Dual-colour (near-infrared/visible) emitting annexin V for fluorescence imaging of tumour cell apoptosis in vitro and in vivo

Indocyanine green (ICG) labelled recombinant annexin V proteins (ICG–EGFP–Annexin V and ICG–mPlum–Annexin V) were synthesized for dual-colour fluorescence imaging of tumour cell apoptosis in vitro and in vivo. The ICG-labelled fluorescent annexin V proteins showed dual (near-infrared and visible) fluorescence emissions with binding ability to phosphatidylserines on the plasma membranes of apoptotic cells. Although several types of fluorescence labelled annexin V (e.g. FITC–annexin V, Cy3- and Cy5-annexin V) have been reported, there are no dual-colour (near-infrared/visible) emitting apoptosis-detection probes which can be used in vitro and in vivo. In this paper, the utilities of the dual-colour fluorescent annexin V are demonstrated for in vitro and in vivo fluorescence imaging of the apoptosis of human breast tumour cells induced by an antibody–drug conjugate, Kadcyla. The results suggest that the present annexin V probes will be useful to visualize the action of anti-cancer drugs against tumours both at the cellular and whole-body level.


Introduction
Recent development of antibody-drug conjugates (ADCs) enables specic binding of anti-cancer drugs to tumour cells for apoptosis-based cancer therapy. 1 In vivo visualization of tumour apoptosis induced by ADCs is crucial for the evaluation of action of the drugs in cancer therapy and for the development of cancer therapeutics. 2 For in vivo imaging of ADC-induced apoptosis, we can use several imaging modalities such as single photon emission computed tomography (SPECT), positron emission tomography (PET), and magnetic resonance imaging (MRI). [3][4][5][6][7] SPECT and PET need to use radioactive tracers for in vivo imaging, where their spatial resolution is not enough to visualize apoptotic cells. 5,8 Although MRI enables in vivo three-dimensional imaging with a high spatial resolution (0.1-1 mm), the detection sensitivity of MRI is relatively low. 4 An alternative of the imaging modality for in vivo detection of apoptotic cells is optical imaging. 4,8 Compared to SPECT, PET and MRI, optical imaging has a high tempo-spatial resolution (ms mm À1 ). 8 Herein, we present non-invasive optical imaging of ADC-induced apoptosis of breast tumour cells in vitro and in vivo using near-infrared (NIR) and visible (VIS) emitting uorescent annexin V.
Apoptosis can be induced in cancer cells through intrinsic and extrinsic pathways without inducing an inammatory response. 9 During the early process of apoptosis, morphological changes in the plasma membrane induce the asymmetrical distribution of phosphatidylserine (PS) on the outer monolayer of the membrane. 10 Annexin V, an endogenous protein with binding ability to PS in the presence of Ca 2+ ions, is widely used as an optical probe for the detection of apoptosis. 11 To date, a variety of apoptosis probes based on annexin V modied with uorescence, luminescence, magnetic resonance contrast, and radioisotopes have been developed. [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30] Among these probes, uorescein conjugated annexin V (FITC-Annexin V) 9 is a most popular probe for the detection of apoptotic cells. FITC-Annexin V has been widely used as a uorescent probe for the apoptosis detection by uorescence microscopy and uorescence-activated cell sorting methods. 31 However, FITC-Annexin V is not suitable for in vivo imaging of apoptotic cells due to its VIS uorescence, which is strongly absorbed and scattered by tissues and organs.
Although a few types of NIR-uorescence labelled annexin V (e.g. Cy5-annexin V 29,30 and 800CW-annexin V 27 ) have been reported, there are no NIR and VIS emitting dual-colour probes which can be used for the detection of apoptosis both at the cellular (in vitro) and whole-body level (in vivo). In this work, we developed indocyanine green (ICG) labelled recombinant uorescent annexin V 32 probes (ICG-EGFP 33 -Annexin V and ICG-mPlum 34 -Annexin V, Scheme 1) that enable both in vitro and in vivo uorescence imaging of breast-tumour cell apoptosis induced by an ADC, trastuzumab emtansine 35 (T-DM1, Kadcyla). Kadcyla is a conjugate between a humanized monoclonal anti-HER2 (human epidermal growth factor receptor 2) 36 antibody and maytansinoid (DM1), a highly potent microtubule polymerization inhibitor. 35 Kadcyla acts as an anti-cancer drug against HER2 positive cancer. 37 By using the dual-colour (NIR/VIS) emitting annexin V probes, we achieved in vitro and in vivo imaging of apoptosis of HER2-positive breast tumour cells (KPL-4) treated with Kadcyla. Our ndings suggest that the present dual-colour (NIR/VIS) emitting annexin V can be used for the evaluation of the action of ADCs against cancerous tumours both at the cellular and wholebody level.

Probe design and characterization
For in vivo optical imaging of apoptotic cells, NIR probes emitting over 700 nm are desirable. VIS uorescence (400 to 700 nm) is strongly absorbed and scattered by intrinsic chromophores and organelles in cells. In contrast, NIR uorescence shows higher permeability and lower scattering in living tissues, leading to clearer deep-tissue images. To prepare NIR apoptosis-imaging probes, we labelled recombinant annexin V proteins with ICG ( Fig. 1a, S1 and S2 †). Annexin V is a 39.8 kDa molecule with 22 lysine residues, and it is easily labelled with N-succinimidyl ester of ICG (ICG-NHS). The molecular weight of EGFP-Annexin V and mPlum-Annexin V is 66.7 kDa and 65.5 kDa, where EGFP and mPlum have 19 and 20 lysine residues, respectively. EGFP and mPlum can also be labelled with ICG-NHS ester (Fig. S3 †). From the absorption spectra of ICG-labelled recombinant proteins (Fig. 1b), the number of ICG molecule binding to one recombinant protein molecule was estimated to be 3, 4, and 5 for ICG-Annexin V, ICG-EGFP-Annexin V and ICG-mPlum-Annexin V, respectively. ICG-Annexin V shows an emission peak at 820 nm, while ICG-EGFP-Annexin V and ICG-mPlum-Annexin V show NIR/ VIS dual-emission peaks resulting from ICG and uorescent proteins (Fig. 2a). ICG-EGFP-Annexin V has emission maxima at 515 nm (for EGFP) and 820 nm (for ICG), where the quantum yield of EGFP and ICG emission was estimated to be 35% and ca.1% (Table S1 †), respectively. ICG-mPlum-Annexin V has emission maxima at 640 nm and 820 nm, where the quantum yield of mPlum and ICG emission was 7% and ca.1% (Table S1 †), respectively. To visualize the NIR/VIS dual-emission nature of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V, we measured their uorescence images by excitation at 470 nm for EGFP, 590 nm for mPlum, and 760 nm for ICG. ICG-EGFP-Annexin V and ICG-mPlum-Annexin V show green/NIR dual uorescence and far-red/NIR dual uorescence, respectively (Fig. 2b). This NIR and VIS dual emission property suggests that ICG-EGFP-Annexin V and ICG-mPlum-Annexin V can be used for the VIS and NIR uorescence imaging of apoptotic cells both in vitro and in vivo.
Although ICG is a widely used NIR dye for in vivo uorescence imaging, the drawback of ICG is low photo-and chemicalstabilities in physiological buffer solutions. For instance, ICG degrades in PBS resulting in a loss of absorption and uorescence because of the formation of reduced non-uorescence species. 38 To check the practical usefulness of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V, we examined their photostability in PBS. The time course of uorescence intensity of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V in PBS showed that the uorescence emissions of ICG as well as EGFP and mPlum were not drastically changed for one week (Fig. 3). In addition, we observed that the irradiation of excitation lights (3 mW cm À2 ) to the aqueous solutions of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V did not result in signicant photobleaching ( Fig. S4 †).

Fluorescence imaging of apoptotic breast tumour cells in vitro
For the uorescence imaging of tumour cell apoptosis, we used a human breast tumour cell line (KPL-4), 39 which overexpresses HER2 on its cell surface. To observe the apoptosis of KPL-4 cells, we rst used a widely used FITC-Annexin V as an apoptosisdetection probe. Apoptosis of KPL4 cells was induced by Kadcyla. Aer the treatment of KPL-4 cells with Kadcyla for three days, uorescence imaging of KPL-4 cells was performed by using FITC-Annexin V (Fig. 4a). The KPL-4 cells treated with Kadcyla showed the signicant emission of uorescein (green emission) from the plasma membrane of KPL-4 cells, indicating that the apoptosis of KPL-4 cells was induced by Kadcyla ( Fig. 4a and S5 †). In contrast, the control KPL-4 cells (with no Kadcyla) showed only a weak uorescein emission from the cell membrane Fig. 4a. Flow cytometric analysis showed that the apoptosis of KPL-4 cells was induced by a factor of 76% by treated with 10 nM of Kadcyla for 72 h (a right graph in Fig. 4a).
We next checked whether the apoptosis of KPL-4 cells can be detected by the uorescence of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V. In this case, we performed VIS and NIR dual-colour uorescence imaging of the apoptosis of KPL-4 cells. Fluorescence images of KPL-4 cells by incubated with ICG-EGFP-Annexin V and ICG-mPlum-Annexin V showed the   VIS and NIR intense emissions from the cells treated with Kadcyla, but not from the cells treated with no Kadcyla (Fig. 4b  and c). This result shows that ICG-EGFP-Annexin V and ICG-mPlum-Annexin V can be used as VIS/NIR dual-colour uorescent probes for the detection of apoptosis in vitro. Blocking experiments using unlabelled Annexin V supported the practical usefulness of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V as uorescent probes for the detection of apoptotic cells (Fig. S6 †).

Breast-tumour bearing mice
For in vivo imaging of tumour apoptosis, we used breast-tumour bearing mice. Breast-tumour bearing mice were prepared by implanting of KPL-4 cells to nude mice. 39c Tumour apoptosis was induced by an intravenous injection of Kadcyla to the nude mouse. First, we checked whether Kadcyla can accumulate to a HER2-positive breast tumour in the nude mouse. For in vivo uorescence imaging of Kadcyla, we labelled Kadcyla with ICG, and ICG-labelled Kadcyla was intravenously injected to the mouse via its tail vein. Immediately aer the injection of ICGlabelled Kadcyla, uorescence emission of ICG was observed from the liver (Fig. 5a). Three days aer the injection, intense uorescence emission of ICG was observed from the tumour, indicating the accumulation of Kadcyla to the HER2-positive breast tumour. Ex vivo uorescence imaging of a breast tumour and organs clearly showed the specic accumulation of Kadcyla to the breast tumour (Fig. 5b).

Fluorescence imaging of apoptotic breast tumour cells in vivo
To image tumour apoptosis, we injected ICG-EGFP-Annexin V and ICG-mPlum-Annexin V to the breast tumour-bearing mice treated with Kadcyla for three days (Fig. 6a). Fluorescence imaging of the mice was started three days aer the injection of Kadcyla (Fig. 6a). Fluorescence emission of ICG was observed from a breast tumour one day aer the injection of ICG-EGFP-Annexin V (Fig. S7 †). Aer three days aer the injection of the ICG-EGFP (or mPlum)-Annexin V, signicant ICG uorescence from the breast tumour was clearly observed (Kadcyla (+) in Fig. 6b and c), indicating the accumulation of the Annexin V probes to the tumour. In the VIS region, EGFP or mPlum emission from the tumour was not clearly observed due to the strong emission of tissue autouorescence (Fig. S8 †). ICG-Annexin V also accumulated to the breast tumour treated with Kadcyla (Fig. S9 †). For the control mice treated with no Kadcyla, we observed only weak uorescence emissions resulting from the accumulation of ICG-EGFP-Annexin V and ICG-mPlum-  Annexin V to the tumour (Kadcyla (À) in Fig. 6b and c). The emission intensity ratio between the breast tumours treated with and without Kadcyla was 6.3 for an ICG-EGFP-Annexin Vinjected mouse and 2.4 for an ICG-mPlum-Annexin V-injected mouse (lower images in Fig. 6b and c).
To conrm the specic accumulation of ICG-EGFP-Annexin V and ICG-mPlum-Annexin V to the apoptotic tumour, we performed a control experiment using ICG labelled EGFP (ICG-EGFP) and ICG-labelled mPlum (ICG-mPlum), which have no annexin V moieties.
When ICG-EGFP and ICG-mPlum with no annexin V moieties were injected to breast-tumour bearing mice treated with Kadcyla, signicant uorescence emissions were not observed from the tumours (Fig. S10 †). The control experiment clearly showed that ICG-EGFP-Annexin V and ICG-mPlum-Annexin V accumulated to the tumour by binding of its annexin V moiety to PS molecules on the plasma membrane of apoptotic tumour cells.

Conclusion
In this paper, we present ICG labelled recombinant annexin V probes (ICG-EGFP-Annexin V and ICG-mPlum-Annexin V) for NIR and VIS uorescence imaging of tumour cell apoptosis. The objective of this work is to develop new uorescent annexin V probes which can be used for the detection of apoptosis both at the cellular and whole-body level. Although several types of uorescent annexin V have been reported, there are no uorescent annexin V probes which can be used for the optical detection of apoptotic cells both in vitro and in vivo. By using the present annexin V probes, we demonstrate the capability of VIS and NIR dual-colour imaging for the detection of apoptosis. Deregulation in apoptosis induces numerous diseases such as cancer, cardiovascular, and neurodegenerative diseases. We believe that the present VIS/NIR dual-colour emitting annexin V probes will greatly contribute to the study of apoptosis-related diseases at the whole-body level as well as the cellular level.

Materials and instruments
N-Succinimidyl ester derivative of ICG (ICG-NHS) was purchased from Goryo Chemicals (Japan). Kadcyla was purchased from Chugai Pharmaceutical Co., Ltd (Tokyo, Japan). Breast tumour cells were kindly provided by Dr Kurebayashi (Kawasaki Medical School). Nude mice (ve-week-old female BALB/c nu/nu mice) were purchased from Nihon SLC Inc (Japan).
(b) Expression of the recombinant proteins in E. coli. The pRSET-Annexin V, pRSET-EGFP-Annexin V and pRSET-mPlum-Annexin V plasmids were transformed into E. coli KRX competent cells (Promega). Transformed E. coli were grown as a preculture (2 mL) in LB medium containing ampicillin (100 mg mL À1 ) at 37 C overnight. For large-scale cultures, the overnight culture (2 mL) was grown in 200 mL of LB medium containing ampicillin (100 mg mL À1 ) at 37 C, until they approached to 0.6 of OD 600 (absorbance). To induce production of the targeted protein, isopropyl b-D-1-thiogalactopyranoside (0.2 mM) and Lrhamnose (0.1%) were added to the LB medium, and then incubated with shaking gently for 16 h at 18 C.
(c) Extraction and purication of the recombinant proteins. The cells were harvested by centrifugation at 5000 Â g for 10 min. The pellet was resuspended in 5 mL of binding buffer (50 mM Tris-HCl, 500 mM NaCl, 20 mM imidazole, pH 8.0). Before cell lysis, complete EDTA-free protease inhibitor cocktail tablets (1Â, Roche) were added as a protease inhibitor. The suspension was sonicated on ice. Bursts of 10 seconds with intermediate intensity are repeated 7-10 times with a 10 second cooling period between each burst. The lysate was claried by centrifugation at 20 000 Â g for 30 min to eliminate cell debris. The supernatant was then puried by Ni Sepharose 6 Fast Flow (GE Healthcare). 0.5 mL of Sepharose media equilibrated with binding buffer was added to each 5 mL of lysed sample, and incubated with gentle agitation at 4 C for 60 min. Aer the solution was transferred to an empty column, it was washed with binding buffer ve column volumes. Lastly, the recombinant proteins were drained from the column by the addition of elution buffer (50 mM Tris-HCl, 500 mM NaCl, 500 mM imidazole, pH 8.0). The eluted fractions were further puried by a size-exclusion column (PD-10, GE Healthcare) to exchange buffer.

ICG labelling of Kadcyla and recombinant proteins
ICG-NHS ester (1 mg) was resolved to 1 mL of anhydrous dimethyl sulfoxide. To an aqueous solution (0.01 M Na 2 CO 3 ) of Kadcyla (1 mg mL À1 ) or recombinant proteins (1 mg mL À1 ), 30 mL of a dimethyl sulfoxide solution of ICG-NHS was dropwisely added. The coupling reaction was performed for 30 min at room temperature. The ICG conjugates were puried by a sizeexclusion column (PD-10, GE Healthcare) to remove unreacted dyes.

SDS PAGE
The proteins were run on a 5-20% polyacrylamide gel (Extra PAGE one Precast gel, Nacalai Tesque) in Tris-glycine-SDS buffer, 200 V for 40 min and stained with Coomassie Brilliant Blue (CBB Stain One Super, Nacalai Tesque). The sizes of Annexin V, EGFP-Annexin V and mPlum-Annexin V calculated from the amino acid sequence were 39.8 kDa, 66.7 kDa and 65.5 kDa, respectively. Precision Plus Protein Standard (BIO-RAD) was used as a size marker.
Cellular imaging and ow cytometric analysis KPL-4 cells were plated to 35 mm cell culture dishes (3 Â 10 5 cells per dish) and incubated in Dulbecco's Modied Essential Medium supplemented with 10% fetal bovine serum, 100 mg mL À1 penicillin, and 10 mg mL À1 streptomycin in 5% CO 2 at 37 C for overnight. Then, KPL-4 was exposed to Kadcyla (10 nM or none) for 72 h. All the cells oating in the medium and the cells that had detached during the PBS washing were collected, and the cells attached to the dish were carefully detached by trypsinization and added to the cell suspension. The cells were washed once with PBS and then resuspended in 1 mL of Annexin V binding buffer (Nacalai Tesque).
The cell suspension was divided into 100 mL aliquots, added with FITC-Annexin V (5 ml; FITC-Annexin V Apoptosis Detection Kit, Nacalai Tesque), ICG-EGFP-Annexin V (nal concentration, 0.7 mM) or ICG-mPlum-Annexin V (nal concentration, 0.7 mM) and incubated at room temperature for 15 minutes in the dark. As a blocking control, cell suspensions were preincubated with unlabeled recombinant Annexin V (nal concentration 4 mM) for 15 minutes at room temperature prior to this staining step. Then, 400 mL of binding buffer was added to the stained cell suspension, which was passed through a 35 mm cell strainer before use in the next experiment. Quantication by ow cytometry was performed using the MACSQuant Analyzer (Miltenyi Biotec Inc.). FITC-Annexin V was collected through a FL2 lter (ex: 488 nm, em: 525 AE 25 nm). Fluorescence images were acquired with a uorescence microscope (BZ-X700, Keyence Corp., Japan). The lter set for FITC and EGFP was ex: 470 AE 20 nm, em: 525 AE 25 nm. The lter set for mPlum was ex: 560 AE 20 nm, em: 590 nm LP. The lter set for ICG was ex: 769 AE 20 nm, em: 832 AE 19 nm.

Preparation of tumour-bearing mice
A suspension of KPL-4 cells ($10 7 cells per mouse) was transplanted to the ventral side of 5 week old female BALB/c nu/nu mice. Aer several weeks, we selected a mouse bearing a tumour less than 10 mm in diameter for imaging.

In vivo uorescence imaging of tumour-bearing mice
For the uorescence imaging of a breast-tumour, an aqueous solution (200 mL) of ICG-labelled Kadcyla (1 mg mL À1 ) was intravenously injected via a tail vein of the mouse. NIR uorescence images of the tumour were taken 0, 1, 3, and 5 days aer the injection of ICG labelled Kadcyla. Five days aer of the probe, ex vivo images of a breast tumour and organs were taken.
For the uorescence imaging of tumour-apoptosis, Kadcyla (200 mL, 1 mg mL À1 ) was intravenously injected via a tail vein of breast tumour-bearing mice. Three days aer the injection of Kadcyla, ICG-EGFP (or Plum)-Annexin V or ICG-Annexin V (200 mL, 1 mg mL À1 ) was injected to the mice. Three days aer the probes, uorescence images of a mouse treated were taken.
Fluorescence images were taken using an in vivo uorescence imaging system (Bruker, MS FX PRO). NIR uorescence of ICG was observed at 830 AE 20 nm by excitation at 760 nm. Exposure time of the excitation light was 30 s. VIS uorescence of EGFP and mPlum was observed at 515 AE 20 nm (ex: 470 nm) and 670 AE 20 nm (ex: 590 nm), respectively. Exposure time of the excitation light was 1 s for EGFP emission and 10 s for mPlum emission. Excitation light (400 W Xenon lamp) power was 30 mW cm À2 at a ventral side of a mouse.

Ethical statement
Mice maintenance and animal experiments were performed in accordance with the Guidelines for Care and Use of Laboratory Animals of RIKEN and approved by the Animal Ethics Committee of RIKEN.

Conflicts of interest
There are no conicts to declare.