1. Introduction
Photodynamic therapy (PDT) is a cancer treatment modality that was discovered many decades ago. The first patients were treated in 1975 by Dougherty
et al. [
1], who successfully eradicated skin cancer with a hematoporphyrin derivative (HpD) in 98 out of 113 patients.
PDT requires three components: light, dioxygen and a photosensitizing agent. Ideally, the photosensitizer (PS) should selectively accumulate into tumor tissue. In reality, relative selectivity is observed due to leaky vasculature, acidic pH of the tumor and also high metabolism of tumor cells. Increasing research is focusing on the design of third generation of PS, which consists of a PS covalently attached to a tumor-targeting moiety or encapsulated within nanoparticles. These tumor-targeting moieties could be biomolecules such as peptides [
2,
3], monosaccharides [
4], low-density lipoprotein (LDL) [
5,
6] or antibodies [
7,
8,
9], for example. Furthermore, PS can play another major role in the treatment of cancers. In fact, after preferential uptake by malignant cells, PS can act as image-guidance diagnostic tool due to their emission in the near-infrared [
10].
Peptides are an attractive choice for targeting strategies because of their small size and ease of synthesis [
11]. They have become of interest as tumor-targeting molecules in the field of photodynamic therapy, especially in improving the selectivity of PS towards tumor tissues or neo-vessels [
12,
13,
14]. Active targeting using peptides was found to be useful in helping the delivery of substances [
15]. Highly expressed enzyme receptors in cancer cells such as the matrix metalloproteinase enzymes (MMP) [
16] and vascular endothelial growth factor receptors (VEGFR) [
17] are examples of useful targets.
The peptide-conjugated PS first developed in our group was TPC–Ahx–ATWLPPR. This conjugate consisted of a PS (5-(4-carboxyphenyl)-10,15,20-triphenyl chlorin (TPC)), a spacer (6-aminohexanoic acid (Ahx)) and a peptide sequence (ATWLPPR) [
12,
13,
18,
19], which was designed to target the neuropilin-1 receptor (NRP-1), a receptor of vascular endothelial growth factor 165 (VEGF
165). The conjugate successfully demonstrated enhanced uptake and comparable photodynamic properties to those of the non-conjugated TPC [
12], which indicated that the selectivity of PS accumulation in tumor tissue could be enhanced by conjugation with peptide moieties. However, further studies found that this conjugate was degraded
in vivo into TPC–Ahx–A, leading to loss of selectivity of the peptide moiety [
14].
In a submitted paper, we report the binding affinity of nine peptides for NRP-1 [
20]. We first performed molecular docking studies to predict the binding interactions of selected peptides with the targeted NRP-1 and subsequently confirmed the molecular affinity
in vitro by performing competitive binding experiments (ELISA) using recombinant NRP-1 protein. Among the peptides tested, only five were able to displace the binding of VEGF
165, a physiological ligand of the NRP-1 receptor. These were DKPRR, DKPPR, TKPRR, TKPPR and CDKPRR. Three of these peptides (TKPPR [
21], DKPRR [
22] and CDKPRR [
23]) have already been described in previous works. Two novel pentapeptides DKPPR and TKPRR were chosen to be further conjugated with a chlorin molecule. To this aim, three different spacers were used: Ahx (1-aminohexanoic acid), PEG9 (1-amino-3,6-dioxaoctanoic acid) and PEG18 (1-amino-9-aza-3,6,12,15-tetraoxa-10-on-heptadecanoic acid). These spacers were selected to characterize the influence of spacer lengths and hydrophobicity on the peptides’ affinity towards NRP-1 receptor, as well as on the conjugates’ solubility and polarity profiles. In this paper we report on the
in vitro conjugates’ competitive binding study and on the
in vivo stability profiles and make a comparison with results obtained for our previous TPC–Ahx–ATWLPPR conjugate.
3. Discussion
The synthesis of novel peptides-conjugated photosensitizer started with the synthesis of peptides (DKPPR and TKPRR) using the Fmoc strategy. The spacers (Ahx, PEG9 or PEG18) were then attached to the peptide. Before coupling to the peptide–spacer, TPC was converted into TPC–COOSu and subsequently conjugated to the peptide with DIEA as the coupling agent. This step gave the novel TPC–peptide conjugates TPC–Ahx–DKPPR, TPC–PEG9–DKPPR, TPC–PEG18–DKPPR, TPC–Ahx–TKPRR, TPC–PEG9–TKPRR and TPC–PEG18–TKPRR, which were then purified using reverse phase HPLC to obtain very pure compounds. The HPLC chromatogram produced two peaks, which derive from the isomeric chlorins of TPC.
The photophysical properties of the novel TPC–peptide conjugates were evaluated and the absorption spectra obtained shows the typical chlorin derivatives spectra. This was also in accordance with the findings of Tirand
et al. [
12] and Thomas
et al. [
18]. Not all the conjugated compounds exhibited significant variations in molar extinction coefficients, fluorescence and singlet oxygen quantum yields regardless of the spacers (Ahx, PEG9 or PEG18) and peptides (DKPPR or TKPRR) and in fact showed similar spectra as the non-conjugated TPC.
A competitive binding study comparing the non-conjugated and conjugated TPC showed that conjugated TPC (whether conjugated with TKPRR or DKPPR) succeeded in binding to the NRP-1 receptor while free TPC did not show any significant binding at all. This is indeed in agreement with one previous study, which showed that TPC–ATWLPPR conjugate had managed to displace the binding of VEGF
165, whereas the non-conjugated TPC failed to do so [
14]. This showed that the conjugation of a PS with a peptide may well increase the targeting of the PS towards cancer cells and hence improve the overall photodynamic therapy performance levels. Another study found that the accumulation of TPC–Ahx–ATWLPPR conjugate in NRP-1 and KDR-expressing HUVEC was higher than for non-conjugated TPC [
12]. We previously demonstrated that TPC–Ahx–ATWLPPR was a much more potent photosensitizer
in vitro than TPC, in NRP-1-expressing cells [
13]. Indeed, for
in vitro experiments, human umbilical vein endothelial cells (HUVEC), pooled from several donors, were used and were incubated with either of the photoactive compounds (TPC or TPC–Ahx–ATWLPPR) and irradiated by red light. Whereas the control photosensitizer TPC displayed little photodynamic activity in HUVEC, conjugation with ATWLPPR significantly enhanced photodynamic activity (10.4-fold). Light doses 50 values, after incubation with either TPC–Ahx–ATWLPPR or TPC, were 0.47 ± 0.23 and 4.9 ± 0.64 J/cm
2, respectively. This wide difference in photocytotoxicity was consistent with the dissimilar accumulation of the photoactive compounds in HUVEC.
The binding of biotinylated VEGF
165 to NRP-1 was displaced by the novel conjugates synthesized in this study in a concentration-dependent manner. This showed that the spacer length and its nature have no influence on the affinity of the peptides towards its receptor but instead the spacers were however found to increase the solubility of the conjugates. The novel conjugates also showed better affinity towards NRP-1 compared with TPC–Ahx–ATWLPPR (IC
50 = 171 µM [
12]).
The conjugated peptides were found to have lower binding affinity than free peptides which could be due to the presence of TPC close to the peptide moiety which may cause steric hindrance and hence difficulty in binding. The aim of using a spacer was to ensure the presence of space between TPC and peptide so that these were individualized and separated. Several studies have reported that the presence of a spacer may bring flexibility to the molecule and that its length as well as the nature of the molecule attached to it may have an impact on receptor affinity [
26,
27]. In one study, folic acid was conjugated to 4-carboxyphenylporphyrin via two short spacers (hexane-1,6-diamine or 2,2′-(ethylenedioxy)-bis-ethylamine), which were different in nature but similar in size [
28]. Both conjugated PSs showed improved intracellular uptake and photodynamic activity in human oropharyngeal epidermoid carcinoma KB cells as compared with the non-conjugated PS. This indicates that the type of spacers have no effect on the activity of the conjugated compounds. A study by Tirand
et al. [
12] compared the IC
50 values of the ATWLPPR peptide with the Ahx–ATWLPPR conjugate and found that there was no significant change on the IC
50 values (19 and 22 µM, respectively), which demonstrated that the presence of Ahx did not interfere with the binding of the ATWLPPR peptide moiety on the NRP-1 receptor.
In vivo studies were performed to characterize the distribution of the conjugates in plasma and in healthy tissues (skin, muscle, liver, kidney and spleen) as a function of time. The main degradation products in plasma and liver were characterized by MALDI-TOF mass spectroscopy and were evaluated after one, four and 24 h after injection. The plasma concentration of the new conjugates TPC–PEG18–DKPPR and TPC–PEG18–TKPRR were significantly different from that of TPC–Ahx–ATWLPPR [
13]. While TPC–Ahx–ATWLPPR could be detected in the plasma for 48 h or more, the two new conjugates were found to be present in the plasma for only 4 h. This may be due to the difference of hydrophobicity between Ahx and PEG18 and the nature of the peptides. TPC was observed at one hour after injection of TPC–PEG18–DKPPR. Another peak at
m/
z 1605 corresponding to a glycosylated parent drug was also observed. TPC–PEG18–TKPRR behaves in the same way as the DKPPR analogous with a TPC peak, which appears to a significant degree after 24 h.
A tissue distribution study of the TPC–PEG18–DKPPR and TPC–PEG18–TKPRR conjugates showed high accumulation into the organs of the reticulo-endothelial system in comparison with the peripheral muscle and skin tissues. TPC–PEG18–DKPPR was present in all tissues even after 24 h. A maximum concentration was observed in kidney after 4 h. In liver, the main cleavage occurs at an early stage of the kinetic (from 1 h on) between aspartic acid (D) and lysine (K) in the peptide moiety. TPC–PEG18 appears from 4 h on and its detection is unambiguous at 24 h.
TPC–PEG18–TKPRR was found to have a high accumulation in the spleen and even more in the liver. It was almost nonexistent in the kidney, muscle and skin. In liver, the main degradation product after injection is TPC–PEG18–T. Another cleavage between PEG18 and T occurred after four hours and reached a significant level at 24 h.
4. Experimental Section
4.1. Synthesis of TPC–Peptide Conjugates
5-(4-carboxyphenyl)-10,15,20-triphenylporphyrin was successfully synthesized through modification of the method described by Bonnett
et al. [
29]. It was obtained in the form of a purple solid with a yield of 43%. Subsequently, 5-(4-carboxyphenyl)-10,15,20-triphenylchlorin (TPC) was obtained through diimide reduction via the Whitlock method [
30], with a final yield of approximately 25%. TPC was purified on a C18 semi-preparative column (150 × 10 mm, Apollo, Alltech, Lokeren, Belgium) using a 0.1% (
v/
v) TFA-water/methanol gradient, with a flow rate of 4.0 mL/min and monitored by absorbance at 415 and 650 nm on a SPD-10A UV-Visible detector (Shimadzu, Marne la Vallée, France). During purification by HPLC, TPC was eluted as double peaks at 23.90 and 24.60 (±1 min), which corresponded to the presence of two isomers of this molecule, due to the asymmetric reduction of double bonds on either one side of the tetrapyrrolic macrocycle. TPC-NHS was prepared in the dark under a nitrogen atmosphere by the action of one equivalent of both
N-hydroxysuccinimide (NHS) and dicyclohexylcarbodiimide (DCC) in DCM overnight at room temperature. The reaction was monitored by TLC carried out on Merck silica gel 60 F
254 plates (Merck Chimie S.A.S., Fontenay-sous-Bois, France) and the spots were visualized under UV light. The solvent was evaporated and the crude product was then purified by HPLC on the same column and in the same conditions as TPC.
All peptides were synthesized by using the Fmoc/tBu method with HBTU activation using an automated ResPep XL peptide synthesizer (Intavis AG, Bioanalytical Instruments, Köln, Germany) and operated with a Multiple-Parallel Peptide Synthesis Program. The amounts of reagents are given in equivalents (eq.). Double coupling was carried out by using a 3-fold excess of N-Fmoc amino acid and activation reagents 2-(1H-benzotriazol-1-yl)-1,1,3, 3-tetramethyl-uronium tetrafluoroborate (HBTU) (3 eq.), 1-hydroxybenzotriazole (HOBt) and N,N-diisopropylethylamine (DIEA) (9 eq.) in dimethylformamide (DMF). Unless otherwise stated, reagents were purchased from chemical companies and used without prior purification. The resins and Fmoc-amino acids were from Iris Biotech GmbH, Marktredwitz, Germany. The three linkers were then attached with Fmoc-Ahx-OH, Fmoc-PEG9-OH and Fmoc-PEG18-OH like normal amino acids. TPC (1.1 equivalents was then coupled to the N-terminus of the each peptide-spacer on the resin via its NHS ester in a solution of DMF/DCM (1:1, v/v) and DIEA (9 eq.). The resulting TPC–spacer–peptides were cleaved from the resin by treating with a mixture of trifluoroacetic acid/triisopropylsilane/water (TFA/TIPS/H2O; 92.5/5.0/2.5). The resulting crude TPC–spacer–peptides were purified by HPLC (column Altech, Apollo C18 reverse-phase (5 µm; 250 mm × 4.6 mm)) on a Shimadzu LC-10ATvp, monitored by a UV/Visible detector at 214 and 415 nm on a SPD-10A UV-Visible detector (Shimadzu, Marne la Vallée, France). Pure TPC–peptide conjugates obtained were analyzed accordingly by mass spectroscopy and 1H NMR, COSY and TOCSY).
4.2. Chemical Characterization of TPC and the Novel TPC–Peptide Conjugates
Verification of the relative atomic mass of the conjugates was carried out using Liquid Chromatography Mass Spectrometry (LCMS) on a LCMS-2010 EV (Shimadzu Corporation, Marne-la-Vallée, Paris, France) equipped with a diode array detector SPD-M20A, a column oven CT0-20AC and a DGU-20A3degasser.1H NMR spectra were recorded on a Bruker Avance 300 MHz spectrometer (Bruker, Wissenburg, Germany). All samples were prepared in deuterated dimethyl sulfoxide (DMSO-d6) with a concentration of about 10 mM.
Hydrophilic and hydrophobic characteristics were also evaluated using the “shake flask” method to determine the solubility of the free TPC and the conjugates. The distribution coefficients D of the compounds between the non-polar (octanol) and polar phases (PBS
pH7.4) were hence determined by using Equation (1):
where,
Coctanol is the concentration of compound species in octanol,
CPBS is the concentration of compound species in PBS, while
If(octanol) and
If(PBS) are the areas under fluorescence spectra between 600 and 750 nm of the species in octanol and PBS, respectively. The distribution coefficient,
DpH7.4 was then expressed as log (mean ± S.D.,
n = 6).
4.3. Photophysical Characterization of Novel TPC–Peptide Conjugates
The photophysical properties of TPC and the TPC–peptide conjugates were characterized using three different techniques; UV/visible absorption (Perkin Elmer Lambda 2 UV-Vis spectrophotometer, Courtaboeuf, France), fluorescence emission (SPEX Fluorolog-3 spectrofluorometer, Jobin Yvon, Longjumeau, Paris, France) and singlet oxygen emission. For UV-Vis measurements, the samples of TPC and the six conjugates were diluted in ethanol and their spectra were determined in the range of 300 to 800 nm. For fluorescence emission spectra, samples with an absorption at 415 nm equal to 0.200 (±0.02) were prepared. Solvent refractive index and absorption efficiencies of fluorescence quantum yields (
Φf) were calculated based on a TPP solution in toluene as a fluorescence standard reference (
Φf(ref) = 0.11) [
24] as described by Equation (2):
where
A and
Aref correspond to the absorbance at 415 nm of the sample and TPP, respectively, while
I and
Iref are the areas under the fluorescence spectra between 600 and 750 nm of the sample and TPP, respectively.
n and
nref are the refractive indices of solvents used for the sample and TPP, respectively (
n/nethanol = 1.361,
nref/
ntoluene = 1.496).
Singlet oxygen emission spectra were recorded on SPEX 1680 with 0.22 mm double monochromator with a Xenon arc lamp as excitation source. Singlet oxygen generation was specifically detected at 1270 nm and the emission was monitored by using liquid nitrogen-cooled germanium-detector (EO-817P, North Coast Scientific, North Ridgeville, OH, USA). A solution of rose Bengal in ethanol was used as the reference,
ΦΔ(ref) = 0.68 [
25]. The singlet oxygen quantum yield was then determined using Equation (3):
where
A and
Aref are the UV visible absorbance at 415 nm of the sample and rose Bengal (the reference), respectively, while
I and
Iref are the peak intensities of luminescence at 1270 nm of the sample and rose Bengal, respectively.
4.4. Binding of Peptides and Conjugates to Recombinant KDR and NRP-1 Proteins
HUVEC, pooled from several donors, were used (Cambrex, Verviers, Belgium). These cells were routinely grown in endothelial growth medium (EGM-2), containing 2% fetal bovine serum (FBS), growth factors and supplements, and maintained according to the manufacturer’s instructions. Only cells from passages 3–7 were used for our experiments. The binding affinities of the conjugates (TPC–Ahx–DKPPR, TPC–PEG9–DKPPR, TPC–PEG18–DKPPR, TPC–Ahx–TKPRR, TPC–PEG9–TKPRR and TPC–PEG18–TKPRR) were evaluated through the enzyme-linked immunosorbent assay, the ELISA study. The conjugates were first dissolved in DMSO and subsequently diluted in methanol. Tween-20 was added in the blocking buffer to assist in the solubilization of the conjugate. The samples were prepared in ten different concentrations ranging from 3 to 960 µM. Affinities were expressed as IC50, which means the concentration of competitor (TPC–peptide conjugates) that has the ability to displace 50% of biotinylated VEGF165 binding on the NRP-1 receptor, using the medium effect method.
4.5. In Vivo Study
Female athymic Swiss nude/nude mice (7–9 weeks, weight range 20–25 g) were obtained from Harlan (Gannat, France). The in vivo study we conducted was based on the results obtained from the ELISA test conducted earlier. The conjugates of TPC–Ahx–ATWLPPR, TPC–PEG18–DKPPR and TPC–PEG18–TKPRR (2.8 mg/kg) were injected intravenously via the tail vein in ethanol–PEG 400–water (2:3:5, v/v/v). The mice were kept in the dark after the injection. After 1, 4 and 24 h, blood samples were collected and plasma was separated from the blood. The mice were sacrificed and tissues were carefully excised. All samples were kept in the dark at −80 °C in polypropylene tubes until further analysis.
4.5.1. Preparation of Samples for HPLC Analysis
The surface blood from the tissues was removed by rinsing in physiological saline, blotted dry and weighed. The tissue samples were then homogenized in 500 µL of Tris–EDTA–molybdate buffer pH 7.4 before being spiked with 100 μL of 5,10,15,20-tetrakis(m-hydroxyphenyl) porphyrin (mTHPP) (500 ng/mL in methanol) as an internal standard. The next step involved solvent precipitation using methanol and DMSO (5:0.1, v/v).
Samples were then vortexed and homogenized for 30 min followed by sonication for 10 min (Branson 1200, Roucaire Instruments Scientifiques, Les Ulis, France). Tissue and cellular debris were separated by centrifugation (2500× g; 15 min). The PS-containing organic phase was concentrated by evaporation with Speedvac apparatus (Fisher Bioblock Scientific, Illkirch, France). The residue was then reconstituted in methanol (200 µL). Calibration curves were prepared by mixing plasma and organs with appropriate concentrations of conjugates in the range of 50–1000 ng/mL. For these control and calibration samples, the conjugate extraction procedure was identical to the method described above.
4.5.2. In Vivo Stability Analysis by HPLC Technique of TPC–Ahx–ATWLPPR, TPC–PEG18–DKPPR and TPC–PEG18–TKPRR in Plasma and Tissue Distribution of the Conjugates
The in vivo stability analyses were performed by HPLC equipped with programmable software GOLD version 1.6 (Beckman Coulter, Fullerton, CA, USA), an auto sampler injector (507e, System Gold, Beckman Coulter) and a fluorescence detector (RF-10A XL, Shimadzu, Kyoto, Japan). Analyses were performed by RP-HPLC on a C18 analytical column (250 × 4.6 mm, YMC, Interchim, Montluçon, France). Isocratic elution condition was applied with a mobile phase of methanol/water (95:5, v/v) and a flow rate of 1.0 mL/min. Fluorescence excitation and emission wavelengths were detected at 416 and 652 nm, respectively. All the solvents used were of the analytical grade quality. The conjugate levels in plasma and organs were determined from the peak areas and normalized based on standard calibration curves constructed earlier. The data were calculated as the mean value of three mice.
4.5.3. Statistical Analysis
Unless otherwise indicated, all data are mean values ± standard deviations (S.D.) calculated from at least three independent data experiments.
4.5.4. Mass Spectrometry
MALDI-TOF MS analyses were carried out on a Bruker Reflex IV time-of-flight mass spectrometer (Bruker-Daltonic, Bremen, Germany) equipped with the SCOUT 384 probe ion source fitted with a nitrogen pulsed laser (337 nm, VSD-337ND model, Laser Science Inc., Boston, MA, USA). The laser output energy was 400 μJ/pulse. Positives ions were accelerated with a 200 ns extraction delay. The reflector voltage was 23 kV.
Mass spectra were manually acquired using Flex-control software (Bruker Daltonic, Bremen, Germany) by accumulating four series of 100 laser shots (at 45% of total laser energy).
MALDI-TOF MS analyses were performed using 2,5-dihydroxybenzoic acid (DHB) (Sigma-Aldrich, Saint-Quentin-Fallavier, France) as the matrix. This matrix was prepared at a concentration of 1 M in an acetonitrile-water mixture (50/50, v/v) acidified with 0.1% trifluoroacetic acid (TFA) (Merck, Darmstadt, Germany). All deposits were carried out using the dried-droplet method with 1 μL of both analyte and matrix. The detection of pure products by MALDI-TOFMS was checked using 10−4 M solutions of analyte. External calibration was carried out using DHB matrix peaks and a mixture of standard peptides (Bruker Daltonic, Bremen, Germany). The following monoisotopic peaks were used: m/z 155.034 ([DHB + H]+), m/z 757.400 (bradikynin), m/z 1046.542 (human angiotensin II) and m/z 1533.860 (P14R). The calibration was considered successful if the rms error was in the range of ±10–15 ppm. Results were accepted when their rms error were less than 50 ppm. In order to highlight mass peaks resulting from the hydrolysis of the targeted photosensitizers in the different organs after 1, 4 and 24 h, peaks coming from the 2,5-DHB and from the tissue were removed from the list. All samples were analyzed in duplicate.