Edited by: Jianliang Shen, Wenzhou Medical University, China
Reviewed by: Haihua Xiao, Chinese Academy of Sciences, China; Qian Cao, SYSU, China
This article was submitted to Supramolecular Chemistry, a section of the journal Frontiers in Chemistry
†These authors have contributed equally to this work and share first authorship
‡Present address: Duy-Khiet Ho, Department of Bioengineering, School of Medicine, University of Washington, Seattle, WA, United States
Xabier Murgia, Kusudama Therapeutics, Parque Científico y Tecnológico de Gipuzkoa, Donostia-San Sebastián, Spain
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Limited drug loading capacity (LC), mostly below 5% w/w, is a significant drawback of nanoparticulate drug delivery systems (DDS). Squalenoylation technology, which employs bioconjugation of squalenyl moiety and drug, allows self-assemble of nanoparticles (NPs) in aqueous media with significantly high LC (>30% w/w). The synthesis and particle preparation of squalenoylated prodrugs are, however, not facile for molecules with multiple reactive groups. Taking a different approach, we describe the synthesis of amphiphilic squalenyl derivatives (SqDs) as well as the physicochemical and biopharmaceutical characterizations of their self-assembled NPs as DDSs. The SqDs included in this study are (i) cationic squalenyl diethanolamine (ii) PEGylated SqD (PEG 750 Da), (iii) PEGylated SqD (PEG 3,000 Da), and (iv) anionic squalenyl hydrogen sulfate. All four SqDs self-assemble into NPs in a size range from 100 to 200 nm in an aqueous solution. Furthermore, all NP derivatives demonstrate appropriate biocompatibility and adequate colloidal stability in physiological relevant pH environments. The mucoprotein binding of PEGylated NPs is reduced compared to the charged NPs. Most importantly, this technology allows excellent LC (at maximum of 45% w/w) of a wide range of multifunctional compounds, varying in physicochemical properties and molecular weight. Interestingly, the drug release profile can be tuned by different loading methods. In summary, the SqD-based NPs appear as versatile drug delivery platforms.
Nano-sized drug delivery systems (DDS) having the size range from 10 to 1,000 nm have been investigated intensively to improve the treatment efficacy of severe diseases (Bobo et al.,
However, it is challenging to bio-conjugate squalene and drug molecules with multiple functional groups or without functional groups (Ralay-Ranaivo et al.,
All chemicals were purchased from Sigma-Aldrich unless otherwise specified. Tetrahydrofuran (HPLC grade) (THF), ethanol, absolute (HPLC grade) (EtOH), ethyl acetate (analytical grade reagent), and Formic Acid Optima LCMS were purchased from Fisher Scientific. Acetonitrile and methanol (MeOH) were obtained from VWR Chemicals. Yeast extract was obtained from Fluka. Bacto™ Tryptone was obtained from BD Biosciences. Luria Bertani (LB) agar was obtained from Carl Roth. Gibco® HBSS (1x) Hanks' Balanced Salt Solution and Gibco® PBS was obtained from Life Technologies. Purified water was prepared by a Milli-Q water purification system (Merck Millipore, Billerica, MA, USA) (called water in the manuscript).
The preparation of 1,1′,2-trisnorsqualenic aldehyde (compound 2) and 1,1′,2-trisnorsqualenol (compound 6) (
Synthetic scheme of squalenyl derivatives (SqDs) [cationic squalenyl diethanolamine (cSq) (compound 3), PEG3000Sq (compound 4), PEG750Sq (compound 5), squalenyl derivative (aSq) (compound 7)], and schematic sketches of nanoparticles (NPs) upon nanoprecipitation in aqueous solution. The insert tables indicate abbreviation and molecular weight of the corresponding SqD.
cSq (compound 3), PEG3000Sq (compound 4), and PEG750Sq (compound 5) were synthesized using the same procedure from 1,1′,2-trisnorsqualenic aldehyde (compound 2) (
aSq was synthesized from 1,1′,2-trisnorsqualenol as described by Ho et al. (
All drug-free self-assembled SqD–NPs in this study were prepared by nanoprecipitation in aqueous solution as described previously (Fessi et al.,
cSq–NPs, which was prepared as follows: 0.15 ml of cSq solution in THF at a concentration of ~7 mg/ml was dropped into 1 ml of water under stirring. THF was then removed under reduced pressure resulting in a cSq–NPs suspension in water at a concentration of 1 mg/ml. PEG750Sq–NPs and PEG3000Sq–NPs, which were prepared using the same protocol as follows: 0.5 ml PEGylated SqD in THF:water (1:1 v/v) mixture at a concentration of 1 mg/ml was dropped into 0.75 ml of water. THF was then removed under reduced pressure resulting in the PEGylated SqD–NPs in water at a concentration of 0.5 mg/ml. aSq–NPs, which were prepared as follows: 0.1 ml of aSq solution in THF at a concentration of 10 mg/ml was dropped into 1 ml of water under stirring. THF was then removed under reduced pressure resulting in an aSq–NPs suspension in water at a concentration of 1 mg/ml.
The NP characteristics, especially size and polydispersity index (PDI), of the drug-free cSq–NPs and aSq–NPs were studied by varying the initial SqD concentration in THF and the final NP concentration in water. The detailed information is reported in the supplementary information (
The intensity-based hydrodynamic size (reported as
The morphology of each drug-free SqD–NPs was investigated by cryogenic transmission electron microscopy (cryo-TEM). In brief, after plotting 3 μl of SqD–NPs suspension on a holey carbon grid (S147-4, Plano Wetzlar, Germany) for 2 s, the sample was frozen by plunging into −165°C liquid ethane, then transferred to the sample holder under liquid nitrogen conditions. All samples were examined using a JEOL (Akishima, Tokio, Japan) JEM-2100 LaB6 TEM equipped with a Gatan model 914 cryo-TEM sample holder (Pleasonton, CA, USA) and a Gatan Orius SC1000 CCD camera to gain bright-field images. Sample analysis was done at −170°C, under low-dose conditions, meaning conservative settings of ~10 μa/cm2 radiation level to avoid sample destruction.
The colloidal stability of the drug-free SqD–NPs was studied in physiological relevant pH milieus, including pH 2, pH 5 (acetate buffer solution), and pH 7.4 (HBSS buffer solution). The tested samples were prepared by adding 25 μl of SqD–NPs suspension into 975 μl of buffer solution. NP characteristics, including size, PDI, and zeta-potential, were determined by DLS and ELS using a Zetasizer after 1, 3, and 24 h at 25°C.
As surrogate for biocompatibility, the determination of cytotoxicity by MTT assay on A549 cells was chosen. Briefly, prior to the assays, 104 cells were seeded in each well of the 96-well plates and grown until reaching 80% cell confluence. The SqD–NPs were suspended in HBSS at concentrations ranging from 0.0652 to 1 mg/ml and incubated with cells for 4 h at 37°C and 5% CO2. After the incubation time, cells were washed twice with PBS, and MTT reagent (0.5 mg/ml in HBSS) was added. The cells were then incubated for an additional 4 h, at 37°C, and 5% CO2 allowing the formation of formazan crystals intracellularly (Mosmann,
The interaction between fluorescent SqD–NPs and mucin glycoproteins was studied
To study the protein–SqD–NP interaction, the Brownian motion of fluorescent NPs in water was compared to their movement in water containing non-fluorescent 0.1% mucin solution (mucin from porcine stomach, type II, Sigma). For the measurements, used dilutions of the NPs, either in 0.1% mucin solution or water, were 1:100 (aSq–NPs), 1:200 (cSq–NPs), and 1:50 (PEG750Sq–NPs). To detect only the movement of the SqD–NPs, the particles were fluorescently labeled, and the fluorescence mode of the NTA was used. Therefore, SqD–NPs were loaded with 0.5% Nile red by coprecipitation; a detailed description of the method is in the Preparation of Drug-Loaded SqD–NPs section. As a vehicle control sample, the 0.1% aqueous mucin solution was studied with and without the fluorescence filter, to show the ability of the filter to exclude optical interferences of the proteins (
The LC of the synthesized SqDs was investigated using compounds representing different physicochemical properties, namely:
hydrophobic compounds: cholesteryl BODIPY, Nile red, and dexamethasone, hydrophilic and charged compounds: isoniazid, colistin, tigecycline, and FITC-albumin.
With the aim to maximize the LC while maintaining the NP stability, appropriate preparation methods for generating drug-loaded NPs were chosen depending on the drug properties. In brief, we explored three NP preparation methods as follows:
The “solvent evaporation method” and “dropping method,” which allowed the loading of drugs in both compartments, core and shell, of the NPs could be used in multiple drug-loading purposes. As an example, Nile red and FITC-albumin co-loaded cSq–NPs are presented in
Schematic illustration of drug-loaded SqD–NPs preparation methods and model compounds with corresponding molecular weight and clog
The loading capacity (LC) and encapsulation efficiency (EE) of the drugs loaded to SqD–NPs were determined indirectly using the amount of non-loaded drug in the supernatant of the SqD–NPs suspension. The hydrophobic compound (dexamethasone, cholesteryl BODIPY, or Nile red) was extracted from the supernatant using ethyl acetate (Ho et al.,
The release studies of selected drug-loaded SqD–NPs were performed using the same procedure in PBS (pH 7.4) at 37°C and constant shaking at 250 rpm. Briefly, the optimal LC sample of the drug-loaded SqD–NPs was concentrated and then diluted in PBS to have a final concentration of the corresponding drug at 10% w/w. The cumulative drug release in percent was evaluated over a 24 h period. Samples were collected after 1, 2, 4, 6, 8, 16, and 24 h, while the release acceptor volume was always kept constant. The drug amount in the acceptor fluid was analyzed. The hydrophobic drugs were extracted using ethyl acetate before further analysis. The drug quantification was done by plate reader or HPLC (
If not stated otherwise, all procedures were conducted at least in three independent experiments and measured in technical triplicate. Results are presented as mean ± standard deviation (SD). Calculations were done using either Excel, Microsoft 2016 and 2019, or GraphPad Prism 8.0. Physicochemical information of molecules was predicted by ChemDraw Professional 16.0.
The cSq and PEGylated SqDs were straightforwardly obtained by simple reductive amination reaction from 1,1′,2-trisnorsqualenic aldehyde. The successful synthesis and purification of all SqDs were confirmed by 1H-NMR and 13C-NMR (
Overall, the introduction of anionic, cationic, or PEG moiety into squalene enhanced the amphiphilic properties of the synthesized SqDs and enabled the facile self-assembly into uniform and stable NPs in aqueous solution. Regardless of the SqD, the drug-free NPs had mean sizes ranging from 100 to 200 nm and a narrow size distribution (PDI < 0.25) (
Characterization of SqD–NPs, from left to right: cSq–NPs, PEG3000Sq–NPs, PEG750Sq–NPs, aSq–NPs.
We prepared two PEGylated SqDs having PEG with different chain length and terminal groups. The PEG750Sq terminates with a methyl group, while the PEG3000Sq terminates with a hydroxyl group. The self-assembling of these PEGylated SqDs allowed the formation of NPs with a dense surface layer of PEG. Regardless of the recorded zeta-potential, this PEG layer is expected to stabilize the NPs and minimize surface interactions between NPs and other molecules (Suk et al.,
Different initial concentrations of aSq in THF could slightly tune the size of aSq–NPs that were prepared by nanoprecipitation in water at a final concentration of 1 mg/ml (
The colloidal stability of drug-free SqD–NPs was investigated in different pH values representing gastric acid (pH 1–3), skin (pH ~5), and common physiological pH environments at 7.4 (Schmid-Wendtner and Korting,
Colloidal stability of SqD–NPs in different physiological relevant pH milieus (pH 2, 5, 7.4) at room temperature studied by DLS and ELS. Hydrodynamic size, PDI, and zeta-potential are reported as mean ± SD. Measurements were taken at three different time points: 1 h (light blue bars), 3 h (blue bars), 24 h (dark blue bars).
The hydrogen sulfate groups on aSq-NPs surface were strongly acidic and deprotonated at all tested pH values resulting in a low zeta-potential (around −50 and −25 mV in pH 2/5 and 7.4, respectively), which helped stabilizing the NP suspension (Bhattacharjee,
The cSq–NPs, in turn, exposing the diethanolamine moiety on its surface, in which pKa-value is estimated at 8.44 (ChemDraw Professional 16.0), showed positive zeta-potential of 71 ± 11 and 33 ± 4 mV at pH 2 and 5, respectively, reflecting the protonation of the amine groups. The size and PDI of stable cSq–NPs after 24 h of incubation at pH 5 were 120.5 ± 1.5 nm and 0.19 ± 0.01 nm, respectively. Interestingly, a slightly smaller particle size (76.9 ± 17.1 nm) and a larger PDI (0.30 ± 0.07) were recorded when dispersing cSq–NPs at pH 2. The cSq–NPs could be fully charged at pH 2, which enhanced the amphiphilicity of cSq molecules. Consequently, this induced intermolecular forces and hydrophobic interactions between the squalenyl moieties leading to a denser particle packing (Ho et al.,
PEGylation offers plenty of advantages to nano-sized drug carriers especially to improve the
The particles holding appropriate colloidal stability in physiological relevant pH milieus were used in further investigations, including aSq–NPs, cSq–NPs, and PEG750Sq–NPs.
The biocompatibility of the SqD–NPs was tested on A549 cells
Biocompatibility study
Exemplary for a protein interaction study, the physicochemical interactions between fluorescent SqD–NPs and mucin glycoproteins were studied using NTA. The hydrodynamic size of the SqD–NP suspension was compared either in (i) water (blue line in
Fluorescent SqD–NPs interaction study with mucin glycoproteins. Interactions detected by size shift determined by nanoparticle tracking analysis (NTA) measurements. Blue lines: measurement of fluorescent SqD–NPs in water. Green lines: measurement of fluorescent SqD–NPs in non-fluorescent 0.1% mucin glycoprotein aqueous solution. Results are presented as mean ± SE.
In detail, cSq–NPs were found to present a highly positive surface net charge enabling interaction with the negatively charged mucin glycoproteins. These strong interactions are demonstrated by a size shift to a larger size from ~180 to ~960 nm (mean of number–weighted distribution) (blue and green lines in
Our results, showing that interactions between mucin glycoproteins and charged NPs, cSq–NPs, and aSq–NPs are more pronounced than interactions with neutral PEG750Sq–NPs, are in good concordance with previous findings (Crater and Carrier,
We investigated the LC of the SqD–NPs using a variety of compounds owning multiple functional groups as well as representing different physicochemical characteristics and Mw. The suitable SqD and drug-loading method to obtain optimal LC for each model compound are reported in
Summary of preparation methods used to load model compounds to squalenyl derivative–nanoparticles (SqD-NPs).
aSq | Cholesteryl BODIPY | Coprecipitation | 254.3 ± 2.4 | 0.094 ± 0.011 | −18.6 ± 0.1 | 87.6 ± 8.2 | 10.35 ± 0.66 |
Dexamethasone | Coprecipitation | 140.2 ± 25.7 | 0.203 ± 0.061 | −40.0 ± 7.1 | 90.6 ± 1.2 | 32.75 ± 0.33 | |
Isoniazid | Solvent evaporation | 198.9 ± 8.6 | 0.249 ± 0.029 | −29.5 ± 1.0 | 42.6 ± 1.7 | 27.43 ± 0.79 | |
Dropping | 112.9 ± 2.0 | 0.125 ± 0.017 | −35.6 ± 2.3 | 07.4 ± 1.2 | 06.17 ± 0.97 | ||
Colistin | Solvent evaporation | 325.0 ± 7.1 | 0.152 ± 0.041 | 22.9 ± 1.1 | 90.1 ± 2.3 | 45.31 ± 0.73 | |
Dropping | 195.6 ± 2.6 | 0.184 ± 0.002 | 25.1 ± 1.0 | 83.6 ± 4.1 | 35.83 ± 0.51 | ||
Tigecycline | Dropping | 198.6 ± 4.8 | 0.014 ± 0.010 | 28.6 ± 1.3 | 85.3 ± 6.7 | 44.26 ± 2.03 | |
cSq | Nile Red | Coprecipitation | 208.0 ± 3.2 | 0.185 ± 0.069 | 38.6 ± 3.7 | 60.2 ± 6.7 | 09.03 ± 0.91 |
Dexamethasone | Coprecipitation | 146.9 ± 8.9 | 0.183 ± 0.027 | 44.1 ± 5.7 | 87.9 ± 6.6 | 32.09 ± 1.64 | |
Fluorescent Albumin | Dropping | 205.6 ± 2.8 | 0.174 ± 0.010 | 17.4 ± 1.7 | 92.1 ± 3.8 | 03.52 ± 0.33 | |
PEG750Sq | Dexamethasone | Coprecipitation | 162.95 ± 59.45 | 0.255 ± 0.075 | 00.1 ± 9.2 | 92.4 ± 2.5 | 23.49 ± 0.48 |
In this study, the model compounds dexamethasone, cholesteryl BODIPY, and Nile red—representing different Mw and hydrophobicity—could be loaded in the core of any SqD–NPs. The drug-loaded NPs had a size range of 130–250 nm and a narrow size distribution (PDI < 0.3). Notably, the optimal LC of dexamethasone was ~33, ~32, and ~24% in aSq–NPs, cSq–NPs, and PEG750Sq–NPs, respectively, while the EE values in all cases were ~90% (
Isoniazid, colistin, tigecycline, and FITC-albumin, representing different Mw and hydrophilicity, were loaded in either aSq–NPs or cSq–NPs. Accordingly, the loading of these compounds was investigated using solvent evaporation method and/or dropping method (
Isoniazid—a positively charged and small molecule (Mw 137.14 Da)—was loaded in aSq–NPs. The solvent evaporation method resulted in a significantly higher LC of isoniazid at 27.43 ± 0.79% compared to that of the dropping method at 6.17 ± 0.97%. Using the solvent evaporation method, the preparation of isoniazid and aSq in the solvent mixture (THF:water 1:1 v/v) allowed the maximum charged interaction between the two reagents, thus, increased the EE and LC. In contrast, the preformed aSq–NPs and dropping method limited the interaction with isoniazid molecules. The optimal isoniazid-loaded aSq–NPs had the size of 198.9 ± 8.6 nm and PDI below 0.25.
The molecules with multiple functional groups—tigecycline (Mw 585.66 Da), colistin (Mw 1,155.46 Da), and FITC-albumin (Mw ~60,000 Da)—could be loaded in NPs using the dropping method. The optimal LC and EE of stable tigecycline-loaded aSq–NPs (198.6 ± 4.8 nm, PDI 0.014 ± 0.010) were 44.26 ± 2.03 and 85.3 ± 6.7%, respectively. The LC and EE of colistin in aSq–NPs were significantly high at 35.83 ± 0.51 and 83.6 ± 4.1%, respectively, which also produced a stable DDS having the size of 195.6 ± 2.6 nm and PDI lower than 0.2. The reasonable LC and EE of FITC-albumin—a representative of negatively charged protein molecules—in cSq–NPs were 3.52 ± 0.33 and 92.1 ± 3.8%, respectively. The stable protein-loaded cSq–NPs had the size of 205.6 ± 2.8 nm and PDI below 0.2. As shown in
The positively charged colistin contains a lipophilic moiety and could interact with aSq molecules
The dual loading capacity of the SqD–NPs was illustrated by loading both hydrophobic Nile red and hydrophilic FTIC-albumin in cSq-NPs. The Nile red-loaded cSq–NPs (LC ~0.5%) were prepared by the coprecipitation, and then further loaded with hydrophilic FTIC-albumin (LC~0.5%) by the dropping method. The colocalization of both fluorescent compounds in NPs was confirmed by confocal laser scanning microscopy images (
In short, the SqD–NPs are capable of loading a broad range of drug molecules differing in physicochemical characteristics, thus, are versatile drug delivery platforms.
Cumulative release study of selected model drug-loaded SqD–NPs at pH 7.4 and 37°C for 24 h in PBS.
The release profiles of hydrophobic compounds were determined for the ones with the highest and lowest Mw, cholesteryl BODIPY, and Nile red, respectively. The release profiles for both compounds are shown in
The release profiles of hydrophilic compounds were conducted for tigecycline (highest achieved LC), FITC-albumin (protein representative) and colistin (comparison of two preparation methods), respectively. The association of hydrophilic drugs loaded onto the NP surface by the dropping method is not as strong as a covalent conjugation, leading to the assumption of a complete drug release. Cumulative release of tigecycline, colistin, and FITC-albumin reached 88 ± 13, 75 ± 11, and 83 ± 13%, respectively, after 24 h in PBS at 37°C (
Furthermore, IC90 values of tigecycline and colistin-loaded aSq–NPs were determined by the minimum inhibitory concentration (MIC) assay performed on
In this study, we explored the use of self-assembling amphiphilic SqDs as DDS, which had started with the aSq. We have further described the straightforward synthesis of cSq and PEGylated SqDs. In addition to the ability of self-assembling to supramolecular colloids, the latter demonstrated that SqD–NPs showed excellent stability in physiological relevant media and biocompatibility. The application of different methods for the preparation of drug-loaded SqD–NPs allowed to modulate not only the LC but also the release rate, as desired. In further studies of carrier properties, the SqD–NPs showed significantly high LC of various cargos. Our findings postulate the self-assembled SqD–NPs as versatile drug delivery platforms.
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
D-KH, BL, DD, PC, and C-ML conceptualized and initiated the study. PC and C-ML acquired funding and provided resources. XM, US, BL, DD, PC, and C-ML were supervisors. D-KH did chemical synthesis and characterization. D-KH and RC investigated NP self-assembly and drug-loading capacities. CDR, RC, and D-KH validated the drug quantification methods. D-KH conducted
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors would like to thank the kind support from Dr. Sangeun Lee, Petra König, Jana Westhues, Dr. Arnaud Peramo, and Sandrine Gouazou.
The Supplementary Material for this article can be found online at:
drug loading capacity
squalenyl derivative
anionic squalenyl derivative
cationic squalenyl derivative
PEG750 squalenyl derivative
PEG3000 squalenyl derivative
nanoparticle
nanoparticle tracking analysis
dynamic light scattering
electrophoretic light scattering
cryogenic transmission electron microscopy
mass spectrometry
polydispersity index
high performance liquid chromatography
standard deviation
standard error.