Poor aqueous solubility or dissolution of active ingredients is one of the most significant problems hindering effective drug delivery. More than 40% of new drugs under investigation have poor water solubility and hence poor bioavailability [1
]. A number of solubility-enhancing strategies including co-solvents, micronization and nanonisation, amorphous solid dispersions (ASDs) [3
], co-crystal formation [4
], surfactants [5
], complexation utilizing cyclodextrins [6
] and the use of polymers [7
] have been utilized to address this problem.
Amphiphilic polymers have been developed as a first-rate alternative to low molecule weight surfactants for drug solubilisation. This is due to the lower excipient:drug ratios required for solubilisation and a higher degree of stability due to the decreased critical aggregation concentrations [11
]. The most unique characteristic of amphiphilic polymers is their wide array of structures and architectures [11
]. Amphiphilic polymers may exist as block copolymers [13
], graft polymers [11
], dendrimers [14
] and star shaped polymers [15
]. Each architecture has different physical properties, but all can be used to solubilise hydrophobic compounds.
Amphiphilic poly(allylamine) (PAA) derivatives have been explored for their potential in drug delivery [16
]. Hoskins and colleagues assessed the in vitro and in vivo pancreatic anticancer action of a nano-sized novel poly(allylamine) derivative grafted with 5% cholesteryl pendant groups (CH5
-PAA) for the formulation of a novel bisnaphthalimide compound [16
]. Bisnaphthalimide based drugs act as DNA intercalators and have shown huge potential in pancreatic cancer therapy [16
]. However, their activity is such that with increased drug potency, decreased drug solubility is observed. Hence careful formulation strategies are required in order to render these potent anticancer agents as clinically useable. Hoskins reported that after formulation into the CH5
-PAA amphiphiles, bis(naphthalimidopropyl)diaminooctane (BNIPDaoct) resulted in solubility enhancement up to 0.3 mg mL−1
]. The CH5
-PAA did not demonstrate any significant toxicity, yet the formulation demonstrated strong in vitro and in vivo anticancer action [16
Here we report the synthesis and evaluation of a novel poly(allylamine) (PAA) derivative which is grafted with 5% naphthalimide moieties (Figure 1
a). The hydrophobic moieties used, are envisaged to allow for greater insertion and stability within the hydrophobic core through π-π interactions between the BNIPDaoct drug molecules (Figure 1
b) and the planar naphthalimide pendant groups. Additionally, the use of the naphthalimide pendant groups will confer some inherent anti-cancer activity to the polymer itself. Ideally, for most drug delivery vehicles, biocompatibility is vital. However, in diseases such as pancreatic cancer, combined treatment and innovative therapies are required in order to overcome drug resistance and poor drug penetration. Nanotechnologies have shown to be useful for this. As such, we propose that a synergistic reduction in cell viability will occur after formulation in this system. It is proposed that these drug carriers if successful can be further functionalised with targeting moieties in order to bring the payload cargo to the exact site hence reducing likelihood of systemic toxicity.
The aggregation ability in aqueous environments and the ability to act as a drug solubilising agent was determined using novel BNIPDaoct and 5-fluorouricil (5-FU). 5-FU is an anti-cancer drug, which exerts its effects via inhibition of thymidylate synthase (TS) and the incorporation of its metabolites into RNA and DNA [21
]. Numerous studies have demonstrated the potential use of 5-FU in pancreatic cancer as single therapy or in combination with the former gold standard treatment gemcitabine [22
]. This study will evaluate whether the novel formulations formed, are capable of enhanced drug efficacy in vitro compared with the free drugs.
This study highlights the potential of a naphthalimide grafted polyallylamine as a dual functioning platform for increased efficacy in pancreatic cancer cell lines. Here, we reported the successful synthesis and characterisation of the comb shaped polymer. We have demonstrated its ability to form nano-self assemblies capable of incorporation of hydrophobic drug entities into their lipophilic core. Poly(allyamine) (PAA)-based comb polymers have been reported previously to act as universal drug solubilising agents [27
]. In particular to formulate novel bisnaphthalimide molecules. One previous study reported 0.3 mg mL−1
solubility of BNIPDaoct within PAA-Ch5
]. In another study the PAA was modified with a hydrophobic oxadiazole pendant group which resulted in 9.88 mg mL−1
BNIPDaoct solubilisation [18
]. This highly potent drug has proven effective in vitro as a potential chemotherapy agent, however, clinical usage is severely hindered by its almost negligible solubility. Previous solubility studies showed some enhancement; however, the relatively low solubilisation potential results in higher excipient:drug ratios and, hence, more expensive therapies [16
]. In this study we have shown that by substituting the bulky cholesteryl and oxadiazole moieties used previously with a planar naphthalimido moiety, greater quantity of drug compound can be introduced into the hydrophobic core in a ‘like dissolves like’ manner. Additionally, we believe that the planar nature of functionalities within the formulation may make the resultant nano-aggregates highly stable in solution as observed in our stability measurements. Interestingly, we observed that loading both anticancer drugs 5-FU and BNIPDaoct into the core resulted in particle compaction. This may be due to the planar nature of both the drug and formulation allowing for a less sterically hindered system, whereby drug molecules sit closer together forming a hydrophobic ‘strong-hold’. Such size phenomena has previously been reported elsewhere [28
]: polymeric micelles experienced size reduction upon drug loading, which was explained by a lowering in aggregation number, erosion as well as hydrolysis [28
]. Using our system, we were able to solubilise up to 2.4 mg mL−1
of BNIPDaoct which was an 8-fold improvement on previous findings.
Previous reports on the use of aromatic pendant groups attached to PAA had demonstrated a level of intramolecular and intermolecular aggregations [27
]. In this study, we only observed one aggregation mechanism. This may be due to the smaller size of the naphthalimido moieties not harnessing enough hydrophobicity to initiate intramolecular aggregation so, therefore, relying solely on intermolecular aggregation where more than one polymer strand is required for nano self-assembly.
The zeta potential measurement after 5-FU loading was expectedly similar to the unloaded vehicle. This finding agrees with the principle that if the drug inserts fully into the hydrophobic core, it will become shielded from the exterior environment—which includes all of its charges and ability to interact. However, for BNIPDaoct the zeta potential measurement obtained after sonication was curious. Here, the data appeared to show a dramatic shift in zeta potential. Previously, PAA amphiphiles have been shown to be effective complexing agents for biological molecules such as insulin and salmon calcitonin [30
]. Here, the charge-charge interactions initiate complex formation and hold the constituents tightly into the nanoparticles. It is possible that complex formation occurred between the BNIPDaoct, however, we do not think that is likely due to the polyamine chain resulting in a net positive charge on the molecule, a phenomena which we have previously been able to exploit for attachment onto negatively charged gold nanoparticle surfaces [20
]. Hence, it may be the case that the linear drug molecules are either experiencing one of the naphthalimido moieties in their structure anchoring into the lipophilic self-assembly core with the other directed towards the surface. Alternatively, the naphthalimido moieties are anchored into different cores of two self-assemblies hence stitching them together—therefore the drug molecules are sitting more towards the surface of the macromolecules. Looking at the irregular morphology of these aggregates in the TEM (Figure S2c
) compared with those loaded with 5-FU (Figure S2b
) this may be likely. However, further investigation is required to confirm such a hypothesis.
This polymer was intentionally designed not only to allow greater insertion of drug molecules into its core compared with our previous system [16
], but also to possess its own inherent cytotoxic nature. This may seem counterintuitive for a drug delivery vehicle. However, in diseases such as pancreatic cancer, extremely harsh therapies are required in order to eradicate the rapidly proliferating cells. We believe that the more stable particulates formed will result in less breakdown, premature release of drug and also exposure of hydrophobic pendant group to the external environment. This study serves as a proof of concept to elucidate whether the naphthalimide moieties could result in an increase in therapeutic effect in combination with the bisnaphthalimide drugs. This study resulted in a drug formulation exhibiting an IC50
of 3 µg mL−1
against pancreatic cancer cells. In previous studies [16
] with the same drug we had achieved slightly lower IC50
values of 0.7 µg mL−1
, however, this was after 48 h drug exposure and not 24 h as in this work. Additionally, in previous reports we have demonstrated no observable IC50
for the previous gold standard treatment gemcitabine in BxPC-3 cells after 24 h [20
], showing that this formulation is notably more rapid in its potency. Therefore, we believe there is some potential in our findings and currently further investigations are underway in order to determine whether this formulation may be modified in a manner which will allow for safe nanoparticle travel in the bloodstream before actively seeking out the pancreatic tumours. We must first address the timeliness of drug release in the context of application and time for intracellular tracking. We will do this by exploiting active targeting mechanisms alongside exploitation of the highly specific flora within the tumour microenvironment in order to render a stimuli responsive system. In addition, further in vitro evaluation is required in non-cancerous cell lines in order to estimate the extent of toxicity to health tissue which will help inform the surface engineering of non-toxic polymers anchored onto the nano-aggregate surface to render more effective biocompatibility.
4. Materials and Methods
4.1. General Information
Poly(allylamine) (PAA) (MW:17,500), 1,8-naphthalic anhydride, 3-aminopropan-1-ol, para-toluenesulfonyl chloride and caesium carbonate were purchased from Alfa Aesar (Heysham, UK). Chloroform, diethyl ether, ethanol and anhydrous pyridine were purchased from Fisher Scientific (Loughborough, UK). Deuterated chloroform, deuterated dimethyl sulfoxide and deuterium oxide were purchased from Apollo Scientific Limited (Stockport, UK). Human pancreatic adenocarcinoma (BxPC-3) cells were obtained from ATCC (Manassas, VA, USA). Foetal bovine serum, Roswell Park Memorial Institute Medium (RPMI) 1640, trypsin and penicillin streptomycin were purchased from Life Technologies (Runcorn, UK).
4.2. Synthesis and Characterisation of Polymer
4.2.1. Reaction of 1,8-Naphthalic Anhydride with 3-Aminopropan-1-ol: Synthesis of N-(3-Hydroxy-propyl)Naphthalimide (1)
1,8-Naphthalic anhydride (3.96 g, 0.02 mmol) was suspended in ethanol (100 mL) in a 250 cm3 round bottom flask. The suspension was stirred using a magnetic stirrer and bar. 3-Aminopropan-1-ol (1.5 g, 0.02 mmol) was added to the mixture dropwise. The mixture was heated to reflux for 15 h and was then allowed to cool. A small amount of ethanol was removed using a rotary evaporator until a solid precipitate began to form. The mixture was then cooled in the refrigerator (3–5 °C) until no more precipitate formed. The mixture was filtered under vacuum (Buchner funnel and flask) and the pale cream crystalline precipitate was dried on the filter. The dried precipitate was crystallised from ethanol, filtered and the crystalline product dried in vacuum at 60 °C to give pale cream-coloured needles.
4.2.2. Tosylation of N-(3-Hydroxypropyl)Naphthalimide (1): Synthesis of 3-tosyloxypropyl) naphthalimide (2)
N-(3-Hydroxypropyl)naphthalimide (1, 2.55 g, 10 mmol) was dissolved in anhydrous pyridine (dried over solid KOH) (80 mL) in a dry 250 cm3 round bottom flask. The reaction solution was left to stir at 0 °C for 10 min. After this time para-toluenesulfonyl chloride (2.86 g, 15 mmol) was added dropwise over 30 min. The mixture was then stored at 4 °C in a refrigerator overnight. The refrigerated mixture was poured slowly onto 200 mL icy water in a 500 mL beaker, followed by vigorous stirring to deliver a viscous liquid that solidified quickly on cooling. The solid was filtered through a Buchner funnel and washed with water several times. The dried precipitate was crystallised from ethanol, filtered and the crystalline product dried in vacuum at 60 °C to give a white powdery precipitate.
4.2.3. Liberation of PAA Free Base
PAA-HCl salt (10 g) was dissolved in deionised water. Sodium hydroxide (8 g) was added into the solution until pH 13 was achieved. The mixture was subsequently stirred for 1 h. After this time the PAA solution was exhaustively dialysed against water using 7000 Dalton Visking membrane for 24 h. The resultant solution was freeze-dried to obtain a white solid.
4.2.4. Synthesis of PAA-Propylnaphthalimide (3)
PAA (1 g) was dissolved in dissolved in 1:1 (v/v) methanol: chloroform (30 mL) with stirring. N-(3-tosyloxypropyl)naphthalimide (0.605 g) was added to solution followed immediately by caesium carbonate (1.2 g) and the mixture was stirred at room temperature for 24 h. The solvent was removed using a rotary evaporator and the polymer residue was washed three times with diethyl ether and subsequently dried. The product was purified by dissolving in water and exhaustive dialysis (Visking membrane, molecular weight cut-off = 12–14 kDa) against deionized water (3 L) with six changes over 24 h. The resultant solution was freeze-dried.
4.3. Characterisation of Self-Assemblies
Self-assemblies were formed by dissolving solid polymer in water. The solutions were probe sonicated for 5 min before filtration through a 0.45 μm syringe filter.
4.3.1. Characterisation Using Photon Correlation Spectroscopy and Zeta Potential Measurement
Nano-aggregate solutions/formulations were formed in water and analysed for their size and surface charge. The size was monitored over a 4-week period. Hydrodynamic diameter, polydispersity index (PDI) and zeta (ζ) potential measurements were performed using PCS on a Zetasizer Nano-ZS (Malvern Instruments, Malvern, UK). The data presented are the averaged values of three successive measurements.
4.3.2. Transmission Electronic Microscopy (TEM) Imaging
Transmission electron microscopy (TEM) was used to visualise the polymer formulas. Imaging was performed and processed using a JEOL 1200 EX-FDL5000 microscope (Jeol, Tokyo, Japan) transmission electron microscope. TEM samples were prepared by placing one drop of the formulations prepared as described above, onto formvar/carbon-coated 200 mesh nickel grids and dried under a heat lamp for 3 h.
4.3.3. Surface Tension Measurement
The surface tension of polymer formulation was investigated by a torsion balance (Torsion Balance Supplies, Weston-super-Mare, UK). In brief, various concentrations (0.00195–3 mg mL−1) of naphthalimide-PAA were prepared in deionised water. Samples were sonicated for 5 min using a probe sonicator the samples were cooled to room temperature, after which the surface tension was measured. All experiments were run in triplicate. Critical aggregation concentration was indicated by dramatic reduction in surface tension, as observed by inflection in the graph.
4.4. Drug Loading
Polymer was dissolved in deionised water (1 mg mL−1) and probe sonicated for 10 min using a Soniprep 150 (Wolflabs, Pocklington, UK). The hydrophobic drug (5 mg mL−1) of was added individually at 5:1 initial drug:polymer weight ratio and the drug-polymer solutions were probe sonicated for a further 10 min. The polymer solutions were filtered using a 0.45 µm syringe filter. The drug content was quantified: 5-FU by UV-visible spectroscopy at 256 nm (in DMSO) and BNIPDaoct using the reverse phase high performance liquid chromatography (HPLC) coupled to a fluorescent detector using a RP Zorbax ODS 250 mm × 46 mm × 5 µm HPLC column (Hichrom, Lutterworth, UK). The mobile phase was 55:45 (v/v) buffer:acetonitrile and the flow rate was 1 mL min−1. The buffer for the mobile phase was made up of 0.432 g octane sulfonic acid and 1.64 g anhydrous sodium acetate made up to 200 mL with deionised water, which the solution was subsequently pH adjusted to pH 4.5. The flow rate was set at 1 mL min−1, 20 µL injection volume and excitation and emission wavelengths set to 234 nm and 294 nm respectively. Control samples of polymer alone were run and subtracted from the data in order to eliminate any fluorescence due to the naphthalimide residues within the intrinsic polymer structure (for the BNIPDaoct formulation samples). The solutions were compared to calibrations of their respective drugs (R2 = 0.999 for 5-FU and R2 = 0.998 for BNIPDaoct), all experiments were run in triplicate. Encapsulation efficiencies (EE) were calculated to be: % EE = (drug concentration determined by HPLC/original drug concentration) × 100%.
4.5. Drug Release
Formulations (2 mL) were pipetted into Visking membrane of 12–14 KDaltons and dialysed against deionised water for 24 h. At set time intervals (e.g., 1 min, 5 min, 10 min etc.), 1 mL of water was removed and drug content analysed as previously described. The cumulative drug release profile was calculated in respect to drug loading concentration. All experiments were run in triplicate.
4.6. Cytotoxic Activity
For evaluation of the cytotoxic potential of PAA-N: drug aggregates, an MTT assay was performed in comparison with PAA-N alone and free drugs, in equivalent weight as in PAA-N: drug aggregates) at 24 h incubation time. In brief, human pancreatic adenocarcinoma (BxPC-3) cells were cultured in RPMI medium containing 10% foetal bovine serum (FBS) and 1% penicillin streptomycin (P/S). Cells were sub-cultured into a 96-well plate at, and then the plates were incubated in the incubator at 37 °C in humidified 5% CO2 atmosphere. Cells were then treated with PAA-N alone, PAA-N-BNIDaoct, PAA-N-5FU, and free drugs BNIPDaoct and 5-fluorouracil (dissolved in DMSO—Stock solutions of 20 mg mL−1 were prepared and diluted with cell culture media). Once 70% cell confluence was achieved the media was removed and replaced with the drugs or formulations (0.00001–0.1 mg mL−1). Following this, the media was removed and replaced with a 100 μL solution of 10% MTT in media and incubated for 4 h. After which, the absorbance of the formazan solution is read spectrophotometrically at 570 nm and cell viability was calculated with respect to the controls. PBS and Triton X (80 µL each) were used as the negative and positive controls respectively. All experiments were run independently in triplicate.
4.7. Cellular Uptake of Drug
4.7.1. Intracellular Drug Concentration
BxPC-3 cells were seeded at a density of (50,000 cells/well) into 6-well plates and incubated for 24 h. The media was then replaced with 50 µg mL−1 solution of formulation, or free drug as well as the polymer alone (at equivalent concentration) diluted in media. The plates were incubated for 4 h and 24 h at 37 °C. Following incubation, the media was removed and the cells were washed three times with PBS and trypsinised. The cell suspension was counted for viable cells and 100,000 cells transferred into an Eppendorf tube. Deionised water (1 mL) was added to lyse the cells and the tube was centrifuged at 500 rpm for 5 min in a Z-323 centrifuge (Hermule, Wehingen, Germany) to remove the cell debris. The supernatant was removed and diluted in DMSO for 5-FU and 55:45 (v/v) buffer: acetonitrile (previously described) for BNIPDaoct. Control samples of polymer alone were run and subtracted from the data in order to eliminate any fluorescence due to the naphthalimide residues within the intrinsic polymer structure (for the BNIPDaoct formulation samples). Drug content was measured as previously described and calculated per cell. All experiments were carried out in triplicate.
4.7.2. Fluorescence Microscopy
Cell lines were seeded onto glass cover slip in 6-well plates (50,000 cells/well) and treated as described above. After washing, the cells were fixed with 2% paraformaldehyde for 10 min and further washed three times with PBS. The glass slips were fixed on the glass slides and visualised by fluorescence microscopy (Invitrogen, EVOSTM, Runcorn, UK).