Daniela V.
Tomasino
a,
Ashfaq
Ahmad
b,
Tauseef
Ahmad
c,
Golestan
Salimbeigi
c,
Jennifer
Dowling
c,
Mark
Lemoine
cd,
Ruth M.
Ferrando
e,
Alan
Hibbitts
cd,
Ruairí P.
Branningan
f,
Mathew I.
Gibson
bg,
Luigi
Lay
*e and
Andreas
Heise
*adh
aDepartment of Chemistry, RCSI University of Medicine and Health Sciences, 123 St Stephen's Green, Dublin 2, Ireland. E-mail: andreasheise@rcsi.ie
bDepartment of Chemistry, University of Warwick, Gibbet Hill Road, CV4 7AL, Coventry, UK
cTissue Engineering Research Group, Dept. of Anatomy and Regenerative Medicine, RCSI University of Medicine and Health Sciences, 123 St. Stephen's Green, Dublin 2, Ireland
dAMBER, The SFI Advanced Materials and Bioengineering Research Centre, RCSI, Dublin D02, Ireland
eDepartment of Chemistry and CRC Polymeric Materials (LaMPo), University of Milan, via Golgi 19, 20133 Milan, Italy. E-mail: luigi.lay@unimi.it
fSchool of Chemical Sciences, Dublin City University, Glasnevin, Dublin 9, Ireland
gDivision of Biomedical Sciences, Warwick Medical School, University of Warwick, Gibbet Hill Road, CV4 7AL, Coventry, UK
hScience Foundation Ireland (SFI) Centre for Research in Medical Devices (CURAM), RCSI, Dublin 2, Ireland
First published on 9th February 2024
The synthesis of spherical polymeric nanoparticles containing alkyne surface functionalities for post polymerisation glycosylation is described. The nanoparticles were obtained by a polymerisation induced self-assembly (PISA) inspired methodology in dispersed media by Cu(0) mediated polymerisation. A water soluble poly(ethylene glycol methacrylate-stat-propargyl methacrylate), poly(PEGMA18-stat-PgMA5), macroinitiator was first synthesised and chain extended with 2-hydroxypropyl methacrylate (HPMA) in water using a copper wire catalyst. It was found that irrespective of the macroinitiator to HPMA ratio and the reaction time the desired spherical morphologies (<100 nm) were obtained while the absence other morphologies suggest a deviation from the classical PISA process due to chain termination in the nanoparticle's core. The obtained nanoparticles contained alkyne functionalities in the shell, which were successfully reacted by copper mediated click chemistry with fluoresceine azide and mannosides with hydrophobic and hydrophilic spacers of different lengths. The obtained mannosylated nanoparticles displayed no significant cytotoxicity against human alveolar basal epithelial adenocarcinomic (A549) cells at any dose <0.5 mg mL−1. Preliminary binding studies confirm the ability of the mannosylated nanoparticles to bind to human lectin dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN). The methodology reported here is a convenient route to well-defined spherical and shell-functionalisable nanoparticles to create libraries of bio-active nanomaterials.
Several examples of glycosylated polymer NP have been reported in the literature, mostly micelles or vesicles readily obtained by self-assembly of amphiphilic block copolymers.11,13–16 These are typically synthesised by either controlled polymerisation of glycosylated monomers or by post polymerisation glycosylation. Direct NP formation by emulsion and dispersion polymerisation techniques is an attractive alternative as they are carried out in an aqueous medium omitting or minimising the use of organic solvents. For instance, we used glycosylated amphiphilic block copolymers as surfactants in mini emulsion polymerisations, where the glycosylated hydrophilic block formed the NP shell.17–19 Similarly, dispersion polymerisation techniques such as polymerisation-induced self-assembly (PISA) are suitable for the in situ preparation of block copolymer nano-objects of controlled size, morphology, and surface chemistry in aqueous media.20–28 Moreover, the post PISA modification of the NP shell, including bioconjugation, utilising different coupling chemistries has been demonstrated.29–36
Here we devise a PISA inspired methodology by which a hydrophilic statistical copolymer comprising poly(ethylene glycol) methyl ether methacrylate (PEGMA) and propargyl methacrylate (PgMA) is first synthesised. This copolymer acts as macroinitiator for the polymerisation of 2-hydroxypropyl methacrylate (HPMA) in water thereby triggering block copolymer self-assembly due to the water insolubility of the polyHPMA block. This gives rise to NP with PEG side chains as well as reactive sites for mannose conjugation by copper mediated click reaction. A single-electron transfer living-radical polymerization (SET-LRP) approach was selected for the dispersion polymerisation as it enables low catalyst loading by using copper wire as a catalyst thus facilitating biocompatibility.37–40 While metal mediated approaches are still underutilised in PISA, low catalyst loading atom transfer radical type polymerisations (ATRP) such as initiator for continuous activator regeneration (ICAR ATRP) or activator regenerated by electron transfer (ARGET ATRP), have been successfully employed.41–44 To date, SET-LRP PISA was only disclosed by Neufeld et al. for spherical particles with diameters of 35–350 nm applying a partially depolymerized alginate macroinitiator and poly(methyl methacrylate) (PMMA) as the core block in a water/methanol mixture.45
1H-NMR spectra were recorded on a Bruker Avance 400 (400 MHz) in DMSO-d6 or CDCl3 as solvents. 13C NMR spectra were measured at 100 MHz and 298 K with a Bruker Avance III spectrometer; δC values are reported in ppm relative to the signal of CDCl3 (δC = 77.0, CDCl3). NMR signals were assigned by homonuclear and heteronuclear 2-dimensional correlation spectroscopy (COSY, HSQC). All chemical shifts are reported as δ in parts per million (ppm) and are relative to sodium 2,2-dimethyl-2-silapentane-5-sulfonate (DSS, δ = 0 ppm) or residual solvent peaks (CDCl3δ = 7.26 ppm, DMSO δ = 2.50 ppm). Mass of mannosides were measured by Electron Spray Ionization (ESI) or Matrix-Assisted Laser Ionization (MALDI) mass spectrometry. Dynamic light scattering (DLS) was done at 20 °C using a Malvern Zetasizer Nano ZSP instrument (Malvern Instruments, Malvern UK) with a detection angle of 173° and a 3 mW He–Ne laser operating at a wavelength of 633 nm. Gel permeation chromatography (GPC) was measured using a PSS SECurity GPC system equipped with a SDV 7 μm 8 × 50 mm pre-column, a SDV 100 Å, 7 μm 8 × 300 mm and a SDV 1000 Å, 7 μm 8 × 300 mm column in series and a differential refractive index (RI) detector at a flow rate of 1.0 mL × min−1 (THF). The system was calibrated against Agilent Easi-Vial linear poly(methyl methacrylate) (PMMA) standards and analysed by PSS winGPCUniChrom. All GPC samples were prepared using a concentration of 2 mg × mL−1, and were filtered through a 0.2 μm millipore filter prior to injection. Dry-state TEM imaging of the NPs was performed on a Hitachi H-7650 instrument at 150–50 K magnification. The samples (5 μL) were dropped on a Cu grid coated with SiO and Formvar and were wiped off after 10 minutes and let dry overnight. ImageJ software was used for particles size analysis. Attenuated total reflection (ATR) FTIR spectra were recorded using PerkinElmer Spectrum 100 in the region of 4000–650 cm−1. Eight scans were completed with a resolution of 2 cm−1. Mannose modification reactions were monitored by thin-layer chromatography (TLC) on Silica Gel 60 F254 (Sigma Aldrich) or with high performance thin-layer chromatography (HPTLC); compounds were visualised by heating with 10% (v/v) ethanolic H2SO4. Column chromatography was performed using Silica Gel 200–400 mesh or Biotage SNAP Ultra.
The safety profile of starting and mannosylated NP was assessed using Human alveolar basal epithelial adenocarcinomic (A549) cells. Experimental design was similar to prior studies.46,47 Briefly, 24 hours prior to treatment, A549 cells were seeded in 96 well plates at a density of 1.25 × 104 cells per well and incubated overnight at 37 °C 5% CO2. Cells were then treated with synthesized NPs at a range of concentrations and returned to the incubator. Cells were then assessed for changes in metabolic activity using the WST-1 cell viability reagent (Merck, Ireland). Cytotoxicity was assessed using the CyQUANT™ LDH Cytotoxicity Assay (Thermo Fisher Scientific, Ireland). Both kits were used according to manufacturer's instructions. In both cases, treated cells were compared against those of untreated controls and cells treated with media spiked with 0.1% Triton-X detergent (Merck, Ireland) as a positive control and background adjusted against cell medial only wells. Studies were undertaken with samples in triplicate and repeated three independent times.
Bio-layer Interferometry (BLI) was performed on an Octet® RED96 Bio-Layer Interferometry system (Forte Bio, USA) with Octet® Streptavidin (SA) used for lectin binding studies at 30 °C with shaking at 1000 rpm. For SA biosensors, the lectin DC-SIGN was first biotinylated using the EZ-Link™ Sulfo-NHS-LC-Biotin kit as per the manufacturer's instruction, in the assay buffer (10 mM HEPES, 100 mM NaCl, 10 mM CaCl2, pH 7.4). The biosensors were pre-hydrated in 200 μL of BLI assay buffer for at least 10 min in the biosensor's plate to remove the protective sucrose coating. Flat bottom black 96-Well microplates were used and loaded with 200 μL of liquid per well. The assay plate was prepared as follows: column 1 (assay buffer), column 2 (200 μg mL−1 of the DC-SIGN in assay buffer), column 3 (assay buffer), column 4 (nanomaterials at concentration NP-M1 = 2.7 mg mL−1, NP-M2 = 3.1 mg mL−1, NP-M3 = 6.4 mg mL−1, NP-M4 = 5.2 mg mL−1, NP-M5 = 3 mg mL−1 in assay buffer), and column 5 (assay buffer). Furthermore, the BLI assay was carried out as follow, Baseline 1 in column 1 (Equilibration), loading in column 2 (immobilization of the lectin on the biosensor), baseline 2 in column 3 (wash off loosely bound lectins), binding/association in column 4 (immobilized lectin binding to nanomaterials in solution), and finally dissociation in column 5 (wash off loosely bound complexes).
The monomer/macroinitiator solution was added to the Schlenk tube containing the catalytic copper catalyst via cannula transfer and the reaction was allowed to proceed for 40 min. Under a nitrogen flux, aliquot of the reaction mixture was withdrawn via degassed syringe at 10 min intervals and monitored by DLS analysis (10 μL dissolved in 1 ml of H2O) and 1H NMR spectroscopy (in DMSO-d6 with a DSS internal standard). The final solution was diluted with water, purged with air and dialysed against water for 72 h. Monomer consumption was determined by comparing the integrals at time t (At) the integral at time t = 0 (A0) of the R-CH2 HPMA signal at 6.03–6.01 ppm, following the equation:
For each targeted ratio, the polymerisation was repeated in triplicate and samples withdrawn every 10 min for 40 min. The initially transparent reaction solutions turned increasingly opaque with reaction time indicating the formation of polymer particles (Fig. S6†). Optical inspection revealed the highest colloidal stability at a DP = 300. We hypothesize that at a ratio of DP = 150 the HPMA block is too short to sufficiently stabilise the nanoparticles, while at DP = 450 gelation around the copper wire catalyst was observed (Fig. S12†). This is a common limitation of aqueous SET-LRP and is due to the ability of amphiphilic growing chains to compete with the ligand in complexation to the hydrophobic catalyst surface.48,491H NMR analysis confirmed the presence of HMPA signals in the final product as well as signals assigned to the macroinitiator (Fig. S5†). However, regardless of the MI:HPMA ratio, only spherical morphologies were observed in TEM images (Fig. S11†) with no morphological transition even at longer reaction times. Morphological transitions from spheres to worms or vesicles are a typical characteristic of PISA due to the continued polymer chain extension within the NP.21,50 The absence of it suggests a deviation from the classical PISA mechanism. To obtain more insight, HPMA conversions were calculated from 1H NMR analysis for samples withdrawn at time intervals. It was found that HPMA conversions reached a plateau between 46 and 60% at the first timepoint at 10 min with longer reaction times not resulting in any increase (Fig. 1B and Fig. S7–S9†). This suggests that the disproportionation equilibrium of the polymerisation is disrupted due to the slow rate at which Cu/ligand complexes diffuse into the particle core or by the presence of NaAsc producing additional Cu(I) activator. Ultimately both effects compromise reaction control leading to bimolecular termination. GPC analysis corroborates these results displaying broad dispersities and identical elution times for all samples (Fig. S10†). Therefore, the process does not strictly follow a PISA mechanism. It is, however, suitable to produce the sought-after spherical NP evident from the TEM images of the lead sample from DP = 300 (Fig. 1C). Software-based size analysis revealed a mean size of 90 nm (standard deviation 20 nm, n = 54).
One motivation for using Cu(0) mediated polymerisation was the low copper concentration rendering the materials cyto-compatible. Indeed, when exposing human alveolar basal epithelial adenocarcinomic A549 cells to the purified NP from DP = 300 there were no significant decreases in cell metabolic activity as measured by two assays. The WST-1 assay monitors the conversion of water-soluble tetrazolium salt (WST) by metabolically active cells into a formazan dye. As seen in Fig. 2a, their metabolic activity remained high at any of the doses examined compared to untreated cells. The Lactate dehydrogenase (LDH) assay confirmed these results. LDH is released by damaged cells and a convenient marker to determine the level of cell death. The low LDH levels as well as the high WST readings, as compared to the positive control samples (Triton-X), confirm that the NP were very well tolerated, which is in keeping with classical PEGylated NP highlighting the potential of these NP for safe in vitro/in vivo application.51–53
Fig. 3 Functionalisation of the NP surface by fluoresceine and mannose azides with different spacer lengths. |
Fig. 4 (a) Photograph of a NP suspension functionalised with fluorescein azide (DPHPMA = 300); (b) Comparison of the hydrodynamic diameter before and after the shell functionalization of the NP with fluorescein (correlograms see Fig. S15†). |
The mannose derivatives were co-clicked with fluorescein azide to the NP shell at a ratio of mannose to fluoresceine of 3:2, resulting in five sets of mannosylated fluorescein-conjugated NP (NP-M1 to NP-M5). While the total number of clickable sites per hydrophilic chain is about 5 (PgMA repeating units), the exact number of PgMA units per NP is unknown. Consequently, no stoichiometric ratio of the clickable moieties (R–CCH) to azide compounds could be calculated and a mass ratio of 1:0.6 mg of NPs to azide was used. After extensive dialysis to remove unreacted mannosides and fluoresceine, the five distinct NP solutions exhibit fluorescence when exposed to UV light, whilst appearing yellow under visible light (Fig. 5A). DLS measurements revealed a reproducible hydrodynamic diameter increase of the NPs from 183 nm (PDI = 0.13) to around 200 nm (PDI < 0.2) and a change in zeta potential from −8 mV to around −14 mV upon shell modification (Fig. 5B). TEM images confirmed similar spherical morphologies (Fig. 5C) for all samples although some aggregation was observed after extended storage time of >2 weeks.
1H NMR spectroscopy performed on the dispersed NP provided further qualitative evidence for the successful shell modification for all samples (Fig. S69†). Characteristic triazole peaks are present around 8.2 ppm and 4.8 ppm as well as fluorescein peaks in the region of 6.6 ppm and at 9.5 ppm while the signals from 5 to 4 ppm are attributed to the mannose ring. FTIR spectroscopy corroborated the NMR results (Fig. S70†). A band at 3657 cm−1 corresponding to the –CC–H stretching vibration is clearly detectable in the alkyne-functional NP but is no longer present after co-clicking with fluorescein azide and the mannoside azides of M1, M3, M4 and M5 suggesting a near quantitative reaction of the alkyne groups. In case of M2, the band corresponding to the alkyne group was still visible hence a second click reaction was attempted. Although the band appeared to have reduced in intensity it was still detectable. The reason for the incomplete functionalisation of this sample is currently unclear. The change in surface polarity of the NP caused by the shell functionalisation was further demonstrated by solvent tests. Unfunctionalised NP and NP functionalised with fluorescein only resulted in a clear solution in tetrahydrofurane (THF) while they precipitate in hexafluoroisopropanol (HFiP). Functionalisation with fluorescein and mannosides inversed solvent dispersibility in these solvents attributed to the presence of OH groups from the mannose thus providing further evidence of the altered shell properties.
To demonstrate that the NP mannose shell residues were available for binding, preliminary biolayer interferometry (BLI) studies were carried out. DC-SIGN (dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin), a mannose-binding human lectin, was used at concentration of 0.2 μg mL−1 and immobilised to the surface of the BLI sensor, and nanoparticles were tested at concentrations of 2.7–6.4 mg mL−1. As illustrated in Fig. 7, a strong binding signal was observed for most of the nanoparticles used in the study, NP-M1, NP-M2, NP-M4, and NP-M5 qualitatively confirming binding availability of the mannose with the DC-SIGN. No signal was observed for the negative control, buffer alone, which validated the positive binding signal obtained from the mannosylated nanoparticles. No signal was observed for the NP-M3, which might suggest a linker effect in the binding availability of mannose, but it would require further in-depth studies to conclusively validate this.
Footnote |
† Electronic supplementary information (ESI) available: Additional characterisation of nanoparticles, synthesis and characterisation of mannosides, additional characterisation of functional nanoparticles. See DOI: https://doi.org/10.1039/d3py01361h |
This journal is © The Royal Society of Chemistry 2024 |