Lead Structures for Applications in Photodynamic Therapy. 6. Temoporfin Anti-Inflammatory Conjugates to Target the Tumor Microenvironment for In Vitro PDT

Due to the ongoing development of clinical photodynamic therapy (PDT), the search continues for optimized photosensitizers that can overcome some of the side effects associated with this type of treatment modality. The main protagonists being: post-treatment photosensitivity, due to only limited cellular selectivity and post-treatment tumor regrowth, due to the up-regulation of pro-inflammatory agents within the tumor microenvironment. A photosensitizer that could overcome one or both of these drawbacks would be highly attractive to those engaged in clinical PDT. Certain non-steroidal anti-inflammatory drugs (NSAIDs) when used in combination with PDT have shown to increase the cytotoxicity of the treatment modality by targeting the tumor microenvironment. Temoporfin (m-THPC), the gold standard chlorin-based photosensitizer (PS) since its discovery in the 1980’s, has successfully been conjugated to non-steroidal anti-inflammatory compounds, in an attempt to address the issue of post-treatment tumor regrowth. Using a modified Steglich esterification reaction, a library of “iPorphyrins” was successfully synthesized and evaluated for their PDT efficacy.


Introduction
Photodynamic therapy (PDT) is a still developing treatment modality for solid tumors of the head, neck, throat, peritoneal cavity, skin as well as non-oncologic maladies such as psoriasis or age-related macular degeneration [1,2]. The treatment involves the use of a three component system comprising of a photosensitizer, light of specific wavelength and singlet oxygen. The light causes the photochemical generation of cytotoxic oxygen species, among them most prominent 1 O 2 by the photosensitizer within the target tissue. These reactive oxygen species aromatic anhydride 2, it was possible to selectively functionalize the porphyrin periphery to the desired degree of substitution by monitoring the reaction time and equivalents used (Fig 1) [18].
In an attempt to avoid any potential cardiovascular pitfalls caused by coxib based anti-inflammatories, a library of NSAIDs was chosen for the synthesis of these iPorphyrin conjugates [16]. As the carboxylic acid functionality is present in most NSAIDs, and following on from our work on the functionalization of known PSs, 5,10,15,20-tetrakis(3-hydroxyphenyl)porphyrin (m-THPP) and 5,10,15,20-tetrakis(3-hydroxyphenyl)chlorin (m-THPC), it seemed a logical step to attempt the synthesis of these conjugates through esterification reactions. The ester functionality is a bond that readily undergoes enzymatic cleavage and can be easily installed into molecules. There are also wide arrays of NSAIDs that are commercially available and therefore provide a large scope for variation for a library of conjugates.
We wished to use this newly developed methodology to functionalize Temoporfin (m-THPC) [19,20,21], the gold standard of meso-substituted chlorin-based photosensitizers with known molecules used to treat inflammation. Fenamic acid serves as a parent structure to a number of different NSAIDs, e.g., tolfenamic acid and flufenamic acid. However, for an initial screen of the utility of this approach we wished to optimize our synthetic pathway with a cheap and readily available synthetic precursor, i.e. 5-hydroxyanthranilic acid, 4. This molecule contains many of the functional groups present in fenamic acid and other NSAIDs and provides an initial increase in complexity from the simple anhydride used in previous work [18].
The initial screen of conditions began with the use of a simple Fischer esterification reaction to yield the chlorin ester product. A catalytic amount of H 2 SO 4 was used with dry CH 2 Cl 2 and left to heat at reflux overnight. This reaction proved unsuccessful, even at increased temperatures, and it seems that the chlorin macrocycle may be interfering with the electronics of the phenol group by reducing the nucleophilicity of the hydroxy group and therefore inhibiting the reaction with the carboxylic acid.
A Steglich reaction [22] using N,N'-dicyclohexylcarbodiimide (DCC) and 4-dimethylaminopyridine (DMAP) as coupling partners was attempted in DMF at room temperature. Initial results from this reaction appeared promising, however, the yield for the reaction was low and the presence of an insoluble N-acyl urea made purification of the product problematic. A suitable substitute for DCC is 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), a preferred alternate cross-coupling agent due to its solubility in water [23]. This was used in the presence of a catalyst, DMAP, in DMF and at room temperature with no observable esterification occurring. The reaction was reattempted at elevated temperatures and no reaction product was observed by TLC and NMR.
Hydroxybenzotriazole (HOBt) may also be used in conjunction with EDC as a carboxyactivating agent [24,25]. It reacts with the O-acylurea formed from the reaction between the carboxylic acid group of the NSAID and the EDC, to form an activated ester, which can be isolated or used in situ [26,27]. Generation of this activated ester and subsequent addition of the chlorin photosensitizer allowed for the formation of the ester conjugates 8a and 8b. By monitoring the reaction time and equivalents, one is able to maximize the amount of monofunctionalized chlorin in the reaction mixture in comparison to the higher order derivatives. Thin-Layer chromatography showed a distinctive degree of substitution pattern, with the tetrafunctionalized derivative having the highest R f and the unsubstituted chlorin having the lowest. Following on from the success of these reaction conditions, the scope of the synthesis was investigated by using a variety of NSAIDs to create a small library of conjugates. Ibuprofen, (S)-2-(4-(2-methylpropyl)phenyl)propanoic acid 5 and fenamic acid, i.e. 2-(phenylamino)benzoic acid 6 were chosen as suitable initial candidates for the screen due to similarities in structure to anthranilic acid and also their prolific record as agents used in the treatment of numerous inflammatory diseases [28,29,30]. Both NSAIDs readily conjugated to the chlorin scaffold in a controlled and predictable fashion and altering the reaction time and equivalents of reagents used could maximize the degree of functionalization and achieve conjugates 9a, 9b, 10a and 10b in good yields as per the Fig 2. The final section of this library of conjugates was completed with the introduction of Nicotinic acid 7 onto the chlorin periphery. Nicotinic acid, whilst traditionally clinically implemented to influence cholesterol levels, has been shown to have quite promising anti-inflammatory properties in cases of atherosclerosis [31,32]. Unfortunately, nicotinic acid did not readily undergo conjugation with the chlorin photosensitizer under these optimized conditions. However, after investigation into a number of different bases and modified procedures, it was found that pyridine and EDC worked to yield the mono-(11a) and tetrafunctionalized (11b) derivatives in yields comparable to the previous conjugates.
This library has shown that m-THPC is tolerable of standard reaction conditions, workups and chromatography without subsequent oxidation to the parent porphyrin derivative; evidence that contradicts historical perceptions of this molecule. NMR assignment of the mono-functionalized compounds proved extremely difficult due to the presence of a statistical mixture of the possible regiochemical isomers. These compounds were prepared for comparative purposes to gain some initial insight in the influence of the substituent pattern and amphiphilicity on the uptake. These regioisomers are formed because of the reduced double bond located on the pyrrole A ring and they cannot be resolved using standard chromatographic procedures. This mixture is consistent with literature reports of similar systems, however, where possible, the β and inner pyrrolic protons were used as diagnostic peaks [33,34,35]. Thus, high-resolution mass spectrometry (HRMS) analysis acted as our main analytical tool for confirmation of product formation and NMR was used for the assignment of the dominant regioisomer being presented in the experimental section.

Singlet Oxygen Studies
For any molecule to be considered a potential clinical PS, they must possess a high quantum yield of 1 O 2 . To evaluate the 1 O 2 production capability of our library of conjugates, compounds were subjected to a crude assay in order to estimate their 1 O 2 generation [36]. Their production of 1 O 2 was determined through the monitoring of the change in the absorption at 410 nm of the dye 1,3-diphenylisobenzofuran (DPBF) in solution. DPBF has an absorbance maximum at 410 nm and readily undergoes a ring opening reaction in the presence of 1 O 2. One can monitor the reduction of this band over time and graph an estimate of 1 O 2 production. Compounds 9b and 10a were taken as representative samples of the library and screened for their singlet oxygen production capabilities. The results obtained for compounds 9b and 10a from this assay were then compared to results obtained with the known 1 O 2 producer m-THPC to give a relative indication of their 1 O 2 producing potential.
Compounds 9b and 10a were dissolved in an aerated solution of DPBF in DMF and irradiated with white light and the intensity of the DPBF's absorption band was monitored over time using UV-vis spectroscopy. Both compounds showed good production of 1 O 2 , as per Fig 3, however not to the same extent as m-THPC.

Biological Testing
Compounds 8a-11b underwent biological testing to ascertain their localization and cytotoxic properties using esophageal carcinoma OE33, esophagus adenocarcinoma, and well-differentiated SKGT-4 human cell lines (Fig 4). All compounds were successfully taken up into the cell and in accordance with previous research conducted with the same PS core, appear to localize in the endoplasmic reticulum (ER) and Golgi apparatus. The compounds were then subjected to a colorimetric MTS assay to assess their cytotoxicity. The compounds were tested over a range of concentrations with an illumination period of 2 min [37]. Unfortunately, none of the conjugates exhibited any phototoxic effect, which may be due to the relatively short illumination period. As per the graph in Fig 3, m-THPC has a considerably large yield of 1 O 2 after 2 min of illumination, whereas the conjugates tested only begin to become comparable to m-THPC after an illumination period of 12 min.

Conclusions
We have successfully synthesized a small library of chlorin conjugated NSAIDs through the controlled functionalization of the chlorin core through a modified Steglich reaction. This methodology proved to be simple, robust and produced mono-and tetra-functionalized conjugates in reasonable yields and few synthetic steps. This new reaction pathway also highlights the ability of m-THPC, a molecule historically perceived to be difficult to work with, to undergo standard chemical manipulations without subsequent oxidation to its parent porphyrin derivative. All conjugates successfully entered into the cells and whilst no cytotoxicity was observed, this work has served as an initial entry point into the field of PDT and post-treatment inflammation. We hope to fine-tune the cytotoxicity of these porphyrins by synthesizing a larger library of conjugates using more regulatory approved anti-inflammatories. These will then undergo a thorough assessment of their PDT efficacy over a range of irradiation times as suggested by the 1 O 2 production assay data we have obtained above. It is hoped that this approach will further progress the reduction of post-PDT inflammation and tumor regrowth, and therefore improve the overall efficiency of PDT. This library of conjugates could also undergo in vivo experiments in mice in conjunction with researchers in the biochemists. Ideally, these potential collaborators would administer one of the conjugates intravenously and through the use of Western-blot analysis, observe reduced levels of the pro-inflammatory molecules associated with this tumor regrowth.

Materials and Methods
All commercial chemicals used were of analytical grade and were supplied by Sigma Aldrich, Frontier Scientific, Inc. and Tokyo Chemical Industry (TCI) and used without further purification unless otherwise stated. Anhydrous THF and diethyl ether were obtained via distillation Anti-Inflammatory Porphyrin Conjugates over sodium/benzophenone and dichloromethane dried was obtained via distillation over P 2 O 5 . Cell lines were obtained from 1 H and 13 C NMR spectra were recorded on a Bruker dpx 400 (400 MHz for 1 H NMR; 100.6 MHz for 13  The human esophageal squamous cell carcinoma cell line OE21 and the human adenocarcinoma cell lines SKGT-4 and OE33, derived from Barrett's esophagus, were purchased from the European Collection of Cell Cultures (ECACC) [38,39].

Biological Testing
General procedure for cell cultures and the (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) (MTS) cell proliferation assay: Cell lines were seeded at a concentration of 8 × 10 4 cells per mL into sterile 96-well plates, left to attach overnight and treated. To previously prepared 96-well assay plates containing cells in 100 μL of culture medium, the test compounds at different concentrations and appropriate controls were added. After incubation for 24 h the medium was removed and changed for fresh one, dark controls were left in the dark for next 24 h. To assess the phototoxicity, the rest of the plates were illuminated for 2 min and incubated for 24 h. Finally, 20 μL of MTS dye solution was added to each well of the dark controls and illuminated plates and these were incubated for 3 h and the absorbance was recorded at 470 nm using a 96-well plate reader. Cell lines were seeded at a concentration of 3 × 10 4 cells per mL into sterile 96-well plates leaving them for 24 h to attach. For imaging experiments, the cell culture medium was removed, replaced with freshly prepared solutions of the porphyrins 8-11 of various concentrations in the medium and incubated at 37°C under 5% CO 2 for 24 h. After that the medium was removed and the cells were fixed with 4% paraformaldehyde (PFA) in medium and then washed with phosphate-buffered saline (PBS). Fluorescent images were collected and analyzed by high content screening and imaging technique (IN Cell 1000 instrument, GE Healthcare). Full details on the cell cultures and treatment, cell proliferation assay, high content imaging and analysis, statistical treatment, and illumination protocol are as given in Vaz et al. [37].

General Synthetic Procedure A
Monofunctionalization of m-THPC: EDC (2 eq.), HOBt (2 eq.), K 2 CO 3 (2 eq.) were added to a 50 mL round bottomed flask and charged with DMF (10 mL). The appropriate anti-inflammatory agent (2-3 eq.) was added and the reaction was allowed to stir for 60 minutes under argon. m-THPC (1 eq.) was added to the mixture and the reaction was monitored by TLC and once significant conversion to the mono-substituted species was achieved, the reaction was terminated through the addition of CH 2 Cl 2 (50 mL). The crude reaction mixture was washed with distilled water (1×30 mL), sat. aq. NaHCO 3 (1×30 mL), brine (1×30 mL), and distilled water (2×30 mL). The organic phase was dried over anhydrous sodium sulfate, filtered and evaporated under reduced pressure.

General Synthetic Procedure B
Tetrafunctionalization of m-THPC: m-THPC (1 eq.) was added to a round-bottomed flask (50 mL) and charged with DMF (10 mL) and K 2 CO 3 (5-20 eq.). The reaction was left to stir for 60 minutes under argon before the appropriate functionalization reagent (10-20 eq.) was added. The reaction mixture was monitored by TLC and once full conversion to the desired degree of substitution was observed the reaction was terminated through the addition of CH 2 Cl 2 (50 mL). The work-up proceeds as per general procedure A.