1. Introduction

Peryleneiimide (PDI) derivatives have outstanding electronic and photonic properties such as broad band absorption, high quantum efficiency as well as thermal stability [1]. They have been extensively studied in fluorescence sensing devices [2], organic light emitting diodes (OLED) [3], biosensors [4], as well as in vitro and in vivo bioimaging [5]. Each application in which PDI derivatives are utilized requires specific chemical and physical properties. The possibility of easy substitution reactions have also contributed to the interest in PDI derivatives. The precedent PDI derivatives have substituents at bis-N-imide positions and/or four bay regions. This prior research has drawn some plausible structure-property inferences. The substitution of alkyl chain at N-imide position induced organic solubility and reduced quantum yield compared to N,N'-bis(2,6-diisopropylphenyl)-1,6,7,12-tetrachloroperylene-3,4:9,10-tetracarboxdiimide (PDI core) [6]. Further, bay substitution of electron accepting aromatic groups induced red shift in the maximum emission wavelength, and a slight increase in the fluorescence behavior, due to potential steric effects between the core and bay substituents [7].

PDI derivatives are desirable as candidates of fluorescent probes for the bioimaging due to their efficient red fluorescence activity. A challenge in using PDI dyes in bio applications is making them water-soluble to prevent their aggregation in aqueous media whilst retaining their exceptional fluorescent properties. Müllen and associates reported incorporation of ionic moieties onto fluorescent PDIs while Shen and associates incorporated hyperbranched polyglycerols around PDIs [8,9] to try to achieve the above goal. However, these modifications were accompanied by poor solubility due to π-π stacking of PDI core as well as diminished fluorescence signal.

In this study, we report a new fluorescence probe based on a PDI core. To induce a good biocompatibility and high fluorescence of the PDI derivatives in human blood environment, we replaced one chlorine in bay region of N,N'-bis(2,6-diisopropylphenyl)-1,6,7,12-tetrachloroperylene-3,4:9,10-tetracarboxdiimide with oligomeric ethylene glycol moiety. We have investigated optical properties in both aqueous and organic media. The cytotoxicity was evaluated in various cancer cell lines. Following which in vivo and ex vivo fluorescence imaging and organ specific targeting ability of the chromophore was investigated. Our studies verily demonstrate the excellent biocompatibility, stability and fluorescence bioimaging capability of the reported material.

2. Experimental

2.1 Synthesis of flurophore

N,N'-Bis(2,6-diisopropylphenyl)-1,6,7,12-tetrachloroperylene-3,4:9,10-tetracarboxdiimide (PDI core) as a fluorophore molecule was synthesized following reported methods [10]. The oligo(ethylene glycol)methyl ether (OEGME) of number average molecular weight 550 Da was obtained from Sigma-Aldrich. The solution of the PDI core (1 g, 1.18 mmol) in anhydrous tetrahydrofuran (THF) (50 mL) was stirred under N₂ atmosphere. A mixture of OEGME (0.75 mL, 1.485 mmol) and sodium hydride (35.64 mg, 1.485 mmol) in THF was added to the above solution and stirred at 22~24 °C in a round bottom flask under N₂ atmosphere for 2 days isolated from the light. The resulting reaction mixture was evaporated under vacuum. The crude product was purified by silica gel chromatography using 4:1 volume mixture of ethyl acetate:n-hexane as an eluent. The product (PDI-OEG) is obtained in 45% yield (45 mg) as a wine colored solid. ¹H-NMR (300 MHz, CDCl₃, ppm) of PDI-OEG: 8.75 (s, 2H), 8.68 (d, 1H), 8.53 (d, 1H), 7.18 (m, 6H), 3.81-3.72 (m, 4H), 2.82-2.08 (br, Ethylene glycol CH), 1.67-1.42 (br, 24H). MALDI-TOF, m/z: 1362 (100%, M⁺). The uncertainty in the molecular weight of OEGMA has to be taken into account when considering the NMR and mass spectra (Fig. 1).

 

Fig. 1 Synthetic route for PDI-OEG.

Download Full Size | PDF

2.2 Spectral measurements

¹H-NMR spectrum was recorded on an Varian 300 (300 MHz, Agilent Technology, USA).The UV-vis absorption was measured on an UV-3600 (Shimadzu, Korea) and an Agilent 8453 (Agilent Technology, USA). Photoluminescence spectra were measured on an F-7000 fluorescence spectrophotometer (Hitachi, Japan) and an FluoroMate FS-2 spectrophotometer (Scinco, Korea). The fluorescence quantum efficiency (ɸ_FL) of PDI-OEG was measured by using Rhodamine 6G with a quantum efficiency of 0.95 in ethanol following previous reports [11]. Phosphate buffered saline PBS (Welgene ML 008-01, in pH 7.4), fetal bovine serum FBS (12105C) and human serum albumin (HSA) stock solution (A1653) were purchased from Sigma Aldrich and used without further purification. For UV-vis measurements, PBS and FBS were mixed in 1 to 1 ratio. HSA stock solution was diluted with PBS at pH 7.4 to the concentration of 3.07 × 10⁻⁵ M.

2.3 In vitro cytotoxicity test

The cell viability of samples was estimated by 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. MDA-MB 231 cells (human breast adenocarcinoma cell lines) were seeded into the 96-well plate and incubated with 150 μL cell culture media at 37 °C with 5% CO₂ for 2 days. After the incubation, the culture media was carefully removed and the cells were washed one time with phosphate-buffered saline (PBS). MTT solution of 0.4 mL was added to wells containing 3.6 mL of fresh cell culture media, and 50 μL PDI solution was filled into each of the 96-well plates. Then, the cultured cells were incubated at 37 °C with 5% CO₂ for 2 hr. The formazan produced by mitochondrial reductase of the living cells was solubilized by the addition of 150 μL of dimethyl sulfoxide (DMSO) and shaking for 20 min. The values of the plate were measured on a micro-plate reader at 570 nm.

2.4 In vitro cellular uptake studies with tumor cells

In vitro cellular imaging was carried out by three distinct cell lines namely human cervical epithelial carcinoma (HeLa), human breast adenocarcinoma cell lines (MDA-MB 231), and squamous cell carcinoma (SCC-7) cells. HeLa cells were cultured in Dulbecco’s modified Eagle’s medium and MDA-MB 231 and SCC-7 cells were cultured in Rosewell Park Memorial Institute 1640 (RPMI 1640), 10% fetal bovine serum (FBS) and 1% antibiotic-antimycotic (AA). The cells were cultured at approximately 1 × 10⁵ cells in 35 mm culture plate and 2 mL of the applicable cell media was added. The culture plate was placed in incubator at 37 °C with 5% CO₂ for 48 hr. After the incubation for 48 hr, the cells were washed with PBS and replaced with fresh media. For cell staining, 20 μL of the sample was added to 1.98 mL of fresh cell culture media. Culture plate was returned to the incubator. After the incubation for 2 hr with sample, the media was carefully removed and the cells rinsed with PBS. A culture plate was added with fresh PBS and then placed for cellular imaging under a Nuance FX multispectral imaging system (Cambridge Research & Instrumentation Inc., USA) that was set to BP515-560 nm excitation filter.

2.5 In vivo and ex vivo studies for organ targeting with tumor cells

All animal studies were performed by the Animal Care and Use Committee of the Korea Institute of Science and Technology and all handling of mice was performed in accordance with institutional regulations. In vivo and ex vivo experiments were conducted by intravenous injection of PDI-OEG labeled 5 × 10⁶ SCC-7 cells in RPMI 1640 cell culture media in 3-week-old-male Blab/C nude mice (Orient Bio Inc., Korea). A subcutaneous injection of 1 × 10⁷ SCC-7 (squamous cell carcinoma) cells suspended in RPMI1640 cell culture media in 3-week-old male BALB/c nude mice (Orient. Korea) was used to induce xenografts on the mice. In vivo and ex vivo images were taken with IVIS Spectrum Preclinical In Vivo Imaging System (PerkinElmer, USA). After in vivo imaging for 1 day, ex vivo images of resected organs were taken by IVIS Spectrum imaging system with the same condition as used for in vivo imaging. We obtained the wavelength spectrum from fluorescence image on a Nuance 2.10.

3. Results and discussion

3.1 Optical properties

3.1.1 Optical properties in organic media

The UV-vis absorption and emission spectra of the PDI core and PDI-OEG were measured in solvents of increasing polarity in the order of toluene, THF and DMSO. The PDI core has a twisted structure with two halves of the perylenediimide moeities at 42° due to the repulsion of peri-chlorine atoms in the bay regions [1]. The absorption maximum of PDI core was observed around 520 nm, meanwhile that of PDI-OEG was bathochromically shifted to around 550 nm. The absorption peaks of PDI core exhibited profound vibrational structure with respect to those of PDI-OEG. It is obvious that the mono-OEGylation of a chlorine on the bay site of PDIs results in the less twisted molecular geometry and better π-conjugation leading to a change in the molecular structure of PDI core.

The electron donating oligo(ethylene glycol) (OEG) groups change the electronic properties of PDI-OEG with respect to the PDI core. The emission maxima of the PDI core barely exhibited any solvatochromic behavior between toluene and THF, however, in the highly polar solvent of DMSO the fluorescence intensity of the PDI core was dramatically quenched (Fig. 2(a)). This quenching is well-known phenomenon from the J-type self-aggregations of the PDI core units through π-π stacking [2]. This feature has been a key sticking point in its use as a fluorescent probe in in vitro and in vivo bioimaging .

 

Fig. 2 Optical properties of PDI derivatives depending on solvent polarity. (a) Fluorescence spectra of PDI core in different solvents (inset is the enlarged emission in DMSO), and (b) those of PDI-OEG in toluene (black solid line), THF (red solid line) and DMSO (blue solid line). The concentration of each fluorophore is 7.27 × 10⁻⁷ M.

Download Full Size | PDF

In contrast to the PDI core, the emission of PDI-OEG is strongly solvent dependent and shows positive solvatochromism. The emission maxima of PDI-OEG steadily moves to longer wavelengths with increasing solvent polarity (594 nm in toluene, 598 nm in THF, and 619 nm in DMSO) as shown in Fig. 2(b). This feature is responsible for the stabilization of the excited state by the polar solvent compared to that of the ground state of PDI-OEG. It should be noted that relatively PDI-OEG shows a higher fluorescence intensity in DMSO than PDI core alone in DMSO. The fluorescence quantum efficiencies ɸ_FL for PDI-OEG was measured as 0.75 in THF and 0.60 in DMSO. These implies that the less twisted molecular structure of PDI-OEG compared to PDI core leads to lesser fluorescence quenching due to restricted π-π stacking.

In electronic terms the OEGylation of the PDI core induces the change in the electronic properties, due to the electron donating nature of OEG moietiy. In the UV-vis spectra, the absorption cut off of PDI-OEG reveals a smaller band gap than that of PDI core. Additionally OEG chains play an active role in separating the PDI-OEG fluorophores. Further the introduction of hydrophilic OEG chains onto hydrophobic PDI renders PDIs compatible for bioimaging with their high quantum yields in polar and hydrophilic medium.

3.1.2 Optical properties in vitro

The UV-vis absorption and fluorescence spectra of the PDI core and the PDI-OEG were measured in the aqueous phosphate-buffered saline (PBS) as shown in Figs. 3(a) and 3(b), respectively. The fluorescence of the PDI core exhibited bathochromically shifted intense emission with maximum at 669 nm. PDI-OEG was found to be non-fluorescent in PBS accompanied with bathchromic shift compared to that in DMSO (Fig. 2(a)). The enhanced fluorescent band of the PDI core is may be due to its partial self-aggregation to minimize contact with a protic solvent. The fluorescence quenching of PDI-OEG in PBS is distinct from previously observed fluorescent behavior in the polar organic solvent. Such kind of effects may be due to micelle generation. However this possibility can be discarded because the critical micelle concentration (CMC) required for OEG to start forming micelles were far above the concentrations at which the optical properties were studied. We inferred that the observed quenching of fluorescence in PDI-OEG might be due to the increased interactions between the charged phosphate groups in buffer solution and the OEG moieties in PDI-OEG leading to fluorescence quenching.

 

Fig. 3 Optical behaviors of PDI derivatives in PBS, in the FBS in PBS 50:50 buffer, and in PBS containing HSA. (a) Fluorescence spectra of PDI core in PBS. (b) Fluorescence spectra of PDI-OEG in FBS in PBS 50:50 buffer under at a concentration 2.02 × 10⁻⁷ M. (c) Fluorescence spectra of PDI-OEG solution with different concentrations in PBS containing HSA. (d) Sonication time dependent fluorescence spectra of a 2.02 × 10⁻⁷ M solution of PDI-OEG prepared with HSA in PBS buffer. The excitation wavelength for recording the fluorescence spectra was 510 nm.

Download Full Size | PDF

To examine PDI core and PDI-OEG in a solution analogous to blood further studies were carried out in a PBS solution containing 50% fetal bovine serum (FBS). The absorption and emission spectra can be seen in Fig. 3(b). At a glance, their photoluminescent intensities in the blood-like environment were found to counter the trend observed in PBS alone (Fig. 3(a)). The fluorescence of PDI-OEG was far enhanced and that of PDI core was reduced. This can be attributed to the amphiphilicity of PDI-OEG holding both the lipophilic and hydrophilic structures in solution, consequently each part interact with the corresponding favorable domain of proteins on FBS.

The FBS generally consist of peptides which are able to interact with both domains of the fluorescent probe, PDI-OEG. Based on the examination, we expect that the hydrophilic and hydrophobic parts of PDI-OEG to be interacting with corresponding components in the peptides. This phenomenon was further investigated with spectroscopic examination of PDI-OEG in PBS with the human serum albumin (HSA). Human serum albumin is a major component in FBS. Increasing the concentration of the fluorophore PDI-OEG with respect to the PBS/HSA media, the emission intensity gradually enhanced and their maxima slightly shifted to a little longer wavelength (Fig. 3(c)). The fluorescence intensity of PDI-OEG in PBS in presence of HSA increased considerably compared to PDI-OEG solution in PBS. The enhancement of fluorescence of PDI-OEG prepared with HSA in PBS buffer with increased time of sonication can be seen in Fig. 3(d) enhanced fluorescence is an indication of the noncovalent interactions of the amphiphilic PDI-OEG chromophore with hydrophobic and liphophilic domains of the HSA protein.

3.2 Cytotoxicity

Cytotoxicity of the PDI derivative was evaluated by carrying out an enzyme linked immunosorbent assay (ELISA) test using MDA-MB 231 cell lines. MDA-MB 231 cells were treated with PDI derivatives for different time intervals spanning 1 to 24 hours. The viability was tested by MTT assay. The values from the assay were evaluated by a ELISA plate reader at 570 nm. The viable percentage graphs summarizing the cell viability are obtained by comparing MDA-MB-231 cells treated with PDI core and PDI-OEG to control sample containing PDI core/PDI-OEG untreated MDA-MB-231 sample. The viable percentage graphs can be seen in Fig. 4(a). Observations were made for treatment times of 1, 2, 6, 12 and 24 hours. More than 80% of MDA-MB 231 cells treated with the PDI derivatives were maintained during the analysis time indicating low cytotoxicity. Fluorescence images of cells treated with PDI core and PDI-OEG for 2 hours can be seen in Fig. 4(b). Cells treated with PDI-OEG shows brighter fluorescence compared to those treated with PDI core. This can be attributed to the greater cell penetration of the former on account of the OEG group attached to it. The high cell viability and biocompatibility of OEG leads to better staining of cells.

 

Fig. 4 (a) Cell viability of MDA-MB 231 cells treated with PDI derivatives of time interval. (b) Comparison of cellular images of HeLa cells treated with PDI core and PDI-OEG.

Download Full Size | PDF

3.3 Cell staining

The variability of staining across various cell lines was evaluated by incubating PDI-OEG with MDA-MB 231, HeLa and SCC-7 cells for 2 hours. The fluorescence images from these samples can be seen in Figs. 5(a)-5(c). The cells were treated with 7.27 μM concentration of PDI-OEG, rinsed and imaged using filter BP515 at an excitation wavelength of 560 nm. The flurophore showed very good cell penetrability in case of all the tested cell lines. This could be discerned from the bright red fluorescence in the images. The red fluorescence of the dye could be seen in the cytoplasm of all the cell lines, in MDA-MB 231 and SCC-7 cell lines fluorescence was detected in the nuclear region third image to the left in Fig. 5(a) and 5(c). This means that PDI-OEG is capable of crossing into nuclear space in these cell lines. The variability in the activity of PDI-OEG in the studied cell lines may be due to differences in physiological interactions between PDI-OEG and the cell lines [12].

 

Fig. 5 Fluorescence cellular images of (a) MDA-MB 231, (b) HeLa and (c) SCC-7 tumor cells cultured with PDI-OEG dissolved in DMSO for 2 hours.

Download Full Size | PDF

3.4 Organ targeting

In vivo and ex vivo studies were carried out in nude mouse to assess the possibility of delivering, and imaging using PDI-OEG. An intravenous injection of PDI-OEG stained SCC-7 cells (5 × 10⁶/200 μL) was used to introduce the flurophore into the sample mouse. The results were compared with control mouse which did not undergo any treatment. After the injecting PDI-OEG stained SCC-7 cells, the sample as well as the control mice were monitored through in vivo imaging (Fig. 6(a)). During 1 day of recording in vivo images no difference were observed between the sample and control.

 

Fig. 6 In vivo and ex vivo real time fluorescence images using IVIS spectrum imaging system of PDI-OEG. (a) In vivo fluorescence images of Balb/C nude mice intravenous injection of PDI-OEG labeled SCC-7 cells (200 μl/5 × 10⁶). (b) Ex vivo fluorescence images of major organs obtained after in vivo images for 1 day. (c) In vivo fluorescence images of Balb/C nude mice intravenous injection of PDI-OEG labeled SCC-7 cells (200 μl/1 × 10⁷). (d) Ex vivo fluorescence images of major organs obtained after in vivo images for 1 h.

Download Full Size | PDF

The ex vivo fluorescence images of the resected major organs of the sample and the control can be seen in Fig. 6(b). The graph in this figure compares the relative ex vivo fluorescence of the liver from the sample and the control. The sample shows very high fluorescence compared to the control indicating the high efficacy of PDI-OEG staining. The in vivo and ex vivo fluorescence imaging of a mouse injected with PDI-OEG stained SCC-7 cells 1 × 10⁷/200 μL after one hour can be seen in Figs. 6(c)-6(d). There is not much difference between the control and the in vivo images, however the ex vivo images shows a strong localization of red fluorescence in the lung. It is then clear from Figs. 6(b) and 6(d) that the lung gets stained by the fluorophore ahead of the other organs. This demonstrates the potential of PDI-OEG in preferential staining of lung.

The SSC-7 stained with PDI-OEG was introduced into the mouse by an intravenous injection to the tail. The SSC-7 was used to achieve localization of the dye to the xenografted tumor site on the mouse. The ex vivo studies however revealed the dye to be localizing on liver and lung which are internal organs rather than at the subcutaneous SCC-7 xenograft where it was expected to end up. This phenomenon might be leading to the absence of external fluorescence during in vivo imaging. Introduction of the dye into the mouse in solution without using SCC-7 cells as carrier resulted in high local fluorescence at the injection site indicating the non-specificity of staining by the fluorophore. We believe the localization and imaging on tumor can be improved by complexing the molecule with chemical or biological entities capable of targeting specific cancer cells.

4. Conclusions

In conclusion the new fluorescence probe PDI-OEG was synthesized and characterized. They showed a considerable difference in their fluorescence property and cellular imaging capability compared to the PDI core. In a PBS buffer solution containing 50% HSA the PDI-OEG exhibited enhanced fluorescence and stability. We infer that this is derived from noncovalent interactions between its amphiphilic structure with the hydrophilic and lipophilic parts of proteins in HSA. The cellular staining images showed clear difference between the PDI core and the OEG containing PDI-OEG emphasizing the role of OEG group. The fluorophore PDI-OEG showed efficient staining of MDA-MB 231, HeLa and SCC-7 cancer cell lines. The nucleii of MDA-MB 231 and SCC-7 cell lines showed fluorescence indicating the penetration of dye into that location. Cytotoxicity studies on MDA-MB 231 showed an 80% cell viability for up to 24 hours. Ex-vivo studies carried out on a Balb/C nude mice showed localization of PDI-OEG specifically in the lung during the first hour, with proliferation into other organs at longer durations after injections. Through this study we have demonstrated the synthesis and application of a stable amphiphilic red emitting fluorescent dye PDI-OEG and demonstrated its applications in bioimaging.