Polycationic Nanofibers for Nucleic Acid ScavengingJennifer G. Jackman,† Hemraj Juwarker,‡ Luke P. Poveromo,‡ Howard Levinson,‡ Kam W. Leong,*,†,§
and Bruce A. Sullenger*,‡
†Department of Biomedical Engineering and ‡Department of Surgery, Duke University, Durham, North Carolina, United States§Department of Biomedical Engineering, Columbia University, New York, New York, United States
ABSTRACT: Dying cells release nucleic acids (NA) and NA-containing complexes that activate inflammatory pathways ofimmune cells. Sustained activation of these pathways contributes to chronic inflammation frequently encountered in autoimmuneand inflammatory diseases. In this study, grafting of cationic polymers onto a nanofibrous mesh enabled local scavenging ofnegatively charged pro-inflammatory molecules in the extracellular space. Nucleic acid scavenging nanofibers (NASFs) formedfrom poly(styrene-alt-maleic anhydride) conjugated with 1.8 kDa bPEI resulted in nanofibers of diameters 486 ± 9 nm. NASFsinhibited the NF-κB response stimulated by the negatively charged agonists, CpG and poly(I:C), in Ramos-blue cells but notPam3CSK4, a nonanionic agonist. Moreover, NASFs significantly impeded NF-κB activation in cells stimulated with damage-associated molecular pattern molecules (DAMPs) released from doxorubicin killed cancer cells. In vivo application of NASFs toopen wounds demonstrated nucleic acid scavenging in wounds of diabetic mice infected with Pseudomonas aeruginosa, suggestingthe in vivo efficacy of NASFs. This simple technique of generating NASF results in effective localized anti-inflammation in vitroand local nucleic acid scavenging in vivo.
■ INTRODUCTION
Extracellular nucleic acids can elicit pro-inflammatory responsesby activating the innate immune system. Pattern recognitionreceptors (PRRs) such as the toll-like receptors (TLRs) areactivated by pathogenic nucleic acids (e.g., viral RNAs, bacterialDNAs) as well as nucleic acids released by necrotic cells.1−3
Hindered clearance of extracellular nucleic acids from the bloodhas been linked to inappropriate activation of TLRs byendogenous nucleic acids, which is observed in variousinflammatory and autoimmune diseases including systemiclupus erythematosus (SLE), rheumatoid arthritis (RA), andmultiple sclerosis (MS). In the case of autoimmune diseases,antibodies to extracellular DNA (eDNA) lead to the formationof immune complexes that increase cellular uptake, allowing forTLR activation to occur, leading to cytokine production andprolonged immune activation. Studies in lupus-prone MRL/lprmice confirm that activation of TLR7 and TLR9, nucleic acidreceptors, can mediate the pathogenesis of SLE.4−6 Currenttherapies (e.g., monoclonal antibodies) can reduce thesymptoms of lupus, but fail in many patients due to seriousside effects from widespread immunosuppression.7 In light ofthese side effects, researchers have shifted their focus toblocking antibody−DNA interactions using oligonucleotides,
peptides, and small molecules to block the antibody bindingsite on DNA. However, these approaches are limited due to theexpression of DNA antibodies that interact with diverse sites onthe DNA molecule, therefore, limiting the efficacy of antibodyblocking. Previous work from our laboratory has demonstratedthe effectiveness of certain nucleic acid-binding polymers (e.g.,PAMAM-G3, CDP, HDMBr, protamine, polyethylenimine) atinhibiting nucleic acid-mediated activation of nucleic acid-sensing PRRs, irrespective of whether they recognize ssRNA,dsRNA, or hypomethylated DNA.8
An alternative approach for nucleic acid scavenging of freelycirculating NAs would be useful, as such NAs may presentissues with regard to cytotoxicity and nonspecific cellularuptake. Here we show that a nucleic acid-scavenging nanofiber(NASF) serves as a highly cationic mesh to scavengeextracellular nucleic acids in a controlled and localizedenvironment without compromising TLR responses to non-nucleic acid, pathogenic stimulators.
Received: August 17, 2016Revised: October 10, 2016Published: October 14, 2016
The NASFs are synthesized by electrospinning of poly-(styrene-alt-maleic anhydride) (PSMA). The combination ofpolystyrene and maleic anhydride provides tunable mechanicalproperties9−12 and ease of chemical modification. ElectrospunPSMA nanofibers have been applied for aptamer and gassensing,12,13 while microdiameter fibers have been investigatedfor hydrogel formation and enzyme stabilization.14−16 In thisstudy, we present the covalent modification of PSMAnanofibers with amine-containing polycations, namely,branched polyethylenimine (bPEI), for nucleic acid scavenging.
■ MATERIALS AND METHODSPoly(styrene-alt-maleic anhydride) Electrospun Nanofiber
Formation. Preparation of neutral fibers was optimized by testingvarious concentrations of poly(styrene-alt-maleic anhydride) (PSMA)with solvent combinations of one or more of the following:tetrahydrofuran, acetone, and dimethylformamide. Effective electro-spinning occurred using PSMA (0.6 or 1 g) with an average molecularweight of 1900 Da (Sigma-Aldrich, St. Louis, MO) dissolved in a 1:1:1(v/v/v) mixture of tetrahydrofuran/acetone/dimethylformamide (10mL; Sigma-Aldrich) by constant mixing for 24 h at room temperature.PSMA nanofibers were electrospun using 2 mL of polymer solution ina 2 cm3 glass syringe (Cadence Science, Staunton, VA) at a dispensingrate of 1 mL/h, achieved by insertion of the syringe into an automatedsyringe pump, with an applied voltage of +15 kV. The polymer fiberswere collected on a grounded cylindrical mandrel (∼6.4 cm wide witha ∼21.6 cm circumference) spinning at ∼130 rpm at a distance of 10cm away from the tip of the syringe needle. The neutral electrospunPSMA fibers were soaked in a solution of 1.8 kDa branchedpolyethylenimine (bPEI; 0.1 M; Polysciences, Inc., Warrington, PA)for 72 h at room temperature with constant shaking to form NASF. Toform PAMAM-NASF, the neutral fibers were soaked in PAMAM-G3(0.004 M; Sigma-Aldrich) at 4 °C for 72 h with constant shaking.Following conjugation, the NASF were washed for 10 min withdeionized water a total of 5 times. NASF were sterilized for 30 min in70% ethanol, the ethanol was removed, and the NASF were allowed toair-dry in a sterile environment. PAMAM-NASF were washed fivetimes for 10 min each in sterile water and allowed to air-dry in a sterileenvironment, skipping the ethanol sterilization step. The levels of 1.8kDa bPEI conjugated onto the fiber were determined by ninhydrinassay.Scanning Electron Microscopy (SEM) of Nanofibers. Dry
fibers were placed on aluminum foil and mounted on an SEM stub.The fibers were gold sputter-coated for 250 s using the DentonVacuum Desk IV sputter unit (Denton Vacuum, Moorestown, NJ) andimaged using a FEI XL30 SEM-FEG (FEI, Hillsboro, OR). Imageswere analyzed in Scandium (ResAlta Research Technologies, Golden,CO).X-ray Photospectroscopy (XPS) of Nanofibers. XPS measure-
ments were taken on a Kratos Analytical Axis Ultra XPS (Kratos,Manchester, UK) using a monochromated Al Kα source. The sourcewas operated at 15 kV and 10 mA (150 W). Electron collection wasmade at 90° to the sample surface. Survey scans were taken with a passenergy of 160 eV and region scans with a pass energy of 20 eV.Nucleic Acid Adsorption of NASF. Alexa Flour 488 labeled CpG
ODN 1668 (3.33 × 10−4 to 1 × 10−3 μg/mL; IDT, Coralville, IA) wasincubated with 0.08 mg of NASF for 4 h at room temperature underconstant shaking. The NASFs were washed three times with deionizedwater, placed on a microscope slide, and mounted with SlowFadeDiamond reagent (Life Technologies, Carlsbad, CA). Fluorescentimages of adsorbed DNA onto NASFs were captured with an UprightAxioImager A1 microscope powered by a Zeiss HBO100 power supplyand lamp housing. To generate the DNA adsorption curve, salmonsperm DNA (25−100000 ng; Life Technologies) in 1xTE buffer wasadded to 0.08 mg of NASF for 4 h at room temperature (RT) withconstant shaking. Total salmon sperm DNA concentration remainingin the 1xTE solution was determined using a PicoGreen assay (LifeTechnologies).
Cell Culture. All cell experiments were performed in completegrowth media unless specifically stated otherwise, with cells incubatedat 37 °C (5% CO2). STO (ATCC, Manasses, VA) and RAW (ATCC)cells were cultured in DMEM (Gibco 11960−044) supplemented with10% FBS and 1% Pen-Strep. Ramos-Blue cells (InvivoGen, San Diego,CA) were cultured in IMDM (Gibco 12440−053) supplemented with10% FBS, 1% Pen-strep, and 100 μg/mL Zeocin every 4 passages.NHDF (Lonza, Basel, Switzerland) cells were cultured in DMEM(Gibco 11960−044) with 10% FBS, 1% Pen-Strep, 1% nonessentialamino acids, 1% sodium pyruvate, 1% Glutamax, and 0.1% β-mercaptoethanol.
Cell Viability. STO cells (40000 cells/well) were plated 18−24 hbefore experiments to allow for cellular adhesion. Ramos-Blue cellviability experiments were performed by plating cells (200000 cells/well) immediately prior to adding the NASF into the medium. Bothcell types were incubated with 0.107 mg of NASF for 4 h and cellviability was measured using CellTiter-Glo (Promega, Madison, WI).Proliferation studies were performed by putting a 1.13 mg piece ofNASF, cut to fit into the well, onto the bottom of a nontissue culturetreated 48-well plate (Greiner Bio-One, Kremsmunster, Austria) using10 μL of PBS to aid in NASF adhesion to the well. After the PBSdried, Normal Human Dermal Fibroblast cells (NHDF; 100000 cells/well) were added to the top of the NASF. Proliferation wasdetermined using LIVE/DEAD viability/cytotoxicity kit (LifeTechnologies) at 24 and 48 h.
Inhibition of Nucleic Acid Driven TLR Activation UsingPolycationic NASFs. For scavenging studies, NASFs or neutralPSMA fibers (0.107 mg) were incubated in media for 10 min with oneof the following TLR agonists at 10 μg/mL in 250 μL of completemedia: CpG 1668 (TLR 9), Poly(I:C) (TLR 3), or Pam3CSK4 (TLR1/2; InvivoGen). The fiber was removed and the fiber treated media(62.5 μL) was added to Ramos-Blue cells (200000 cells in 137.5 μL ofcomplete media) for 18−24 h at 37 °C. For comparison, TLR agonists(2.5 μg/mL) were also applied directly to Ramos-Blue cells. For directcontact studies, NASFs were added directly to Ramos-Blue cells(200000 cells/well), followed by incubation with agonists (2.5 μg/mL) for 18−24 h at 37 °C. After incubation, the supernatant (40 μL)from the treated Ramos-Blue cells was added to the serum alkalinephosphatase colorimetric indicator Quanti-Blue (160 μL; InvivoGen),incubated for 5 h at 37 °C, and absorbance was measured at 650 nm.
Inhibition of TLR Activation by DAMPs Using NASFs.Doxorubicin (DOX) was used to cause cell death of RAW cells andrelease of damage-associated molecular pattern molecules (DAMPs) asdescribed.17,18 Different doses of DOX were administered to RAWcells and the subsequent TLR activation was measured in Ramos-Bluereporter cells and the doses that generated DAMPs capable ofactivating the Ramos-Blue cells potently were used in subsequentstudies. The DOX doses were incubated for 48 h in RAW cells (40000cells/well, DOX (3, 6, or 9 μg/mL; Sigma-Aldrich)) and thesupernatant from the DOX-treated cells (100 μL) was added to a1.41 mg piece of NASF. The NASF and supernatant were incubatedfor 30 min and the entire volume was added to Ramos-Blue cells(200000 cells in 100 μL IMDM). After incubation for 18−24 h,Ramos-Blue supernatant (40 μL) was added to Quanti-blue (160 μL;InvivoGen) and the absorbance was read at 650 nm at 3 and 5 h.
NASF Application to Open Wound Mouse Model Infectedwith Pseudomonas aeruginosa. The murine diabetic wound modelwas created according to the protocol described by Zhao et al.19
Briefly, diabetic female mice, 8−12 weeks old (BKS.CgDock7m+/+Leprdb/J) from the Jackson Laboratory were given an 8 mm diameterexcisional wound between the shoulder blades that was covered withTegaderm Film (3M). NASF treatment was started 72 h followingwounding in diabetic mice. All mice received 1.13 mg of NASF placeddirectly inside the open wound, which was then covered by Tegadermfilm. Diabetic mice received NASF treatment for 7 consecutive daysand the NASF sheets were changed every 24 h over the 7 day period.Following treatment, the NASF was removed from the mouse wound,placed in PBS, and vortexed in pulses at the highest level for 30 s toremove any mechanically adhered components. The NASF was thenplaced in 10 mg/mL of heparin sodium salt (Sigma) for 30 min with
intermittent vortexing. After 30 min, the NASF was discarded and aTRIzol (ThermoFisher Scientific) extraction was performed on thesolution to separate the RNA, DNA, and protein componentsextracted from the NASF.For Pseudomonas aeruginosa infected wound studies, diabetic mice
with an 8 mm diameter excisional wound had 50 μL of Pseudomonasaeruginosa (PA01, ATCC 15692) at a 104 CFUs administered to theopen wound 72 h after initial wounding. NASF was administered andprocessed as described above.DNA Gels and PCR. DNA extracted using the TRIzol reagent was
resuspended in 8 mM NaOH. 200 ng of DNA was run on a 1%agarose gel at 100 V for 90 min and visualized using GelStar NucleicAcid Stain (Lonza). Two primer sets for mitochondrial cytochrome coxidase were used for PCR. Primer set 1 included primers: 5′-ACCAAG GCC ACC ACA CTC CT-3′ and 5′-ACG CTC AGA ATCCTG CAA AGA A- 3′, while primer set 2 included primers: 5′-TCCAAG TCC ATG ACC ATT AAC TG-3′ and 5′-TAT TGG TGA GTAGGC CAA GGG-3′, leading to fragment sizes of ∼101 and ∼115 bp,
respectively. Taq 2x Master Mix (New England BioLabs) and 100 ngof DNA was used for all PCR reactions. Primer set 1 was done with anannealing temperature of 54 °C and an extension time of 10 s, whileprimer set 2 was done with an annealing temperature of 50 °C and anextension time of 10 s. All PCR products were run on a 3% agarose gelat 70 V for 60 min and visualized using GelStar Nucleic Acid Stain.
RNA Sequencing. RNA library preparation was done using theStrand RNA-Seq Kit (Kapa) with 10 ng of RNA. Following librarypreparation, the libraries were pooled in equimolar amounts, andsequenced with a 50 bp SR run on the Illumina HiSeq 2500 sequencer.RNA-seq data was processed using the TrimGalore toolkit,20 whichemploys Cutadapt21 to trim low quality bases and Illumina sequencingadapters from the 3′ end of the reads. Only reads that were 20 nt orlonger were kept for further analysis. Reads were mapped to a customgenome and transcriptome that contained the mouse NCBIM38r7322
data as well as the Pseudomonas aeruginosa PAO1 data using the STARRNA-seq alignment tool.23 Reads were kept for subsequent analysis ifthey mapped to a single genomic location. Gene counts were compiled
Figure 1. (A) SEM image of neutral PSMA nanofibers; (B) SEM image of 1.8 kDa bPEI modified PSMA nanofibers; (C) (i) Survey scan of neutralPSMA fibers, (ii) High-resolution XPS scan for C 1s of neutral PSMA fibers, (1) is CO, (2) is C−O, (3) is C−C, (iii) Survey scan of bPEIconjugated NASF, (iv) High-resolution XPS scan for C 1s of bPEI conjugated NASF showing the presence of (1) CO, (2) C−O, (3) C−N, (4)C−C.
using the HTSeq tool.24 Only genes that had at least 10 reads in anygiven library were used in subsequent analysis. Normalization anddifferential expression was carried out using the DESeq225
Bioconductor26 package with the R statistical programming environ-ment.27
Statistical Analysis. All data were evaluated for statistical analysiswith One Way Analysis of Variance (ANOVA) followed by secondaryTukey-HSD analysis in JMP. The average values and standard errors ofthe mean are presented with p < 0.05 being considered statisticallysignificant.
■ RESULTS AND DISCUSSION
Unmodified PSMA nanofibers electrospun from two concen-trations (6% and 10% w/v) resulted in randomly alignednanofibers with average diameters of 297 ± 13 and 737 ± 26nm, respectively (Figure 1A). Although electrospun PSMAfibers and polystyrene-PSMA fiber sheets can be found in theliterature,12,14,28 there is no report on functionalizing thesestructures with a cationic moiety. The process used to form theNASF has some distinct advantages because it involves modularsteps that allow conjugation of any amine-containing polymeronto the neutral fiber. This modular approach has been utilizedfor development of microfibers,16 but here we employ it toproduce nanofibers to generate a higher surface area forcationic polymer attachment that are subsequently used fornucleic acid scavenging. Nucleic acid scavenging using fibershas been previously reported;29 however, in the aforemen-tioned study, a biodegradable poly(ε-caprolactone) and bPEIblock copolymer was used for electrospinning and involved acomplicated synthesis step followed by electrospinning. Thisprocess would require reoptimization of electrospinningparameters for each different block copolymer synthesis, asopposed to our neutral fiber modification technique that allowsfor modification of already formed nanofibers.The PSMA nanofibers electrospun from 6% and 10% w/v
polymer solutions demonstrated different physical character-istics after conjugation with 1.8 kDa bPEI. The 10% NASFs(NASFs made with 10% PSMA nanofibers) were brittle anddifficult to handle whereas the 6% NASFs demonstrated amalleable tissue paper-like texture. After initial proof-of-conceptin vitro experiments with both 6% and 10% NASFs, 6% PSMAfibers were chosen for preparation of the NASF sheets as theywere more durable, easier to cut into various sizes and did notshow reduced efficacy as compared to 10% NASFs. Uponconjugation of 1.8 kDa branched bPEI to the 6% PSMAnanofibers, the resulting polycationic nanofibers (NASFs) hadan increased fiber diameter of 486 ± 9 nm (Figure 1B) ascompared to the original at 297 ± 13 nm (Figure 1A) andcontained ∼11 μM bPEI per 1 mg of nanofiber. Successful
conjugation of bPEI onto the nanofibers was confirmed bycontact angle measurements, X-ray photoelectron spectroscopy(XPS), and DNA-binding affinity. The contact angle of neutralPSMA nanofibers was 122°, indicating high hydrophobicity,while the NASFs were so hydrophilic that the contact anglecould not be determined, validating the conversion fromhydrophobic neutral nanofibers to hydrophilic cationic nano-fibers. Additionally, XPS data confirmed the abundance ofnitrogen on the surface of the NASFs at an atomic percent of10.57%, which indicated successful conjugation of aminogroups onto the nanofiber surface. The unmodified PSMAfibers showed a surface nitrogen abundance of atomic percent<0.01% (Figure 1C).Functionality of cationic NASFs to bind nucleic acids was
validated using CpG ODN 1668 DNA and salmon spermDNA. Alexa-Fluor 488 labeled CpG was used to demonstratebinding of nucleic acids by the NASFs (Figure 2A,B). Asexpected, increasing amounts of labeled CpG resulted inincreased fluorescence as compared to background nanofiberfluorescence. The increased fluorescence of the NASFsfollowing soaking with labeled DNA and subsequent washesto remove any unattached DNA confirmed the ability of NASFsto scavenge nucleic acids. Through adsorption analysis usingsalmon sperm DNA (Figure 2C), the maximum adsorptioncapacity of the NASFs was ∼35.7 μg DNA/0.08 mg of fiber.SEM images showed that the initial modification of neutralPSMA nanofibers with bPEI resulted in swelling of the fibersand coalescing of the fibers at points where they overlapped(Figure 1A,B). However, interaction with salmon sperm DNAdid not change the fiber morphology, thereby maintaining theNASF mechanical structure following DNA interaction (datanot shown).Previous studies indicate that the presence of highly charged
moieties (e.g soluble cationic polymers) lead to highcytotoxicity in some cell lines.30−32 A benefit of this insolubleNASF strategy is that scavenging activity happens entirely inthe extracellular space, avoiding cellular uptake of the boundpolycation-NA complex. In an in vivo application, this wouldallow the systemic circulation to be void of circulatingpolycations, bypassing the toxicity issue associated with solublepolycations. Additionally, the NASF utilized is not biodegrad-able, so there are no concerns of components slipping intocirculation and causing systemic toxicity. When cell viability wastested with STO and Ramos-Blue cells in the presence ofNASF, the cell viability remained over 70% (Figure 3A). Todetermine the effects of direct cellular interaction with theNASF, a LIVE/DEAD stain was performed after 24 and 48 h ofplating NHDF cells directly onto NASFs (Figure 3B). TheLIVE/DEAD assay in NHDF showed that viability remained
Figure 2. Fluorescent microscope images of polycationic nanofibers after 4 h interaction with (Ai) no AlexaFlour488-CpG or AlexaFlour488-CpG at(Aii) 3.33 × 10−4 μg/μL, (Aiii) 6.6 × 10−4 μg/μL, and (Aiv) 1 × 10−3 μg/μL; (B) quantification of average fluorescence after interaction withAlexaFlour488-CpG normalized to polycationic nanofiber alone and compared to the initial amount of CpG added (n = 3); (C) Salmon sperm DNAabsorption onto 0.08 mg of polycationic nanofiber (n = 3), where the x-axis represents the amount of DNA added to the solution and the y-axisrepresents how much DNA the polycationic nanofiber removed from the solution.
above 80% at all times when cells were grown directly on top ofNASF. Taken together, these results indicate that the NASFsdisplay minimal toxicity.Next, the NASF were tested to determine if they could
effectively reduce inflammatory cytokine production byscavenging NAs from solution. Initial tests in TLR reportercells, Ramos-Blue cells, demonstrated the NASFs’ ability toscavenge CpG, prevent TLR9 activation, and concomitantlyreduce NF-κB/AP-1-inducible secreted alkaline phosphatase(SEAP) levels down to baseline. To show that this inhibitionwas due to electrostatics-driven scavenging and not bynonspecific adsorption of fluid and the components in it,comparative experiments were done with neutral fibers (Figure4A). Neutral, unmodified PSMA fibers have no effect on CpGTLR stimulation, therefore showing that nonspecific fluidinteraction was not responsible for the scavenging of CpG.Beyond CpG scavenging as was done by Kang and Yoo,29 the
NASF were also tested for their ability to scavenge Poly(I:C), asynthetic double-stranded RNA molecule. Specificity fornegatively charged TLR agonists is demonstrated in Figure4B by comparing the ability of the NASFs to return SEAP levelsto baseline after preincubation with CpG, Poly(I:C), andPam3CSK4. CpG and Poly(I:C) are both nucleic acid agonists,while Pam3CSK4 is a synthetic triacylated lipoprotein agonist.NASFs blocked TLR3 and TLR9 activation by Poly(I:C) andCpG, respectively, both negatively charged nucleic acidagonists. However, they were not able to block activation bynon-nucleic acid agonist, Pam3CSK4 (TLR2/1). To demon-strate that the NASFs can effectively scavenge NAs in thepresence of cells and serum, Ramos-Blue cells were coincubatedwith the fibers, followed by administration of the TLR agonists(Figure 5). Co-incubation experiments showed the same resultsas preincubation with NASFs; both CpG and Poly(I:C)stimulation effects were reduced to baseline SEAP levels,while stimulation by Pam3CSK4 remained unaffected. Todemonstrate the functional versatility of this scavengingstrategy, a polyamido(amine) cationic dendrimer (PAMAM-G3) was grafted onto the neutral fibers to make PAMAM-
NASF and tested for efficacy in the coincubation studies withRamos-Blue TLR reporter cells. As expected, the PAMAM-NASFs mimicked the results seen with bPEI-grafted NASFs;nucleic acid TLR agonists CpG and Poly(I:C) were scavenged,therefore, significantly reducing SEAP expression; whereas,SEAP expression caused by non-nucleic acid agonistPam3CSK4 was not affected. Together, these results provethat the NASFs can effectively prevent TLR activation by CpGand Poly(I:C) and present relevant evidence that this blockingby NASF can be extended to other stimulatory nucleic acids.Additionally, the nucleic acid scavenging effectiveness of NASFsis not limited to using bPEI as the grafted polycation. Theneutral PSMA fibers can also be modified with differentpolycations (e.g., PAMAM-G3) while maintaining scavengingcapabilities.To test the NASF in a biologically relevant in vitro study,
DOX-induced cell death was used to generate cell debris anddamage-associated molecular patterns (DAMPs). Studies show
Figure 3. (A) Cell viability of STO and Ramos-Blue cells after 4 h ofinteraction with polycationic nanofibers (n = 3); (B) Live/Dead assayperformed on NHDF cells, where Live cell % was determined at 24and 48 h.
Figure 4. Secreted alkaline phosphate levels from Ramos-Blue cells.(A) SEAP levels after preincubation of neutral fibers (6/10% PSMAFiber) or NASFs with CpG as compared to CpG alone and baselinelevels of SEAP (untreated) in serum-free media. Significance is shownas compared to Untreated Ramos-Blue cells. (B) SEAP levels afterpreincubation NASFs with agonists compared to agonists alone anduntreated SEAP levels of Ramos-Blue cells. Significance is shown ascompared to the agonists alone, * indicates p < 0.001 (n = 3).
Figure 5. Co-incubation of NASFs with Ramos-Blue cells and agonistsin serum containing media. 6% NASF and PAMAM-NASFsignificantly block TLR activation by CpG and Poly(I:C) and maintainSEAP levels comparable to baseline, Untreated; * denotes p < 0.0001,# denotes p < 0.05 (n = 3).
that DOX-, a commonly used chemotherapeutic, induced celldeath leads to transient NF-κB expression,18 which is associatedwith nucleic acid fragments that are released from dead anddying cells. Increased amounts of circulating nucleic acidsresulting from excessive cell death caused by chemotherapy orradiation therapy initiates inflammatory responses. High dosesof DOX cause apoptotic cell death resulting in abnormal DNAfragmentation,17,33 these endogenous DNA and RNA frag-ments can subsequently be uptaken into healthy cells leading toTLR activation.34 To model increased nucleic acid amounts asreleased from dying cells and to show the application of theNASFs in a biologically relevant inflammatory situation,nanofiber scavenging was tested using DOX-killed cell debrisas the pro-inflammatory TLR stimulant. Administration ofDOX to RAW 264.7 cells induced cell death that releasedvarious TLR and/or NOD1 agonists that stimulate NF-κB/AP-1 and subsequent SEAP secretion from Ramos-Blue Blymphocyte cells. RAW 264.7 cells were chosen for thisapplication because DOX treatment for 48 h resulted in cell-death debris that promoted high TLR/NOD1 activation. Whenthe cell-death debris was scavenged by the NASFs, a decrease inNF-κB/AP-1 secretion, as illustrated by SEAP amounts, wasobserved (Figure 6). The maximum blocking of 41.4% of TLR/
NOD1 activation by NASFs was achieved after interaction withthe DOX-treated cell debris administered from the lowest DOXdose (3 μg/mL). The percent blocking by NASFs decreased to26.9% and 28.3% after treatment with DOX-treated cell debriswith initial DOX doses of 6 and 9 μg/mL, respectively. Wehypothesize that this modest blocking of TLR activation is dueto apoptosis-associated protein complexes that stimulateNOD1, which also induce NF-κB/AP-1 secretion.35 Yetanother explanation is that the higher DOX concentrationsresulted in more cell death consequentially causing saturationof the nanofibers and therefore reduced blocking. Exploringincreased doses of NASF may allow for increased blocking ofthe immune response caused by the DOX-treated cell debris.Nevertheless, these results indicate that the NASFs arefunctional in a clinically relevant application.
An open wound mouse model with and without aPseudomonas aeruginosa infection was used to validate theNASFs nucleic acid scavenging capabilities in an in vivo setting.NASFs were placed on uninfected and infected wounds for 24h, the contents on the NASF were competed away using aheparin competition assay, and the components were separatedinto DNA and RNA extracts. In both infected and uninfecteddiabetic mice wounds treated with NASF, DNA extracted fromthe NASF and DNA gels run on the samples demonstrated twoprominent DNA bands, an unidentified high base pair band andanother band around ∼16000 bp (data not shown).Amplification of mitochondrial DNA (mtDNA) with cyto-chrome-c oxidase primers validated that the ∼16000 bp bandwas mtDNA. Figure 7A,B shows a sample of the PCR
amplicons from the DNA samples extracted from the NASF.mtDNA is important because increased circulating mtDNAleads to increased TLR9 activation36 resulting in a prolongedinflammatory response.The RNA extracts were analyzed using RNA sequencing.
The Venn diagram in Figure 8B shows the number of detectedRNAs that were extracted from mouse wounds that were eitherinfected with P. aeruginosa or not infected. Overall, the NASFthat was not exposed to P. aeruginosa bacteria had highernumbers of RNA at 18448 as compared to 12693 that weredetected from the NASF extract used on P. aeruginosa infectedwounds. An overlap in 11983 RNAs was detected. The RNaseq
Figure 6. Blocked SEAP production from Ramos-blue cells where 0%SEAP blocking represents Ramos-Blue cells that are not stimulatedwith any DOX cell debris and 100% SEAP blocking represents Ramos-Blue cells that are treated with DOX cell debris without NASFscavenging. Initial DOX dose to Raw cells describes the amount ofDOX used to treat Raw cells 48 h prior to using the Raw cell debris foractivation of Ramos-blue cells. Polycationic nanofiber blockingdemonstrates the polycationic nanofiber’s ability to prevent NF-κBproduction by scavenging immune-stimulating cell debris from themedia (n = 3); * denotes p < 0.0001 as compared to cells withoutDOX cell debris and without NASF, ε denotes p < 0.03 as comparedto DOX dose at 9 μg/mL with no NASF. Figure 7. Representative gel of (A) PCR amplicon results using primer
set 1 and (B) PCR amplicon results using primer set 2 on openwounds in diabetic mice treated with NASF for 24 h as compared to aGeneRuler Low Range DNA Ladder.
heatmap in Figure 8A shows how the RNAs extracted fromuninfected wounds and P. aeruginosa infected wounds differ.Approximately 20−24% of the total RNAs detected in theinfected wounds were found to be of bacterial origin, whereasin the uninfected wounds, all of the RNAs were of mouseorigin. This demonstrates that the NASF can scavenge bothbacterial and mouse RNA and that over a 24 h period, a greaterpercentage of the RNA scavenged is of mouse origin.
■ CONCLUSION
In conclusion, we have built upon our previous findings thatscavenging of extracellular nucleic acids can reduce inflamma-tion.8 We report an insoluble, minimally toxic cationicnanofiber, NASF, which selectively scavenges polyanionicpro-inflammatory species (DNA, RNA) and blocks improperactivation of nucleic-acid sensing TLRs without compromisingthe ability to respond to non-nucleic acid agonists. We alsoreport that this nanofiber scavenging strategy may be useful inreducing pro-inflammatory side effects associated with chemo-therapy induced cell death, which releases large amounts of pro-inflammatory nucleic acids and complexes into circulation.Finally, we validate the scavenging capacity of the NASF in vivousing an open wound model. If the utility of soluble polycationsis limited due to the toxicity associated with cellular uptake ofhighly positively charged molecules, these insoluble, functionalNASFs prevent cell internalization and minimize cytotoxic sideeffects, presenting an attractive strategy to reduce chronicinflammation due to extracellular nucleic acids.
Author ContributionsThe manuscript was written through contributions of allauthors. All authors have given approval to the final version ofthe manuscript.FundingNational Science Foundation Graduate Research Fellowship,Duke Skin Disease Research Center Grant, and support by theGlobal Research Laboratory Program (GRL; 2015032163)through the National Research Foundation of Korea (NRF) isalso acknowledged.NotesThe authors declare no competing financial interest.
■ ACKNOWLEDGMENTSWe would like to acknowledge the Duke Center for Genomicand Computational Biology, the Duke School of MedicineCore Facility Voucher Program, and the National Institutes ofHealth.
■ REFERENCES(1) Kawai, T.; Akira, S. Antiviral Signaling Through PatternRecognition Receptors. J. Biochem. 2006, 141 (2), 137−145.(2) Lotze, M. T.; Zeh, H. J.; Rubartelli, A.; Sparvero, L. J.; Amoscato,A. A.; Washburn, N. R.; Devera, M. E.; Liang, X.; Tor, M.; Billiar, T.The grateful dead: damage-associated molecular pattern molecules andreduction/oxidation regulate immunity. Immunol. Rev. 2007, 220, 60−81.(3) Palm, N. W.; Medzhitov, R. Pattern recognition receptors andcontrol of adaptive immunity. Immunol. Rev. 2009, 227 (1), 221−33.(4) Berland, R.; Fernandez, L.; Kari, E.; Han, J. H.; Lomakin, I.;Akira, S.; Wortis, H. H.; Kearney, J. F.; Ucci, A. A.; Imanishi-Kari, T.Toll-like receptor 7-dependent loss of B cell tolerance in pathogenicautoantibody knockin mice. Immunity 2006, 25 (3), 429−40.(5) Christensen, S. R.; Shupe, J.; Nickerson, K.; Kashgarian, M.;Flavell, R. A.; Shlomchik, M. J. Toll-like receptor 7 and TLR9 dictateautoantibody specificity and have opposing inflammatory andregulatory roles in a murine model of lupus. Immunity 2006, 25 (3),417−28.
Figure 8. (A) RNaseq heat map from RNA extracts of db/db wounds without infection (untreated, NASF) and with PA01 infection (infected,NASF). (B) Venn diagram of the number of RNAs that were extracted from NASF on PA01 infected db/db wounds (infected, NASF, purple) vsNASF on db/db wounds with no bacterial infection (untreated, NASF, yellow).
(6) Nickerson, K. M.; Christensen, S. R.; Shupe, J.; Kashgarian, M.;Kim, D.; Elkon, K.; Shlomchik, M. J. TLR9 regulates TLR7- andMyD88-dependent autoantibody production and disease in a murinemodel of lupus. J. Immunol. 2010, 184 (4), 1840−8.(7) Stearns, N. A.; Lee, J.; Leong, K. W.; Sullenger, B. A.; Pisetsky, D.S. The Inhibition of Anti-DNA Binding to DNA by Nucleic AcidBinding Polymers. PLoS One 2012, 7 (7), e40862.(8) Lee, J.; Sohn, J. W.; Zhang, Y.; Leong, K. W.; Pisetsky, D.;Sullenger, B. A. Nucleic acid-binding polymers as anti-inflammatoryagents. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (34), 14055−60.(9) Sun, T.; Shaoming, M.; Norton, D.; Haycock, J. W.; Ryan, A. J.;Macneil, S. Self-Organization of Skin Cells in Three-DimensionalElectrospun Polystyrene Scaffolds. Tissue Eng. 2005, 11 (7−8), 1023−33.(10) Uyar, T.; Besenbacher, F. Electrospinning of uniformpolystyrene fibers: The effect of solvent conductivity. Polymer 2008,49 (24), 5336−5343.(11) Kang, M.; Jung, R.; Kim, H.-S.; Jin, H.-J. Preparation ofsuperhydrophobic polystyrene membranes by electrospinning. ColloidsSurf., A 2008, 313−314 (0), 411−414.(12) Lee, S. J.; Tatavarty, R.; Gu, M. B. Electrospun polystyrene−poly(styrene-co-maleic anhydride) nanofiber as a new aptasensorplatform. Biosens. Bioelectron. 2012, 38 (1), 302−307.(13) Tatavarty, R.; Hwang, E. T.; Park, J.-W.; Kwak, J.-H.; Lee, J.-O.;Gu, M. B. Conductive quantum dot-encapsulated electrospunnanofibers from polystyrene and polystyrene-co-maleic anhydridecopolymer blend as gas sensors. React. Funct. Polym. 2011, 71 (2),104−108.(14) Tang, C.; Ye, S.; Liu, H. Electrospinning of poly(styrene-co-maleic anhydride) (SMA) and water-swelling behavior of crosslinked/hydrolyzed SMA hydrogel nanofibers. Polymer 2007, 48 (15), 4482−4491.(15) Stoilova, O.; Ignatova, M.; Manolova, N.; Godjevargova, T.;Mita, D. G.; Rashkov, I. Functionalized electrospun mats fromstyrene−maleic anhydride copolymers for immobilization of acetyl-cholinesterase. Eur. Polym. J. 2010, 46 (10), 1966−1974.(16) Ignatova, M.; Stoilova, O.; Manolova, N.; Mita, D. G.; Diano,N.; Nicolucci, C.; Rashkov, I. Electrospun microfibrous poly(styrene-alt-maleic anhydride)/poly(styrene-co-maleic anhydride) mats tailoredfor enzymatic remediation of waters polluted by endocrine disruptors.Eur. Polym. J. 2009, 45 (9), 2494−2504.(17) Muller, I.; Jenner, A.; Bruchelt, G.; Niethammer, D.; Halliwell,B. Effect of Concentration on the Cytotoxic Mechanism ofDoxorubicinApoptosis and Oxidative DNA Damage. Biochem.Biophys. Res. Commun. 1997, 230 (2), 254−257.(18) Eom, Y. W.; Kim, M. A.; Park, S. S.; Goo, M. J.; Kwon, H. J.;Sohn, S.; Kim, W. H.; Yoon, G.; Choi, K. S. Two distinct modes of celldeath induced by doxorubicin: apoptosis and cell death throughmitotic catastrophe accompanied by senescence-like phenotype.Oncogene 2005, 24 (30), 4765−77.(19) Zhao, G.; Hochwalt, P. C.; Usui, M. L.; Underwood, R. A.;Singh, P. K.; James, G. A.; Stewart, P. S.; Fleckman, P.; Olerud, J. E.Delayed wound healing in diabetic (db/db) mice with Pseudomonasaeruginosa biofilm challenge: a model for the study of chronic wounds.Wound Repair and Regeneration 2010, 18 (5), 467−477.(20) Babraham Bioinformatics; http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/.(21) Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.journal 2011, 17 (1), 10−12.(22) Kersey, P. J.; Staines, D. M.; Lawson, D.; Kulesha, E.; Derwent,P.; Humphrey, J. C.; Hughes, D. S. T.; Keenan, S.; Kerhornou, A.;Koscielny, G.; Langridge, N.; McDowall, M. D.; Megy, K.; Maheswari,U.; Nuhn, M.; Paulini, M.; Pedro, H.; Toneva, I.; Wilson, D.; Yates, A.;Birney, E. Ensembl Genomes: an integrative resource for genome-scaledata from non-vertebrate species. Nucleic Acids Res. 2012, 40 (D1),D91−D97.(23) Dobin, A.; Davis, C. A.; Schlesinger, F.; Drenkow, J.; Zaleski, C.;Jha, S.; Batut, P.; Chaisson, M.; Gingeras, T. R. STAR: ultrafastuniversal RNA-seq aligner. Bioinformatics 2013, 29 (1), 15−21.
(24) Anders, S. HTSeq, Analysing high-throughput sequencing datawith Python; EMBL Heidelberg (Genome Biology Unit), 2010.(25) Love, M. I.; Huber, W.; Anders, S. Moderated estimation of foldchange and dispersion for RNA-seq data with DESeq2. Genome Biol.2014, 15 (12), 550.(26) Robinson, M. D.; McCarthy, D. J.; Smyth, G. K. edgeR: aBioconductor package for differential expression analysis of digitalgene expression data. Bioinformatics 2010, 26 (1), 139−40.(27) The R Foundation, The R Project for Statistical Computing;www.r-project.org.(28) Ignatova, M.; Stoilova, O.; Manolova, N.; Markova, N.; Rashkov,I. Electrospun mats from styrene/maleic anhydride copolymers:modification with amines and assessment of antimicrobial activity.Macromol. Biosci. 2010, 10 (8), 944−954.(29) Kang, J.; Yoo, H. S. Nucleic Acid-Scavenging ElectrospunNanofibrous Meshes for Suppressing Inflammatory Responses.Biomacromolecules 2014, 15 (7), 2600−2606.(30) Sarkar, K.; Debnath, M.; Kundu, P. P. Preparation of low toxicfluorescent chitosan-graft-polyethyleneimine copolymer for genecarrier. Carbohydr. Polym. 2013, 92 (2), 2048−2057.(31) Filion, M. C.; Phillips, N. C. Toxicity and immunomodulatoryactivity of liposomal vectors formulated with cationic lipids towardimmune effector cells. Biochim. Biophys. Acta, Biomembr. 1997, 1329(2), 345−356.(32) Goodman, C. M.; McCusker, C. D.; Yilmaz, T.; Rotello, V. M.Toxicity of Gold Nanoparticles Functionalized with Cationic andAnionic Side Chains. Bioconjugate Chem. 2004, 15 (4), 897−900.(33) Eom, Y.-W.; Kim, M. A.; Park, S. S.; Goo, M. J.; Kwon, H. J.;Sohn, S.; Kim, W.-H.; Yoon, G.; Choi, K. S. Two distinct modes of celldeath induced by doxorubicin: apoptosis and cell death throughmitotic catastrophe accompanied by senescence-like phenotype.Oncogene 2005, 24 (30), 4765−4777.(34) Guiducci, C.; Tripodo, C.; Gong, M.; Sangaletti, S.; Colombo,M. P.; Coffman, R. L.; Barrat, F. J. Autoimmune skin inflammation isdependent on plasmacytoid dendritic cell activation by nucleic acidsvia TLR7 and TLR9. J. Exp. Med. 2010, 207 (13), 2931−42.(35) Iwasaki, A.; Medzhitov, R. Toll-like receptor control of theadaptive immune responses. Nat. Immunol. 2004, 5 (10), 987−95.(36) Oka, T.; Hikoso, S.; Yamaguchi, O.; Taneike, M.; Takeda, T.;Tamai, T.; Oyabu, J.; Murakawa, T.; Nakayama, H.; Nishida, K.; Akira,S.; Yamamoto, A.; Komuro, I.; Otsu, K. Mitochondrial DNA thatescapes from autophagy causes inflammation and heart failure. Nature2012, 485 (7397), 251−255.
