PET/PDT Theranostics: Synthesis and Biological Evaluation of a Peptide-Targeted Gallium Porphyrin

The development of novel theranostic agents is an important step in the pathway towards personalised medicine, with the combination of diagnostic and therapeutic modalities into a single treatment agent naturally lending itself to the optimisation and personalisation of treatment. In pursuit of the goal of a molecular theranostic suitable for use as a PET radiotracer and a photosensitiser for PDT, a novel radiolabelled peptide-porphyrin conjugate targeting the α 6 β 1 -integrin has been developed. 69/71 Ga and 68 Ga labelling of an azide-functionalised porphyrin has been carried out in excellent yields, with subsequent bioconjugation to an alkyne-functionalised peptide demonstrated. α 6 β 1 integrin expression of two cell lines has been evaluated by flow cytometry, and therapeutic potential of the conjugate demonstrated. Evaluation of the phototoxicity of the porphyrin-peptide theranostic conjugate in comparison to an untargeted control porphyrin in vitro , demonstrated significantly enhanced activity for a cell line with higher α 6 β 1 -integrin expression when compared with a cell line exhibiting lower α 6 β 1 -integrin expression. development of prep-HPLC methodology to isolate the pure purification by Sephadex G15 as on 69/71 Ga conjugate for facile isolation of the pure theranostic conjugate for HPLC purification. assays was carried out using an Oriel lamp system with a Schott 06515 Long Pass Optical Filter, allowing irradiation with 95-98% of light above 515 nm. RP-HPLC analyses were performed on a Agilent 1100 with a Lablogic NaI gamma detector. The separations were performed on a Gemini C1 8 column, 5µ, 150 x 4.6 mm, 110 Å (Phenomenex, UK) at a flow rate of 1 mL/min, with a mobile phase consisting of 0.1% TFA in water (solvent A) and 0.1% TFA in acetonitrile (solvent B). Gradient: 0.0–12.0 min 0-40% solvent B, 10.2–13.9 min 95% solvent B, 13.9–14.0 min at 95-5% solvent B, 14.0–16.0 min 5% solvent B. 5-(4-azidophenyl)-10,15,20-tris(4-methylpyridiniumyl)porphyrinato trichloride, and 5-[4-2-(2-(2-azidoethoxy)ethoxy)


Introduction
The goal of developing personalised medicine is an emerging trend within both clinical and scientific areas of research, and is of particular interest in the treatment of neoplastic conditions. Tumour heterogeneity in patient populations can be vast and can lead to significant differences in patient outcomes when using standardised treatments. It is unsurprising, that the development of novel theranostics is also a growing trend in research, with the combination of diagnostic and therapeutic modalities into a single agent naturally lending itself to the optimisation and personalisation of treatment. While the majority of work in the area of theranostics utilises nanoparticles either for conjugation, or encapsulation, of two or more distinct treatment and imaging agents (see Xie et al for a comprehensive review 1 ), progress has also been seen in the area of molecular theranostics, allowing the combination of therapeutic and diagnostic abilities without the need for synthesis of complex nanostructures.
Porphyrins naturally lend themselves to the area of molecular theranostics, with many examples in the literature of their separate use in both therapeutic and diagnostic capacities. The therapeutic abilities of porphyrins are undisputed; their favourable photophysical properties, and relative ease of both synthesis and functionalization, has led to their dominance in the field of photodynamic therapy; with their potent and selective cytotoxic action on tumour tissue being extensively documented 2, 3 . In addition, while both endogenous and exogenous free-base porphyrins have cemented their position as field-leaders in the area of fluorescent imaging 4 , the utility of porphyrins as chelators for a host of metals suitable for imaging applications allows the use of these metallated porphyrins as positron emission tomography (PET), single-photon emission computed tomography (SPECT) or magnetic resonance imaging (MRI) agents. 5 PET imaging is of particular interest for theranostic applications, offering functional imaging for the detection of metabolic changes in neoplastic tissues, rather than the mainly structural imaging offered by MRI. This provides good quality images while minimising the need for invasive biopsies to determine tumour malignancy. While the clinically approved 18 F radiolabelled 2fluorodeoxyglucose (FDG) is currently considered to be the "gold standard" of PET-imaging agents, uptake of FDG is an untargeted process, relying solely on the increased metabolism of the neoplastic tissue to promote tumour localisation. Development of alternative radiotracers which target tumour-associated receptors allows characterisation of cell biomarker expression and can give an indication of prognosis as well as allowing personalised planning of treatment. In particular, the radiolabelling of therapeutic agents such as photosensitisers also provides valuable information regarding the uptake and localisation of the molecule, as well as providing an indication of likely responses to PDT treatment.
The vast majority of porphyrin radiolabelling strategies in the literature have utilised the porphyrin as a chelating agent for radioisotopes such as Cu-64 6-8 , Ga-68 [9][10][11] , and Nd-140 12 (see Waghorn et al. for a comprehensive review of radiolabelled porphyrins 13 ). However, chelation of paramagnetic isotopes quenches the therapeutic action of the porphyrin, requiring use of the unmetallated porphyrin as the treatment agent. This is a major limitation in utilising radioisotopes such as 64 Cu, since it necessitates the administration of therapeutic and diagnostic agents in separate doses, with an appropriate delay between these doses to allow clearance of the radiolabelled species. In addition, it has previously been demonstrated that alteration in tumour uptake can occur upon metallation of porphyrin photosensitisers 14,15 , limiting the usefulness of any imaging data generated in this way. As gallium (III) is diamagnetic, chelation of 68 Ga by the porphyrin photosensitiser can be utilised to produce a theranostic which maintains functionality as both a PDT and a PET agent.
Despite the potential utility of 68 Ga porphyrins as diagnostic agents, to date there have been few published examples of the development of these metallo-porphyrins as radiotracers for PET. 68 Ga radiolabelling of both exogenous 11 , and targeted 9 and untargeted 10 endogenous porphyrins has been carried out in radiochemical yields (RCY) of between 22-73%, with the utility of these radiotracers as both PET 10,11 and fluorescent 9 imaging agents demonstrated. However, to date no examples of galliumlabelled porphyrins have undergone biological evaluation as therapeutic agents, and no examples of 68 Ga radiolabelling of exogenous, water-soluble porphyrins have been published.
It was therefore envisaged that the development of novel targeted theranostic agents could be achieved through the combination of the demonstrated radiolabelling potential of porphyrins with our current interest in the mild bio-orthogonal conjugation of porphyrin photosensitisers to targeting moieties [16][17][18] through use of the Copper-Catalysed Azide-Alkyne Cycloaddition (CuAAC) reaction. This reaction is mild and generally high yielding 19,20 and has been shown to be highly compatible with both porphyrin bioconjugation and radiochemistry, being largely insensitive to steric hindrance and operating well in aqueous conditions without the need for high temperatures or long reaction times.

Evaluation of cell receptor expression
Conjugation of a targeting peptide to the radiolabelled porphyrin was selected as this strategy is highly compatible with short-lived radioisotopes; the small size of peptides allows accumulation in target tissue in a short time period, and rapid clearance from the bloodstream. Imaging of the cell-surface integrin receptors utilising PET has previously been demonstrated utilising both 68 Gaand 18 F-21, 22 labelled peptides and has been shown to be particularly effective in the determination of both metastasis and angiogenesis associated with tumours.
The dodecapeptide TWYKIAFQRNRK was first characterised in 1996 by Nakahara et al. 23 , and exhibits a good affinity for the α6β1-integrin 24 . While this integrin is involved in cellular migration and adhesion in normal tissue, it is also upregulated in multiple cancers, including breast carcinomas and glioblastomas, in which it is associated with the facilitation of tumorigenesis and promotion of metastasis. 25 Despite this, the use of this peptide in imaging applications is controversial, as its binding has in some cases demonstrated stimulation of the invadopodial activity of the integrin. 26 However, use in targeted theranostic conjugates would allow treatment immediately following imaging, limiting the deleterious effects of the integrin signalling activation.
In order to confirm retention of the α6β1-integrin targeting ability of the peptide following synthesis of the porphyrin-peptide theranostic, two cell lines exhibiting a differential in α6β1-integrin expression were required. However, due to the essential roles played by the α6β1-integrin in normal cells some basal expression is always present, and therefore knock out of integrin expression in order to generate a "negative" control cell line does not produce an accurate model of the in vivo environment.. For this reason, we evaluated two cell lines to ascertain the differential in their natural integrin expression via relative fluorescent intensity. As illustrated in panel A of figure 1, a large dynamic range of expression was observed, with the cell lines displaying natural integrin receptor quantities over two orders of magnitude.
In order to quantify how these levels of expression could be correlated to the number of receptors, fluorescently calibrated beads were used in order to provide quantification of the number of receptors on the cell surface (panel B of Figure 1). In particular, the U87 cell line was shown to have extremely low levels of integrin expression, while in contrast the HeLa cell line was found to be integrin high expressing.

Synthesis of 69/71 Ga porphyrins
Two water-soluble cationic porphyrins bearing azide functionalities were selected for synthesis in this work. These porphyrins have previously been shown to exhibit potent cytotoxic action even when metallated with zinc, 18 and also undergo click reactions rapidly without the need for reaction intensification 17 , important for the rapid click conjugation of a thermally-sensitive peptide following radiolabelling. Synthesis of the porphyrins was carried out according to a previously described method 17 , with the free base porphyrins obtained in excellent overall yield. Metallation with "cold" 69/71 Ga was then carried out to produce the gallium complex for use as HPLC standard and for use in biological evaluation (figure 2).
In previously described literature synthesis of gallium porphyrins 10 , microwave irradiation has been utilised in order to facilitate metal chelation, showing considerably improved rates of reaction and yields in comparison to conventional heating. Microwave irradiation also allows for solvents to be heated under pressure above their boiling points, allowing the use of water as a solvent at temperatures above 100°C. Initially, microwave heating of porphyrin 1 to 160 °C in aqueous conditions with gallium (III) chloride was shown to be extremely effective, with complete metallation observed by TLC after 1.5 minutes. However, formation of a second metallated porphyrin by-product was also observed as a result of thermal degradation of the azide functionality.
Optimisation of reaction time and temperature demonstrated complete cessation of thermal azide degradation below 100°C, however metallation below this temperature was also slow, taking in excess of 40 minutes to achieve complete conversion to product 3. For this reason, a reaction temperature of 110°C was selected; while some degradation was observed at this temperature with reaction times longer than 15 minutes, TLC monitoring of the reaction displayed complete conversion of the starting material to 3 after 7 minutes heating, with no appreciable formation of by-products. These optimised reaction conditions were also utilised to produce 4; in both cases the desired product was obtained in near-quantitative yield following the described workup, with no further purification required.

Click conjugation of 69/71 Ga porphyrin to peptide
Following the successful metallation of porphyrins 3 and 4, click conjugation to the alkyne dodecapeptide was then attempted (figure 3). Click conjugation to the porphyrin prior to gallium insertion was not attempted as the high temperatures required for metallation are incompatible with the peptide. In addition, click conjugation prior to metallation would necessitate the inclusion of a protective metallation step to protect the porphyrin central cavity from copper insertion during the click reaction. Reaction of both porphyrins 3 and 4 with the peptide was attempted in an aqueous system at room temperature, with the reaction monitored by HPLC. Despite the similarity between the two structures, reaction rates were markedly different, with porphyrin 3 demonstrating no appreciable formation of the bioconjugate after 3 hours.
This poor reactivity was attributed to the high steric hindrance of the system, with the addition of the linker chain in porphyrin 4 significantly increasing the rate of reaction. Product formation was evident after 20 minutes reaction time, with no further reaction observed after 1 hour. Interestingly, this reaction rate is considerably slower than that observed in previous examples of the click conjugation of this porphyrin 17 , with the longer reaction time attributed both to the increased steric bulk of the peptide..
A one-pot, two-step metallation and click methodology was also investigated, and while neutralisation of the solution was required before addition of the peptide, no other workup was required between steps and no effect on yield or reaction time was observed. Following reaction completion, the product was purified on a Sephadex G15 column to remove unreacted starting materials and reagents, and the product lyophilised overnight to produce the desired product as characterised by HPLC and MS.

Evaluation of phototoxicity
Confirmation of the potential of the targeted theranostic to differentiate between high and low integrin expressing cells was carried out through evaluation of the phototoxicity of conjugate 5 in both a highly integrin expressing cell line (HeLa) and a cell line displaying minimal integrin expression (U87). In order to confirm that the selectivity of this conjugate was as a result of the peptide targeting, synthesis of a control, untargeted porphyrin 6 was carried out through the end-capping of porphyrin 4 with propargyl alcohol to prevent potential singlet oxygen quenching by the azide functionality ( figure 4). The phototoxic action of this control conjugate was then assessed in the same cell lines. Irradiation of both conjugates 5 and 6 was carried out using a light system filtered to remove light below 515nm, and the results compared to a non-irradiated control. The irradiating light chosen aligns well with the position of the two Q-bands in the gallium porphyrin spectra allowing good light absorption, however excludes absorption at the porphyrin Soret band, allowing greater applicability of the results to an in vivo clinical setting. Under these conditions it can be seen that conjugate 5 exhibited excellent ability to eradicate the highly integrin expressing HeLa cell line (ca. 80% kill), while at the same concentration showing considerably lower cell killing in the U87 control cell line ( Figure 5). Minimal dark toxicity was observed for this conjugate, with both cell lines displaying >75% cell survival at all concentrations in the absence of irradiation. In contrast, at the same concentrations the control porphyrin 6 displayed minimal cytotoxicity in both cell lines, with less than 20% cell kill observed in both irradiated and non-irradiated conditions for both cell lines. The limited cytotoxicity of 6 was attributed to the lack of cellular uptake; while the cellular targeting and subsequent internalisation of the α6β1-integrin allows uptake of conjugate 5 into target cells, no uptake of the untargeted control 6 was detected by fluorescence microscopy.

Synthesis of 68 Ga radiolabelled conjugate
Following the successful synthesis of the 69/71 Ga conjugate 5, development of a methodology utilising "hot" 68 Ga was carried out. Radiolabelling of porphyrin 2 was carried out under microwave irradiation at 110 °C due to the thermal instability of the azide functionality above this temperature. The radiolabelling was found to proceed well in 0.6M HCl at this temperature, removing the need to alter the pH or buffer the solution of 68 Ga prior to chelation. Increasing reaction time was not found to significantly improve radiolabelling, with reaction times of longer than five minutes showing increased formation of unlabelled by-products with little improvement in RCY, however reaction optimisation demonstrated improving radiochemical yields (RCY) with increasing quantities of porphyrin 2, with >95% RCY obtained with addition of 20 mg of the chelating porphyrin ( figure 6).
Following radiolabelling, the remaining unlabelled porphyrin was metallated with 69/71 Ga to create a carrier-added radiotracer. While the generation of such carrier-added systems is generally not optimal for PET imaging, it was preferred in this instance due to the large difference in the quantity of administered theranostic required for PET and PDT; while PET imaging is sensitive to low nanomolar quantities of radiotracers, PDT requires high nanomolar to micromolar quantities of the photosensitiser.

6-NIR
Administration of the carrier-added radiotracer in this way allows treatment following imaging without the need for administration of a second dose of the theranostic. 69/71 Ga metallation was carried out under microwave heating as for the synthesis of porphyrin 4. HPLC analysis showed that addition of the GaCl3 directly to the solution of radiolabelled porphyrin 4-Ga68 gave poor yields, with the radiochromatogram showing formation of significant quantities of the thermal degradation by-product. Dilution of the mixture with 200 μl of water before addition of gallium (III) chloride improved solubilisation of the porphyrin in the solution, and produced essentially quantitative complexation of all unlabelled porphyrin after 3 minutes of heating with less than 10% of the thermal degradation product observed. Heating beyond this time also increased degradation of the azide and did not improve yields. Click conjugation of the peptide to the radiolabelled porphyrin was then attempted utilising the one-pot methodology previously developed on porphyrin 4. In this case, the dodecapeptide and the Cu (I) catalyst system were added to the radiolabelled porphyrin solution following neutralisation with sodium hydrogen carbonate, and the mixture stirred at room temperature. RadioHPLC analysis was carried to ascertain reaction completion, with the formation of the conjugate confirmed with a new peak at 8:40 minutes. A RCY of 19% was obtained after 20 minutes, with no formation of radiolabelled by-products observed. Although residual precursor porphyrin 4-Ga68 was also evident, the large difference in Rf values observed between the porphyrin 4-Ga68 and conjugate 5-Ga68 would allow for the development of a prep-HPLC methodology to isolate the pure conjugate. Alternatively, purification by Sephadex G15 as demonstrated on 69/71 Ga conjugate 5 would allow for facile isolation of the pure theranostic conjugate without the need for HPLC purification.

Materials and methods
1 H and 13 C NMR spectra were recorded on JEOL Eclipse 400 and JEOL Lambda 400 spectrometers (operating at 400 MHz for 1 H and 100 MHz for 13 C). Chemical shifts (δ) are reported in parts per million (ppm), referenced to DMSO. Coupling constants (J) are recorded in Hz and significant multiplicities described by singlet (s), doublet (d), multiplet (m). MALDI mass spectra were performed by the EPSRC National Mass Spectrometry Facilities, Swansea, UK. UV−visible spectra were recorded on a Varian Cary spectrophotometer. Chemical reagents and solvents were purchased from Sigma-Aldrich and Alfa Aesar, and were used as received unless otherwise stated. Peptide was purchased from PeptideSynthetics. Gel filtrations were performed on Sephadex ® G-15 medium (GE Healthcare, UK), using deionised water as the eluent. 68 Ga was produced on a 68Ge/68Ga iThemba LABs gallium generator, and eluted with 3 ml 0.6M HCl. Activity was measured using a Capintec Inc CRC-55t dose calibrator. Microwave reactions were carried out in a CEM Discover Benchmate microwave reactor controlled using Synergy software. In all cases, maximum stirring, maximum pressure of 200 bar and 2 minute maximum heating step were used. Reaction temperatures were monitored using an external IR probe and carried out in a 10ml sealed reaction vessel. Irradiation of cells during phototoxicity assays was carried out using an

Synthetic procedures
General procedure for synthesis of gallium porphyrins: To a microwave tube was added the porphyrin (0.012 mmol) and water (5 ml). Gallium (III) chloride (10 mg, 0.06 mmol) was added and the mixture heated to 110°C (100W, MW) for 5 minutes. The mixture was neutralised with saturated sodium bicarbonate, and ammonium hexafluorophosphate added. The precipitated product was collected by filtration and redissolved in acetone. Tetrabutylammonium chloride was added and the precipitated product was collected by filtration. The product was precipitated from diethyl ether over methanol to yield the product as a red-purple solid.

Cytotoxicity assays:
The Gallium-porphyrin control dye was formulated in 5%DMSO/MEM and further diluted in MEM medium (+2mM L-glutamine but no FCS) to give a range of concentrations (between 1x10 -4 to 7.5 x10 -6 M). The Gallium-porphyrin-peptide dye was formulated in 5% DMSO/H2O and further diluted as above to give a range of concentrations (between 3x10 -5 to 3.75 x10 -6 M). The two cell types (HeLa and U87) were adjusted to a concentration of 1x10 6 cells /ml, added to the dilutions and incubated in the dark for an hour at 37°C and 5% CO2 after which they were washed in a 3x excess of medium to eliminate any unbound dye. The pellets of cells and porphyrin were re-suspended in 1ml medium and 4x100µl of each concentration put in two 96 wells plates. One plate