Research Article www.acsami.org
Multifunctional Hyaluronic Acid and Chondroitin Sulfate Nanoparticles: Impact of Glycosaminoglycan Presentation on Receptor Mediated Cellular Uptake and Immune Activation Oommen P. Oommen,*,†,§ Claudia Duehrkop,‡ Bo Nilsson,‡ Jöns Hilborn,† and Oommen P. Varghese† †
Department of Chemistry, Ångström Laboratory, Science for Life Laboratory, Uppsala University, S-75121 Uppsala, Sweden Department of Immunology, Genetics and Pathology, Rudbeck Laboratory, Uppsala University, S-75121 Uppsala, Sweden § BioMediTech - Institute of Biosciences and Medical Technology, Bioengineering and Nanomedicine Group, Tampere University of Technology, 33520 Tampere, Finland ‡
S Supporting Information *
ABSTRACT: Hyaluronic acid (HA) and chondroitin sulfate (CS) polymers are extensively used for various biomedical applications, such as for tissue engineering, drug delivery, and gene delivery. Although both these biopolymers are known to target cell surface CD44 receptors, their relative cellular targeting properties and immune activation potential have never been evaluated. In this article, we present the synthesis and characterization of novel selfassembled supramolecular HA and CS nanoparticles (NPs). These NPs were developed using fluorescein as a hydrophobic component that induced amphiphilicity in biopolymers and also efficiently stabilized anticancer drug doxorubicin (DOX) promoting a near zero-order drug release. The cellular uptake and cytotoxicity studies of these NPs in different human cancer lines, namely, human colorectal carcinoma cell line HCT116 and human breast cancer cell line MCF-7 demonstrated dose dependent cytotoxicity. Interestingly, both NPs showed CD44 dependent cellular uptake with the CS−DOX NP displaying higher dose-dependent cytotoxicity than the HA−DOX NP in different mammalian cells tested. Immunological evaluation of these nanocarriers in an ex vivo human whole blood model revealed that unlike unmodified polymers, the HA NP and CS NP surprisingly showed platelet aggregation and thrombin−antithrombin complex formation at high concentrations (0.8 mg/mL). We also observed a clear difference in early- and late-stage complement activation (C3a and sC5b-9) with CS and CS NP triggering significant complement activation at high concentrations (0.08−0.8 mg/mL), unlike HA and HA NP. These results offer new insight into designing glycosaminoglycan-based NPs and understanding their hematological responses and targeting ability. KEYWORDS: Hyaluronic acid, chondroitin sulfate, nanoparticles, cancer, immune activation, drug delivery
■
INTRODUCTION Designing drug delivery systems (DDSs) that can prolong drug release at the target site and simultaneously minimize toxicity by reducing off-target accumulation is highly desirable for cancer therapy. The target specific delivery is generally achieved by hijacking cell surface receptor−ligand interactions.1 Such targeted DDSs significantly improve efficacy and reduce toxicity by providing control over the drug biodistribution and pharmacokinetics.1 However, the primary hurdle for the success of any delivery system is to overcome innate immunity as a result of blood−NP interaction, which could compromise the potency and efficacy of the drug. Several chemotherapeutic drugs trigger thromboinflammatory responses such as the activation of complement or coagulation cascade resulting in adverse thrombotic complications, sepsis, acute or chronic inflammation, and renal or pulmonary dysfunction.2 It is also important to note that several nanomedicines currently used in © XXXX American Chemical Society
the clinic also trigger complement activation, which in fact stimulates tumor growth and progression.3 Several poly(ethylene glycol) (PEG) coated nanocarriers are in clinical use (to increase the blood circulation time); it elicits antibody formation in humans 4 and also triggers complement activation.5,6 Thus, designing nanocarriers that can deliver potent drugs to cancer cells, suppress complement activation, and avoid detection by the innate immune system is of paramount importance. These systems could prevent premature drug elimination and mitigate side effects associated with thromboinflammation.7 Biopolymers such as hyaluronic acid (HA) or chondroitin sulfate (CS) are the ideal scaffolds for designing DDS, because Received: June 8, 2016 Accepted: July 28, 2016
A
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
complement, and the kinin systems, the key cascade systems used by the body to recognize and opsonize foreign materials and trigger inflammation. The drug-loaded NPs were subsequently tested in two different human cell lines, and cellular uptake, nuclear translocation, dose-dependent cytotoxicity, and dependence of caspase 3/7 for apoptosis were evaluated. To the best of our knowledge, our study is the first report on the direct comparison of hemocompatibility and anticancer properties of HA- and CS-based materials.
they possess biocompatibility and biodegradability properties and provide a large polymeric backbone for chemical modification. These are natural glycosaminoglycans (GAGs) that are present in the extracellular matrix and play a unique role in modulating cellular functions. Also, they are known to target a transmembrane cluster of differentiation 44 (CD44) receptor and HA receptor for endocytosis (HARE) that facilitates efficient receptor-mediated endocytosis.8 We have recently reported CS nanoparticles9−11 and HA/CS molecular conjugates12,13 as CD44 specific drug delivery vehicles. CD44 is a known biomarker of cancer stem cells and ∼4−5-fold overexpression of this receptor is an early sign for cancer metastasis.14 Unlike HA, CS is a natural substrate for P-selectin, which is expressed on endothelial cells, platelets, and some cancer cells.7 Since P-selectin plays a critical role in tumor metastasis,15 CS based nanoparticles could have dual targeting capabilities, particularly for metastatic cancer. These polymers degrade in vivo by a ubiquitous enzyme, hyaluronidase, that is expressed in high concentrations at the tumor site.16 Although the existence of chondroitinase in mammals is still not established, it is believed that CS is also degraded by the hyaluronidase.17 The clearance of HA and CS from circulation depends on its molecular weight. The half-life of native high molecular weight HA in humans is 2.5−5.5 min,18 whereas CS (50 kDa) has a half-life of 12−15 min.19 The chemical modification of HA and CS increases the blood circulation time due to hindered enzymatic recognition.20 For example, thiolation or conjugation with amino acids is known to improve the enzymatic stability; however, chemical modification of the carboxylate moiety of these polymers hampers the biochemical functions and recognition with cell surface receptors.21,22 Since HA and CS based materials are used for several biomedical applications, such as hydrogels for tissue engineering, (regeneration of organs such as bone, liver, heart, skin, vocal folds, etc.), restoration of brain function following ischemic stroke, coating of material surfaces, and drug delivery,23−25 it is extremely important to understand the impact of chemical modifications of ECM polymers on activation of complement and coagulation cascade in blood. For designing polysaccharide based polymeric NPs, general strategies include conjugation of small hydrophobic molecules such as fatty acids or ceramides,26 cholesterol,27 nicotinamidebased hydrotropic agents,28 5β-cholanic acid,29 deoxycholic acid,30 or hydrophobic block polymers31 to develop amphiphilic polymers that self-assemble to form particles. In such a particle design, the hydrophobic moiety constitutes the core, while the hydrophilic unit forms the shell.32 The hydrophobic molecules used for inducing amphiphilicity do not contribute to any other function. Though several HA- and CS-based NPs are developed as DDSs, they generally show burst release because of poor drug stabilization.33 Thus, the development of targeted delivery systems with a sustained drug release profile that can overcome innate immune reactions could have major implications in the field of NP-based DDSs. Herein, we have designed a strategy to develop theranostic NPs using fluorescein as the planar aromatic moiety that induces amphiphilicity to HA and CS. Such a modification fostered self-assembly to form NPs that not only gave an imaging possibility but also improved the binding of aromatic drug molecules using π−π interactions resulting in controlled drug release. We examined the biopolymers and their NPs in a non-anticoagulated human whole blood model to study the activation of innate immunity by measuring the coagulation,
■
EXPERIMENTAL SECTION
Materials and Methods. Chondroitin sulfate-A (CS-A, 54 kDa from bovine trachea) was purchased from Sigma-Aldrich (Sweden). Hyaluronan (51 kDa) was purchased from Lifecore Technologies, USA. Doxorubicin (DOX)·HCl was purchased from Tocris Bioscience, UK. Monoclonal anti-CD44−FITC antibody, fluorescein-5thiosemicarbazide (FTSC), N-hydroxybenzotriazole (HOBt), and 1ethyl-3-(3-dimenthylaminopropyl) carbodiimide hydrochloride (EDC· HCl) were purchased from Sigma-Aldrich (Sweden). Hyaluronidase from bovine testes (Type IV-S) was obtained from Sigma-Aldrich. Dialysis membranes used for purification were purchased from Spectra Por-6 (MWCO 3500). Fetal bovine serum (FBS) was purchased from Hyclone, Perbio Scientific, Sweden. All other chemicals were purchased from Sigma-Aldrich. All solvents were of analytical quality. All the cell lines were obtained from American Type Culture Collection (ATCC)-LGC standards, Sweden. The 1H NMR experiments (δ scale; J values in Hz) were carried out on Jeol JNM-ECP series FT NMR system at a magnetic field strength of 9.4 T, operating at 400 MHz for 1H. Spectroscopic analyses were carried out on PerkinElmer instruments, namely, Spectrum One ATFTIR, Lambda 35 UV−vis spectrophotometer, and LS 45 luminescence spectrophotometer, and DLS measurement was carried out in Zetasizer Nano-ZS from Malvern. Synthesis of HA−FTSC (HA NP) and CS−FTSC (CS NP) Conjugates. HA NPs and CS NPs were synthesized using carbodiimide-coupling chemistry as follows. For optimization of nanoparticle size, the amount of EDC·HCl was varied (0.01 to 0.1 mmol). This resulted in different degrees of FTSC conjugation resulting in different sizes of self-assembled particles. The final optimized particle size was obtained as follows. Briefly, 0.5 mmol of HA or CS (with respect to disaccharide units) was dissolved in 60 mL of deionized water. Thereafter, FTSC (21 mg, 0.05 mmol) dissolved in 15 mL of DMSO was added followed by HOBt (77 mg, 0.5 mmol). The reaction mixture was stirred for 30 min (until it becomes homogeneous), and the pH of the reaction mixture was adjusted to 5.0 by careful addition of 1 M NaOH. Finally, EDC·HCl (19.2 mg, 0.1 mmol) was added, and the mixture was stirred overnight. The reaction mixture was loaded into a dialysis bag (Spectra Por-6, MWCO 3500 g/ mol) and dialyzed against dilute HCl (pH = 3.5) containing 100 mM NaCl (4 × 2L, 48 h) and then dialyzed against deionized water (2 × 2L, 24 h). The solution was lyophilized, and fluffy yellow material was obtained in nearly quantitative yield. This product was finally washed with methanol to remove any traces of unreacted FTSC. Percentage of FTSC conjugation was estimated by UV measurement (at pH 8.5 in water) using the FTSC extinction coefficient of 78 000 M−1 cm−1 at 492 nm. Preparation of DOX Loaded NPs. DOX was loaded on the HA NPs and CS NPs by nanoprecipitation method. Briefly, 100 mg of HA NPs or CS NPs was dissolved in 25 mL of phosphate buffer saline (PBS) at pH 7.4. Thereafter, 5 mg of DOX·HCl (in 1 mL of DMSO) was added dropwise under magnetic stirring (1000 rpm). The pH was adjusted again to 7.4, and the reaction mixture was stirred overnight. Thereafter, the solution was loaded into a dialysis bag (Spectra Por-6, MWCO 3500) and dialyzed against 100 mM NaCl (2 × 2 L, 24 h), followed by dialysis against deionized water (2 × 2 L, 24 h). This solution was lyophilized yielding ∼96 mg of orange-yellow fluffy product (96% yield). In Vitro Release Experiment. The amount of DOX loading and rate of DOX release from HA NPs and CS NPs was performed by B
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces dialysis method, and release medium was analyzed by UV/vis spectroscopy at 485 nm (Lambda 35 UV/vis spectrometer, PerkinElmer) using the extinction coefficient of 11 500 M−1 cm−1. Briefly, 10 mg of the sample was dissolved in 1 mL of PBS buffer (pH 7.4) and transferred to Slide-A-Lyzer MINI dialysis device, (3500 MWCO, 2 mL). This device was placed in a glass tube with 18 mL buffer, instead of falcon tube, to prevent DOX adhesion to plastic surfaces. The samples were placed on a shaker (at 100 rpm) at room temperature, and the UV absorbance of the released media was measured at various time points at 485 nm. The release medium was placed back into the dialysis vial after each measurement. The percentage of release in each case was plotted using Excel software. The release experiment was carried out in duplicate. The samples were protected from light throughout the release operation to prevent DOX degradation. Determination of the Rate of Drug Release. The rate of DOX release was determined by dialysis method following pseudo-first-order kinetics using the equation Y = Ymax(1 − e−kt). Here k is the pseudofirst-order rate constant and t is time (min). The half-life of the kinetics was determined by the equation 0.693/k. Particle Size Measurement. The particle size distribution was carried out using Malvern laser granulometer (Zetasizer Nano ZS, Malvern, United Kingdom). Freeze-dried CS NPs were dissolved in PBS, pH 7.4, at 1 mg/mL concentration, stirred at room temperature for 2 h, and filtered through 0.45 μm filter before performing the DLS measurement. Cell Culture Experiments. All the cells used in this study were cultured in Dulbecco’s modified Eagle medium (DMEM), supplemented with 10% fetal bovine serum (FBS) (Gibicol) and 1% antibiotics (10.000 U penicillin and 10 mg/mL streptomycin, Sigma). Cells were maintained in a humidified incubator at 37 °C and 5% CO2. Estimation of CD44 Expression Levels by Flow Cytometry. Cells in the log phase were seeded onto a culture plate with appropriate cell number and incubated at 37 °C overnight. After 24 h incubation, the medium was removed and washed with 1× PBS. The cells were treated with enzyme free dissociation buffer (Gibco cell dissociation buffer, Invitrogen) and incubated at 37 °C for 30 min until they are detached. Thereafter an equal amount of DMEM medium was added. The cells were then centrifuged at 1000 rpm for 5 min and subsequently washed with 3 mL of cold PBS containing 10% FBS. Then 5 × 105 cells per mL were freshly suspended in cold FACS buffer (10% FBS in PBS). To this cell suspension, 4 μL of FITC− CD44 Ab or 2 μL of FITC−IgG Ab (2 μg/mL) was added into each FACS tube and was incubated at 4 °C for 30 min. The cells were washed once with PBS and then resuspended in FACS buffer. FACS measurement was performed under FITC channel using CyAn ADP Analyzer (Beckman Coulter). CD44 Blocking Studies with Low Molecular Weight HA. Cells (50 000) were plated per well in 24 well plates and allowed to adhere overnight. Next day, the medium was changed with fresh medium without serum, which may or may not contain LMW HA (7.5 kDa, 10 mg/mL) as blocking agent and incubated at 37 °C for 1 h. The medium was subsequently replaced with fresh medium containing 4 mM HA NPs or CS NPs and incubated at 37 °C for another 1 h. The cells were trypsinized, washed twice with cold FACS buffer (PBS with 10% FBS), and resuspended in 500 μL of FACS buffer. Samples were acquired on BD Accuri C6 FACS instrument (Beckton Dickinson) and analyzed by BD Accuri C6 software. For each analysis, 10 000 events were counted. Confocal Microscopy. Monolayers of HCT116 were seeded in 8well chamber plates (2000 cells in 200 μL/well) overnight at 37 °C. After 24 h incubation, cell culture medium (DMEM containing 10% FBS) was replaced with fresh medium containing 4 mg/mL of HA NPs or CS NPs. After 4 h at 37 °C, the medium was removed and washed twice with blocking buffer (PBS, 0.5% FCS). Thereafter, cells were fixed and permeabilized by treating with a 4% (w/v) solution of paraformaldehyde in phosphate buffer for 20 min at 4 °C and washed twice with blocking buffer. Nuclei were stained with DAPI, and cells were analyzed by confocal laser scanning microscopy. Images were
obtained at 63× magnification with a Zeiss LSM 510 META confocal microscope. Cytotoxicity Studies and Caspase Activity. Cell viability and caspase activity were measured using ApoTox-Glo Triplex assay kit following manufacturer’s protocol. Briefly, HCT116 and MCF-7 cells were seeded in 384-well BD Falcon black microplates (1000 cells in 50 μL/well) using automated Biomek FX pipetting workstation and incubated at 37 °C for 24 h for cell attachment. A stock solution of DOX (in DMSO), HA−DOX NPs, and CS−DOX NPs (in cell culture medium, DMEM) was prepared separately. After cell attachment, a different volume of stock solution was added to each well using noncontact acoustic dispenser (Echo 555) to obtain a gradient concentration ranging from 25 nM to 2 μM, and plates were incubated for additional 48 h at 37 °C. After 48 h, 20 μL of medium was treated with 5 μL of cell viability assay reagent, containing both GF-AFC substrate and bis-AAF-R110 substrate and incubated for 30 min at 37 °C. Fluorescence values were recorded at two wavelength sets: 400Ex/505Em (viability) and 485Ex/520Em (cytotoxicity) using Envision multilabel plate reader, and the cell viability was obtained as a percentage of the untreated control (100% cell viability). Caspase activity was also measured by adding 25 μL of Caspase-Glo 3/7 reagent to all wells and incubating for 30 min at room temperature, followed by measuring its luminescence with plate reader (PerkinElmer). All the cell experiments were performed in triplicate. The ex Vivo Chandler Loop Model. The ex vivo Chandler loop model was described in detail earlier.34 In brief, the entire material coming in contact with blood was heparin-coated (Corline Systems AB, Uppsala, Sweden) in advance to ensure that no surface-induced activation takes place. Fresh human non-anticoagulated whole blood samples were obtained from three healthy volunteers who were medication-free for at least 10 days prior to donation. Blood was freshly drawn with an 18-G needle connected to silicon tubing allowing the blood to run into collection containers. This blood (1.8 mL) was transferred into the loops, and 0.2 mL of CS NPs, HA NPs, or native polymers was added (giving the final concentration of 0.008, 0.08, and 0.8 mg/mL of polymer). Of note, the loops were not filled completely because the occurring air bubble assures a constant movement of blood. The loops were closed using specially designed metal connectors and were incubated at 37 °C for 1 h under rotating conditions. After the experiment, the blood samples were taken and supplied with 10 mM EDTA to stop the reaction. Blood samples were centrifuged to obtain EDTA plasma, which was stored at −80 °C until use. Determination of Plasma Cascade Activation Using ELISA. In order to analyze CS or HA NP mediated activation of the plasma cascades, EDTA plasma was scrutinized for C3a, the membrane attack complex (sC5b-9), factor XIa−antithrombin (AT) complexes, factor XIa−C1 esterase inhibitor (C1 INH) complexes, factor XIIa−AT complexes, factor XIIa−C1 INH complexes, kallikrein−AT complexes, kallikrein−C1 INH complexes, and thrombin−AT complexes by ELISA. In brief, PBS containing 0.05% Tween 20 (Sigma-Aldrich Inc., St. Louis, MO, USA) was used as a washing buffer. The working buffer was based on the washing buffer and additionally contained 1% (w/v) bovine serum albumin (BSA; Sigma-Aldrich) and 10 mM EDTA. Zymosan-activated serum was used as the calibration standard. 3,3,5,5Tetramethylbenzidine (TMB; Invitrogen) was used as a substrate in all assays. For determination of concentrations of the different parameters in EDTA plasma, DeltaSoft (BioMetallics Inc., Princeton NJ, USA) software was used. Statistical Analysis. Data are expressed as mean ± standard deviation (SD). Statistical significance was determined by one-way ANOVA followed by Dunn’s post hoc test for significance vs baseline, using GraphPad Prism 6.0e software for Mac OS X, GraphPad Software, San Diego California USA, www.graphpad.com.
■
RESULTS AND DISCUSSION To design theranostic NPs, we induced amphiphilicity on HA and CS by hydrophobically modifying these polymers with fluorescein, a fluorescent aromatic molecule. Of note, we C
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces Scheme 1. Schematic Representation of Doxorubicin Stabilized Amphiphilic HA and CS Polymers
size below 150 nm (Table 2). These FTSC functionalized GAGs were extensively dialyzed against 100 mM NaCl and 3.5
utilized HA and CS of similar molecular weights (approximately 50 kDa) to facilitate direct comparison in this study. The fluorescein molecule is considered to be nontoxic, and several fluorescein-tagged drugs are approved by the Food and Drug Administration (FDA) for diagnostic angiography or angioscopy of the retina and iris vasculature (e.g., Fluorescite 10%, AK-Fluor 10%).35 We hypothesized that the planar undistorted π-core of fluorescein would enable favorable π−π stacking interactions with an aromatic hydrophobic drug such as doxorubicin hydrochloride (DOX) (Scheme 1). The amine residue of DOX, on the other hand, will favor electrostatic binding with the carboxylate/sulfate moiety of GAGs. Therefore, these interactions may synergistically stabilize the encapsulated aromatic drug in the nanocarrier. Such interactions could, therefore, minimize the burst release of the drug and regulate its release into the extracellular space, while maximizing its concentration at the tumor site. Synthesis and Characterization of HA and CS NPs. To tailor the nanosized particles, we first optimized the degree of chemical modification26 with HA and CS carboxylates to induce amphiphilicity using carbodiimide coupling chemistry that was previously optimized in our group.36 The degree of chemical modification was determined by ultraviolet (UV) spectroscopy at pH 8.5 in water using the FTSC extinction coefficient of 78 000 M−1 cm−1 at 492 nm.37 Notably, 1H nuclear magnetic resonance (NMR) based quantification of FTSC was unsuccessful because of the core−shell assembly of polymer−fluorescein conjugates in water as observed previously by other groups.38,39 The degree of FTSC functionalization was carefully optimized to achieve size range below 150 nm. As the % of FTSC functionalization increased from 0.9% to 4.06% (with respect to HA disaccharide repeat units), the average size reduced from 603 to 123 nm as observed from dynamic light scattering (DLS) measurement (Table 1). This resulted in a gradual increase in the ζ potential of these conjugates (from −50.5 to −23.9 mV). Similarly, by varying the degree of functionalization of CS, we could obtain CS NPs of a
Table 2. Optimization of the Synthesis of CS NPs
average size (nm)
0.9 1.076 1.267 2.22 4.06
603.7 578.9 486.6 440.1 123
± ± ± ± ±
0.65 0.775 1.40 1.51 2.53
185.5 172.7 165.8 163.1 146.1
ζ potential (mV) −52.3 −48.8 −47.3 −47.1 −39.5
± ± ± ± ±
0.354 0.919 0.134 1.41 1.48
Degree of FTSC conjugation in CS as determined by UV spectroscopy.
pH water (adjusted with 1 M HCl) to ensure the complete elimination of unreacted FTSC before lyophilization. This optimized method of purification is based on our previous observation of removing electrostatically bound DOX from HA.13 The final optimized percentage of FTSC conjugation is estimated to be 4.06% and 2.5% for HA and CS, respectively (with respect to the disaccharide repeat units) with the average hydrodynamic diameter of 123.4 ± 2.1 nm and 146.1 ± 1.65 nm, respectively (Figure 1a,b). These FTSC conjugated HA and CS complexes are designated as HA NPs and CS NPs, respectively. In order to evaluate the role of hyaluronidase in the degradation of HA and CS NPs, we incubated HA NPs and CS NPs in the presence of 1000 U/mL hyaluronidase and evaluated the time dependent particle degradation (Figure S1 in Supporting Information or SI). Interestingly, we found that the hydrodynamic size of HA NPs increased from ∼123 to 345 nm in 10 min, which subsequently increased to 650 and 955 nm in 30 and 120 min (Figure S1). However, CS NPs did not show any signs of degradation in the presence of the bovine hyaluronidase. This degradation pattern suggests that HA and CS undergo different dissociation mechanisms that might result in different drug release kinetics in vivo from different nanocarriers. Designing DOX-Loaded NPs. We then loaded DOX, a clinically used antineoplastic agent, to the HA NPs and CS NPs by the nanoprecipitation method. After drug loading, the sample was purified by dialysis in 100 mM NaCl solution (to eliminate any electrostatically bound DOX) followed by deionized water. The addition of NaCl did not destabilize the micelle, as we did not observe any aggregation of NPs. The DOX-loaded HA NPs and CS NPs are designated as HA− DOX NPs and CS−DOX NPs, which rendered particles of 196 and 140.6 nm, respectively, with a narrow size distribution (Figure 1c,d). The drug-loaded NPs were subsequently
ζ potential (mV) −50.5 −41.8 −38.3 −35.7 −23.9
average size (nm)
a
Table 1. Optimization of the Synthesis of HA NPs % conjugation of FTSCa
% conjugation of FTSCa
0.141 0.283 1.13 0.354 0.11
a
Degree of FTSC conjugation in HA as determined by UV spectroscopy. D
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 1. Particle size distribution of (a) HA NPs, (b) CS NPs, (c) HA−DOX NPs, and (d) CS−DOX NPs in water (1 mg/mL) as determined by dynamic light scattering measurements.
Figure 2. Photophysical properties and drug release profile of HA NPs, CS NPs, HA−DOX NPs and CS−DOX NPs. (a) UV spectra of components (75 μg/mL), (b) fluorescence spectra of components (4 μg/mL), (c) recovery of UV absorbance of DOX from HA−DOX NPs upon temperature dependent disruption of π−π stacking interaction, and (d) cumulative drug release kinetics: free DOX (in blue), DOX physically mixed with HA (in red), HA−DOX NPs (in green), and CS−DOX NPs (in purple).
lyophilized and stored at −20 °C until use. It is important to note that these lyophilized NPs retained their hydrodynamic size when resuspended in water because of the strong anionic repulsion between the particles. Such stability is not generally observed in nanoformulations, as they tend to be susceptible to aggregation resulting in macroparticles upon lyophilization. Characterization of DOX-Loaded NPs. The 1H NMR analysis of HA−DOX NPs and CS−DOX NPs in D2O showed only carbohydrate signals without any aromatic signals. The absence of the aromatic signals of DOX and FTSC in these NPs clearly suggests the formation of core−shell structure, consistent with literature reports.38,39 In order to verify the core−shell assemblies, we performed 1H NMR analysis of HA NPs and CS NPs in DMSO-d6 solvent (Figure S5 and S6 in SI). The three methylene proton signals of the acetamido moiety in HA (N-acetyl-D-glucosamine) and in CS (N-acetyl-Dgalactosamine) were integrated as 3.0. The 1H NMR of HA NPs and CS NPs in D2O showed the characteristic spectra of
native HA and CS without any aromatic FTSC molecules, indicating that the FTSC molecules form the core and remain invisible.39 Interestingly, in the organic solvent, the selfassembly is disrupted exposing the FTSC molecules, which resulted in the recovery of aromatic signals at δ 6.51−6.83 ppm (Figure S5 and S6 in SI). The UV−visible (vis) spectroscopy analysis of HA−DOX NPs and CS−DOX NPs exhibited pronounced hypochromicity upon DOX complexation (Figure 2a) confirming efficient π−π stacking interactions between the drug and the fluorescent probe. However, when UV absorbance was measured at elevated temperatures, a partial increment in absorbance was observed (Figure 2c) indicating the disruption of π−π-stacking interactions. This interaction was also evident from the fluorescence emission spectroscopy studies because these drug-loaded NPs show quenching of fluorescence upon complexation with DOX confirming efficient stacking interactions (Figure 2b). E
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 3. Flow cytometry histogram showing different CD44 expression levels in HCT116 and MCF-7 cell lines.
Figure 4. Flow-cytometric analysis of the uptake of (a) HA NPs and (b) CS NPs by HCT116 and MCF-7 cell lines in the presence and absence of 7.5 kDa HA (10 mg/mL).
Figure 5. Confocal laser scan microscopy images of HCT 116 cells treated with 4 mg/mL concentration of (a) HA−FTSC and (b) CS−FTSC for 4 h. Panels i−iv represents dark field, nuclear staining with DAPI, overlay image, and bright field image, respectively, in panels a and b.
Evaluation of Drug Release from DOX-Loaded NPs. We performed the drug release efficiency test by dialysis method to ascertain the drug release under physiological conditions (Figure 2d). Free drug and drug physically mixed with HA polymer were used as controls, to understand the influence of electrostatic interactions on drug release kinetics. The HA−DOX NPs and CS−DOX NPs displayed slow and sustained release of DOX with near zero-order drug release for 96 h without any burst release and displayed an overall pseudofirst-order release profile in the 14 day release experiment. Such sustained drug release (zero-order for over 96 h) without burst release is generally not observed in any self-assembled HA or CS based nanoformulation16,26,28,33 suggesting the role of π−π stacking interactions. The pseudo-first-order kinetics showed rate constants of (13.6 ± 0.8) × 10−3 h−1 and (10.6 ± 0.3) × 10−3 h−1, which corresponded to drug release half-lives of 51
and 65 h for HA−DOX NPs and CS−DOX NPs, respectively (Figure S2 in SI). The free DOX and DOX mixed with HA showed burst release properties. Because DOX and fluorescein are known to form a compact FRET pair,40 DOX loading could not be ascertained by UV spectroscopy. In addition, both DOX and FTSC have characteristic peaks in the same UV range (λmax 485 and 494 nm, respectively). Attempts to measure DOX concentration by disrupting the NPs in 100% DMSO, followed by centrifugation at 15 000 rpm was also not successful. Finally, we performed the release experiments (14 days, until no increment in UV absorbance was observed; Figure S2 in SI) in order to quantify the DOX loading to HA and CS NPs. The UV from the final release media accounted for the total DOX loading, which was estimated using the extinction coefficient of 11 500 M−1 cm−1 at 485 nm.41 The percentage of drug loading was estimated to be 3% and 4% w/w, which corresponds to F
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 6. Dose dependent cytotoxicity of DOX and DOX-loaded nanoparticles in HCT116 and MCF-7 cell lines.
loading efficiency of 60% and 80% for HA−DOX NPs and CS−DOX NPs, respectively. However, free DOX displayed fast release kinetics with an approximately 90% release within 24 h (Figure 2d). Role of Cell Surface CD44 Receptors for Cellular Uptake. In order to elucidate the role of CD44 in the cellular uptake of HA NPs and CS NPs, we estimated the levels of CD44 receptors in two human cell lines, namely, human colorectal carcinoma cell line HCT116 and human breast cancer cell line MCF-7 by flow cytometry. We found that HCT116 expressed high levels of CD44 receptors compared with MCF-7 cells (Figure 3). Thereafter, we performed receptor blocking studies using low molecular weight HA (7.5 kDa)33 by FACS to evaluate the role of HA receptors on cellular uptake of HA- and CS NPs. These experiments revealed that CD44 positive HCT116 cells showed higher uptake of HA NPs compared with MCF-7 cells. Pretreatment of cells with free HA reduced uptake of HA NPs in both HCT116 and MCF-7 cells. For CS NPs, there was equal uptake of NPs in both cell lines; however addition of freeHA reduced the uptake in HCT116 cells but not in MCF-7 cells (Figure 4 and Figure S3 in SI). Comparison of HA NPs and CS NPs clearly suggested that HA uptake is indeed HA receptor dependent but CS NPs have an alternate uptake mechanism. This is consistent with previous observation that CS NPs undergo clathrin-mediated endocytosis in addition to the CD44-mediated endocytosis mechanism.42 Cellular uptake of HA and CS NPs was also confirmed by confocal laser scanning microscopy in HCT116 cells, which further validated the cellular uptake of HA NPs and CS NPs. The confocal study clearly revealed that both HA NPs and CS NPs underwent efficient cellular uptake, as well as nuclear localization as confirmed by DAPI staining (Figure 5 and Figure S4 in SI). Cytotoxic Evaluation of DOX-Loaded NPs. We then evaluated the cytotoxicity and caspase activity induced by HA− DOX NPs and CS−DOX NPs using the ApoTox-Glo Triplex assay following the manufacturer’s protocol. In this assay, we measure live-cell protease activity of viable cells. This live-cell protease becomes inactive upon the loss of cell membrane integrity. The IC50 estimation (50% inhibitory concentration) for HA−DOX NPs, CS−DOX NPs, and free DOX was performed by logarithmic curve fitting of cell viability (%) using Graphpad Prism software against DOX equivalents (Figure 6). These experiments clearly showed a dose-dependent cytotoxicity for both HA−DOX NPs and CS−DOX NPs (Table 3). Comparison of HA−DOX NPs and CS−DOX NPs showed that HA−DOX NPs showed similar relative cytotoxicity (∼1.8-
Table 3. Cytotoxicity of HA−DOX NPs and CS−DOX NPs in Different Human Cell Lines cell lines
DOXa
HA−DOX NPsa (fold change)b
CS−DOX NPsa (fold change)b
HCT116 MCF7
223.87 363.07
407.38 (−1.82) 680.30 (−1.87)
239.88 (−1.07) 493.51 (−1.36)
a
IC50 value in nM concentration. bFold decrease in activity with respect to DOX.
fold lower with respect to free DOX) in both cell lines. This observation is contrary to our previous study using HA−DOX conjugates that showed 3-fold higher cytotoxicity in HCT116 compared with that in MCF-7 cells line. We believe this discrepancy is due to the size of HA (51 kDa) that was used in this study compared with our previous HA−DOX conjugate (150 kDa).13 It is known that HA uptake is size dependent with size above 130 kDa to be ideal for CD44 mediated endocytosis.43 CS−DOX NPs on the other hand showed ∼35% higher cytotoxicity in HCT116 cells compared with MCF-7 cells, which could be attributed to CD44 and clathrin dependent uptake. Caspase 3/7 Activities and Confocal Microscopy. To understand the mechanism underlying cytotoxicity, we measured the caspase 3/7 activity in these cell lines using the Caspase-Glo assay, which provides a luminescent signal proportional to the amount of the caspase activity in the cell lysate. The basal caspase activity of the cell was taken as 100%. We found that HCT116 cells showed 5−7-fold higher caspase activity in a concentration-dependent manner. However, in MCF-7 cells, we did not show any caspase 3/7 activity (Figure 7). This indicates that DOX induced apoptosis in HCT116 cells in a caspase-3/7 dependent manner, whereas in MCF-7, apoptosis occurs in a caspase 3/7 independent manner (Figure 7). Hematological Evaluation of HA and CS NPs. We further evaluated the hemocompatibility of our NPs using the ex vivo whole blood loop model.34 Specifically, we assessed the activation of the inflammatory response of native biopolymers (HA and CS) and the corresponding NPs (HA NPs and CS NPs) by the enzyme-linked immunosorbent assay (ELISA) using fresh non-anticoagulated human whole blood. The use of anticoagulant in these types of experiments could distort the results, and consequently they will not be reliable. We investigated different hematological parameters of the complement, the coagulation, and the kinin systems, which are important components of the innate immune system. The complement pathway is an integral part of the innate immune G
DOI: 10.1021/acsami.6b06823 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 7. Estimation of caspase 3/7 activities in two cell lines treated with drug- or drug-loaded nanoparticles. The baseline caspase 3/7 levels in each cell line are considered as 100%.
coagulation by evaluating the generation of contact activation markers in blood when in contact with NPs or biopolymers. For this purpose, we investigated the activation of the coagulation system by determining the levels of thrombin− antithrombin (TAT) complexes, which is the parameter of choice to assess coagulation activation (Figure 9). At lower
system, responsible for eradicating invasive pathogens and foreign particles. The coagulation and kinin pathways are responsible for the generation a blood clot and contribute to inflammation, blood pressure control, and pain. Initial analysis of the coagulation activation revealed that the formation of factor XIa−AT complexes, factor XIa−C1 INH complexes, factor XIIa−AT complexes, factor XIIa−C1 INH complexes, kallikrein−AT complexes, and kallikrein−C1 INH complexes was minimal and hardly detectable via ELISA (data not shown). These results clearly prove that the native biopolymers possessed excellent blood compatibility. Platelet Consumption. We then analyzed the concentration-dependent platelet aggregation of native biopolymer and NPs. As anticipated the native HA or CS did not result in significant platelet consumption at the different concentrations tested. However, the HA NPs and CS NPs showed platelet aggregation when incubated at higher concentrations (0.8 mg/ mL) leading to a significant reduction of the platelet count from 100% baseline to 78% ± 4.6% (P < 0.01) for HA NPs and 80% ± 6.2% (P < 0.05) for CS NPs (Figure 8). This indicates that the self-assembled biopolymer (with less than 5% chemical modification) behaves differently from the native polymer when in contact with human whole blood at a higher concentration and induces platelet aggregation. Activation of the Coagulation System. We further studied the intrinsic pathway or the contact system of
Figure 9. Coagulation system activation was analyzed by determining the concentration of formed thrombin−antithrombin complexes in EDTA plasma via ELISA (N = 3). One-way ANOVA followed by Dunn’s post hoc test for significance vs baseline was used. Error bars indicate mean ± SD: *P < 0.05; **P < 0.01; ***P < 0.001.
concentrations, we observed that neither native biopolymers nor their NPs showed any activation of the coagulation system compared with that at baseline values. However, at a high concentration (0.8 mg/mL) self-assembled HA NPs and CS NPs triggered the coagulation cascade when incubated with blood. The corresponding TAT levels of HA NPs and CS NPs were 209.3 ± 116.1 μg/L (P < 0.05) and 297.7 ± 176 μg/L (P < 0.01), respectively. This suggests that HA NPs and CS NPs could be utilized for drug delivery applications when used at lower concentrations (
