Photophysical Properties of Pheophorbide-a Derivatives and Their Photodynamic Therapeutic Effects on a Tumor Cell Line In Vitro

1 Department of Chemistry, Yonsei University, 50 Yonsero, Seodaemun-Gu, Seoul 120-749, Republic of Korea 2Department of Microbiology, College of Medicine, Yonsei University, Seoul 120-752, Republic of Korea 3 Department of Chemistry, Kangwon National University, Chuncheon 200-701, Republic of Korea 4Department of Laboratory Medicine and Research Institute of Bacterial Resistance, College of Medicine, Yonsei University, 211 Eonjuro, Gangnam-Gu, Seoul 135-720, Republic of Korea


Introduction
Photodynamic therapy (PDT), a minimally invasive treatment method for a multitude of diseases including cancer, uses reactive oxygen species (ROS) that are generated from photoexcited photosensitizers [1]. After the cancer cells localize the photosensitizer, this photodynamic compound absorbs an appropriate wavelength of light and generates ROS through both energy and charge transfer processes [2]. Among the generated ROS, singlet oxygen is one of the important species that have the high oxidation capacity that damages the cancer cells [3]. The singlet oxygen can be produced through the energy transfer process from the triplet state of the photosensitizer to the triplet oxygen molecules [2]. Therefore, the triplet state quantum yield of a photosensitizer is one of the major factors influencing singlet oxygen generation [4]. There are number of desirable photophysical properties for the photosensitizer agent of PDT application. Not only does the photosensitizer need to have a good cellular affinity [5,6], but also it should possess the high singlet oxygen generation efficiency [7]. In addition, since the red light penetrates the tissue more deeply with a minimal light loss in biological media [8], the photosensitizer that absorbs the red light is also preferred. In that aspect, pheophorbide-a derivatives (PaDs), hematoporphyrin derivatives (HpDs), and indocyanine green (ICG) are considered to be the efficient photosensitizers for the diagnosis and the treatment of tumors [8,9]. Previous studies suggest that PaDs are potentially valuable photosensitizers due to their favored optical properties that include the long wavelength absorption and the high singlet oxygen quantum yields, compared with HpDs and ICG [10][11][12].
In this study, we investigated the photophysical properties of pheophorbide-a derivatives (PaDs) and their photodynamic biological activities in vitro. In particular, the photophysical factors influencing singlet oxygen generation were studied in terms of the substitution effect of the hydroxyl groups and the coil length difference of the alkyl groups of PaDs. Also, for the possibility of PDT application, cellular affinity and photodynamic tumoricidal effect of PaDs were investigated with mouse breast cancer cell line (EMT6) in vitro.

Materials.
Pheophytin-a (PA) and 10-hydroxylpheophytin-a (HPA) were extracted from silkworm excreta. Other PaDs including pheophorbide-a propyl ester (PPE), 10-hydroxylpheophorbide-a propyl ester (HPPE) pheophorbide-a methyl ester (PME), and 10-hydroxylpheophorbide-a methyl ester (HPME) were synthesized and characterized as previously reported [13][14][15][16]. As shown in Figure 1, PA, PPE, and PME were modified by changing the coil length of the substituted alkyl group (PaD-alkyls), and HPA, HPPE, and HPME were modified with the substitution of the hydroxyl group (PaD-OHs). Toluene (Merck, HPLC grade) was used as a solvent for all samples without further purification. Degassed toluene was prepared by ten times of freeze-thaw-pump cycle. Air-saturated toluene was prepared by bubbling air gas for 40 minutes. Concentration of the samples was controlled to be in the range of 10 −5 M.

Spectroscopy
Measurements. Steady-state absorption and emission spectra were obtained using a UV-Vis spectrophotometer (Shimadzu, UV-160A) and a spectrofluorimeter (Hitachi, F-4500), respectively. A N 2 laser-pumped dye laser (Laser Photonics, LN1000, 600 ps FWHM pulse) was used as an excitation source for time-resolved fluorescence. Fluorescence from the samples was collected perpendicular to the excitation beam and detected with a monochromator (Acton Research, SpectraPro) and photomultiplier tube (PMT, Hamamatsu, H5783-04) at the magic angle. The signals were processed by a 500 MHz digital oscilloscope (Hewlett Packard, 54520A) and transferred to a computer for data analysis [17]. Time-resolved triplet-triplet absorption spectra were measured by using a xenon lamp as a probe source. The probe light from the xenon lamp was shaped by an iris and then perpendicularly focused on the sample that was excited by the pump beam from Nd-YAG (BMI series, 7 ns FWHM pulse) pumped optical parametric oscillator (OPO) laser (BMI OP901-355, 5 ns FWHM pulse). The transmitted probe light was collected by focusing optics and then detected with a monochromator (Jobin Yvon, H20) and a PMT (Hamamatsu, R-928). The signals were processed by a 500 MHz digital oscilloscope (Hewlett Packard, 54520A) and then transferred to a computer [18,19]. Triplet quantum yield was determined with various excitation powers by the powerdependent comparative method [18,19].
The Nd-YAG-pumped OPO laser was utilized as an excitation source for the detection of time-resolved singlet oxygen phosphorescence. The phosphorescence signal was collected perpendicular to the excitation beam and detected with a germanium photodiode (EG&G, Judson). An interference bandpass (1270 nm, spectrogon) and cut-off filters (<1000 nm, CVI) were inserted between the sample and the detector window to remove the stray scattering light of laser excitation and the sample fluorescence that interfered with the phosphorescence signal. The signal was acquired by a 500 MHz digital oscilloscope and transferred to a computer. The singlet oxygen quantum yield was directly determined from the intensity of the singlet oxygen phosphorescence signal [19].

Biological Assay.
For evaluation of cell uptake efficiency of photosensitizer, mean fluorescence intensity (MFI) of photosensitizer which was loaded to the cell was measured [20]. The EMT6 cells (1 × 10 6 in a 35 mm tissue culture dish) were incubated with 1.5 mL of DMEM with photosensitizers (0.1 g/mL) for 1 hour at 37 ∘ C. These experiments were performed at the concentration range of PaDs where the nontoxicity was confirmed from previous reports [21,22]. After incubation with photosensitizers, the cells were harvested from 35 mm tissue culture dish. The cells were washed twice with fresh phosphate buffered saline (PBS). Flow cytometry measurement was then performed using a fluorescence-activated cell sorting (FACS) flow cytometer (Becton Dickinson, Canto II) equipped with a single 488 nm argon laser. A minimum of 10,000 cells per sample was acquired in list mode and analyzed using LYSYS II software.
For the photodynamic tumoricidal activity assay in vitro, the photosensitizers were dissolved in pure N, N-dimethyl formamide (DMF) for the stock solution (10 g/mL) and the stock solution was stored in the dark at −20 ∘ C. A light source of irradiation was a 200 W halogen lamp (MVI Micro Video Instruments Inc., MA, USA) with a 515 nm cut-off filter. Total power output of the irradiation was measured using a laser power meter (Metrologic Instruments, Inc., Blackwood, NJ, USA). The irradiation energy was controlled to be 1.2 J/cm 2 . The mouse breast cancer cell line (EMT6 cells, 1 × 10 6 in a 35 mm tissue culture dish) was incubated with 1.5 mL of Dulbecco's modified Eagle's medium (DMEM) and the photosensitizers (0.1 g/mL) for 1 hour at 37 ∘ C in dark condition. After incubation, the cells were exposed to light for PDT. Then, the cells were further incubated at 37 ∘ C for 24 hours in a dark humidified incubator [21]. After the incubation, the cells were harvested from the 35 mm tissue culture dish and 100 L of MTT solution (1 mg/mL) was added to the cell pellets. The sample was incubated at 37 ∘ C in a dark humidified incubator. After 4 hours, the cells were centrifuged and the supernatant was discarded,

Results and Discussion
The steady-state absorption and fluorescence spectra of PaDs are presented in Figure 2. The coil length of the substituted alkyl group in PaD-alkyls (PA, PPE, and PME) did not significantly affect the absorption spectra (not shown in the Figure). However, PaD-OHs (HPA, HPPE, and HPME) show slightly different spectral characteristics. The two peaks at 506 nm and 537 nm correspond to the transitions and the other two peaks at 610 nm and 667 nm are responsible for the transitions. The bands of PaD-OHs are shifted to the shorter wavelengths more than those of PaD-alkyls as shown in Figure 2(a), while the bands of PaD-OHs are consistent with those of PaD-alkyls [24][25][26]. It implies that the symmetric degeneracy of PaDs is influenced by hydroxyl group in the -direction. The fluorescence spectra in Figure 2(b) presents the peaks at 677 and 725 nm and peaks are slightly red shifted due to the structural deformation by the substituted hydroxyl group.
The peak position of fluorescence maximum and the fluorescence quantum yields are listed in Table 1. The fluorescence quantum yields were independent of the concentration and the excitation wavelength, and the aggregation effect was not observed. The fluorescence spectra of all PaDs were similar to each other (data not shown). The fluorescence quantum yields were measured by a comparative method using  a standard reference with the reported value (Φ of H 2 TPP = 0.09 in toluene) as follows [19]: In (1), the subscripts of and represent sample and reference, respectively, while is the refractive index of the solution. is the intensity of the excitation source at the given wavelength. and are the absorbance and the integrated fluorescence area, respectively [27]. The fluorescence quantum yield was reduced when the hydroxyl groups were added. However, the coil length of the substituted alkyl group in PaD-alkyls and PaD-OHs did not influence the fluorescence quantum yield.
Time-resolved fluorescence decay was measured at the peak position of emission band with the excitation wavelength of 507 nm (Figure 3). Fluorescence decay signals were well fitted to a single exponential with an iterative nonlinear deconvolution fitting method. The fluorescence decay rate did not show the detection wavelength dependence for all samples, implying that the emission state was homogeneity [28]. The fluorescence lifetime and the quantum yield of PaD-OHs were faster and smaller than those of PaD-alkyls as shown in Table 1. The fluorescence decay rate was determined using the following relationship: , ic , and isc are radiative, internal conversion, and intersystem crossing rates of 1 state, respectively, and they can be estimated as follows: The distortion in the -conjugated macrocycle ring resulted in low fluorescence quantum yield due to the high intersystem crossing rates. And, therefore, the fluorescence lifetime became fastened. The triplet state quantum yield was determined by the power dependent comparative method [25]: In (4), is a proportionality constant and equals Φ (Φ is a triplet quantum yield, is an absorption coefficient of the transition, and is an instrumental constant). With the absorption coefficient of the ground state and the values for both reference and sample, the triplet state quantum yields of PaDs are estimated as follows [27]: International Journal of Photoenergy 5   In (5), and represent the sample and the reference, respectively. Triplet state quantum yields (Φ ) of PaDs were estimated using the H 2 TPP reference value (Φ = 0.72) in toluene solution (listed in Table 2) [29]. Estimated triplet state quantum yields of PaD-OHs were higher than those of PaD-alkyls. The enhanced triplet state quantum yield of the hydroxyl-substituted PaDs is due to the increased intersystem crossing rate as mentioned above.
Singlet oxygen phosphorescence was measured at the detection wavelength of 1270 nm under air-saturated condition. As shown in Figure 4, the singlet oxygen lifetime was estimated to be 29 s in toluene. It correlates with the reported values of 27∼29 s [29]. Singlet oxygen quantum yields (Φ Δ ) were obtained using the following equation with the reference value of H 2 TPP (Φ Δ = 0.68) [29]: is the phosphorescence signal at time zero and is the absorbance at the excitation wavelength. The indexes and correspond to the sample and the reference, respectively. The singlet oxygen quantum yields and lifetimes are shown in Table 2. The singlet oxygen quantum yields of PaD-OHs were higher than those of PaD-alkyls. The photophysical factors governing the relationship between the triplet state and the singlet oxygen were determined using the following expression [29]: where Φ is the triplet quantum yield of the sample in degassed condition, O 2 is the fraction of the triplet states quenched by oxygen, and Δ is the efficiency of singlet oxygen generation induced by the energy transfer during the oxygen quenching of the triplet state. is the oxygen quenching rate constant of the triplet state and and are the air-saturated and degassed conditions, respectively. The oxygen concentration of the air-saturated solution ([O 2 ] ) of toluene at room temperature was estimated to be 2.1 × 10 −3 M as in [29].
was determined using the measured lifetime of the triplet state in the air-saturated condition. Since the intrinsic decay rate constant of the triplet state ( 0 ) for toluene was reported to be 1.0 × 10 3 s −1 in the literature [29], the oxygen quenching rate constant of the triplet state ( ) was determined to be 1.6 × 10 9 s −1 from (9). Therefore, O 2 with the estimated values of , 0 , [O 2 ] , and (8) resulted in the value of ∼1. Therefore, the enhanced singlet oxygen quantum yields of the substituted hydroxyl groups are originated to increased intersystem crossing rate and the triplet state quantum yield [30,31]. To demonstrate the efficacy of PaDs for the application of the photodynamic therapy agent, PaD-OHs (HPA, HPPE, and HPME) that have higher singlet oxygen quantum yield than PaD-alkyls (PA, PPE, PME) were studied for the cellular affinity and the photodynamic inactivation with mouse breast cancer cells. The cells were treated with PaD-OHs (0.1 g/mL) for one hour of uptake time. The mean fluorescent intensity (MFI) expressed by the cells was examined to determine the cellular affinity of photosensitizers [20]. The shorter coil length of the substituted alkyl group in PaD-OHs significantly enhanced cellular uptake efficiency of pheophorbide-a as shown in Table 3. The similar results were also reported in Rancan et al.'s work [32]. They suggested that, in the substitution group of photosensitizers, more polar groups with less alkyl chains showed more uptake efficiency. As the plausible explanation for this uptake efficiency, the less aggregation of photosensitizers and more efficient penetration through the membranes with the photosensitizer of high polarity were considered. Our results agreed with their work along with the same explanation. Photodynamic tumoricidal effect in vitro was determined using the 3-(4, 5-dimethylthiazol-2yl)-2, 5-diphenyl-2H-tetrazolium bromide (MTT) assay method. As shown in Table 3, HPME that has the short alkyl chain shows higher tumoricidal effect than HPA and HPPE do as expected.

Conclusion
The wavelength-shift in the absorption and emission spectra of PaD-OHs is due to the structural deformation of theconjugated macrocycle ring. Such structural deformation of PaD-OHs results in larger isc values and the reduced fluorescence quantum yield compared with PaD-alkyls. Therefore, it appears that the hydroxyl group substitution in the ring V of pheophorbide-a affects the electronic states of theconjugated macrocycle ring. However, the nonplanarity does not significantly change the ic values that are not sufficiently enough to enhance the nonradiative decay. The reduced fluorescence quantum yields and the fast lifetime are due to the enhanced intersystem crossing rate. On the other hand, the coil length of the substituted alkyl group does not significantly influence the spectroscopic properties of PaDs, suggesting that the coil length of PaDs does not influence the energy states of the photosensitizer.
The results on the cellular affinity and the photodynamic tumoricidal effect indicate that they are affected by the coil-length of the substituted alkyl group in PaD-OHs. It concludes that the hydroxyl substitution group improves the photophysical efficiencies of the photosensitizers and the shorter coil-length of the substituted alkyl group enhances the cellular affinity, and these are the photodynamic tumoricidal effects.