Delivery of streptomycin to the rat colon by use of electrospun nanofibers

Drug-loaded electrospun nanofibers are potential drug carrier systems that may optimize disease treatment while reducing the impact on commensal microbes. The feasibility of streptomycin-loaded pullulan nanofibers fabricated from a green electrospinning procedure using water as the solvent was assessed. We conducted a rat study including a group treated with streptomycin-loaded nanofibers (STR-F, n = 5), a group treated with similar concentrations of streptomycin in the drinking water (STR-W, n = 5), and a non-treated control group (CTR, n = 5). Streptomycin was successfully loaded into nanofibers and delivered by this vehicle, which minimized the quantity of the drug released in the ileal compartment of the gut. Ingested streptomycin-resistant E. coli colonized of up to 106 CFU/g feces, revealing a selective effect of streptomycin even when given in the low amounts allowed by the nanofiber-based delivery. 16S amplicon sequencing of the indigenous microbiota revealed differential effects in the three groups. An increase of Peptostreptococcaceae in the cecum of STR-F animals may indicate that the fermentation of nanofibers directly or indirectly promoted growth of bacteria within this family. Our results elucidate relevant properties of electrospun nanofibers as a novel vehicle for delivery of antimicrobials to the large intestine.


Results and discussion
Streptomycin encapsulation preserves antimicrobial properties. Disk-diffusion assays, performed by the Kirby-Bauer method, confirmed that the streptomycin antimicrobial activity remained after loading into electrospun fibers. Thus, streptomycin-loaded nanofibers inhibited the growth of sensitive E. coli and S. aureus strains (Fig. 1). Overall, we conclude that encapsulation of the drug did not affect the antimicrobial properties of the drug. Thus, electrospun nanofibers were suitable for maintaining the optimum therapeutic level of the antimicrobial drug in vitro. A similar study performed with amoxicillin has demonstrated that the nanofiber antimicrobial embedment does not compromise the release and activity of the drug 24 . Therefore, electrospun nanofibers may be relevant in various applications in the field of drug delivery.

Streptomycin delivered in electrospun nanofibers allow proliferation of streptomycin-resistant E. coli in rats.
A rat study was performed to investigate the effects of streptomycin-carrying nanofibers on the degree of colonization of streptomycin-resistant E. coli MG1655-K12. Before the inoculation, no background resistance to streptomycin or gentamycin was detected in the fecal samples from these animals (data not shown). On Day 1 and Day 2 after inoculation, significantly more streptomycin-resistant E. coli were found in fecal samples from rats that received streptomycin either encapsulated into the nanofibers (STR-F) or in drinking water (STR-W) than in the control (CTR) group (Fig. 2a). No difference was observed between STR-F and STR-W (Fig. 2a). Higher numbers of resistant E. coli were found in cecal samples of the STR-W animals, while the difference between the STR-F group and the CTR was not significant (Fig. 2b,c). No difference in resistant E. coli numbers between any of the groups was observed in colonic and fecal samples obtained on Day 3, likely because the streptomycin treatment ceased on Day 2 in both STR-W and STR-F. Streptomycin, given as 5 g/L in drinking water, has for decades been used in animal models to suppress facultative anaerobic bacteria, and thereby allow ingested streptomycin-resistant proteobacteria species to overcome the colonization barrier [25][26][27] . In the current experiment, we have applied a much lower concentration (0.25 g/L) in drinking water of the STR-W rats in an attempt to match the absolute amount of 7 mg of streptomycin given as nanofibers to the STR-F rats. Nevertheless, our results demonstrate that it is possible to obtain intestinal proliferation of resistant bacteria at this 20-fold lower concentration than the currently applied standard.
Electrospun nanofibers delivery provides a higher concentration of streptomycin. Measurement of actual water consumption during the experiment revealed that the total streptomycin consumed by  www.nature.com/scientificreports/ STR-W rats was around 12.5 mg, considering an estimated 50% spillage. As mentioned above, streptomycin dosed by nanofibers to STR-F rats during the experiment was approximately 7 mg. The concentration of streptomycin released in the gut from nanofibers was assessed by ELISA of rat intestinal content obtained on Day 3. Interestingly, we observed a significantly (P < 0.005) lower concentration of streptomycin in the ileal samples from the STR-F group, as compared to the STR-W group (Fig. 3). We speculate that the coated gelatin capsules start dissolving in the small intestine and release the electrospun pullulan fibers, which have mucoadhesive properties 9 . The pullulan nanofibers then dissolve and form a jelly-like, mucoadhesive construct, retaining the drug partially through the ileum and moving into the large intestine, where more drug get gradually released.  www.nature.com/scientificreports/ In line with our observations, core-shell fibers, which are produced by coaxial electrospinning, have previously shown successful sustained release of hydrophilic drugs such as streptomycin 28 . Regarding drug delivery applications using pullulan nanofibers, there are several studies focusing on fast dissolving properties of pullulan and recommending it for fast delivery approaches via the buccal route 8,19,29,30 or for wound dressing applications 6,31 . Our findings suggest that antibiotic-loaded pullulan nanofibers are also suitable for delivery to the large intestine, due to the mucoadhesive properties of the dissolved pullulan.
Streptomycin alters the gut microbiota composition. Sequencing of 16S rRNA gene amplicons was performed to assess the effect of streptomycin treatment on the intestinal microbiota. Microbial richness was significantly reduced in fecal samples from STR-F (P = 0.008), and STR-W (P = 0.007) groups, compared with the CTR group (Fig. 4). In the cecum, a slightly higher richness was observed in the CTR group than in STR-F samples.
Beta-diversity was compared between groups based on the Bray-Curtis dissimilarity and visualized with non-metric multidimensional scaling (NMDS). Three distinct clusters represented STR-W, STR-F and CTR samples, respectively (Fig. 5a). Overall, the treatment was more influential than the location in the gut, based on the PERMANOVA model (0.27 [R 2 for Group] vs. 0.09 [R 2 for location], P = 0, permutation test with 100 iterations). The pairwise comparison of all groups at each location was significant in all cases, except for fecal samples from the two STR-treated groups (Fig. 5b).
Together, Bacteroidetes and Firmicutes constituted more than 99% of the ASVs identified in rat samples. We found a significant reduction of the relative abundance of Firmicutes and a corresponding relative increase of Bacteroidetes in STR-W and STR-F groups in samples from cecum, colon and feces ( Fig. 5c).
At family level, lower levels of Lactobacillaceae were found in the cecum of STR-F animals as compared to both STR-W and CTR (Fig. 6), while a higher abundance of Peptostreptococcaceae, specifically Romboutsia, was seen (Fig. 6). Peptostreptococcaceae are commensals in healthy rats 32 , and Romboutsia, which belongs to the Peptostreptococcaceae family, consists of anaerobe microorganisms adapted to utilize a wide range of carbohydrates, and relies on exogenous amino acids and peptides for protein synthesis 33 . We speculate that the observed higher levels of this species in the STR-F group than in the STR-W group results from the fermentation of nanofibers in the colon, which directly or indirectly promotes the growth of bacteria within this family. The indigestible portions of the nanofibers were composed of maltotriose units connected by α-1,6 bonds, which are resistant to mammalian intestinal enzymes but are available for bacterial fermentation in the large bowel 34 .

Conclusion
Streptomycin can be incorporated into electrospun nanofibers without losing the antibacterial activity, and furthermore it can successfully be retained through the small intestine and delivered to the large intestine of rats by dosing with encapsulated nanofibers. Our results suggest that electrospun nanofibers represent an advanced alternative for the delivery of antimicrobials to the large intestine. This approach is likely to allow for reduction in the amount of antibiotics needed to be effective, and thereby to reduce the risk of development of resistance, Colon Feces Observed richness

Methods
Fabrication of antibiotic-loaded nanofibers. Pullulan was kindly provided by HAYASHIBARA CO., LTD (Japan). Water was purified using a Milli-QPlus 185 water purification system (Millipore, Bedford, MA) with resistivity higher than 18 MΩ·cm. Streptomycin standard disks (25 µg per disk) were purchased from Oxoid, Denmark. Hexamethyldisilazane (HMDS) was obtained from Merck (Germany). Streptomycin, Gentamycin, Trehalose, and RSM (skim milk powder) were purchased from Sigma-Aldrich, Denmark. Streptomycin was loaded into pullulan electrospun nanofibers with final concentrations of 1 and 10% respectively (w/w; weight of streptomycin to the weight of polymer). For encapsulation of streptomycin in Pullulan, streptomycin powder was first dissolved in water, and then pullulan powder was added to the solution to achieve the desired concentration. For making 1% or 10% streptomycin to Pullulan constructs (w/w), first, a streptomycin solution of 0.2% or 2% (w/v) in water was made under slight stirring at 25 ℃ for 4 h. Then pullulan powder was added to make a 20% solution (w/v) of Pullulan in water and was stirred overnight at room temperature. For instance, for fabrication of a 1% (w/w) streptomycin: pullulan nanofiber sheet, streptomycin was dissolved in water (conc. 2 mg/mL) under slight stirring at 25 ℃ for 4 h. Then pullulan powder (0.2 g) was added to the 1 mL solution to make the desired final concentration of 20% pullulan in water for optimal electrospinning. Finally, the solution was sterile-filtered through a 25 mm filter (Sigma Aldrich) and used for electrospinning.
For electrospinning, a customized conventional electrospinning system including a high voltage source (Glassman, FJ50P2.4-F22), a syringe pump (Aladdin, AL-1000) connected to a syringe with a 21 G hypodermic needle, and a static collector was used. The optimized spinning parameters were set to a feeding rate of 1 mL/h, a voltage of 14-16 kV, a needle tip-to-collector distance of 15 cm, a humidity of 35%, and a temperature of 25 °C. The electrospinning chamber including the pump and the collector was disinfected by EtOH 70% before starting each electrospinning procedure.  The largest node in the center is the kingdom bacteria. Along the tree branch outward, the next node is the phylum level, then followed by Class, Order, Family, and Genus. The size of the node is proportional to the mean relative abundance in the corresponding phylogenetic level. A node is labelled and colored when it has significantly different relative abundances (permutation test) between groups. Figure was made in R as described in the "Methods" section.

Antibiotic-loaded nanofibers for in vivo study.
To dose rats with antibiotic-loaded nanofibers, the electrospun sheets of 10% streptomycin:pullulan were cut manually to make square particles of around 1 mm in length. Then, hard-shelled gelatine capsules (Torpac size 9) were coated in a 12% (w/v) solution of Eudragit L100 in IPA where dibutyl sebacate was added as a plasticizer in a 5% w/w ratio relative to Eudragit. Finally, capsules were filled with 12 mg of particles (equivalent to 1.2 mg streptomycin per capsule) before dosing to the rats.
Animal experiment. To find out whether the nanofiber electrospun sheet loaded with streptomycin affected the establishment of streptomycin-resistant bacteria in the intestine, we performed a rat study involving oral inoculation of the animals with the streptomycin-resistant mCherry-labelled E. coli MG1655 subst. K12, Taxon Identifier 511145 (https:// www. unipr ot. org/ unipr ot/ A0A4D 6FVK6) in association with streptomycin delivered either as encapsulated nanofibers, in the drinking water or without streptomycin as control (Fig. 7). The fluorescence-labeled strain was chosen to allow easy verification by microscopy. Fifteen male 8-weeks-old Sprague-Dawley rats (Charles River) were allocated to three groups of five and housed individually in scantainers (Scanbur). After an acclimatization period of seven days, the animals were treated with streptomycin either in drinking water (STR-W), inside of particles of electrospun nanofibers (STR-F), or did not receive any antimicrobial drug (CTR). The CTR rats received a placebo hard-shelled gelatine capsule, STR-F rats received a hard-shelled gelatine capsule filled with antibiotic-loaded pullulan particles as described above twice a day for 3 days (Day 0, 1 and 2), while the STR-W rats received an empty placebo hardshelled gelatine capsule and Streptomycin in the drinking water (0.25 g/L) for the same three days (Fig. 7). We aimed at a concentration in drinking water, which would make the total consumption of streptomycin by the STR-W rats match the approximately 7 mg given to the STR-F rats considering that the average water consumption was between ca. 30-50 mL of water/animal/day. For this study, the streptomycin-and gentamycin-resistant E. coli were incubated on LB agar (SSI Diagnostica A/S) for 16 h, as described earlier. A colony was transferred to LB broth supplemented with 20 µg/mL gentamycin and 100 µg/mL streptomycin overnight/16 h at 37 °C while shaking. The culture was centrifuged at 5000×g for 10 min, washed and resuspended in sterile phosphate-buffered saline (PBS) to a cell density of 10 9 CFU/mL. Twenty-four hours after the onset of streptomycin treatment, all Enumeration of resistant E. coli in the rat gut. Fecal samples (approximately 1 g) were collected before inoculation and 24 h, 48 h, and 72 h after inoculation. After 72 h, animals were euthanized by CO 2 asphyxiation followed by decapitation. Rats were dissected and luminal contents were aseptically collected from the distal ileum, cecum, and colon regions. The contents from these sections were weighed and homogenized. Tenfold dilutions were spread on LB agar plates supplemented with streptomycin (100 µg/mL) or gentamycin (25 µg/mL) (Sigma-Aldrich, Denmark), as appropriate. CFU counts were assessed after 24-72 h incubation at 37 °C under anaerobic conditions. Random colonies were analysed by mass spectrometry using Bruker Reflex™ III MALDI-TOF (Bruker-Daltonik, Germany) to ensure that the isolates corresponded to E. coli K-12 MG1655.
ELISA assay. To estimate the quantity of drug released in the intestinal lumen, an ELISA assay was performed to verify the concentration of streptomycin present in fecal content. Fecal pellets from the ileum, cecum, and colon were analyzed. Samples were homogenized and diluted, and the procedure was performed according to the SM Streptomycin ELISA kit (Elabscience, USA) (Catalog # E-FS-E031) manufacturer instructions. The optical density (OD) was measured using a microplate reader Biotek EL800 (Agilent, DK), set at 450 nm and 630 nm. Data were processed using GraphPad Prism Software 8.1.1.
DNA extraction, sequencing, and bioinformatics. DNA extraction, library preparation, and subsequent sequencing on the Ion Torrent platform were performed as previously described 36 . Briefly, samples consisted of fecal samples at 72 h (n = 15), samples from ileum (n = 15) and samples from colon (n = 15). Microbial DNA was extracted using ~ 0.2 g of materials with the DNeasy® PowerLyzer® PowerSoil® isolation kit from Qiagen. The extracted DNA was stored at − 20 °C before use. Amplification of the V3-region of the 16S rRNA gene was performed as follows: PCR mastermix consisted of 0.2 µL Phusion High-Fidelity DNA polymerase (ThermoFisher Scientific F-553L), 4 µL HF buffer, 0.4 µL dNTP (10 mM of each base), 1 µM forward primer (PBU; 5ʹA-adapter-TCAG-barcode-CCT ACG GGA GGC AGCAG-3ʹ) and 1 µM reverse primer (PBR; 5ʹ-trP1-adapter-ATT ACC GCG GCT GCTGG-3ʹ) and 5 ng community fecal DNA in 20 µL total reaction volume. Both primers (TAG Copenhagen A/S) were linked to sequencing adaptors and the forward primer additionally contained a unique 10 bp barcode (Ion Xpress™ Barcode Adapters) for each sample. The PCR program consisted of initial denaturation for 30 s at 98 °C, followed by 24 cycles of 98 °C for 15 s and 72 °C for 30 s, and lastly, 72 °C for 5 min to allow final extension before cooling to 4 °C. No-template controls were included for each PCR run, all resulting in less than 0.05 ng/µL. The PCR products were purified using HighPrepTM PCR Magnetic Beads (MAGBIO®, AC-60005) with a 96-well magnet stand (MAGBIO®, MyMag 96), according to the manufacturer recommendation. DNA quantity was measured using Qubit® dsDNA HS assay (InvitrogenTM, Q32851) and sequenced on chip using the Ion OneTouchTM and Ion PGM systems with a 318-Chip v2 incorporating the Hi-Q chemistry in 200 bp runs.
Raw sequences were demultiplexed with barcodes and primers removed using QIIME2 37 . The obtained sequences were further filtered to range from 125 to 180 bp. These pre-processed reads were analyzed with the DADA2 pipeline in R 38 , using the parameters recommended for Ion Torrent. The obtained amplicon sequence variants (ASVs) were thereafter taxonomically annotated using the Ribosomal Database Project (RDP) database (release 11.5).
Statistics and reproducibility. Statistical analysis was performed with QIIME2, R, and GraphPad Prism Software 8.1.1. Samples were rarefied to an even depth of reads before calculating richness. CFU counts, streptomycin concentration, microbial richness, and taxa abundances were compared with a two-sided permutational test. ASVs (Amplicon sequencing variant) with less than 0.01% relative abundances were removed before further analysis. Beta diversity was assessed based on the Bray-Curtis distance matrix and compared with the permutational multivariate analysis of variance (PERMANOVA). The taxonomic composition shown with trees was generated with R-package "metacoder" 39 . The significance level was set at 0.5. Comparison of ELISA-assessed streptomycin levels in ceca of STR-F and STR-W animals was done by t-test assuming equal variances.

Animal ethical statement. Danish Animal Experiments Inspectorate approved the animal trial under
license number 2020-15-0201-00484. The experiment protocol was pre-registered by DTU's BioFacility. Animals were monitored daily during the experiment, and the experiment was carried out in accordance with Danish national guidelines and overseen by the Institute's Animal Welfare Committee for animal care and use. Methods are reported in accordance with ARRIVE guidelines.