Chlorophyllin Derivatives as Photosensitizers : Synthesis and Photodynamic Properties

Two new photosensitizers (PSs) derived from copper-chlorophyllin were designed to have excitation wavelengths appropriate for the use in photodynamic therapy (PDT) and to have amphiphilic character with positive charge, which favors binding to cell membranes and walls and the intracellular localization in mitochondria. Herein we describe the synthesis and characterization of several properties of these two new PS, i.e., photophysical (absorption, fluorescence and singlet oxygen emission quantum yields, Φf and ΦΔ, respectively), physical-chemical (aggregation) and photobiological (binding, incorporation and cell killing). As expected, the aggregation affected not only the absorption spectra but also lowered considerably the values of Φf and ΦΔ, which could be controlled by the interaction of the PS with aqueous micelles. In vitro studies were performed in cells, mitochondria, and vesicles to determine uptake, membrane binding, cytotoxicity, phototoxicity, and intracellular localization. The positively charged derivatives showed to be considerably more efficient for cell killing than methylene blue.


Introduction
Copper-chlorophyllin (Cu 2+ -Chl) is a commercially available chlorin, non-symmetric and meso-substituted with three carboxyl groups at positions 2 3 , 18 1 and 20 2 .It has several applications, including anti-microbial (against human immunodeficiency virus-HIV and bacteria), antitumoral and food coloring. 1,2It is obtained from chlorophyll by replacing the methyl and phytyl ester groups with alkali and replacing the magnesium with copper. 3he spectroscopic (electronic spectrum and infrared) of Cu 2+ -Chl is very similar to that of chlorophyll.It has an intense Q band at 630 nm, which is ideal for use in photodynamic therapy (PDT), a technique that has been used for the treatment of cancer and other diseases and that is based on the damage of living tissues by visible light in the presence of a photosensitizer (PS) and molecular oxygen. 4,58][9] However, Cu 2+ -Chl does not work as a PS, since the paramagnetic Cu 2+ avoids the formation of long-lived excited states. 10e process of photosensitization in PDT leads to the generation of singlet oxygen and of several radicals (R .) and reactive oxygen species (ROS). 11,12][15] The PS should be non-toxic in the absence of light and should be selectively up-taken by the target tissues. 4,5everal features are being considered in the development of new PSs: molecular targeting to diseased tissues, optimization of the bioavailability and of the photophysical and photochemical properties, decrease in photobleaching and self-aggregation. 7,16,17The efficiency of the PDT is also driven by the main site of intracellular localization, which is related to the structure of the PS. 5,7,17,18n this work, we describe a simple synthetic route to obtain new non-symmetric PSs derived from Cu 2+ -Chl, containing either three tertiary amines or three quaternary ammonium groups (see Figure 1).These new compounds were characterized chemically and their PDT-relevant properties such as photo-physical, intracellular sub-location and photo-toxicity in human adenocarcinoma cell line (HeLa) cells were studied.Our results demonstrate that these compounds have the potential to become improved PDT PSs.
200 mg (0.3 mmol) of compound 2, was dissolved in 4 mL of CHCl 3 and added to 20 mL aqueous HCl 2 mol L -1 .After 4 h, the solvents were distilled off under reduced pressure, thus obtaining the compound 3 (122 mg, 0.2 mmol, 66%) 19 without further purification.After that, 100 mg (0.17 mmol) of the compound 3 was refluxed under a dry nitrogen atmosphere with an excess of fresh oxalyl chloride (2 mL) for 2 h, followed by the distillation of the remaining oxalyl chloride, thus yielding a green residue as a film. 17Finally, this last obtained compound was stirred with an excess of 3-dimethylamino-1-propylamine (2 mL) for 2 h, under a dry nitrogen atmosphere and at room temperature.The amine excess was distilled off under reduced pressure, and the solid was dissolved in DCM and washed with an aqueous solution of ammonium hydroxide (0.5 mol L -1 ).The organic phase was separated and the solvent distilled off to obtain a green solid.The product was purified by column chromatography over basic alumina, (Merck: Brockmann Grade V) using DCM/MeOH = 10:1 (v/v) as eluent, and the obtained precipitate was crystallized with DCM and hexanes to give compound 4 (108 mg, 0.12 mmol) in 76% yield.

Spectroscopic studies
Ground-state absorption spectra were recorded using a SPEX UV-2401 PC Scanning Dual Beam Spectrometer by measuring the sample against its background from 350 nm to 700 nm.Extinction coefficients were determined by measuring absorption as a function of concentration in MeOH.In order to evaluate the state of aggregation, equal volumes of aqueous solutions of compounds 4 and 5 were diluted in a volumetric flask in deionized water, and dissolved with surfactant stock solution to give 10 mmol L -1 sodium dodecyl sulfate (SDS), or 10 mmol L -1 certimoniumn bromide (CTAB).
Fluorescence studies were conducted using a SPEX Fluorolog 1934D spectrofluorimeter.Fluorescence quantum yields (Φ f ) were determined by measuring the area under the curve against a known standard (methylene blue-MB, Φ f = 0.03; excitation 532 nm, emission 555-900 nm) with optical densities of solutions under 0.1.Singlet oxygen measurements were performed using a phosphorescence detection method, utilizing a Continuum Surelite III Nd:YAG laser, exciting at 532 nm.The radiation emitted at 1270 nm by the sample was detected by a Hamamatsu R5509 photomultiplier.To obtain singlet oxygen quantum yields (Φ Δ ), a standard of hematoporphyrin IX (Φ Δ = 0.74) was used in the same solvent (methanol) and excitation as the compounds 4 and 5. [17][18][19][20] n-Octanol-water partition coefficient studies Equal volumes of buffer of specified pH and n-octanol were shaken vigorously for 20 minutes.The phases were allowed to separate until there was mutual saturation between the solvents.The compound was dissolved in the pH buffer or in n-octanol, whichever was more soluble, to an optical density of 1 at the Soret band, complimented with an equal volume of either buffer or n-octanol, and added to the saturated solvent system.The mixture was agitated vigorously for 2 hours followed by centrifugation at 4000 rpm at room temperature for 10 minutes.The absorbance at the Soret band of both phases was determined and analyzed to determine the n-octanol/water partition coefficient. 21tochondrial uptake of photosensitizers Mitochondria were isolated from the livers of rats as described previously. 17Protein concentration was determined by the biuret method in buffer containing 0.250 mol L -1 saccharrose, 10 -2 mol L -1 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 1 mmol L -1 ethylene glycol tetraacetic acid (EGTA), 2 × 10 -3 mol L -1 succinate, 10 -6 mol L -1 sodium phosphate, and 10 -6 mol L -1 of both rotenone and oligomycin.A suspension of 13 mg mL -1 of mitochondria was mixed with 2 × 10 -6 mol L -1 compounds 4 and 5 each and mixed in phosphate buffer for 10 minutes, after which, the suspension was centrifuged for 10 minutes.[18] Liposomal uptake of photosensitizers 46 mg of phosphatidylglycerol was dissolved in 1 mL of CHCl 3 and evaporated by a flow of argon. 2 mL of phosphate buffer was added to the test tube, and the system was shaken vigorously for 3 minutes.The tube was centrifuged for 3 minutes at 14,000 rpm.The supernatant was discarded, and the process was repeated in triplicate.To the precipitate, 1 mL of phosphate buffer was added, and shaken vigorously for 3 minutes.Of this solution, 25 µL was taken and added to 1 mL of a 2 × 10 -5 mol L -1 solution of either compounds 4 and 5.The system was then subjected to mechanical agitation for 30 minutes, followed by rest for 30 minutes.The solution was centrifuged again as above.Concentration incorporated was determined by comparing a control solution of 2 × 10 -5 mol L -1 of the compounds unexposed to the liposomal solution. 17,21,22ll culture studies Human adenocarcinoma cell line (HeLa) was grown according to protocol in 75 cm 2 plastic culture tissue flasks in Dulbecco's minimum eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S).The flasks were maintained with constant humidity at 37 °C with 5% CO 2 . 17

Cellular uptake in HeLa cells
Cells were plated in 1.8 cm diameter Petri dishes at a concentration of 10 6 cells per dish.The cells were allowed to adhere to the dish overnight in fresh media supplemented with 10% FBS and 1% P/S.The media was removed, and fresh media containing human adenocarcinoma cell line (HeLa) was grown according to protocol in 75 cm 2 plastic culture tissue flasks in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S).The flasks were maintained with constant humidity at 37 °C with 5% CO 2 , either compounds 4 and 5 was added.The cells were incubated with the photosensitizer for 3 hours, the medium was removed, and the difference of absorbance versus unexposed DMEM with compounds 4 and 5 was used to compare uptake by HeLa cells. 17

Cellular toxicity studies
HeLa cells were seeded in 1.8 cm diameter Petri dishes and incubated with 2 × 10 -5 mol L -1 compounds 4 and 5, or methylene blue (MB) as above.After 3 hours, the medium with sensitizer was removed and the cells were washed with fresh phosphate-buffered saline (PBS) in triplicate.Cells were then exposed to laser light from a Morgotron laser of 532 nm (for MB, 20 mW) or 650 nm laser (for compounds 4 and 5, 20 mW) in fresh PBS.The laser sheaf was held to the culture dish cover and cells were exposed to the light for 1 minute on, 1 minute off, for a total of 7 minutes of irradiation.The laser dose measured at the surface of the culture dish was 2.7 mJ cm -2 and power of 4 mW for 532 nm laser, and 2.5 mJ cm -2 and power of 3.6 mW for the 650 nm laser.After irradiation, the PBS was removed, and the cells were placed in fresh DMEM with 10% FBS and 1% P/S and maintained overnight.Cellular viability was determined 24 hours later by standard (3-(4,5-dimethylthizol-2-yl)-2,5diphenyltetrazolium bromide (MTT) assay.

Synthesis of compounds 4 and 5
We have synthesized two new porphyrins with an easy synthetic pathway and using reagents that are simple and cost-effective.Compared to other possible sources of lead compounds for this synthesis such as chlorophyll a, Cu 2+ -Chl ( 1) is commercially available, stable, and has low-cost. 3,4The removal of the central copper of Cu 2+ -Chl is crucial, because the paramagnetic nature of copper avoids the formation of the long-lived triplet states. 2 After removing the copper center, the spectral properties of the demetallated product (compound 2) are very similar to those of chlorophyll a, since both compounds have the same chromophoric group (Scheme 1). 1,2The elimination of copper was found to be most effective by the use saturated HCl in methanol, and yielded a shift in the Q4 band from 630 nm to 650 nm in the demetallated, compound 2.Although minor changes in the Soret band remained, the major changes in the Q4 band that resulted from the removal of Cu 2+ , could be accurately attributed to metal complex (Figure 1).We were interested in producing an efficient and low-cost synthesis.Reagents and conditions were optimized from several experimental conditions and only the best conditions are described.As such, we chose reagents that do not produce harmful by-products allowing simple purification schemes and that would also be simple to replicate on a larger scale.To produce the  chlorophyllin acid chloride (compound 3, Scheme 1), we chose oxalyl chloride. 3The bis-amine substituent reagent (N,N-dimethylpropane-1,3-diamine) was chosen to facilitate the nucleophilic attach in the acid chloride and to allow the product with exposed amine side group. 4

Spectroscopic properties
The UV-Vis and fluorescence spectra for both (compounds 4 and 5) in methanol are typical for chlorophyll derivatives, with a strong Soret band at 400 nm and four minor "Q" bands, with Q4 at 650 nm (Figure 1) being the band of major interest for PDT. 2 Both compounds exhibit hyperchromic displacement of 4 nm for Soret band in comparison to Cu 2+ -Chl.Compound 4 showed a Soret band at 400 nm, Q4 at 651 nm and emission at 657 nm.Compound 5 showed a Soret band at 400 nm, Q4 at 650 nm and emission at 653 nm (Figure 1, Table 1).
Chlorophyll and porphyrin derivatives have a large tendency to aggregate, distorting the absorption spectra and decreasing severely the photophysical properties relevant to PDT and micelles are known to control aggregation. 17We analyzed the effects of surfactants on the absorption spectra of both compounds 4 and 5.There is a 7-and 5-fold increase in absorbance at the Soret band of compound 5 in SDS and CTAB, respectively, compared with the absorbance in water (SDS and CTAB are anionic and cationic surfactants, respectively) (Figure 2a).The increase in absorbance was higher for compound 5 in SDS and for compound 4 in CTAB.There was also a substantial increase in fluorescence and decrease resonance light scattering in the presence of micelles indicating the lower photophysical parameters in water due to aggregation and the improvement of these properties in the presence of micelles.The fact that compound 5 interacts more efficiently with SDS micelles and compound 4 with CTAB micelles (absorbance data) indicates that there are two main forces driving this interaction, i.e., electrostatic and hydrophobic.
The fluorescence quantum yields (Φ f ) were determined using methylene blue (MB) as a standard and were about half of that of MB for both compounds 4 and compound 5 (Φ f = 0.013 and 0.014, respectively).Singlet oxygen quantum yields (Φ Δ ) were determined by phosphorescence detection at 1270 nm against a standard of hematoporphyrin IX.Φ Δ values were 0.57 and 0.56 for compounds 4 and 5, respectively, in methanol.Compound 5 in 2 × 10 -5 mol L -1 SDS in D 2 O exhibited a 37% increase in singlet oxygen emission as compared to 2 × 10 -5 mol L -1 of compound 5 in D 2 O.This confirms that the monomer state of the sensitizer is the only state that is an effective PDT agent.

Water to n-octanol partition and viability studies
Water to n-octanol partition coefficients, which were expressed as logP O/W , were used to estimate the photosensitizer's affinity for cell membranes. 5Figure 4a show logP O/W values for compounds 4 and 5 across the relevant pH range.Compound 5 has negative logP O/W values in the whole pH range, being always more hydrophilic than compound 4, which is explained by its cationic structure.logP O/W becomes more negative with the decrease in pH due to protonation of the chlorophyllin ring nitrogens.Compound 4 exhibits hydrophilicity at low pHs, a condition that allows protonation of both the ring and the side-group nitrogens, nearly equal affinity for the organic or aqueous phase from pH 3 to 8, a range in which only the side-group nitrogens are protonated, and is lipophilic at pH > 8, for loosing protonation of the side-group nitrogens.
Photosensitizer incorporation in mitochondria, liposomes, and HeLa cells were determined and are shown in Figure 4b.For compound 4, incorporation was 58%, 28%, and 10% for mitochondria, liposomes, and HeLa cells, respectively.For compound 5, incorporation was 61%, 48%, and 13% for mitochondria, liposomes, and HeLa cells, respectively.Therefore, both compounds interact well with cells and membranes, and compound 5 interacts a bit more favorably with liposomes.The photodynamic efficiency of compounds 4 and 5 were tested in HeLa cell culture and compared with the photodynamic efficiency of MB.Cells cultured and incubated at the same PS concentration (2 × 10 -5 mol L -1 ), were irradiated at 650 nm.HeLa cell viability was determined to be 11.2% for MB, 4.8% for 7.7% for 4 and 5 in relation to a control group (no light), indicating a similar overall phototoxicity for all three compounds with a bit of better efficiency for compounds 4 and 5.
To determine the intrinsic photototoxicity of these compounds we compared the amount of absorbed photons in each case.It was not possible to directly measure the absorption of PS in cells, but we were able to estimate the absorption factor (fraction of incident photons that are absorbed, equation 1) using e (molar absorption coefficient) of each compound and the percentage of incorporation of the PS in cells at 2 × 10 -5 mol L -1 of PS concentration, which was kept constant.At 650 nm, e values are 68,000 mol L -1 cm -1 for MB, 23 3797 and 3584 mol L -1 cm -1 for compounds 4 and 5, respectively.
The fraction of PS incorporation in HeLa cells is 70% for MB, 58% for compound 4 and 61% for compound 5.By multiplying incorporation concentration by e, one obtains the values of absorption, and by using equation 1, absorption factors are calculated which were 0.9 for cells treated with MB and 0.09 for cells treated with the compounds 4 or 5 (Figure 4b).In other words, cells treated with compounds 4 and 5 absorb 10 times less photons than MB at 650 nm.Therefore, if one calculates the efficiency of cell killing divided by the number of absorbed photons (Figure 4c), one realizes that the new compounds induce cell death much more efficient per absorbed photon.The reason for this difference may be due to several factors, including intracellular sub-location and avoidance of bleaching reactions.It should be noted that MB can be reduced in the cellular environment, 24 having photodynamic activity decreased, which does not happen for the chlorophyllins 4 and 5.

Conclusions
New chlorophyllin non-symmetric and positivelycharged photosensitizers were obtained by using inexpensive copper-chlorophyllin as raw material and a simple chemical route.These compounds  absorb reasonable well in the far-red spectral region (λ > 650 nm), are efficient singlet oxygen generators Φ ∆ ca.0.6 and have a strong tendency to interact with membranes.The overall photodynamic efficiency was similar to that of MB.However, the photodynamic efficiency per absorbed photons was considerably higher than that of MB, suggesting that compounds 4 and 5 have optimized properties for use in PDT.

Scheme 1 .
Scheme 1. Scheme of the synthetic route used to prepare the new derivatives of copper-chlorophyllin.

Table 1 .
Spectroscopic properties of compounds 1