- Email: [email protected]
SPIONs functionalized with small peptides for binding of lipopolysaccharide, a pathophysiologically relevant microbial product
Accepted Manuscript Title: SPIONs functionalized with small peptides for binding of lipopolysaccharide, a pathophysiologically relevant microbial product Authors: Weronika Karawacka, Christina Janko, Harald Unterweger, Marina Muhlberger, ¨ Stefan Lyer, Nicola Taccardi, Andriy Mokhir, Wolfgang Jira, Wolfgang Peukert, Aldo R. Boccaccini, Mikhail Kolot, Richard Strauss, Christian Bogdan, Christoph Alexiou, Rainer Tietze PII: DOI: Reference:
S0927-7765(18)30778-1 https://doi.org/10.1016/j.colsurfb.2018.11.002 COLSUB 9767
To appear in:
Colloids and Surfaces B: Biointerfaces
Received date: Revised date: Accepted date:
1 August 2018 16 October 2018 1 November 2018
Please cite this article as: Karawacka W, Janko C, Unterweger H, Muhlberger ¨ M, Lyer S, Taccardi N, Mokhir A, Jira W, Peukert W, Boccaccini AR, Kolot M, Strauss R, Bogdan C, Alexiou C, Tietze R, SPIONs functionalized with small peptides for binding of lipopolysaccharide, a pathophysiologically relevant microbial product, Colloids and Surfaces B: Biointerfaces (2018), https://doi.org/10.1016/j.colsurfb.2018.11.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
SPIONs functionalized with small peptides for binding of lipopolysaccharide, a pathophysiologically relevant microbial product Weronika Karawacka1, Christina Janko1, *, Harald Unterweger1, Marina Mühlberger1, Stefan Lyer1, Nicola Taccardi2, Andriy Mokhir3, Wolfgang Jira4, Wolfgang Peukert5,6, Aldo R. Boccaccini6,7, Mikhail Kolot8, Richard Strauss9, Christian Bogdan10,11, Christoph Alexiou1, Rainer Tietze1 1
Department of Otorhinolaryngology, Head and Neck Surgery, Section of Experimental Oncology and Nanomedicine (SEON), Else Kröner-Fresenius-Stiftung-Professorship, Universitätsklinikum Erlangen, 91054 Erlangen, Germany 2 Institute
IP T
of Chemical Reaction Engineering, Friedrich-Alexander-Universität (FAU) ErlangenNürnberg, 91058 Erlangen, Germany 3
Department of Chemistry and Pharmacy, Organic Chemistry Chair II, FAU 91058 Erlangen, Germany 4
SC R
Federal Research Institute of Nutrition and Food, Max Rubner-Institut, 95326 Kulmbach, Germany 5
Institute of Particle Technology (LFG), FAU, 91058 Erlangen, Germany
6
Interdisciplinary Center for Functional Particle Systems (FPS), FAU, 91058 Erlangen, Germany
7
N
U
Institute of Biomaterials, Department of Materials Science and Engineering, FAU, 91058 Erlangen, Germany Department of Biochemistry and Molecular Biology, Tel-Aviv University, Tel-Aviv 69978, Israel
9
Deptartment of Medicine 1, Universitätsklinikum Erlangen, 91054 Erlangen, Germany
10
und
Hygiene,
Universitätsklinikum
Medical Immunology Campus Erlangen, FAU Erlangen-Nürnberg, 91054 Erlangen, Germany
ED
11
Immunologie
M
Institut für Klinische Mikrobiologie, Erlangen,91054 Erlangen, Germany
A
8
PT
* Corresponding author; [email protected]; Glückstraße 10a, 91054 Erlangen
A
CC E
Graphical abstract
Highlights
16-mer peptides from binding motives of agglutinating salivary proteins act as specific pathogen scavengers. A reliable orthogonal binding strategy of peptides to stable SPIONs was developed. Specific pathogen binding capacity of peptide-functionalized SPIONs is clearly dependent on the peptide sequence. The results could support the development of new treatment options for acute systemic infectious diseases.
IP T
Abstract
PT
ED
M
A
N
U
SC R
Systemic inflammation such as sepsis represents an acute life-threatening condition, to which often no timely remedy can be found. A promising strategy may be to functionalize magnetic nanoparticles with specific peptides, derived from the binding motives of agglutinating salivary proteins, that allow immobilization of pathogens. In this work, superparamagnetic iron oxide nanoparticles with stable polycondensed aminoalkylsilane layer were developed, to which the heterobifunctional linkers N-succinimidyl 3-(2-pyridyldithio)-propanoate (SDPD) and N-succinimidyl bromoacetate (SBA) were bound. These linkers were further chemoselectively reacted with the thiol group of singularly present cysteines of selected peptides. The resulting functional nanoparticles underwent a detailed physicochemical characterization. The biocompatibility of the primarily coated aminoalkylsilane particles was also investigated. To test the pathogen-binding efficacy of the particles, the lipopolysaccharide-immobilization capacity of the peptide-coated particles was compared with free peptides. Here, one particle-bound peptide species succeeded in capturing 90% of the toxin, whereas the degree of immobilization of the toxin with a system that varied in the sequence of the peptide dropped to 35%. With these promising results, we hope to develop extracorporeal magnetic clearance systems for removing pathogens from the human body in order to accelerate diagnosis and alleviate acute disease conditions such as sepsis.
CC E
Key words: Sepsis, Magnetic Clearence, Salivary Agglutinin Peptide, Pathogen Immobilization, Superparamagnetic Iron Oxide Nanoparticles Number of words/tables/figures: 5990; 1; 7. Introduction
A
Sepsis is a massive systemic reaction to infection, associated with life-threatening organ damage. Patients with severe sepsis and septic shock (a severe sepsis with low blood pressure, associated with high mortality) require complex therapy measures in intensive care medicine of hospitals 1. Pathogenic microorganisms express molecules that are known as pathogen-associated molecular patterns (PAMPs) 2. When those PAMPs enter the body, the host reacts with multiple immune responses involving several molecules and cellular components. During sepsis, these substances released into the bloodstream in order to fight the infection, trigger systemic inflammatory responses, mainly due to dysregulated host reactions to microbial infections 3. Blood cultures are still standard to identify causative pathogens. In up to 50% of cases of bloodstream infections, blood cultures remain negative. This can be due to antibiotic pretreatment,
to low number of circulating pathogens, or to organisms which are not cultivable.4 Pathogen counts in blood vary by infection and over time during sepsis. In most cases, numbers of viable microorganisms in the blood are low: about 1 to 10 Colony forming Units (CFU)/mL blood. Blood concentrations of PAMPs however vary depending on time and type of infection.5
N
U
SC R
IP T
Complex cytokine and coagulation cascades trigger severe inflammatory processes that in many cases cannot be stopped in time to save the patients. Current treatment regimens aim at a reliable early detection and treatment of the infection with antibiotics, as wells as intravenous fluid therapy, ventilatory, and hemodynamic support. Recently, the immobilization of pathogens to suppress the continuous initiation of inflammatory process and for rapid diagnostics has attracted considerable attention. Clinically approved devices for extracorporeal clearence of pathogens already exist (e.g. polymyxin-B columns for hemoperfusion, Toramyxin™). However, these techniques are only suitable for gram-negative germs and otherwise the clinical success seems ambivalent 6. Newer approaches, especially these based on particulate clearance methods, utilize antibodies or aptamers, which can intercept pathogens efficiently but too selectively7,8. In order to circumvent this limitation, fragments of erythrocytes, which have a high non-specific binding affinity to different sepsis-associated pathogens9 can be used, although this approach can potentially give rise to allergic reactions. To overcome the problem of highly selective binding and to avoid immune reactions, other mechanisms for the immobilization of sepsis-associated pathogens are proposed. Human saliva, for example, contains peptides that can absorb a wide range of microbial invaders10,11. This could be a new approach to bind the microbial triggers as closely as possible, but without imposing unnecessary constraints resulting from the too specific complementary binding mechanisms.
CC E
PT
ED
M
A
A useful intrinsic property of superparamagnetic iron oxide nanoparticles (SPIONs) is their magnetization by external magnetic fields, which is of great importance in the field of biomedical applications12. The magnetic properties of SPIONs enable magnetic separation of pathogens bound to particles. This serves both to reduce the pathogen load and to accelerate the diagnosis, if pathogens in patient blood samples can be effectively enriched and extracted. To obtain a pathogen-binding nanosystem, the specific peptides must be immobilized on the surface of nanoparticles. As with other macromolecular immobilization strategies13,14, this requires an orthogonal connection to the particle surface. The binding between peptide and pathogen is strongly dependent on the conformation of the binding peptide. Meanwhile, there are various chemoselective strategies for binding of complex molecules, without blocking their actual capturing capacity15. The principle is that the respective functional group is either solitary at the desired binding position of the effective biomolecule, or it is exposed by means of an appropriate protective group strategy. A corresponding heterobifunctional linker can represent the interface between the nanoparticulate drug carrier and the active substance.
A
In our work, we have investigated two selected peptide sequences with regard to their immobilization capacity against a model toxin. The peptides were bonded to the surface of SPIONs, which were coated with aminopropyl triethoxysilane (APTES). We used different protocols16-18, and developed our own synthesis process on their basis. Further, we tested the suitability of two different linkers, N-succinimidyl 3- (2-pyridyldithio)-propanoate (SPDP) and Nsuccinimidyl bromoacetate (SBA), for binding peptides. The peptides each have terminal cysteine, which is solitary in the sequences used here. This enables the peptides to chemoselectively bind to the nanoparticles via thiol groups. Materials and methods Four batches of NPs were synthesised with two methods of purification – either magnetic washing (batches1-3) or dialysis (batch 4). Particles were afterwards dried and kept as powder (batches 1-
3), or as a water suspension (batch 4). Batch 4 was divided into 2 sub-batches with respect to its reaction to external magnetic field. Results for batches 1, 2 and 3 were averaged, as they were prepared in the same way, to gain insight into uniformity of this SPIONs’ fabrication technique. They are further referred to as 1/3. Further information on materials and methods is provided in the Supplemental files. Peptide Functionalization of the Nanoparticle Surface
N
U
SC R
IP T
3 different peptides were used in the procedures: Fluorescein isothiocyanate (FITC)-bearing peptide with the sequence FITC-TVCCC, purity 81.47%, MW 1030.21 Da, stock solution 2 mM in corresponding buffer prepared prior to use. Peptide A: CVTGWSGRYLVEVRGQ, purity 92.97%, MW 1810.07, stock solution 1 mg/mL in corresponding buffer. Peptide B: RKQGRVEVLYRASWGTVC, purity 90.44%, MW 2108.46, stock solution (2 mM) in corresponding buffer prepared prior to use. Prior to peptide functionalization, the binding of the two different linker molecules was performed as follows: Linking SPDP to Fe3O4@APTES: SPDP was synthesized as described in supplemental files. Fe3O4@APTES (stock solutions volumes corresponding to 1 mg Fe) were diluted with phosphate buffer (10 mM at pH 7.5) and SPDP stock solution was added (25 μL, 20 mM in DMFdry). Suspensions were mixed (1 h at room temperature (RT); thermomixer, 1000 rpm). Supernatants were magnetically decanted. Afterwards, nanoparticles were washed with phosphate buffer and magnetically decanted resulting in Fe3O4@APTES@SPDP nanoparticles.
ED
M
A
Linking SBA to Fe3O4@APTES: SBA was obtained as described in supplement. Samples were mixed at RT for 30 min. Fe3O4@APTES (stock solutions volumes corresponding to 1 mg Fe) were diluted in the dark with borate buffer (50 mM at pH 8.4), and SBA stock solution was added (25 μL, 20 mM in DMFdry). Supernatants were magnetically decanted. Nanoparticles were washed with borate buffer and magnetically decanted, resulting in Fe3O4@APTES@SBA nanoparticles.
CC E
PT
Peptide coupling to Fe3O4@APTES@linker: Fe3O4@APTES@linker nanoparticles were redispersed in buffer (phosphate for SPDP, borate for SBA). Peptide solution (bearing 0.1 μmol of the peptide) was added to the dispersion, and the samples were mixed at RT (800 rpm, thermomixer) either overnight (SPDP), or for 1 h (SBA).19 Supernatants were magnetically decanted. Nanoparticles were washed with the appropriate buffer (1 mL) and magnetically decanted, resulting in Fe3O4@APTES@linker@peptide nanoparticles. Estimations on the coupling efficiencies Estimation of existing amino groups
A
Estimation through Si to Fe ratio: The concentrations of Si/Fe in μmol/mg were calculated from ICP-AES measurements, which corresponded 1:1 to NH2/Fe in μmol/mg. Active amino groups through 2(1H)-pyridinethione release: Supernatants of Fe3O4@APTES@SPDP were collected to measure the not-bonded SPDP. Phosphate buffer containing SPDP alone (1.25% v/v) was used as control. Dithiothreitol (DTT, 100 mM in phosphate buffer) was added to supernatants and control samples. The samples were mixed for 20 min and absorbance at 343 nm was recorded. The difference of absorbance corresponds to the concentration of 2-pyrydinethione (ε = 5.81·103 dm3/cm·mol, determined experimentally in phosphate buffer). The control sample gave maximal possible concentration of SPDP, whereas the samples showed the amount of SPDP that was not bonded to Fe3O4@APTES. In order to determine the bonded SPDP, nanoparticles were redispersed in phosphate buffer and stock DTT
was added. The samples were mixed for 15 min at 1000 rpm (thermomixer) and centrifuged (5 min, 13400 rpm, MiniSpin, Eppendorf). Supernatants were collected, diluted with phosphate buffer to 2 mL and the absorbance at 343 nm measured. All measurements were performed in triplicates, results were averaged. The amount of bonded SPDP should correspond to the amount of surface amino groups available for coupling reactions. Absorbance at 280 nm was measured to estimate the dependence between 343 nm and 280 nm absorbance originating from 2-pyrydinethione. Estimation of peptide coupling efficiency
SC R
IP T
Peptide coupling efficiency through fluorescence measurements: FITC-TVCCC was used as a model peptide with excellent fluorescence quantum yield and high water solubility. Fluorescence strength is however sensitive to pH changes and relatively easy to quench 20. Measurements were performed on Filter-Max F5 Plate reader (Molecular Devices, Biberach an der Riss, Germany) with excitation at 485 nm and emission at 535 nm. Measurements were performed in triplicates, results were averaged.
A
N
U
Efficiency of peptide coupling to Fe3O4@APTES@SPDP: FITC-TVCCC peptide was coupled to nanoparticles as described above. Control sample was prepared by mixing phosphate buffer, SPDP stock solution and FITC-TVCCC stock solution. Absorbance was measured in UV-Vis spectrometer at 343 nm against pure phosphate buffer to detect the released 2-pyrydinethione as described previously. As control, peptides coupled to Fe3O4@APTES@SPDP were cleaved with DTT. The nanoparticles were redispersed in phosphate buffer and DTT stock solution was added. After 15 min of incubation, supernatants were collected, diluted with phosphate buffer and fluorescence was measured. FITC-TVCCC concentrations were estimated according to a calibration curve in the concentration range 5.0·10-4-12.5·10-4 μmol/mL.
ED
M
Efficiency of peptide coupling to Fe3O4@APTES@SBA: FITC-TVCCC peptide was coupled to nanoparticles as described above and the measurement was performed similar to the SPDP samples, but with borate buffer. FITC-TVCCC concentrations were estimated according to the calibration curve obtained for FITC-TVCCC peptide in borate buffer for concentration range 2.00·10-4-6.25·10-4 μmol/mL.
PT
Through UV-Vis absorbance: The peptide sequences used in this study contain one tryptophan (W) and one tyrosine (Y), showing a characteristic absorption band at 280 nm in UV-Vis spectrum 21 . Calibration curves were determined for each peptide in both phosphate and borate buffer for the concentration range 0.0125-0.1000 μmol/mL.
A
CC E
To Fe3O4@APTES@linker: Peptide A or B was coupled to nanoparticles as described above. The nanoparticles were redispersed in the appropriate buffer and DTT stock solution was added. Mixtures were shaken at RT for 15 mins and centrifuged. Supernatants were collected and diluted with buffer, after which the absorbance was measured at 280 and 343 nm and background absorbance was subtracted. Investigation of cell death and DNA degradation Jurkat cells (2.5x105) were seeded in 1 mL medium into 48 well plates. Cells were treated with Fe3O4@APTES or SEONLA-BSA, derived from a previous study22, at final Fe concentrations of 25, 50, 100 and 200 μg/mL. Nanoparticle-untreated cells and cells treated with H2O served as controls. After 24 and 48 hours of incubation, cells were mixed and 50 µL aliquots were taken and analysed. Staining was performed with freshly prepared staining solution consisting of AnnexinA5-FITC (AxV, 1 µL/mL), Hoechst 33342 (Hoe, 10 µg/mL), propidium iodide (PI, 6.6 ng/mL), 1,1´,3,3,3´,3´hexametylindodicarbocyanine iodide (DiI, 0.4 μL/mL) in Ringer’s solution. Staining solution (200 µL) was added to 50 µL cell suspensions, incubated for 30 min at 4 °C and measured in flow
cytometer.23 Cell cycle and DNA degradation was investigated using PI-Triton staining (0.1% Triton X-100, 1 mg/mL sodium citrate and 50 μg/mL PI in water). For this purpose, 50 µL of cell suspensions were incubated with 400 µL staining solution overnight at 4 °C in the dark and then monitored using flow cytometry. Estimation of particle uptake
SC R
IP T
Jurkat cells (2.5x105) were seeded in 1 mL medium into 48 well plates. Cells were treated with Fe3O4@APTES or SEONLA-BSA at final Fe concentrations of 25, 50, 100 and 200 μg/mL. Nanoparticle-untreated cells and cells treated with H2O served as controls. Cells were spiked with Lucifer Yellow (LY, 10 μL, 10 μg/mL). After 24 and 48 hours incubation the cells were mixed and 50 µL aliquots were analysed. Staining was performed with freshly prepared staining solution consisting of monobromobimane (MBB, 0.3 µL/mL, stock 100 mM) and DiI (0.4 μL/mL) in Ringer’s solution. 200 µL staining solution was added to 50 µL cell suspensions, incubated for 30 min at 4 °C and measured in flow cytometer. Particle uptake was additionally monitored with optical microscopy (magnification 10x and 40x, Axiovert 40 CFC, Carl Zeiss AG, Germany) with cells incubated for 48 h with nanoparticles (NPs). Endotoxin-binding studies
𝑆𝑖𝑔𝑛𝑎𝑙 𝑟𝑒𝑐𝑜𝑣𝑒𝑟𝑦 𝑟𝑎𝑡𝑒 = 100 ∙
M
A
N
U
Endotoxin-binding studies were performed with EndoZyme Recombinant Factor C Assay kit (Hyglos, Germany). Nanoparticles from batch 4A were conjugated with peptides A or B through SBA linker. Binding efficiency was evaluated by UV-VIS (@280 nm) of supernatants and dispersions of 0.050, 0.0250 and 0.0125 μmol peptide/mL in water were prepared. Furthermore, the solutions of peptides A and B in water at the same concentrations were prepared. As a blank nanoparticle sample, nanoparticles coupled with citric acid (CA) and acetylsalicylic acid (ASA) were used (procedure to be found in Supplementray Information) to estimate signal recovery rate, according to Equation 1: (𝐿𝑃𝑆 𝑠𝑝𝑖𝑘𝑒𝑑 𝑖𝑛 𝐶𝐴 𝑎𝑛𝑑 𝐴𝑆𝐴 𝑐𝑜𝑛𝑗𝑢𝑔𝑎𝑡𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒) −(𝐿𝑃𝑆 𝑖𝑛 𝑛𝑜𝑛−𝑠𝑝𝑖𝑘𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒) 𝑎𝑑𝑑𝑒𝑑 𝐿𝑃𝑆
ED
(Equation 1): LPS = Lipopolysaccharide
CC E
PT
Endotoxin (at 5.00, 2.50, 1.25, and 0.0 EU) was added to 200 μL of each nanoparticle dispersion and peptide solution, resulting in peptide end concentration dilutions of factors 1.20, 1.10, 1.05 and 1.00, respectively. LPS-spiked samples and calibration samples were incubated in thermomixer for 1 h (37°C, 1200 rpm) and subsequently centrifuged (1 h, 8700 relative centrifuge force (rcf)). The precipitability of the LPS peptide agglutinate was separately tested by centrifugation at 18000 rcf for 1 h. Supernatants (100 μL) were analysed for endotoxin content according to the manufacturer`s instructions. The reaction was monitored for 90 min at 37°C in 15 min intervals in a Microplate Reader Filter Max F5 (excitation 360 nm/emission 465 nm). Experiments were performed in duplicate and the results were averaged.
A
Results and discussion Properties of the nanoparticles The coprecipitation method used in this study is the fastest, easiest and cheapest way to fabricate SPIONs, also in a large scale, with acceptable control over size and yielding uniform batches when optimised.24 In situ coating by addition of coating agents performed here enables to omit tedious iron oxide purification process between the synthesis and coating steps. Successful APTES coating of produced SPIONs was confirmed with FT-IR spectroscopy (Suppl. Information, Figure S1/S2). Elemental composition and size of obtained nanoparticles are presented in Table 1. Table 1. Elemental composition and hydrodynamic size of Fe3O4@APTES nanoparticles.
Mass % Molar ratio SMD [nm] Fe Si N H C N/Si C/Si 1/3 63.10 0.15 0.72 1.51 1.21 4.8 18 191.8 Δ 1/3 0.64 0.01 0.08 0.58 0.90 0.4 12 13.1 4A 70.36 0.11 0.47 0.89 2.04 8.2 42 100.1 4B 65.90 0.11 0.62 0.69 2.09 11.5 45 169.2 SMD: standard mean diameter, PDI: Polydispersity index, SF: steepness factor Batch
PDI
SF
0.30 0.01 0.34 0.30
1.16 0.01 1.18 1.16
SC R
IP T
APTES and its polymer, 3-aminopropylsiloxane, have theoretically a stoichiometric N/Si ratio of 1/1 and C/Si ratio of 3/1. Here however, all batches had much higher ratios of N/Si and C/Si. Interestingly, batches 1-3, which were magnetically washed with methanol, seemed to be purer than batch 4, which was dialysed. There are several reasons for increased presence of nitrogen and carbon in the particles. Additional nitrogen may come from ammonia being trapped between nanoparticle cores or adsorbed on their surface alongside with APTES. The presence of carbon contamination on APTES-modified silica has been previously reported.25 As a by-product of APTES hydrolysis, ethanol might have been trapped within nanoparticles during fabrication. Theoretically, some ethoxy groups may have not been hydrolysed, however, this is unlikely due to the fact that the reaction was conducted in water.
ED
M
A
N
U
Nanoparticles with mean hydrodynamic diameter between 100 and 200 nm were obtained. All batches had a polydispersity index (PDI) of 0.3 and steepness factor (SF) of 1.2, indicating their uniformity and narrow size distribution. Particles made in the first 3 batches were larger (190 nm), as compared to the dialysed nanoparticles (4A ~100 nm and 4B~170 nm). Particle size plays crucial role in biodistribution and pharmacokinetics. NPs over 100 nm in diameter do not extravasate far beyond the blood vessel as they remain trapped in the extracellular matrix between cells, whereas nanoparticles bigger than 200 nm accumulate in spleen and liver where they are processed by mononuclear phagocyte system.26 Particles smaller than 200 nm can be sterile filtered and thus are easier to prepare for biological applications.27 Because of that aspect, biological investigations within this study were conducted with batch 4A containing nanoparticles of ~100 nm in hydrodynamic diameter.
PT
Single core size was determined with TEM micrographs (Suppl. Information, Figure S 3/4). The mean core size was 8.8 nm (median of 8.0 nm). No cores bigger than 25 nm were observed, meaning that all precipitated nanoparticles were within superparamagnetic size range.28 Diffraction patterns indicated that nanoparticles were composed of magnetite, partially oxidised to maghemite (Suppl. Information, Figure S5).
A
CC E
ζ-potential is a measure of nanoparticles’ surface charge and thus electrostatic stability. Colloids with the absolute value of ζ-potential higher than 30 mV are considered electrostatically stable.26 Here, ζ-potential reached +25 mV at pH < 3.2 and dropped to +10 and +15 mV at physiological pH (7.4). It then remained stable up to pH 9. This value of ζ-potential was too low for the nanoparticles to remain stable in a solution over prolonged time, resulting in agglomeration and disturbing the measurement. However, aggregation is not necessarily disadvantageous when used ex vivo, as it facilitates magnetic separation of NPs. The isoelectric point (IP) of APTES-coated magnetite NPs oscillated between pH 9.1 and 10.2. The observed pH-dependent ζ-potential trends corresponded well to the changes in hydrodynamic size (see Supplementary Information). Above pH 9.5, nanoparticles formed agglomerates, being most profound around pH 10.5. Size did not differ much at lower pH values (Suppl. Information, Figure S6). Higher concentrated (~6 mg Fe/mL) particles were stable enough in acidic solution and little to no precipitation occured, even after a 8 week storage.
All batches exhibited reasonably high volume magnetic susceptibilities (>4·10-3 in SI units) and drop in Xv was registered over time, indicating the propagation of magnetite oxidation to maghemite (Suppl. Information, Figure S7). Amount of existing NH2 groups
IP T
Estimation of the amount of amino groups was important for subsequent coupling reactions. In principle, each surface NH2 group can become a binding point, but a steric hindrance might occur with bulkier substances. ICP-AES composition analysis allowed to assess the maximum possible amount of amino groups for each batch of nanoparticles. This was 0.09 and 0.06 μmol/mg NH2/Fe (Si/Fe) for batches 1/3 and 4A/B, respectively. These results were further referred to as “theoretical” amount of NH2 groups per mg Fe for each batch of Fe3O4@APTES. Determination of reactive amino groups
A
CC E
PT
ED
M
A
N
U
SC R
One molecule of 2(1H)-pyridinethione can be cleaved from each molecule of SPDP with DTT (Figure 1). Here, the 2(1H)-pyridinethione amounts were measured (a) in the supernatant yielding indirectly the amount of NH2 groups available for SPDP bonding, as well as (b) released from nanoparticles, yielding this amount directly. Both values should be consistent with each other. The measured fraction of NH2 groups, compared to theoretical amount, yields a % of active amino groups.
Figure 1. Schematic process of peptide coupling to Fe3O4@APTES@SPDP and its cleavage with DTT for the estimation of reaction efficiency. Prior to the experiment, the correlation between 2(1H)-pyridinethione absorbance at 343 nm and SPDP concentration in phosphate buffer has been established, along with the correlation between signal coming from 2(1H)-pyridinethione at 343 nm and at 280 nm. The amount of 2(1H)-pyridinethione cleaved from SPDP measured in the supernatants was comparable to the amount cleaved from functionalised nanoparticles (Figure 2). It means that both the indirect (from supernatant) and the direct (form NPs) method of measuring, are consistent. The
values obtained using the indirect method in batch 4A and 4Bwere slightly increased, possibly by traces of nanoparticles still present in the supernatant after magnetic separation and, therefore, disturbing the background signal. According to this experiment, most NH2 groups (~80%) present in the Fe3O4@APTES were active and available for binding of SPDP. 0.09 0.07 0.06 0.05
from unbonded SPDP
0.04
released from NPs
0.03
average
0.02
theoretical
SC R
0.01 0.00 1-3 av
4A
IP T
NH2 (μmol/mg Fe)
0.08
4B
Batch
N
U
Figure 2. SPDP-active NH2 groups of Fe3O4@APTES. Experiments were performed in triplicates, shown are the mean values with standard deviations.
A
Peptide coupling efficiency FITC-TVCCC
PT
ED
M
FITC-TVCCC, a fluorescent dye functionalised with a short peptide bearing 3 cysteines at the Cterminal end, was coupled to Fe3O4@APTES through SPDP and SBA. To measure the coupling efficiency, peptide bonded to the nanoparticles via SPDP was cleaved using DTT. NPs were then magnetically separated and the fluorescence of obtained supernatants was measured. Upon coupling via SBA linker, the cleavage of bonded peptide was not possible. Therefore, the supernatants containing unreacted peptide were measured to estimate the bonding capacity indirectly.
A
CC E
As shown in Figure 3, toupling FITC-TVCCC through SBA linker appeared to be more efficient, especially for batches 1/3. While the amount of successfully coupled FITC-TVCCC was fairly constant for SPDP (0.015 μmol/mg Fe), it had deviations up to 25% between batches when SBA was used. But this could be due to the less precise indirect detection methods for the SBA linking. Short peptide bearing a rigid and bulky fluorescein molecule could have encountered spacial difficulties in finding available binding spots, thus only about 1/3 of active animo groups were successfully coupled.
0.06 0.05 0.04 SPDP 0.03
SBA SPDP-active
0.02
IP T
Coupled FITC-TVCCC [μmol/mgFe]
0.07
0.01 0.00 1-3 av.
4A
4B
SC R
Batch
U
Figure 3. The amounts of FITC-TVCCC coupled to different batches of nanoparticles. Experiments were performed in triplicates, shown are the mean values with standard deviations. Peptides A and B coupled with SPDP and SBA
A
CC E
PT
ED
M
A
N
Peptides A and B were coupled to Fe3O4@APTES nanoparticles by SPDP or SBA linker. The amount of coupling was estimated indirectly by measuring the supernatants’ absorbance at 280 nm (maximum absorbance of peptides). For SPDP, the bonding efficiency was also directly measured by cleavage of the coupled peptides from SPDP linker with DTT. In this case, both results were averaged. Peptide concentrations were calculated from calibration curves. The results are depicted in Figure 4.
Figure 4. Peptides coupled to Fe3O4@APTES@SPDP compared to active NH2 groups (left) and the same peptides coupled to Fe3O4@APTES@SBA compared to the theoretical amount of NH2 groups (right). Experiments were performed in triplicates, shown are the mean values with standard deviations.
Using SBA linkers, binding of 0.06 µmol peptide /mg Fe was possible independent of NP batch. Using SDPD liker, this value was reached only for peptide B binding to batches 1-3 and 4A. That value corresponds to the number of active and even of all present amino groups. It must be noted that the results of these experiments might have been disturbed by high background noise of residual nanoparticles, yielding values that appear even slightly higher than theoretically possible. Biocompatibility
A
CC E
PT
ED
M
A
N
U
SC R
IP T
Sterile filtered batch 4A of Fe3O4@APTES nanoparticles was compared with previously reported SEON@LA-BSA nanoparticles obtained according to Zaloga et. al.29. Biocompatibility was determined using AxV-Fitc/PI staining. AxV binds to phosphatidylserine (PS) exposed on the outer plasma membrane layer on apoptotic cells. In necrotic cells, PI intercalates into the DNA after plasma membrane rupture. Ax-PI- cells are considered viable, Ax+PI- cells apoptotic, and PI+ cells necrotic. After 24 and 48 h of incubation with NPs, cell viability was only slightly reduced. APTEScoated SPIONs induced less apoptosis than BSA-coated ones, while causing similar level of necrosis. In all cases, more than 80% of cells remained viable, thus all nanoparticles proved to be cytocompatible (Figure 5A). DiI staining for mitochondrial membrane potential confirmed these findings (data not shown). Cell cycle phases were analyzed using PI-Triton staining. Cells can be differentiated according to their DNA content into cells with diploid DNA (G1 phase), cells with double diploid DNA (G2 phase) and cells with degraded DNA generated by DNAses during cell death (sub G1).30 After 24 h, only minor cell cycle alterations were observed. After 48 h, a dosedependent increase in subG1 DNA was visible for both BSA- and APTES-coated nanoparticle treated cells (Figure 5B). To check whether APTES-coated magnetite induced production of reactive oxygen species (ROS) and concomitant oxidative stress, staining with MBB was performed. In the presence of ROS, cellular glutathione (GSH) is oxidised to glutathione disulphide S. MBB, initially non-fluorescent, reacts with thiols yielding fluorescent adducts. Its fluorescence is therefore the strongest when GSH is present in non-oxidised state.31 After 24 h and 48 h incubation for both SPION types, ROS was dose- and time-dependently produced, as shown by reduced MBB fluorescence (Figure 5C). Since the SPIONs are to be applied in the whole blood, their hemocompatibility, including the effects on erythrocyte hemolysis and plasma coagulation were also investigated, as described in detail in Suppl. Information (Figure S8).
IP T SC R
A
N
U
Figure 5. Biocompatibility of Fe3O4@APTES with Jurkat cells treated: A) Apoptosis and necrosis were determined by AxV/PI staining. B) Cellular DNA content was determined by PIT staining. C) Oxidative stress was evaluated by MBB staining. Numbers indicate Fe concentrations in μg/mL. APTES, APTES-coated SPIONs; BSA, LA/BSA-coated SPIONs. Experiments were performed in triplicates, shown are the mean values with standard deviations.
M
Nanoparticle uptake
PT
ED
The side scatter (SSc) of viable cells measured in flow cytometry was used to estimate the cellular SPION load. We previously showed that the presence of intracellular nanoparticles dosedependently increases the cellular granularity as determined by SSc.32 SSc of Jurkat cells treated with BSA-coated NPs increased only very slightly with the growing concentration of iron and did not further change within 48 h. In contrast, SSc of cells treated with APTES-coated NPs was dramatically increased as compared to control and SEON@LA-BSA -trested cells, showing an iron dose-dependency only at the concentrations below 200 µg/mL. Such high SSc values may suggest adherence of the nanoparticles to the cell surface rather than uptake (Figure 6A).
A
CC E
Whenever a cell phagocytoses a nanoparticle, it co-ingests a small drop of medium. Fluorescent LY in medium will thus be co-ingested when cells take up nanoparticles. A correlation between SSc and LY signal was observed for SEON@LA-BSA (Figure 6B), indicating that NP are rather internalized than only being attached to the cell surface. In cells treated with APTES-coated SPIONs, only a slight increase in LY fluorescence was detected, indicating that either Fe3O4@APTES are not favourably taken up, or that the quenching of LY fluorescence can occur by the massive nanoparticle adhesion to the cell surface.
a)
500
b)
450
2,5
Fluorescence [a.u.]
400 350
SSc [a.u.]
3,0
300 250 200 150 100
2,0 1,5 1,0 0,5
50
24
48
24
APTES 200
APTES 100
APTES 50
APTES 25
BSA 200
BSA 100
BSA 50
BSA 25
H2O
MEDIUM
APTES 200
APTES 100
APTES 50
APTES 25
BSA 200
BSA 100
BSA 50
BSA 25
H2O
0,0
MEDIUM
0
48
SC R
IP T
Figure 6. Determination of nanoparticle uptake. a) Side scatter of viable cells (positive DiI+) analysed by flow cytometry. b) LY staining. Numbers indicate Fe concentrations in μg/mL. APTES, APTES-coated SPIONs: BSA, LA/BSA-coated SPIONs. Experiments were performed in triplicates, shown are the mean values with standard deviations. Endotoxin binding capacity
M
A
N
U
For the future use in sepsis diagnosis and treatment, the designed nanoparticles should efficiently bind LPS and other bacterial toxins in order to remove them from the bloodstream. Here, the two peptides (A and B) were conjugated to Fe3O4@APTES through SBA linker, due to its better peptide binding properties (Figure 4) and their LPS-binding ability was tested and compared to the LPS-binding efficiency of free peptides. Because the free peptides could not be magnetically separated, the samples were centrifuged to provide the same conditions. The presence of both LPS and capture peptide leads to a signal loss in the endotoxin assay due to the formation of an complex consisting of several peptides and LPS molecules11. These complexes can be precipitated by centrifugation at 18000 rcf for 1h (data not shown). Before the experiment, SPION dispersions and peptide solutions were also tested for intrinsic LPS content.
A
CC E
PT
ED
In Figure 7, the effectiveness of LPS-binding by free and particle-bound peptidesis compared. Free peptide A bound between 75% and 85% of spiked LPS (Figure 7a). However, upon conjugation to APTES-coated SPIONs, its efficiency dropped to 35%-75% (Figure 7c). For peptide B, both free and SPION-conjugated peptides removed more than 90% of spiked LPS, even when as little as 0.0125 μmol/mL of peptide was used (Figure 7 b, d). Peptide B was equally efficient when 10 EU/mL of LPS were spiked, as in the presence of 4 times lower concentrations of endotoxins.
IP T SC R
A
N
U
Figure 7. Endotoxin binding capacity. a) Free peptide A, b) Free peptide B, c) Fe3O4@APTES@SBA@A d) Fe3O4@APTES@SBA@B. Numbers in the legend refer to peptide concentrations in μmol/mL. Experiments were performed in triplicates, shown are the mean values with standard deviations.
ED
M
A peptide consisting of the same amino acids as peptide B, but composed in a different sequence, was used as a spiking control in the concentration range 0.02-0.08 µmol/ml. 2 different LPS concentrations (1 EU/mL and 10 EU/ml) were tested with the result that no optical interferences occurred (data not shown). Conclusions
A
CC E
PT
This work shows a promising approach to bind relevant model toxins using magnetic nanoparticles and special peptides derived from the binding motive of agglutinating salivary proteins. The multistage synthesis process of the developed nanoparticles led to workable yields. The end product, peptide-functionalized magnetic nanoparticles, showed a high affinity to LPS as a model toxin, dependent on the bound peptide. Future work must address the variance of the peptide sequences and their resulting efficiency of binding to LPS and other model toxins, with the aim of predicting which construct is suitable for removal of the respective toxins. The developed binding syntheses that lead to an orthogonal binding of the peptides are relevant for other binding strategies in which directed bonds are decisive. The basic particles are cytocompatible and thus open up further areas of application.
Conflicts of interest: The authors declare no conflict of interest. Acknowledgement: This study was supported by FUMIN Bridge Funding appropriations, the Manfred Roth Foundation (Fürth), Unibund, Medizinische Forschungsstiftung and the Emerging Fields Initiative of the University of Erlangen-Nuremberg (BIG-THERA).
References
9. 10.
11.
12. 13.
14.
15. 16. 17.
IP T
A
18.
SC R
8.
U
7.
N
6.
A
5.
M
4.
ED
3.
PT
2.
Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock: 2016. Intens Care Med. 2017;43(3):304-377. Reinhart K, Bauer M, Riedemann NC, Hartog CS. New Approaches to Sepsis: Molecular Diagnostics and Biomarkers. Clinical Microbiology Reviews. 2012;25(4):609-634. van der Poll T, van de Veerdonk FL, Scicluna BP, Netea MG. The immunopathology of sepsis and potential therapeutic targets. Nature reviews Immunology. 2017;17(7):407-420. Opota O, Jaton K, Greub G. Microbial diagnosis of bloodstream infection: towards molecular diagnosis directly from blood. Clin Microbiol Infec. 2015;21(4):323-331. Kritselis I, Tzanetakou V, Adamis G, et al. The level of endotoxemia in sepsis varies in relation to the underlying infection: Impact on final outcome. Immunol Lett. 2013;152(2):167-172. Iba T, Fowler L. Is polymyxin B-immobilized fiber column ineffective for septic shock? A discussion on the press release for EUPHRATES trial. J Intensive Care. 2017;5:40. Herrmann IK, Urner M, Graf S, et al. Endotoxin Removal by Magnetic Separation-Based Blood Purification. Advanced Healthcare Materials. 2013;2(6):829-835. Wang JH, Wu H, Yang YM, et al. Bacterial species-identifiable magnetic nanosystems for early sepsis diagnosis and extracorporeal photodynamic blood disinfection. Nanoscale. 2018;10(1):132-141. Hu CMJ, Fang RH, Copp J, Luk BT, Zhang LF. A biomimetic nanosponge that absorbs pore-forming toxins. Nat Nanotechnol. 2013;8(5):336-340. Cukkemane N, Bikker FJ, Nazmi K, Brand HS, Veerman ECI. Identification and characterization of a salivary-pellicle-binding peptide by phage display. Arch Oral Biol. 2014;59(5):448-454. Bikker FJ, Ligtenberg AJM, van der Wal JE, et al. Immunohistochemical detection of salivary agglutinin/gp-340 in human parotid, submandibular, and labial salivary glands. J Dent Res. 2002;81(2):134-139. Pelaz B, Alexiou C, Alvarez-Puebla RA, et al. Diverse Applications of Nanomedicine. Acs Nano. 2017;11(3):2313-2381. Cavalli S, Carbajo D, Acosta M, et al. Efficient gamma-amino-proline-derived cell penetrating peptide-superparamagnetic iron oxide nanoparticle conjugates via anilinecatalyzed oxime chemistry as bimodal imaging nanoagents. Chem Commun (Camb). 2012;48(43):5322-5324. Mazzucchelli S, Colombo M, De Palma C, et al. Single-Domain Protein A-Engineered Magnetic Nanoparticles: Toward a Universal Strategy to Site-Specific Labeling of Antibodies for Targeted Detection of Tumor Cells. Acs Nano. 2010;4(10):5693-5702. Hackenberger CPR, Schwarzer D. Chemoselective ligation and modification strategies for peptides and proteins. Angew Chem, Int Ed. 2008;47(52):10030-10074. Liu Y, Li YM, Li XM, He T. Kinetics of (3-Aminopropyl)triethoxylsilane (APTES) Silanization of Superparamagnetic Iron Oxide Nanoparticles. Langmuir. 2013;29(49):15275-15282. Yamaura M, Camilo RL, Sampaio LC, Macedo MA, Nakamura M, Toma HE. Preparation and characterization of (3-aminopropyl) triethoxysilane-coated magnetite nanoparticles. Journal of Magnetism and Magnetic Materials. 2004;279(2-3):210-217. Cao HN, He J, Deng L, Gao XQ. Fabrication of cyclodextrin-functionalized superparamagnetic Fe3O4/amino-silane core-shell nanoparticles via layer-by-layer method. Appl Surf Sci. 2009;255(18):7974-7980. Crosslinking Technical Handbook: easy molecular bonding, crosslinking technology, reactivity chemistries, applications and structure references. Thermo Scientific Inc. 2012. ThermoFisherScientific. Fluorescein (FITC). 2018; https://www.thermofisher.com/de/de/home/life-science/cellanalysis/fluorophores/fluorescein.html. BioTek. Quantitation of Peptides and Amino Acids with a Synergy™HT using UV Fluorescence. 2018; https://www.biotek.com/resources/docs/Synergy_HT_Quantitation_of_Peptides_and_Amin o_Acids.pdf.
CC E
1.
19. 20.
21.
25.
26. 27.
28. 29. 30. 31.
A
CC E
PT
ED
M
A
N
32.
IP T
24.
SC R
23.
Zaloga J, Janko C, Nowak J, et al. Development of a lauric acid/albumin hybrid iron oxide nanoparticle system with improved biocompatibility. Int J Nanomed. 2014;9:4847-4866. Poller JM, Zaloga J, Schreiber E, et al. Selection of potential iron oxide nanoparticles for breast cancer treatment based on in vitro cytotoxicity and cellular uptake. International journal of nanomedicine. 2017;12:3207-3220. Gupta AK, Gupta M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials. 2005;26(18):3995-4021. Acres RG, Ellis AV, Alvino J, et al. Molecular Structure of 3-Aminopropyltriethoxysilane Layers Formed on Silanol-Terminated Silicon Surfaces. The Journal of Physical Chemistry C. 2012;116(10):6289-6297. Albanese A, Tang PS, Chan WC. The effect of nanoparticle size, shape, and surface chemistry on biological systems. Annual review of biomedical engineering. 2012;14:1-16. Lyer S, Tietze R, Unterweger H, et al. Nanomedical innovation: the SEON-concept for an improved cancer therapy with magnetic nanoparticles. Nanomedicine (London, England). 2015;10(21):3287-3304. Estelrich J, Escribano E, Queralt J, Busquets MA. Iron oxide nanoparticles for magneticallyguided and magnetically-responsive drug delivery. Int J Mol Sci. 2015;16(4):8070-8101. Zaloga J, Janko C, Nowak J, et al. Development of a lauric acid/albumin hybrid iron oxide nanoparticle system with improved biocompatibility. Int J Nanomedicine. 2014;9:4847-4866. Riccardi C, Nicoletti I. Analysis of apoptosis by propidium iodide staining and flow cytometry. Nat Protoc. 2006;1(3):1458-1461. Cossarizza A, Ferraresi R, Troiano L, et al. Simultaneous analysis of reactive oxygen species and reduced glutathione content in living cells by polychromatic flow cytometry. Nat Protoc. 2009;4(12):1790-1797. Friedrich RP, Janko C, Poettler M, et al. Flow cytometry for intracellular SPION quantification: specificity and sensitivity in comparison with spectroscopic methods. Int J Nanomedicine. 2015;10:4185-4201.
U
22.
S0927-7765(18)30778-1 https://doi.org/10.1016/j.colsurfb.2018.11.002 COLSUB 9767
To appear in:
Colloids and Surfaces B: Biointerfaces
Received date: Revised date: Accepted date:
1 August 2018 16 October 2018 1 November 2018
Please cite this article as: Karawacka W, Janko C, Unterweger H, Muhlberger ¨ M, Lyer S, Taccardi N, Mokhir A, Jira W, Peukert W, Boccaccini AR, Kolot M, Strauss R, Bogdan C, Alexiou C, Tietze R, SPIONs functionalized with small peptides for binding of lipopolysaccharide, a pathophysiologically relevant microbial product, Colloids and Surfaces B: Biointerfaces (2018), https://doi.org/10.1016/j.colsurfb.2018.11.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
SPIONs functionalized with small peptides for binding of lipopolysaccharide, a pathophysiologically relevant microbial product Weronika Karawacka1, Christina Janko1, *, Harald Unterweger1, Marina Mühlberger1, Stefan Lyer1, Nicola Taccardi2, Andriy Mokhir3, Wolfgang Jira4, Wolfgang Peukert5,6, Aldo R. Boccaccini6,7, Mikhail Kolot8, Richard Strauss9, Christian Bogdan10,11, Christoph Alexiou1, Rainer Tietze1 1
Department of Otorhinolaryngology, Head and Neck Surgery, Section of Experimental Oncology and Nanomedicine (SEON), Else Kröner-Fresenius-Stiftung-Professorship, Universitätsklinikum Erlangen, 91054 Erlangen, Germany 2 Institute
IP T
of Chemical Reaction Engineering, Friedrich-Alexander-Universität (FAU) ErlangenNürnberg, 91058 Erlangen, Germany 3
Department of Chemistry and Pharmacy, Organic Chemistry Chair II, FAU 91058 Erlangen, Germany 4
SC R
Federal Research Institute of Nutrition and Food, Max Rubner-Institut, 95326 Kulmbach, Germany 5
Institute of Particle Technology (LFG), FAU, 91058 Erlangen, Germany
6
Interdisciplinary Center for Functional Particle Systems (FPS), FAU, 91058 Erlangen, Germany
7
N
U
Institute of Biomaterials, Department of Materials Science and Engineering, FAU, 91058 Erlangen, Germany Department of Biochemistry and Molecular Biology, Tel-Aviv University, Tel-Aviv 69978, Israel
9
Deptartment of Medicine 1, Universitätsklinikum Erlangen, 91054 Erlangen, Germany
10
und
Hygiene,
Universitätsklinikum
Medical Immunology Campus Erlangen, FAU Erlangen-Nürnberg, 91054 Erlangen, Germany
ED
11
Immunologie
M
Institut für Klinische Mikrobiologie, Erlangen,91054 Erlangen, Germany
A
8
PT
* Corresponding author; [email protected]; Glückstraße 10a, 91054 Erlangen
A
CC E
Graphical abstract
Highlights
16-mer peptides from binding motives of agglutinating salivary proteins act as specific pathogen scavengers. A reliable orthogonal binding strategy of peptides to stable SPIONs was developed. Specific pathogen binding capacity of peptide-functionalized SPIONs is clearly dependent on the peptide sequence. The results could support the development of new treatment options for acute systemic infectious diseases.
IP T
Abstract
PT
ED
M
A
N
U
SC R
Systemic inflammation such as sepsis represents an acute life-threatening condition, to which often no timely remedy can be found. A promising strategy may be to functionalize magnetic nanoparticles with specific peptides, derived from the binding motives of agglutinating salivary proteins, that allow immobilization of pathogens. In this work, superparamagnetic iron oxide nanoparticles with stable polycondensed aminoalkylsilane layer were developed, to which the heterobifunctional linkers N-succinimidyl 3-(2-pyridyldithio)-propanoate (SDPD) and N-succinimidyl bromoacetate (SBA) were bound. These linkers were further chemoselectively reacted with the thiol group of singularly present cysteines of selected peptides. The resulting functional nanoparticles underwent a detailed physicochemical characterization. The biocompatibility of the primarily coated aminoalkylsilane particles was also investigated. To test the pathogen-binding efficacy of the particles, the lipopolysaccharide-immobilization capacity of the peptide-coated particles was compared with free peptides. Here, one particle-bound peptide species succeeded in capturing 90% of the toxin, whereas the degree of immobilization of the toxin with a system that varied in the sequence of the peptide dropped to 35%. With these promising results, we hope to develop extracorporeal magnetic clearance systems for removing pathogens from the human body in order to accelerate diagnosis and alleviate acute disease conditions such as sepsis.
CC E
Key words: Sepsis, Magnetic Clearence, Salivary Agglutinin Peptide, Pathogen Immobilization, Superparamagnetic Iron Oxide Nanoparticles Number of words/tables/figures: 5990; 1; 7. Introduction
A
Sepsis is a massive systemic reaction to infection, associated with life-threatening organ damage. Patients with severe sepsis and septic shock (a severe sepsis with low blood pressure, associated with high mortality) require complex therapy measures in intensive care medicine of hospitals 1. Pathogenic microorganisms express molecules that are known as pathogen-associated molecular patterns (PAMPs) 2. When those PAMPs enter the body, the host reacts with multiple immune responses involving several molecules and cellular components. During sepsis, these substances released into the bloodstream in order to fight the infection, trigger systemic inflammatory responses, mainly due to dysregulated host reactions to microbial infections 3. Blood cultures are still standard to identify causative pathogens. In up to 50% of cases of bloodstream infections, blood cultures remain negative. This can be due to antibiotic pretreatment,
to low number of circulating pathogens, or to organisms which are not cultivable.4 Pathogen counts in blood vary by infection and over time during sepsis. In most cases, numbers of viable microorganisms in the blood are low: about 1 to 10 Colony forming Units (CFU)/mL blood. Blood concentrations of PAMPs however vary depending on time and type of infection.5
N
U
SC R
IP T
Complex cytokine and coagulation cascades trigger severe inflammatory processes that in many cases cannot be stopped in time to save the patients. Current treatment regimens aim at a reliable early detection and treatment of the infection with antibiotics, as wells as intravenous fluid therapy, ventilatory, and hemodynamic support. Recently, the immobilization of pathogens to suppress the continuous initiation of inflammatory process and for rapid diagnostics has attracted considerable attention. Clinically approved devices for extracorporeal clearence of pathogens already exist (e.g. polymyxin-B columns for hemoperfusion, Toramyxin™). However, these techniques are only suitable for gram-negative germs and otherwise the clinical success seems ambivalent 6. Newer approaches, especially these based on particulate clearance methods, utilize antibodies or aptamers, which can intercept pathogens efficiently but too selectively7,8. In order to circumvent this limitation, fragments of erythrocytes, which have a high non-specific binding affinity to different sepsis-associated pathogens9 can be used, although this approach can potentially give rise to allergic reactions. To overcome the problem of highly selective binding and to avoid immune reactions, other mechanisms for the immobilization of sepsis-associated pathogens are proposed. Human saliva, for example, contains peptides that can absorb a wide range of microbial invaders10,11. This could be a new approach to bind the microbial triggers as closely as possible, but without imposing unnecessary constraints resulting from the too specific complementary binding mechanisms.
CC E
PT
ED
M
A
A useful intrinsic property of superparamagnetic iron oxide nanoparticles (SPIONs) is their magnetization by external magnetic fields, which is of great importance in the field of biomedical applications12. The magnetic properties of SPIONs enable magnetic separation of pathogens bound to particles. This serves both to reduce the pathogen load and to accelerate the diagnosis, if pathogens in patient blood samples can be effectively enriched and extracted. To obtain a pathogen-binding nanosystem, the specific peptides must be immobilized on the surface of nanoparticles. As with other macromolecular immobilization strategies13,14, this requires an orthogonal connection to the particle surface. The binding between peptide and pathogen is strongly dependent on the conformation of the binding peptide. Meanwhile, there are various chemoselective strategies for binding of complex molecules, without blocking their actual capturing capacity15. The principle is that the respective functional group is either solitary at the desired binding position of the effective biomolecule, or it is exposed by means of an appropriate protective group strategy. A corresponding heterobifunctional linker can represent the interface between the nanoparticulate drug carrier and the active substance.
A
In our work, we have investigated two selected peptide sequences with regard to their immobilization capacity against a model toxin. The peptides were bonded to the surface of SPIONs, which were coated with aminopropyl triethoxysilane (APTES). We used different protocols16-18, and developed our own synthesis process on their basis. Further, we tested the suitability of two different linkers, N-succinimidyl 3- (2-pyridyldithio)-propanoate (SPDP) and Nsuccinimidyl bromoacetate (SBA), for binding peptides. The peptides each have terminal cysteine, which is solitary in the sequences used here. This enables the peptides to chemoselectively bind to the nanoparticles via thiol groups. Materials and methods Four batches of NPs were synthesised with two methods of purification – either magnetic washing (batches1-3) or dialysis (batch 4). Particles were afterwards dried and kept as powder (batches 1-
3), or as a water suspension (batch 4). Batch 4 was divided into 2 sub-batches with respect to its reaction to external magnetic field. Results for batches 1, 2 and 3 were averaged, as they were prepared in the same way, to gain insight into uniformity of this SPIONs’ fabrication technique. They are further referred to as 1/3. Further information on materials and methods is provided in the Supplemental files. Peptide Functionalization of the Nanoparticle Surface
N
U
SC R
IP T
3 different peptides were used in the procedures: Fluorescein isothiocyanate (FITC)-bearing peptide with the sequence FITC-TVCCC, purity 81.47%, MW 1030.21 Da, stock solution 2 mM in corresponding buffer prepared prior to use. Peptide A: CVTGWSGRYLVEVRGQ, purity 92.97%, MW 1810.07, stock solution 1 mg/mL in corresponding buffer. Peptide B: RKQGRVEVLYRASWGTVC, purity 90.44%, MW 2108.46, stock solution (2 mM) in corresponding buffer prepared prior to use. Prior to peptide functionalization, the binding of the two different linker molecules was performed as follows: Linking SPDP to Fe3O4@APTES: SPDP was synthesized as described in supplemental files. Fe3O4@APTES (stock solutions volumes corresponding to 1 mg Fe) were diluted with phosphate buffer (10 mM at pH 7.5) and SPDP stock solution was added (25 μL, 20 mM in DMFdry). Suspensions were mixed (1 h at room temperature (RT); thermomixer, 1000 rpm). Supernatants were magnetically decanted. Afterwards, nanoparticles were washed with phosphate buffer and magnetically decanted resulting in Fe3O4@APTES@SPDP nanoparticles.
ED
M
A
Linking SBA to Fe3O4@APTES: SBA was obtained as described in supplement. Samples were mixed at RT for 30 min. Fe3O4@APTES (stock solutions volumes corresponding to 1 mg Fe) were diluted in the dark with borate buffer (50 mM at pH 8.4), and SBA stock solution was added (25 μL, 20 mM in DMFdry). Supernatants were magnetically decanted. Nanoparticles were washed with borate buffer and magnetically decanted, resulting in Fe3O4@APTES@SBA nanoparticles.
CC E
PT
Peptide coupling to Fe3O4@APTES@linker: Fe3O4@APTES@linker nanoparticles were redispersed in buffer (phosphate for SPDP, borate for SBA). Peptide solution (bearing 0.1 μmol of the peptide) was added to the dispersion, and the samples were mixed at RT (800 rpm, thermomixer) either overnight (SPDP), or for 1 h (SBA).19 Supernatants were magnetically decanted. Nanoparticles were washed with the appropriate buffer (1 mL) and magnetically decanted, resulting in Fe3O4@APTES@linker@peptide nanoparticles. Estimations on the coupling efficiencies Estimation of existing amino groups
A
Estimation through Si to Fe ratio: The concentrations of Si/Fe in μmol/mg were calculated from ICP-AES measurements, which corresponded 1:1 to NH2/Fe in μmol/mg. Active amino groups through 2(1H)-pyridinethione release: Supernatants of Fe3O4@APTES@SPDP were collected to measure the not-bonded SPDP. Phosphate buffer containing SPDP alone (1.25% v/v) was used as control. Dithiothreitol (DTT, 100 mM in phosphate buffer) was added to supernatants and control samples. The samples were mixed for 20 min and absorbance at 343 nm was recorded. The difference of absorbance corresponds to the concentration of 2-pyrydinethione (ε = 5.81·103 dm3/cm·mol, determined experimentally in phosphate buffer). The control sample gave maximal possible concentration of SPDP, whereas the samples showed the amount of SPDP that was not bonded to Fe3O4@APTES. In order to determine the bonded SPDP, nanoparticles were redispersed in phosphate buffer and stock DTT
was added. The samples were mixed for 15 min at 1000 rpm (thermomixer) and centrifuged (5 min, 13400 rpm, MiniSpin, Eppendorf). Supernatants were collected, diluted with phosphate buffer to 2 mL and the absorbance at 343 nm measured. All measurements were performed in triplicates, results were averaged. The amount of bonded SPDP should correspond to the amount of surface amino groups available for coupling reactions. Absorbance at 280 nm was measured to estimate the dependence between 343 nm and 280 nm absorbance originating from 2-pyrydinethione. Estimation of peptide coupling efficiency
SC R
IP T
Peptide coupling efficiency through fluorescence measurements: FITC-TVCCC was used as a model peptide with excellent fluorescence quantum yield and high water solubility. Fluorescence strength is however sensitive to pH changes and relatively easy to quench 20. Measurements were performed on Filter-Max F5 Plate reader (Molecular Devices, Biberach an der Riss, Germany) with excitation at 485 nm and emission at 535 nm. Measurements were performed in triplicates, results were averaged.
A
N
U
Efficiency of peptide coupling to Fe3O4@APTES@SPDP: FITC-TVCCC peptide was coupled to nanoparticles as described above. Control sample was prepared by mixing phosphate buffer, SPDP stock solution and FITC-TVCCC stock solution. Absorbance was measured in UV-Vis spectrometer at 343 nm against pure phosphate buffer to detect the released 2-pyrydinethione as described previously. As control, peptides coupled to Fe3O4@APTES@SPDP were cleaved with DTT. The nanoparticles were redispersed in phosphate buffer and DTT stock solution was added. After 15 min of incubation, supernatants were collected, diluted with phosphate buffer and fluorescence was measured. FITC-TVCCC concentrations were estimated according to a calibration curve in the concentration range 5.0·10-4-12.5·10-4 μmol/mL.
ED
M
Efficiency of peptide coupling to Fe3O4@APTES@SBA: FITC-TVCCC peptide was coupled to nanoparticles as described above and the measurement was performed similar to the SPDP samples, but with borate buffer. FITC-TVCCC concentrations were estimated according to the calibration curve obtained for FITC-TVCCC peptide in borate buffer for concentration range 2.00·10-4-6.25·10-4 μmol/mL.
PT
Through UV-Vis absorbance: The peptide sequences used in this study contain one tryptophan (W) and one tyrosine (Y), showing a characteristic absorption band at 280 nm in UV-Vis spectrum 21 . Calibration curves were determined for each peptide in both phosphate and borate buffer for the concentration range 0.0125-0.1000 μmol/mL.
A
CC E
To Fe3O4@APTES@linker: Peptide A or B was coupled to nanoparticles as described above. The nanoparticles were redispersed in the appropriate buffer and DTT stock solution was added. Mixtures were shaken at RT for 15 mins and centrifuged. Supernatants were collected and diluted with buffer, after which the absorbance was measured at 280 and 343 nm and background absorbance was subtracted. Investigation of cell death and DNA degradation Jurkat cells (2.5x105) were seeded in 1 mL medium into 48 well plates. Cells were treated with Fe3O4@APTES or SEONLA-BSA, derived from a previous study22, at final Fe concentrations of 25, 50, 100 and 200 μg/mL. Nanoparticle-untreated cells and cells treated with H2O served as controls. After 24 and 48 hours of incubation, cells were mixed and 50 µL aliquots were taken and analysed. Staining was performed with freshly prepared staining solution consisting of AnnexinA5-FITC (AxV, 1 µL/mL), Hoechst 33342 (Hoe, 10 µg/mL), propidium iodide (PI, 6.6 ng/mL), 1,1´,3,3,3´,3´hexametylindodicarbocyanine iodide (DiI, 0.4 μL/mL) in Ringer’s solution. Staining solution (200 µL) was added to 50 µL cell suspensions, incubated for 30 min at 4 °C and measured in flow
cytometer.23 Cell cycle and DNA degradation was investigated using PI-Triton staining (0.1% Triton X-100, 1 mg/mL sodium citrate and 50 μg/mL PI in water). For this purpose, 50 µL of cell suspensions were incubated with 400 µL staining solution overnight at 4 °C in the dark and then monitored using flow cytometry. Estimation of particle uptake
SC R
IP T
Jurkat cells (2.5x105) were seeded in 1 mL medium into 48 well plates. Cells were treated with Fe3O4@APTES or SEONLA-BSA at final Fe concentrations of 25, 50, 100 and 200 μg/mL. Nanoparticle-untreated cells and cells treated with H2O served as controls. Cells were spiked with Lucifer Yellow (LY, 10 μL, 10 μg/mL). After 24 and 48 hours incubation the cells were mixed and 50 µL aliquots were analysed. Staining was performed with freshly prepared staining solution consisting of monobromobimane (MBB, 0.3 µL/mL, stock 100 mM) and DiI (0.4 μL/mL) in Ringer’s solution. 200 µL staining solution was added to 50 µL cell suspensions, incubated for 30 min at 4 °C and measured in flow cytometer. Particle uptake was additionally monitored with optical microscopy (magnification 10x and 40x, Axiovert 40 CFC, Carl Zeiss AG, Germany) with cells incubated for 48 h with nanoparticles (NPs). Endotoxin-binding studies
𝑆𝑖𝑔𝑛𝑎𝑙 𝑟𝑒𝑐𝑜𝑣𝑒𝑟𝑦 𝑟𝑎𝑡𝑒 = 100 ∙
M
A
N
U
Endotoxin-binding studies were performed with EndoZyme Recombinant Factor C Assay kit (Hyglos, Germany). Nanoparticles from batch 4A were conjugated with peptides A or B through SBA linker. Binding efficiency was evaluated by UV-VIS (@280 nm) of supernatants and dispersions of 0.050, 0.0250 and 0.0125 μmol peptide/mL in water were prepared. Furthermore, the solutions of peptides A and B in water at the same concentrations were prepared. As a blank nanoparticle sample, nanoparticles coupled with citric acid (CA) and acetylsalicylic acid (ASA) were used (procedure to be found in Supplementray Information) to estimate signal recovery rate, according to Equation 1: (𝐿𝑃𝑆 𝑠𝑝𝑖𝑘𝑒𝑑 𝑖𝑛 𝐶𝐴 𝑎𝑛𝑑 𝐴𝑆𝐴 𝑐𝑜𝑛𝑗𝑢𝑔𝑎𝑡𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒) −(𝐿𝑃𝑆 𝑖𝑛 𝑛𝑜𝑛−𝑠𝑝𝑖𝑘𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒) 𝑎𝑑𝑑𝑒𝑑 𝐿𝑃𝑆
ED
(Equation 1): LPS = Lipopolysaccharide
CC E
PT
Endotoxin (at 5.00, 2.50, 1.25, and 0.0 EU) was added to 200 μL of each nanoparticle dispersion and peptide solution, resulting in peptide end concentration dilutions of factors 1.20, 1.10, 1.05 and 1.00, respectively. LPS-spiked samples and calibration samples were incubated in thermomixer for 1 h (37°C, 1200 rpm) and subsequently centrifuged (1 h, 8700 relative centrifuge force (rcf)). The precipitability of the LPS peptide agglutinate was separately tested by centrifugation at 18000 rcf for 1 h. Supernatants (100 μL) were analysed for endotoxin content according to the manufacturer`s instructions. The reaction was monitored for 90 min at 37°C in 15 min intervals in a Microplate Reader Filter Max F5 (excitation 360 nm/emission 465 nm). Experiments were performed in duplicate and the results were averaged.
A
Results and discussion Properties of the nanoparticles The coprecipitation method used in this study is the fastest, easiest and cheapest way to fabricate SPIONs, also in a large scale, with acceptable control over size and yielding uniform batches when optimised.24 In situ coating by addition of coating agents performed here enables to omit tedious iron oxide purification process between the synthesis and coating steps. Successful APTES coating of produced SPIONs was confirmed with FT-IR spectroscopy (Suppl. Information, Figure S1/S2). Elemental composition and size of obtained nanoparticles are presented in Table 1. Table 1. Elemental composition and hydrodynamic size of Fe3O4@APTES nanoparticles.
Mass % Molar ratio SMD [nm] Fe Si N H C N/Si C/Si 1/3 63.10 0.15 0.72 1.51 1.21 4.8 18 191.8 Δ 1/3 0.64 0.01 0.08 0.58 0.90 0.4 12 13.1 4A 70.36 0.11 0.47 0.89 2.04 8.2 42 100.1 4B 65.90 0.11 0.62 0.69 2.09 11.5 45 169.2 SMD: standard mean diameter, PDI: Polydispersity index, SF: steepness factor Batch
PDI
SF
0.30 0.01 0.34 0.30
1.16 0.01 1.18 1.16
SC R
IP T
APTES and its polymer, 3-aminopropylsiloxane, have theoretically a stoichiometric N/Si ratio of 1/1 and C/Si ratio of 3/1. Here however, all batches had much higher ratios of N/Si and C/Si. Interestingly, batches 1-3, which were magnetically washed with methanol, seemed to be purer than batch 4, which was dialysed. There are several reasons for increased presence of nitrogen and carbon in the particles. Additional nitrogen may come from ammonia being trapped between nanoparticle cores or adsorbed on their surface alongside with APTES. The presence of carbon contamination on APTES-modified silica has been previously reported.25 As a by-product of APTES hydrolysis, ethanol might have been trapped within nanoparticles during fabrication. Theoretically, some ethoxy groups may have not been hydrolysed, however, this is unlikely due to the fact that the reaction was conducted in water.
ED
M
A
N
U
Nanoparticles with mean hydrodynamic diameter between 100 and 200 nm were obtained. All batches had a polydispersity index (PDI) of 0.3 and steepness factor (SF) of 1.2, indicating their uniformity and narrow size distribution. Particles made in the first 3 batches were larger (190 nm), as compared to the dialysed nanoparticles (4A ~100 nm and 4B~170 nm). Particle size plays crucial role in biodistribution and pharmacokinetics. NPs over 100 nm in diameter do not extravasate far beyond the blood vessel as they remain trapped in the extracellular matrix between cells, whereas nanoparticles bigger than 200 nm accumulate in spleen and liver where they are processed by mononuclear phagocyte system.26 Particles smaller than 200 nm can be sterile filtered and thus are easier to prepare for biological applications.27 Because of that aspect, biological investigations within this study were conducted with batch 4A containing nanoparticles of ~100 nm in hydrodynamic diameter.
PT
Single core size was determined with TEM micrographs (Suppl. Information, Figure S 3/4). The mean core size was 8.8 nm (median of 8.0 nm). No cores bigger than 25 nm were observed, meaning that all precipitated nanoparticles were within superparamagnetic size range.28 Diffraction patterns indicated that nanoparticles were composed of magnetite, partially oxidised to maghemite (Suppl. Information, Figure S5).
A
CC E
ζ-potential is a measure of nanoparticles’ surface charge and thus electrostatic stability. Colloids with the absolute value of ζ-potential higher than 30 mV are considered electrostatically stable.26 Here, ζ-potential reached +25 mV at pH < 3.2 and dropped to +10 and +15 mV at physiological pH (7.4). It then remained stable up to pH 9. This value of ζ-potential was too low for the nanoparticles to remain stable in a solution over prolonged time, resulting in agglomeration and disturbing the measurement. However, aggregation is not necessarily disadvantageous when used ex vivo, as it facilitates magnetic separation of NPs. The isoelectric point (IP) of APTES-coated magnetite NPs oscillated between pH 9.1 and 10.2. The observed pH-dependent ζ-potential trends corresponded well to the changes in hydrodynamic size (see Supplementary Information). Above pH 9.5, nanoparticles formed agglomerates, being most profound around pH 10.5. Size did not differ much at lower pH values (Suppl. Information, Figure S6). Higher concentrated (~6 mg Fe/mL) particles were stable enough in acidic solution and little to no precipitation occured, even after a 8 week storage.
All batches exhibited reasonably high volume magnetic susceptibilities (>4·10-3 in SI units) and drop in Xv was registered over time, indicating the propagation of magnetite oxidation to maghemite (Suppl. Information, Figure S7). Amount of existing NH2 groups
IP T
Estimation of the amount of amino groups was important for subsequent coupling reactions. In principle, each surface NH2 group can become a binding point, but a steric hindrance might occur with bulkier substances. ICP-AES composition analysis allowed to assess the maximum possible amount of amino groups for each batch of nanoparticles. This was 0.09 and 0.06 μmol/mg NH2/Fe (Si/Fe) for batches 1/3 and 4A/B, respectively. These results were further referred to as “theoretical” amount of NH2 groups per mg Fe for each batch of Fe3O4@APTES. Determination of reactive amino groups
A
CC E
PT
ED
M
A
N
U
SC R
One molecule of 2(1H)-pyridinethione can be cleaved from each molecule of SPDP with DTT (Figure 1). Here, the 2(1H)-pyridinethione amounts were measured (a) in the supernatant yielding indirectly the amount of NH2 groups available for SPDP bonding, as well as (b) released from nanoparticles, yielding this amount directly. Both values should be consistent with each other. The measured fraction of NH2 groups, compared to theoretical amount, yields a % of active amino groups.
Figure 1. Schematic process of peptide coupling to Fe3O4@APTES@SPDP and its cleavage with DTT for the estimation of reaction efficiency. Prior to the experiment, the correlation between 2(1H)-pyridinethione absorbance at 343 nm and SPDP concentration in phosphate buffer has been established, along with the correlation between signal coming from 2(1H)-pyridinethione at 343 nm and at 280 nm. The amount of 2(1H)-pyridinethione cleaved from SPDP measured in the supernatants was comparable to the amount cleaved from functionalised nanoparticles (Figure 2). It means that both the indirect (from supernatant) and the direct (form NPs) method of measuring, are consistent. The
values obtained using the indirect method in batch 4A and 4Bwere slightly increased, possibly by traces of nanoparticles still present in the supernatant after magnetic separation and, therefore, disturbing the background signal. According to this experiment, most NH2 groups (~80%) present in the Fe3O4@APTES were active and available for binding of SPDP. 0.09 0.07 0.06 0.05
from unbonded SPDP
0.04
released from NPs
0.03
average
0.02
theoretical
SC R
0.01 0.00 1-3 av
4A
IP T
NH2 (μmol/mg Fe)
0.08
4B
Batch
N
U
Figure 2. SPDP-active NH2 groups of Fe3O4@APTES. Experiments were performed in triplicates, shown are the mean values with standard deviations.
A
Peptide coupling efficiency FITC-TVCCC
PT
ED
M
FITC-TVCCC, a fluorescent dye functionalised with a short peptide bearing 3 cysteines at the Cterminal end, was coupled to Fe3O4@APTES through SPDP and SBA. To measure the coupling efficiency, peptide bonded to the nanoparticles via SPDP was cleaved using DTT. NPs were then magnetically separated and the fluorescence of obtained supernatants was measured. Upon coupling via SBA linker, the cleavage of bonded peptide was not possible. Therefore, the supernatants containing unreacted peptide were measured to estimate the bonding capacity indirectly.
A
CC E
As shown in Figure 3, toupling FITC-TVCCC through SBA linker appeared to be more efficient, especially for batches 1/3. While the amount of successfully coupled FITC-TVCCC was fairly constant for SPDP (0.015 μmol/mg Fe), it had deviations up to 25% between batches when SBA was used. But this could be due to the less precise indirect detection methods for the SBA linking. Short peptide bearing a rigid and bulky fluorescein molecule could have encountered spacial difficulties in finding available binding spots, thus only about 1/3 of active animo groups were successfully coupled.
0.06 0.05 0.04 SPDP 0.03
SBA SPDP-active
0.02
IP T
Coupled FITC-TVCCC [μmol/mgFe]
0.07
0.01 0.00 1-3 av.
4A
4B
SC R
Batch
U
Figure 3. The amounts of FITC-TVCCC coupled to different batches of nanoparticles. Experiments were performed in triplicates, shown are the mean values with standard deviations. Peptides A and B coupled with SPDP and SBA
A
CC E
PT
ED
M
A
N
Peptides A and B were coupled to Fe3O4@APTES nanoparticles by SPDP or SBA linker. The amount of coupling was estimated indirectly by measuring the supernatants’ absorbance at 280 nm (maximum absorbance of peptides). For SPDP, the bonding efficiency was also directly measured by cleavage of the coupled peptides from SPDP linker with DTT. In this case, both results were averaged. Peptide concentrations were calculated from calibration curves. The results are depicted in Figure 4.
Figure 4. Peptides coupled to Fe3O4@APTES@SPDP compared to active NH2 groups (left) and the same peptides coupled to Fe3O4@APTES@SBA compared to the theoretical amount of NH2 groups (right). Experiments were performed in triplicates, shown are the mean values with standard deviations.
Using SBA linkers, binding of 0.06 µmol peptide /mg Fe was possible independent of NP batch. Using SDPD liker, this value was reached only for peptide B binding to batches 1-3 and 4A. That value corresponds to the number of active and even of all present amino groups. It must be noted that the results of these experiments might have been disturbed by high background noise of residual nanoparticles, yielding values that appear even slightly higher than theoretically possible. Biocompatibility
A
CC E
PT
ED
M
A
N
U
SC R
IP T
Sterile filtered batch 4A of Fe3O4@APTES nanoparticles was compared with previously reported SEON@LA-BSA nanoparticles obtained according to Zaloga et. al.29. Biocompatibility was determined using AxV-Fitc/PI staining. AxV binds to phosphatidylserine (PS) exposed on the outer plasma membrane layer on apoptotic cells. In necrotic cells, PI intercalates into the DNA after plasma membrane rupture. Ax-PI- cells are considered viable, Ax+PI- cells apoptotic, and PI+ cells necrotic. After 24 and 48 h of incubation with NPs, cell viability was only slightly reduced. APTEScoated SPIONs induced less apoptosis than BSA-coated ones, while causing similar level of necrosis. In all cases, more than 80% of cells remained viable, thus all nanoparticles proved to be cytocompatible (Figure 5A). DiI staining for mitochondrial membrane potential confirmed these findings (data not shown). Cell cycle phases were analyzed using PI-Triton staining. Cells can be differentiated according to their DNA content into cells with diploid DNA (G1 phase), cells with double diploid DNA (G2 phase) and cells with degraded DNA generated by DNAses during cell death (sub G1).30 After 24 h, only minor cell cycle alterations were observed. After 48 h, a dosedependent increase in subG1 DNA was visible for both BSA- and APTES-coated nanoparticle treated cells (Figure 5B). To check whether APTES-coated magnetite induced production of reactive oxygen species (ROS) and concomitant oxidative stress, staining with MBB was performed. In the presence of ROS, cellular glutathione (GSH) is oxidised to glutathione disulphide S. MBB, initially non-fluorescent, reacts with thiols yielding fluorescent adducts. Its fluorescence is therefore the strongest when GSH is present in non-oxidised state.31 After 24 h and 48 h incubation for both SPION types, ROS was dose- and time-dependently produced, as shown by reduced MBB fluorescence (Figure 5C). Since the SPIONs are to be applied in the whole blood, their hemocompatibility, including the effects on erythrocyte hemolysis and plasma coagulation were also investigated, as described in detail in Suppl. Information (Figure S8).
IP T SC R
A
N
U
Figure 5. Biocompatibility of Fe3O4@APTES with Jurkat cells treated: A) Apoptosis and necrosis were determined by AxV/PI staining. B) Cellular DNA content was determined by PIT staining. C) Oxidative stress was evaluated by MBB staining. Numbers indicate Fe concentrations in μg/mL. APTES, APTES-coated SPIONs; BSA, LA/BSA-coated SPIONs. Experiments were performed in triplicates, shown are the mean values with standard deviations.
M
Nanoparticle uptake
PT
ED
The side scatter (SSc) of viable cells measured in flow cytometry was used to estimate the cellular SPION load. We previously showed that the presence of intracellular nanoparticles dosedependently increases the cellular granularity as determined by SSc.32 SSc of Jurkat cells treated with BSA-coated NPs increased only very slightly with the growing concentration of iron and did not further change within 48 h. In contrast, SSc of cells treated with APTES-coated NPs was dramatically increased as compared to control and SEON@LA-BSA -trested cells, showing an iron dose-dependency only at the concentrations below 200 µg/mL. Such high SSc values may suggest adherence of the nanoparticles to the cell surface rather than uptake (Figure 6A).
A
CC E
Whenever a cell phagocytoses a nanoparticle, it co-ingests a small drop of medium. Fluorescent LY in medium will thus be co-ingested when cells take up nanoparticles. A correlation between SSc and LY signal was observed for SEON@LA-BSA (Figure 6B), indicating that NP are rather internalized than only being attached to the cell surface. In cells treated with APTES-coated SPIONs, only a slight increase in LY fluorescence was detected, indicating that either Fe3O4@APTES are not favourably taken up, or that the quenching of LY fluorescence can occur by the massive nanoparticle adhesion to the cell surface.
a)
500
b)
450
2,5
Fluorescence [a.u.]
400 350
SSc [a.u.]
3,0
300 250 200 150 100
2,0 1,5 1,0 0,5
50
24
48
24
APTES 200
APTES 100
APTES 50
APTES 25
BSA 200
BSA 100
BSA 50
BSA 25
H2O
MEDIUM
APTES 200
APTES 100
APTES 50
APTES 25
BSA 200
BSA 100
BSA 50
BSA 25
H2O
0,0
MEDIUM
0
48
SC R
IP T
Figure 6. Determination of nanoparticle uptake. a) Side scatter of viable cells (positive DiI+) analysed by flow cytometry. b) LY staining. Numbers indicate Fe concentrations in μg/mL. APTES, APTES-coated SPIONs: BSA, LA/BSA-coated SPIONs. Experiments were performed in triplicates, shown are the mean values with standard deviations. Endotoxin binding capacity
M
A
N
U
For the future use in sepsis diagnosis and treatment, the designed nanoparticles should efficiently bind LPS and other bacterial toxins in order to remove them from the bloodstream. Here, the two peptides (A and B) were conjugated to Fe3O4@APTES through SBA linker, due to its better peptide binding properties (Figure 4) and their LPS-binding ability was tested and compared to the LPS-binding efficiency of free peptides. Because the free peptides could not be magnetically separated, the samples were centrifuged to provide the same conditions. The presence of both LPS and capture peptide leads to a signal loss in the endotoxin assay due to the formation of an complex consisting of several peptides and LPS molecules11. These complexes can be precipitated by centrifugation at 18000 rcf for 1h (data not shown). Before the experiment, SPION dispersions and peptide solutions were also tested for intrinsic LPS content.
A
CC E
PT
ED
In Figure 7, the effectiveness of LPS-binding by free and particle-bound peptidesis compared. Free peptide A bound between 75% and 85% of spiked LPS (Figure 7a). However, upon conjugation to APTES-coated SPIONs, its efficiency dropped to 35%-75% (Figure 7c). For peptide B, both free and SPION-conjugated peptides removed more than 90% of spiked LPS, even when as little as 0.0125 μmol/mL of peptide was used (Figure 7 b, d). Peptide B was equally efficient when 10 EU/mL of LPS were spiked, as in the presence of 4 times lower concentrations of endotoxins.
IP T SC R
A
N
U
Figure 7. Endotoxin binding capacity. a) Free peptide A, b) Free peptide B, c) Fe3O4@APTES@SBA@A d) Fe3O4@APTES@SBA@B. Numbers in the legend refer to peptide concentrations in μmol/mL. Experiments were performed in triplicates, shown are the mean values with standard deviations.
ED
M
A peptide consisting of the same amino acids as peptide B, but composed in a different sequence, was used as a spiking control in the concentration range 0.02-0.08 µmol/ml. 2 different LPS concentrations (1 EU/mL and 10 EU/ml) were tested with the result that no optical interferences occurred (data not shown). Conclusions
A
CC E
PT
This work shows a promising approach to bind relevant model toxins using magnetic nanoparticles and special peptides derived from the binding motive of agglutinating salivary proteins. The multistage synthesis process of the developed nanoparticles led to workable yields. The end product, peptide-functionalized magnetic nanoparticles, showed a high affinity to LPS as a model toxin, dependent on the bound peptide. Future work must address the variance of the peptide sequences and their resulting efficiency of binding to LPS and other model toxins, with the aim of predicting which construct is suitable for removal of the respective toxins. The developed binding syntheses that lead to an orthogonal binding of the peptides are relevant for other binding strategies in which directed bonds are decisive. The basic particles are cytocompatible and thus open up further areas of application.
Conflicts of interest: The authors declare no conflict of interest. Acknowledgement: This study was supported by FUMIN Bridge Funding appropriations, the Manfred Roth Foundation (Fürth), Unibund, Medizinische Forschungsstiftung and the Emerging Fields Initiative of the University of Erlangen-Nuremberg (BIG-THERA).
References
9. 10.
11.
12. 13.
14.
15. 16. 17.
IP T
A
18.
SC R
8.
U
7.
N
6.
A
5.
M
4.
ED
3.
PT
2.
Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock: 2016. Intens Care Med. 2017;43(3):304-377. Reinhart K, Bauer M, Riedemann NC, Hartog CS. New Approaches to Sepsis: Molecular Diagnostics and Biomarkers. Clinical Microbiology Reviews. 2012;25(4):609-634. van der Poll T, van de Veerdonk FL, Scicluna BP, Netea MG. The immunopathology of sepsis and potential therapeutic targets. Nature reviews Immunology. 2017;17(7):407-420. Opota O, Jaton K, Greub G. Microbial diagnosis of bloodstream infection: towards molecular diagnosis directly from blood. Clin Microbiol Infec. 2015;21(4):323-331. Kritselis I, Tzanetakou V, Adamis G, et al. The level of endotoxemia in sepsis varies in relation to the underlying infection: Impact on final outcome. Immunol Lett. 2013;152(2):167-172. Iba T, Fowler L. Is polymyxin B-immobilized fiber column ineffective for septic shock? A discussion on the press release for EUPHRATES trial. J Intensive Care. 2017;5:40. Herrmann IK, Urner M, Graf S, et al. Endotoxin Removal by Magnetic Separation-Based Blood Purification. Advanced Healthcare Materials. 2013;2(6):829-835. Wang JH, Wu H, Yang YM, et al. Bacterial species-identifiable magnetic nanosystems for early sepsis diagnosis and extracorporeal photodynamic blood disinfection. Nanoscale. 2018;10(1):132-141. Hu CMJ, Fang RH, Copp J, Luk BT, Zhang LF. A biomimetic nanosponge that absorbs pore-forming toxins. Nat Nanotechnol. 2013;8(5):336-340. Cukkemane N, Bikker FJ, Nazmi K, Brand HS, Veerman ECI. Identification and characterization of a salivary-pellicle-binding peptide by phage display. Arch Oral Biol. 2014;59(5):448-454. Bikker FJ, Ligtenberg AJM, van der Wal JE, et al. Immunohistochemical detection of salivary agglutinin/gp-340 in human parotid, submandibular, and labial salivary glands. J Dent Res. 2002;81(2):134-139. Pelaz B, Alexiou C, Alvarez-Puebla RA, et al. Diverse Applications of Nanomedicine. Acs Nano. 2017;11(3):2313-2381. Cavalli S, Carbajo D, Acosta M, et al. Efficient gamma-amino-proline-derived cell penetrating peptide-superparamagnetic iron oxide nanoparticle conjugates via anilinecatalyzed oxime chemistry as bimodal imaging nanoagents. Chem Commun (Camb). 2012;48(43):5322-5324. Mazzucchelli S, Colombo M, De Palma C, et al. Single-Domain Protein A-Engineered Magnetic Nanoparticles: Toward a Universal Strategy to Site-Specific Labeling of Antibodies for Targeted Detection of Tumor Cells. Acs Nano. 2010;4(10):5693-5702. Hackenberger CPR, Schwarzer D. Chemoselective ligation and modification strategies for peptides and proteins. Angew Chem, Int Ed. 2008;47(52):10030-10074. Liu Y, Li YM, Li XM, He T. Kinetics of (3-Aminopropyl)triethoxylsilane (APTES) Silanization of Superparamagnetic Iron Oxide Nanoparticles. Langmuir. 2013;29(49):15275-15282. Yamaura M, Camilo RL, Sampaio LC, Macedo MA, Nakamura M, Toma HE. Preparation and characterization of (3-aminopropyl) triethoxysilane-coated magnetite nanoparticles. Journal of Magnetism and Magnetic Materials. 2004;279(2-3):210-217. Cao HN, He J, Deng L, Gao XQ. Fabrication of cyclodextrin-functionalized superparamagnetic Fe3O4/amino-silane core-shell nanoparticles via layer-by-layer method. Appl Surf Sci. 2009;255(18):7974-7980. Crosslinking Technical Handbook: easy molecular bonding, crosslinking technology, reactivity chemistries, applications and structure references. Thermo Scientific Inc. 2012. ThermoFisherScientific. Fluorescein (FITC). 2018; https://www.thermofisher.com/de/de/home/life-science/cellanalysis/fluorophores/fluorescein.html. BioTek. Quantitation of Peptides and Amino Acids with a Synergy™HT using UV Fluorescence. 2018; https://www.biotek.com/resources/docs/Synergy_HT_Quantitation_of_Peptides_and_Amin o_Acids.pdf.
CC E
1.
19. 20.
21.
25.
26. 27.
28. 29. 30. 31.
A
CC E
PT
ED
M
A
N
32.
IP T
24.
SC R
23.
Zaloga J, Janko C, Nowak J, et al. Development of a lauric acid/albumin hybrid iron oxide nanoparticle system with improved biocompatibility. Int J Nanomed. 2014;9:4847-4866. Poller JM, Zaloga J, Schreiber E, et al. Selection of potential iron oxide nanoparticles for breast cancer treatment based on in vitro cytotoxicity and cellular uptake. International journal of nanomedicine. 2017;12:3207-3220. Gupta AK, Gupta M. Synthesis and surface engineering of iron oxide nanoparticles for biomedical applications. Biomaterials. 2005;26(18):3995-4021. Acres RG, Ellis AV, Alvino J, et al. Molecular Structure of 3-Aminopropyltriethoxysilane Layers Formed on Silanol-Terminated Silicon Surfaces. The Journal of Physical Chemistry C. 2012;116(10):6289-6297. Albanese A, Tang PS, Chan WC. The effect of nanoparticle size, shape, and surface chemistry on biological systems. Annual review of biomedical engineering. 2012;14:1-16. Lyer S, Tietze R, Unterweger H, et al. Nanomedical innovation: the SEON-concept for an improved cancer therapy with magnetic nanoparticles. Nanomedicine (London, England). 2015;10(21):3287-3304. Estelrich J, Escribano E, Queralt J, Busquets MA. Iron oxide nanoparticles for magneticallyguided and magnetically-responsive drug delivery. Int J Mol Sci. 2015;16(4):8070-8101. Zaloga J, Janko C, Nowak J, et al. Development of a lauric acid/albumin hybrid iron oxide nanoparticle system with improved biocompatibility. Int J Nanomedicine. 2014;9:4847-4866. Riccardi C, Nicoletti I. Analysis of apoptosis by propidium iodide staining and flow cytometry. Nat Protoc. 2006;1(3):1458-1461. Cossarizza A, Ferraresi R, Troiano L, et al. Simultaneous analysis of reactive oxygen species and reduced glutathione content in living cells by polychromatic flow cytometry. Nat Protoc. 2009;4(12):1790-1797. Friedrich RP, Janko C, Poettler M, et al. Flow cytometry for intracellular SPION quantification: specificity and sensitivity in comparison with spectroscopic methods. Int J Nanomedicine. 2015;10:4185-4201.
U
22.