data of the in vivo studies are not presented here, but will be reported in a separate article.

Figure 5: Activatable liposomes accumulate distinctly in HER2-expressing xenograft models. (A) Representative intensity scaled NIR-fluorescence images of mice bearing human breast carcinoma models with high HER2-expression (SKBR3, left flank) and low HER2-expression (MCF-7, right flank) at 0 h - 48 h post intravenous application of the indicated probes. (B) Representative semi-quantitative levels of fluorescence intensities of tumors at the given time points post injection of liposomes. Each point depicts the mean fluorescence of the respective

23

tumors (SKBR3 and MCF-7) or muscle (thigh region) per group (n≥5) ± SEM. P = 0.059 for SKBR3 versus MCF-7, at 2-8 h post injection of Bi-FAP/Tras-IL. P < 0.05 for MCF-7 versus SKBR3 at 8 h -24 h post injection of LipQ.

3.6. Targeted liposomes deliver cargoes into distinct target cells of xenografted human breast cancer models. Macroscopic in vivo NIRF imaging revealed characteristic accumulation of liposomes in both the high- and low-HER2-expressing tumor models (SKBR3 and MCF-7). Therefore, to pinpoint the subcellular tumor components responsible for the uptake and fluorescence activation of the liposomes, confocal microscopy of freshly isolated tumors were performed. Additionally, the SKBR3 tumors of mice bearing both tumor models (SKBR3 and MCF-7) were subjected to live in vivo confocal microscopy directly after injection of the Bi-FAP/Tras-IL. In the confocal microscopy setup, the tumor components can only be visualized based on liposomal fluorescence and minimal tissue autofluorescence. Thus, the tumor sub-cells could be clearly visualized as strong liposomal green, red or yellow fluorescence which enabled their structural demarcation (Figure 6). Both Tras-based liposomes (Bi-FAP/Tras-IL and Tras-IL) strongly accumulated in the HER2expressing tumor cells and released the NIR fluorescent cargo within the nuclei of the cells (Figure 6, Bi-FAP/Tras-IL orange arrows / Supplementary video V1). In addition, there was detectable fluorescence in tumor associated fibroblasts (Figure 6, pink arrows) and also the cells of some perivascular regions (Figure 6, white arrows). At this early time point (70 min post injection), a higher accumulation is observed in the high HER2 expressing (SKBR3) than the low HER2 expressing (MCF-7) tumor model. Furthermore, a faster delivery and activation of the cargo in the nuclei, seen as red fluorescence, was observed for the bispecific Bi-FAP/Tras-IL compared to the monospecific Tras-IL. 24

Interestingly, the FAP-IL selectively accumulated in the tumor fibroblasts and pericytes, as evident in a strong red fluorescence of elongated cells which are abundantly present in both tumor models (Figure 6, FAP-IL, pink arrows). In contrast, the non-targeted LipQ is only seen as diffusedly distributed greenish vesicles or in few spotted cells in both tumor models (Figure 6, LipQ, red arrows).

Figure 6: Post-mortem fresh tissue microscopy pinpoints the localization of liposomes within tumors. The high HER2 (SKBR3) and low HER2 expressing (MCF-7) tumors were excised 70 min after intravenous injection of the indicated liposomes (20 µmol/kg body weight). The tumors were imaged immediately without preprocessing. The strong blue autofluorescence enables partial visualization of tumor components, whereas the strong green and red fluorescence of liposomal NBD-DOPE and DY-676-COOH ease demarcation of the areas of uptake and activation of liposomes. FAP-IL is seen as intact vesicles extravasating from tumor vessels (yellowish green) and as red fluorescence in tumor associated fibroblasts (pink arrows). The Bi-FAP/Tras-IL and Tras-IL) are activated in the tumor cells (orange arrows) and some regions of the tumor vessels of both tumors (white arrows). The control LipQ is seen as diffused greenish vesicles and partially spotted in some distinct cells (red arrows).

25

3.7. Biodistribution and elimination of the liposomes We verified the distribution of the liposomal DY-676-COOH in mice organs and tumors excised 48 h post injection in order to know if this is altered by bispecific targeting. All the liposomes caused high NIR fluorescence of the kidneys, gallbladder and gastro-intestinal tract at this time point (Figure 7). This was evidence for both urinary and hepatobiliary elimination. A comparison of the distribution of the targeted-liposome based fluorescence with that of non-targeted LipQ showed that liposomal targeting does not alter the elimination properties of DY-676-COOH.

Figure 7: Biodistribution of mono- and bispecific liposomes in tumor-bearing mice. Organs and tumors from mice injected with the indicated probes (20 µmol /kg body weight for 48 h) were excised after euthanasia and

26

subjected to NIRF imaging (A), and then the semi-quantitative levels of their fluorescence intensities were deduced (B). Each bar depicts the mean of fluorescence intensities (n=5) ± SEM.

4. Discussion 4.1. Physical parameters of the synthesized probes We designed activatable bispecific FAP / HER2- targeting liposomes loaded with quenched concentrations of the NIRF dye, DY676-COOH and elucidated their feasibility as tools for targeted delivery and optical molecular imaging purposes. The encapsulated DY-676-COOH permits fluorescence imaging solely after release and activation within target cells. Conjugation of both FAP’scFv and Trastuzumab to preformed quenched liposomes (LipQ) by the postinsertion method resulted in bispecific probes, with unique physicochemical properties. Thus their size, zeta-potential, polydispersity indices, morphology and fluorescence activatability were not altered compared to the non-targeted LipQ. In accordance, freeze damaged liposomes released the encapsulated DY-676-COOH into surrounding buffer and caused their activation and a 3-fold increase in fluorescence emission in vitro. We demonstrated in previous studies [19] that the DY-676-COOH self-quenching at high concentrations is based on interaction of the dye molecules with each other, leading to fluorescence resonance energy transfer (FRET). Thus, dilution of intact liposomes in aqueous buffers that have no influence on liposomal stability revealed double absorption peaks which substantiate their quenched state [30]. Interestingly, the interaction of DY-676-COOH with serum proteins influences its spectroscopic properties causing almost a 10-fold increase in fluorescence emission, similar to other known fluorescent dyes [31]. Hence, the stability of the liposomes which was based on the release of the liposomal DY-676COOH in serum could be elucidated by spectroscopic measurement of fluorescence emission. Except for the HER2-IL which revealed a slightly higher instability, all the other liposome 27

formulations showed only a marginal release of the DY-676-COOH in serum, substantiating their stability. The relative serum instability of the HER2-IL despite the 5 mol % PEGylation is probably related to the HER2’scFv. Hence, its presence in the Bi-FAP/Her2-IL was partially compromised, exposing one potential benefit of bispecific targeting. 4.2. Target selectivity of bispecific liposomes in vitro on cultured cells The target selectivity and intracellular processing and activation of the monospecific or bispecific liposomes are strongly related to the individual properties of the targeting ligands used. Hence, a lower binding, but much rapid processing of the HER2’scFv-based probes (HER2-IL and BiFAP/HER2-IL) was evident in low number of the high-HER2 expressing cells detected by flow cytometry, and a predominant red fluorescence of the target cells detected microscopically. In contrary, FAP-based binding of the bispecific Bi-FAP/HER2-IL to high-FAP expressing cells was highly comparable to the monospecific FAP-IL or bispecific Bi-FAP/Tras-IL binding, and revealed a characteristic liposomal green and red fluorescence of the cells. This strengthens the fact that the slight instability is related to the HER2’scFv ligand. Previous studies revealed that the liposomal phospholipids are recycled back to the cell membrane and released from cells [2, 19]. This, coupled with the fact that all the cancer cell lines used herein cannot take up the free DY-676-COOH (see Figure 3C, DY-676-COOH) strongly indicates that the HER2’scFv-based liposomes are taken up partially and processed by HER2 positive cells quite rapidly, leading to a maximum liposomal (green) fluorescence of HER2-target cells after 2 h incubation. This was opposed to the bispecific Bi-FAP/HER2-IL which revealed a maximum fluorescence after 4 h and 8 h incubation on high-HER2 or high-FAP expressing cells respectively. Other researchers showed a rapid uptake of HER2’scFv, which hindered penetration and a homogenous distribution of HER2-targeted probes in tumor models in mice [32]. Opposed to the monospecific probes, the 28

bispecific liposomes, especially the Bi-FAP/Tras-IL combine promising features of the FAP and HER2 binding moieties which would be of advantage for in vivo detection of at least one of the targets in heterogenic tumor types. Furthermore, both Tras-based liposomes could deliver the encapsulated cargo into the nuclei of a subset of the high HER2-expressing cells, and also cause a more significant level of cell death than the free Tras-DY compared to non-treated cells. This observation suggests that lipids or at least liposomes could serve to improve the delivery of Trastuzumab to the nuclei of target cells. Accumulating evidence revealed that membrane bound HER2 translocates to the nuclei of cancer cells [23] and that the inability of Trastuzumab to target this nuclear HER2 is one major causative of Trastuzumab resistance [22] which arises in many patients in the course of therapy. 4.3. Target selectivity and potential applications of bispecific liposomes in vivo Thanks to their fluorescence quenching and activatability by target cells, the potentials of the bispecific liposomes could be reliably validated in in vivo studies. Opposed to non-quenched, always fluorescent probes which would be detected in the extracellular compartments of tumors after their passive accumulation by the EPR effect, the intact quenched liposomes that accumulate in the tumors by EPR effect alone are not detected by near-infrared imaging. The probes need to be taken up by cells, be degraded, whereby they release the NIRF dye which then undergoes fluorescence activation and subsequent detection. In the in vivo situation, many components of the tumor microenvironment such as macrophages, tumor vascular density and fibroblasts contribute to tumor plasticity, and influence their targeting with drugs or imaging probes. Therefore, the bispecific liposomes would simultaneously address the tumor cells (based on HER2 and /or FAP), tumor fibroblasts (based on murine FAP of host mouse) [2] and macrophages (based on phagocytic uptake) [19]. Accordingly, the in vivo data reveals the ability 29

of the bispecific Bi-FAP/Tras-IL to accumulate in tumors leading to fluorescence intensities similar to monospecific probes. However, remarkable differences could be seen in the subcellular distribution of the liposomal fluorescence within excised tumors. Thus, delivery of the encapsulated cargo into the nuclei of tumor cells, tumor fibroblasts and pericytes was more peculiar with the Bi-FAP/Tras-IL than the monospecific probes, whereas the control LipQ revealed slow delivery in few spotted cells but not the cell nuclei. This substantiated the observations made in in vitro studies with the Tras-based probes. Thus, the bispecific BiFAP/Tras-IL could be useful for simultaneous delivery of Trastuzumab and also therapeutic cargoes into the nuclei of heterogenic tumors in cases where active drug delivery to these cell entities is desired. This is advantageous over using the non-targeted or monospecific formulations, which will target only one (e.g. macrophages by LipQ) or two cell types (e.g. fibroblasts or tumor cells and macrophages by FAP-IL or Tras-IL respectively) within the tumors. For example, the FAP-IL specifically localized in tumor fibroblasts and pericytes of both tumor models. The results concretize the antibody cross reactivity between human and murine FAP, and suggests that the level of tumor associated fibroblasts of both tumor models are similar. The Tras-IL accumulated predominantly in the tumor cells and revealed only partial delivery into the nuclei of target cells opposed to the Bi-FAP/Tras-IL. Interestingly, both Tras-based probes showed affinity for some perivascular regions of the tumor vessels, which was partially compromised in the bispecific formulation. Other researchers showed the affinity of Trastuzumab to the tumor neovasculature [33, 34]. However, it remains unclear what protein is bound by the Trastuzumab in our mice models, since it is documented that Trastuzumab does not cross react with mouse HER2 protein [35]. Hence, the underlying results suggest that the use of a bispecific liposome, at least as far as Trastuzumab is concerned, is more beneficial over the application of a cocktail of monospecific liposomes as was reported for other targeted liposomes by other groups 30

[4]. Furthermore, the ability of the Bi-FAP/Tras-IL to more efficiently deliver the cargo into the nuclei of target cells makes it a more suitable setup for delivery of therapeutic drugs which act on nuclear targets. Considering that the free Trastuzumab efficiently targets predominantly membrane-bound HER2 which is thought to contribute to resistance [22, 23], our results further suggest that alternating applications of the Tras-based liposomes and the free Trastuzumab would grant beneficial targeting of both the nuclei- and membrane- localized HER2 which will potentially prevent Trastuzumab resistance. We admit that detection of the tumors with the non-targeted LipQ was possible. However, imaging with the LipQ was biased, such that larger tumors were preferentially detected. The fact that LipQ undergoes predominant phagocytic uptake and activation [19], suggested a higher macrophage burden of the MCF-7 tumors. Interestingly, we could detect higher vascular densities and macrophage abundance in the MCF-7 tumor model than for the SKBR3 model (result not shown). Hence, non-targeted liposomes such as the ones used clinically [3] would address phagocytic macrophages and not the tumor cells. Although we demonstrated that targeted liposomes are also partially taken up by phagocytosis [36], we envision that a cocktail of the bispecific Bi-FAP/Tras-IL reported herein, together with non-targeted liposomes could suffice as a promising approach to address the tumor heterogeneity in the future. Interestingly, the distribution and overall elimination of the targeted liposomes and non-targeted LipQ was comparable, suggesting that the bispecific targeting does not alter the elimination of the liposomal components. Consequently, a high fluorescence of the liver, gallbladder, kidneys and gastrointestinal tract exposed a predominant hepatobiliary elimination route as reported earlier [36, 37], and further supports the future of Bi-FAP/Tras-IL to manage tumor plasticity.

31

5. Conclusion We designed and characterized the potentials of activatable bispecific FAP- and HER2-targeting liposomes. The probe, termed Bi-FAP/Tras-IL demonstrated potentials to deliver the encapsulated cargo into the nuclei of tumor cells and tumor microenvironment cells in vivo. Thus, they can be implemented as strategic tools to deliver both the targeting Trastuzumab as well as encapsulated therapeutic drugs to manage cancers and Trastuzumab resistance in cancer patients.

Conflict of interest None

Acknowledgments This work was supported by the German Research Foundation (DFG) grants HI-698/10-1 and RU-1652/1-1. C.B. was awarded an IZKF-scholarship by the Jena University Hospital. We are grateful to Adrian Press for his assistance with the in vivo microscopic tumor imaging. We thank Susann Burgold, Yvonne Ozegowski and Julia Goering for excellent technical assistance.

References [1] Rizzo LY, Theek B, Storm G, Kiessling F, Lammers T. Recent progress in nanomedicine: therapeutic, diagnostic and theranostic applications. Current opinion in biotechnology 2013;24:1159-66. [2] Rüger R, Tansi FL, Rabenhold M, Steiniger F, Kontermann RE, Fahr A, Hilger I. In vivo near-infrared fluorescence imaging of FAP-expressing tumors with activatable FAP-targeted, single-chain Fvimmunoliposomes. Journal of controlled release : official journal of the Controlled Release Society 2014;186C:1-10. [3] Gabizon A, Shmeeda H, Grenader T. Pharmacological basis of pegylated liposomal doxorubicin: impact on cancer therapy. European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences 2012;45:388-98.

32

[4] Laginha K, Mumbengegwi D, Allen T. Liposomes targeted via two different antibodies: assay, B-cell binding and cytotoxicity. Biochimica et biophysica acta 2005;1711:25-32. [5] Kalluri R, Zeisberg M. Fibroblasts in cancer. Nature reviews Cancer 2006;6:392-401. [6] Bauer S, Jendro MC, Wadle A, Kleber S, Stenner F, Dinser R, Reich A, Faccin E, Godde S, Dinges H, Muller-Ladner U, Renner C. Fibroblast activation protein is expressed by rheumatoid myofibroblast-like synoviocytes. Arthritis research & therapy 2006;8:R171. [7] Garin-Chesa P, Old LJ, Rettig WJ. Cell surface glycoprotein of reactive stromal fibroblasts as a potential antibody target in human epithelial cancers. Proceedings of the National Academy of Sciences of the United States of America 1990;87:7235-9. [8] Klein-Goldberg A, Maman S, Witz IP. The role played by the microenvironment in site-specific metastasis. Cancer letters 2013. [9] Brennen WN, Isaacs JT, Denmeade SR. Rationale behind targeting fibroblast activation proteinexpressing carcinoma-associated fibroblasts as a novel chemotherapeutic strategy. Molecular cancer therapeutics 2012;11:257-66. [10] Brocks B, Garin-Chesa P, Behrle E, Park JE, Rettig WJ, Pfizenmaier K, Moosmayer D. Speciescrossreactive scFv against the tumor stroma marker "fibroblast activation protein" selected by phage display from an immunized FAP-/- knock-out mouse. Mol Med 2001;7:461-9. [11] Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 1987;235:17782. [12] Fiszman GL, Jasnis MA. Molecular Mechanisms of Trastuzumab Resistance in HER2 Overexpressing Breast Cancer. International journal of breast cancer 2011;2011:352182. [13] Jaehne J, Urmacher C, Thaler HT, Friedlander-Klar H, Cordon-Cardo C, Meyer HJ. Expression of Her2/neu oncogene product p185 in correlation to clinicopathological and prognostic factors of gastric carcinoma. Journal of cancer research and clinical oncology 1992;118:474-9. [14] Vermeij J, Teugels E, Bourgain C, Xiangming J, in 't Veld P, Ghislain V, Neyns B, De Greve J. Genomic activation of the EGFR and HER2-neu genes in a significant proportion of invasive epithelial ovarian cancers. BMC cancer 2008;8:3. [15] Cornolti G, Ungari M, Morassi ML, Facchetti F, Rossi E, Lombardi D, Nicolai P. Amplification and overexpression of HER2/neu gene and HER2/neu protein in salivary duct carcinoma of the parotid gland. Archives of otolaryngology--head & neck surgery 2007;133:1031-6. [16] Craft N, Shostak Y, Carey M, Sawyers CL. A mechanism for hormone-independent prostate cancer through modulation of androgen receptor signaling by the HER-2/neu tyrosine kinase. Nature medicine 1999;5:280-5. 33

[17] Hüsemann Y, Geigl JB, Schubert F, Musiani P, Meyer M, Burghart E, Forni G, Eils R, Fehm T, Riethmuller G, Klein CA. Systemic spread is an early step in breast cancer. Cancer cell 2008;13:58-68. [18] Dent S, Oyan B, Honig A, Mano M, Howell S. HER2-targeted therapy in breast cancer: a systematic review of neoadjuvant trials. Cancer treatment reviews 2013;39:622-31. [19] Tansi FL, Rüger R, Rabenhold M, Steiniger F, Fahr A, Kaiser WA, Hilger I. Liposomal encapsulation of a near-infrared fluorophore enhances fluorescence quenching and reliable whole body optical imaging upon activation in vivo. Small 2013;9:3659-69. [20] Peer D, Karp JM, Hong S, Farokhzad OC, Margalit R, Langer R. Nanocarriers as an emerging platform for cancer therapy. Nature nanotechnology 2007;2:751-60. [21] Matsumura Y, Maeda H. A new concept for macromolecular therapeutics in cancer chemotherapy: mechanism of tumoritropic accumulation of proteins and the antitumor agent smancs. Cancer research 1986;46:6387-92. [22] Cordo Russo RI, Beguelin W, Diaz Flaque MC, Proietti CJ, Venturutti L, Galigniana N, Tkach M, Guzman P, Roa JC, O'Brien NA, Charreau EH, Schillaci R, Elizalde PV. Targeting ErbB-2 nuclear localization and function inhibits breast cancer growth and overcomes trastuzumab resistance. Oncogene 2015;34:3413-28. [23] Wang SC, Lien HC, Xia W, Chen IF, Lo HW, Wang Z, Ali-Seyed M, Lee DF, Bartholomeusz G, OuYang F, Giri DK, Hung MC. Binding at and transactivation of the COX-2 promoter by nuclear tyrosine kinase receptor ErbB-2. Cancer cell 2004;6:251-61. [24] Tansi FL, Rüger R, Rabenhold M, Steiniger F, Fahr A, Hilger I. Fluorescence-quenching of a liposomal-encapsulated near-infrared fluorophore as a tool for in vivo optical imaging. Journal of visualized experiments : JoVE 2015:e52136. [25] Messerschmidt SK, Kolbe A, Muller D, Knoll M, Pleiss J, Kontermann RE. Novel single-chain Fv' formats for the generation of immunoliposomes by site-directed coupling. Bioconjugate chemistry 2008;19:362-9. [26] Allen TM, Sapra P, Moase E. Use of the post-insertion method for the formation of ligand-coupled liposomes. Cellular & molecular biology letters 2002;7:889-94. [27] Iden DL, Allen TM. In vitro and in vivo comparison of immunoliposomes made by conventional coupling techniques with those made by a new post-insertion approach. Biochimica et biophysica acta 2001;1513:207-16. [28] Pauli J, Grabolle M, Brehm R, Spieles M, Hamann FM, Wenzel M, Hilger I, Resch-Genger U. Suitable labels for molecular imaging--influence of dye structure and hydrophilicity on the spectroscopic properties of IgG conjugates. Bioconjugate chemistry 2011;22:1298-308. [29] Feldman JP, Goldwasser R, Mark S, Schwartz J, I O. A Mathematical Model for Tumor Volume Evaluation using Two-Dimensions. J Appl Quant Methods 2009;4:8. 34

[30] Stark B, Pabst G, Prassl R. Long-term stability of sterically stabilized liposomes by freezing and freeze-drying: Effects of cryoprotectants on structure. European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences 2010;41:546-55. [31] Uchiyama S, Takehira K, Yoshihara T, Tobita S, Ohwada T. Environment-sensitive fluorophore emitting in protic environments. Organic letters 2006;8:5869-72. [32] Adams GP, Schier R, McCall AM, Simmons HH, Horak EM, Alpaugh RK, Marks JD, Weiner LM. High affinity restricts the localization and tumor penetration of single-chain fv antibody molecules. Cancer research 2001;61:4750-5. [33] Baker JH, Lindquist KE, Huxham LA, Kyle AH, Sy JT, Minchinton AI. Direct visualization of heterogeneous extravascular distribution of trastuzumab in human epidermal growth factor receptor type 2 overexpressing xenografts. Clinical cancer research : an official journal of the American Association for Cancer Research 2008;14:2171-9. [34] Bains LJ, Baker JH, Kyle AH, Minchinton AI, Reinsberg SA. Detecting vascular-targeting effects of the hypoxic cytotoxin tirapazamine in tumor xenografts using magnetic resonance imaging. International journal of radiation oncology, biology, physics 2009;74:957-65. [35] Pegram M, Ngo D. Application and potential limitations of animal models utilized in the development of trastuzumab (Herceptin): a case study. Advanced drug delivery reviews 2006;58:723-34. [36] Tansi FL, Rüger R, Böhm C, Kontermann RE, Teichgraeber UK, Fahr A, Hilger I. Potential of activatable FAP-targeting immunoliposomes in intraoperative imaging of spontaneous metastases. Biomaterials 2016;88:70-82. [37] Kossatz S, Mansi R, Behe M, Czerney P, Hilger I. Influence of d-glutamine and d-glutamic acid sequences in optical peptide probes targeted against the cholecystokinin-2/gastrin-receptor on binding affinity, specificity and pharmacokinetic properties. EJNMMI research 2013;3:75.

35

Activatable bispecific liposomes bearing fibroblast activation protein directed single chain fragment / Trastuzumab deliver encapsulated cargo into the nuclei of tumor cells and the tumor microenvironment simultaneously Felista L. Tansia,*, Ronny Rügerb,*, Claudia Böhma, Frank Steinigerc, Roland E. Kontermannd, Ulf K. Teichgraebera, Alfred Fahrb & Ingrid Hilgera* This work demonstrates the design of activatable bispecific liposomes aimed to target HER2, a poor prognosis tumor marker in many tumor types, and fibroblast activation protein (FAP), a universal tumor marker overexpressed on tumor fibroblasts and pericytes of almost all solid tumors. Encapsulating liposomes with a quenched concentration of a NIRF dye which only fluoresced after cellular degradation and activation enabled reliable visualization of the destination of the cargo in cells and animal studies. Conjugating single chain antibody fragments directed to FAP, together with Trastuzumab, a humanized monoclonal antibody for HER2 resulted in the activatable bispecific liposomes. In animal models of xenografted human breast tumors, the remarkable ability of the bispecific probes to simultaneously deliver the encapsulated dye into the nuclei of target tumor cells and tumor fibroblasts could be demonstrated. Hence, the bispecific probes represent model tools with high significance to address tumor heterogeneity and manage Trastuzumab resistance in the future.

36

37