Biodegradable polymer nanoparticles that rapidly penetrate the human mucus barrier

Abstract

Protective mucus coatings typically trap and rapidly remove foreign particles from the eyes, gastrointestinal tract, airways, nasopharynx, and female reproductive tract, thereby strongly limiting opportunities for controlled drug delivery at mucosal surfaces. No synthetic drug delivery system composed of biodegradable polymers has been shown to penetrate highly viscoelastic human mucus, such as non-ovulatory cervicovaginal mucus, at a significant rate. We prepared nanoparticles composed of a biodegradable diblock copolymer of poly(sebacic acid) and poly(ethylene glycol) (PSA-PEG), both of which are routinely used in humans. In fresh undiluted human cervicovaginal mucus (CVM), which has a bulk viscosity approximately 1,800-fold higher than water at low shear, PSA-PEG nanoparticles diffused at an average speed only 12-fold lower than the same particles in pure water. In contrast, similarly sized biodegradable nanoparticles composed of PSA or poly(lactic-co-glycolic acid) (PLGA) diffused at least 3,300-fold slower in CVM than in water. PSA-PEG particles also rapidly penetrated sputum expectorated from the lungs of patients with cystic fibrosis, a disease characterized by hyperviscoelastic mucus secretions. Rapid nanoparticle transport in mucus is made possible by the efficient partitioning of PEG to the particle surface during formulation. Biodegradable polymeric nanoparticles capable of overcoming human mucus barriers and providing sustained drug release open significant opportunities for improved drug and gene delivery at mucosal surfaces.

Drugs delivered to mucosal surfaces are typically efficiently removed by mucus clearance mechanisms and systemic absorption, precluding prolonged drug presence locally. Encapsulation of drugs and genes in polymeric particles offers the potential for localized and sustained delivery to mucosal tissues (1, 2). However, the luminal surface of mucosal tissues is protected by a layer of highly viscoelastic and adhesive mucus that traps and rapidly removes foreign particles (3), thereby precluding sustained drug delivery (2, 4–6). The reported diffusion coefficient for latex nanoparticles ranging in size from 59–1,000 nm is 0 μm²/s in human cervical mucus (7); particles trapped in this manner are typically cleared from mucosal surfaces within seconds to a few hours depending on anatomical site (3, 8, 9). The mucus barrier has been cited as a critical obstacle to the treatment of a variety of diseases (10–12). To avoid rapid clearance and achieve high local drug dosages that can be sustained over time, biodegradable nanoparticle systems that are capable of penetrating highly viscoelastic human mucus secretions are needed.

We have recently discovered that densely coating non-biodegradable latex particles with PEG, a hydrophilic and uncharged molecule, effectively minimizes adhesive interactions between nanoparticles and mucins (13), thereby allowing nanoparticles to rapidly penetrate highly viscoelastic human mucus by moving through openings between mucin mesh fibers. We also found that the dense surface coating must be accomplished with low MW PEG, which provides an effective shield of the hydrophobic core of the particles while minimizing interpenetration or intermolecular interactions between PEG polymers and the mucin mesh (14). However, polymeric mucus-penetrating particle (MPP) formulations (excluding particles observed in low viscosity ovulatory cervical mucus (15)) with effective PEG coatings to facilitate rapid nanoparticle transport (D_m/D_w >0.05) have thus far been limited to non-degradable latex particles (13, 14, 16, 17). The preparation process required to achieve sufficiently dense PEG coatings on latex nanoparticles takes advantage of the presence of a high density of reactive groups on the surface of the latex particles to which PEG molecules are covalently attached. However, biodegradable particles do not usually have a high density of reactive groups available on their surfaces. The PEG-coating process for latex beads also includes vigorous mixing in aqueous solutions for prolonged durations (>4 h), during which substantial drug loss and polymer degradation could occur, further limiting the practicality of this approach for biodegradable systems. It was not known whether biodegradable particles with sufficiently high PEG density for rapid penetration of highly viscoelastic and adhesive human mucus secretions could be achieved using conventional encapsulation procedures. In this work, we hypothesized that a biodegradable diblock copolymer with hydrophobic (PSA) and hydrophilic (PEG) block segments would allow sufficient partitioning of PEG to the surface of particles to provide a muco-inert coating of the hydrophobic PSA core, and thereby allow PSA-PEG nanoparticles to rapidly penetrate human mucus secretions.

Results

We synthesized PSA-PEG diblock copolymers (M_W = 18 kDa) (Fig. S1) via melt polycondensation of sebacic acid (SA) and methoxy-PEG prepolymers (18). We used 5 kDa PEG based on our previous observations that 2–5 kDa PEG provided a non-mucoadhesive coating on non-degradable latex beads, whereas coatings using 10 kDa PEG resulted in strong particle mucoadhesion (14). We confirmed the incorporation of PEG into the polymer backbone by ¹H NMR, with PSA-PEG exhibiting strong peaks at 3.65 (s, OCH₂CH₂O, PEG), 2.18 (m, 4H, SA) 1.48 (t, 4H, SA), and 1.24 (s, 8H, SA), combined with a gel permeation chromatograph with no peak corresponding to free PEG. Using these polymers, we formulated nanoparticles with an average hydrodynamic diameter of 173 nm (Table 1) using a conventional solvent diffusion method where the continuous phase was water (19). Particles made using PSA-PEG, PSA, and PLGA all possessed a narrow size distribution, as confirmed by particle polydispersity measurements (Table 1) and scanning electron microscopy (Fig. S2). The separation during particle formulation of a dense PEG coating toward the continuous phase and, therefore, the surface of PSA-PEG particles, was confirmed via fluorescence microscopy (Fig. 1A). ζ-potential measurements showed a near neutral surface charge for PSA-PEG nanoparticles and a highly negative surface charge for nanoparticles made of either PSA or PLGA (Table 1), further evidence of dense surface PEG coverage of the PSA-PEG nanoparticles. The surface charge on PSA-PEG nanoparticles, −1 ± 5 mV, was within the range of ζ-potential values we previously showed necessary for PEG to effectively shield non-degradable latex particles and allow them to move rapidly in human mucus (14).

Table 1.

Physicochemical properties of nanoparticles and their diffusivity in human cervicovaginal mucus (D_m) compared to in water (D_w)

Formulation	Diameter, nm	Polydispersity index*	ζ-potential, mV	Dw/Dm†
PS‡	109 ± 3	0.04	−41 ± 2	40,000
PS‡	217 ± 5	0.03	−59 ± 4	2,200
PLGA	150 ± 62	0.07	−57 ± 7	3,300
PSA	198 ± 74	0.12	−39 ± 9	3,400
PSA-PEG	173 ± 79	0.15	−1 ± 5	12§

Samples include particles composed of polystyrene (PS), poly(sebacic acid) (PSA), poly(lactic-co-glycolic acid) (PLGA), and copolymers of poly(sebacic acid) and poly(ethylene glycol) (PSA-PEG). Effective diffusivity in mucus is calculated at time scale of 1 s and nanoparticle diffusivity in water is calculated from the Stokes-Einstein equation using average particle diameter.

*Polydispersity index (PDI) represents the relative variance in the particle size distribution, as further described in the Malvern Zetasizer instrument manual. In general, the more monodisperse the particles, the lower the PDI.

^†D_w/D_m represents the ratio of the average diffusion coefficient of each nanoparticle type in water compared to in human CVM. Thus, the values shown represent the multiple by which particle diffusivity is reduced in CVM compared to in water.

^‡Data obtained from reference (13).

^§Statistically significant faster transport as compared to all other formulations (P < 0.05).

Fig. 1.

Transport of poly(sebacic acid)-co-poly(ethylene glycol) (PSA-PEG) and poly(sebacic acid) (PSA) nanoparticles in human cervicovaginal mucus (CVM). (A) Fluorescent image shows a dense PEG coating on a PSA-PEG particle. (Scale bar, 500 nm.) Sample 20 s trajectories of (B) PSA and (C) PSA-PEG nanoparticles in CVM. Trajectories reflect particles with effective diffusivities within one standard error of the mean of the ensemble average. (D) Ensemble-averaged geometric mean-squared displacements (<MSD>) for PSA and PSA-PEG nanoparticles as a function of time scale. Data represent the ensemble average of three independent experiments, with n ≥ 150 particles for each experiment. Error bars indicate standard error. * denotes statistically significant difference compared to PSA (P < 0.05).

We next studied, using high-resolution multiple particle tracking (1, 13, 20), the transport rates in mucus of nanoparticles composed of PSA-PEG as compared to those composed of conventional biodegradable polymers, PSA and PLGA. We used fresh, undiluted, and unmanipulated human CVM, obtained from female donors with healthy vagina flora (13). PSA and PLGA nanoparticles were strongly immobilized in CVM, as evidenced by the highly constrained, non-Brownian time lapse traces of typical particles (Fig. 1B). In contrast, nanoparticles composed of PSA-PEG diblock copolymers displayed Brownian-like trajectories that probed distances far greater than their diameters in CVM over the course of 20 s movies (Fig. 1C). We quantified the transport rates of different particle formulations in mucus by their time scale-dependent geometric ensemble-averaged mean-squared displacements (<MSD>). At a time scale of 1 s, the <MSD> of PSA-PEG particles was 400-fold higher than that of PSA particles (Fig. 1D), 230-fold higher than that observed for PLGA, and 220-fold higher than that of latex particles (Table 1 and Fig. 2A). The average effective diffusivity for PSA-PEG particles in mucus was only 12-fold slower than that for the same particles in pure water, indicating that the biodegradable PSA-PEG particles penetrated CVM at similar efficiencies as we have previously found for non-degradable PEG-coated latex particles (13, 14). The extent of impediment to particle motion can be determined by fitting <MSD> to the equation, <MSD> = 4D_oτ^α, where D_o is the time scale-independent diffusion coefficient, τ is the time scale, and α ranges from 0 (completely immobile particles) to 1 (unobstructed Brownian diffusion, such as that of particles in water) (13, 21). We previously observed that 200 nm latex particles, which were strongly hindered in mucus, had an α value of 0.36, whereas 200 nm PEG-coated latex particles, which diffused rapidly, had an α value of 0.81 (13). The average α for PSA and PSA-PEG particles was 0.08 and 0.92, respectively, indicating that PSA particles were strongly trapped in CVM, whereas PSA-PEG particles were able to diffuse nearly unimpeded through the dense mesh network of CVM. This implies that PSA-PEG nanoparticles experience primarily the viscous drag of the low viscosity interstitial fluid between mucus mesh elements and, thus, are able to rapidly penetrate the mucus secretion (17). Since CVM is biochemically and rheologically similar to most human mucus secretions (notable exceptions are ovulatory cervical mucus and the ocular mucus tear film, each of which is more dilute than typical secretions) (13), we expect PSA-PEG nanoparticles may also readily penetrate mucus layers covering most mucosal organs.

Fig. 2.

Diffusivity of poly(sebacic acid)-co-poly(ethylene glycol) (PSA-PEG) and poly(sebacic acid) (PSA) nanoparticles in human cervicovaginal mucus (CVM). (A) Geometric ensemble effective diffusivity (<D_eff>) at a time scale of 1 s for polystyrene (PS), poly(lactic-co-glyolic acid) (PLGA), PSA, and PSA-PEG in CVM. × denotes individual sample <D_eff> values (n = 3); — denotes the average. PS values are obtained from literature (13). (B) Distributions of the logarithms of individual particle effective diffusivities (D_eff) for PSA and PSA-PEG particles at a time scale of 1 s. (C) Transport mode distributions of PSA and PSA-PEG particles in CVM. Data represent mean ± standard error of three independent experiments, with n ≥ 150 nanoparticles for each experiment. Immobile particles possess MSD below the microscope detection limit (10 nm) for the entire duration of microscopy. (D) The estimated fraction of particles predicted to be capable of penetrating a 30 μm thick mucus layer over time using Fick's second law and diffusion coefficients obtained from tracking experiments. * denotes statistically significant differences (P < 0.05).

To ensure the observed rapid transport of biodegradable PSA-PEG nanoparticles was not biased by a small fraction of fast-moving outlier particles, we evaluated the heterogeneity in particle transport rates by examining the distribution of particle effective diffusivities at a time scale of 1 s (Fig. 2B). The fastest 70% of PSA-PEG particles exhibited uniformly rapid transport, with an effective diffusivity only approximately 3-fold slower than that in water and an α value of 0.94. In contrast, the fastest 70% of PSA particles had an effective diffusivity approximately 7,000-fold slower than in water, with an α value of 0.083. To investigate the mechanistic reasons that account for the difference in transport between PSA-PEG and PSA, we further assigned particles to three non-overlapping transport modes, immobile (I), hindered (H), and diffusive (D) (with rates of movement D > H > I), using a Monte Carlo method based on time scale-dependent effective diffusion coefficients (13, 22). Over 50% of PSA-PEG were classified as undergoing nearly unhindered diffusion, and 0% were immobilized (defined as having an MSD below the resolution of the microscope; Fig. 2C). In contrast, less than 1% of the PSA particles were classified as diffusive, and 99% were classified as strongly hindered or immobile.

We next assessed the rates at which biodegradable PSA-PEG and PSA particles can penetrate a mucus gel layer by fitting Fick's second law to their measured diffusion rates. In the lung, estimates for the thickness of the luminal mucus gel range from 7-μm in the deep airways to 55-μm in the bronchi (23). Assuming a mucus layer thickness of 30 μm and extrapolating the measured particle diffusion rates, we predicted the amount of PSA-PEG and PSA particles capable of crossing the mucus layer over time (as a fraction of the original dose) (Fig. 2D). By this rough analysis, we found that more than 35% of PSA-PEG nanoparticles would be capable of crossing a 30 μm mucus layer within 10 min. In the absence of PEG, particles composed of PSA would penetrate a 30 μm thick mucus layer at very low efficiencies (<0.6% after 4 h). PSA nanoparticles (and, similarly, PLGA nanoparticles) are thus expected to be efficiently removed from mucosal surfaces as mucus is cleared, which typically occurs on the order of seconds on the surfaces of the eyes, and minutes to a few hours in the lungs, gastrointestinal (GI) tract and vagina. The increased thickness of mucus layers in the GI tract (thickness ∼50–500 μm) and female reproductive tract compared to the lungs is expected to further accentuate the difference in the penetrable fraction of nanoparticles composed of PSA-PEG versus those composed of PSA or PLGA (2).

We further tested the ability of biodegradable PSA-PEG particles to penetrate fresh, undiluted sputum expectorated from cystic fibrosis (CF) patients, which has a bulk viscosity typically approximately 10⁴- to 10⁵-fold higher than that of water under low shear (compared to ∼10³-fold for mucus from healthy volunteers) (24). At a time scale of 1 s, PSA-PEG nanoparticles exhibited a 50-fold greater <MSD> than uncoated latex particles in CF sputum (Fig. 3A), with average α values of 0.32 and 0.89 for latex and PSA-PEG particles, respectively. Nearly 40% of PSA-PEG particles were classified as undergoing diffusive transport, compared to < 6% for latex particles (Fig. 3B). The penetration speeds achieved by PSA-PEG particles are similar to those we reported for approximately 200 nm non-degradable PEG-coated latex beads, which on average are slowed approximately 66-fold in CF sputum compared to in water (16). The particle tracking measurements at short time scales are also in good agreement with the observed motions of biodegradable PSA-PEG nanoparticles monitored over long time scales using confocal microscopy, which showed that PSA-PEG particles had an effective diffusivity of 0.023 μm²/s over 40 min (25). By using the Fick diffusion model and a sputum layer thickness of approximately 10 μm (the thickness of sputum is decreased due to dehydration of airway mucus secretions and collapse of the periciliary layer from the imbalanced ion transport in CF) (26), we again expect significant fractions of biodegradable PSA-PEG nanoparticles to be capable of penetrating CF sputum (31% after 30 min) compared to unmodified latex particles that are strongly immobilized (0.6% after 30 min) (Fig. 3C).

Fig. 3.

Transport of poly(sebacic acid)-co-poly(ethylene glycol) (PSA-PEG) and polystyrene (PS) nanoparticles in human cystic fibrosis sputum (CFS). (A) Averaged geometric mean-squared displacements (<MSD>) of nanoparticles in CFS as a function of time scale. (B) Transport mode of particles in CFS. (C) The fraction of particles predicted to penetrate a 10 μm thick CFS layer over time using Fick's second law and diffusion coefficients obtained from tracking experiments.

Discussion

PEG coatings have been widely used in the development of polymeric drug carriers, including particles composed of biodegradable polyesters (27, 28) and polyanhydrides (29–32). PEG coatings reduce aggregation and enhance the blood circulation times of biodegradable nanoparticles designed for drug delivery by reducing particle detection by the reticuloendothelial system (i.e., “stealth nanoparticles”) (33–35). However, PEG coatings have not typically provided the same “stealth-like” advantages to biodegradable nanoparticles upon their exposure to mucus barriers. Highly viscoelastic mucus secretions usually adhere to, wrap up, and efficiently immobilize nanoparticles, including some previous attempts with biodegradable PEG-coated systems (2, 7). Since particles that are immobilized by mucus barriers are typically efficiently cleared from the mucosal tissue, the design of biodegradable mucus-penetrating nanoparticle systems capable of sustained drug or nucleic acid delivery has become a priority for the improved treatment of numerous diseases that affect mucosal tissues (2, 36).

The rapid mucus penetration by biodegradable PSA-PEG nanoparticles demonstrated here, as directly compared to particles composed of conventional PSA and PLGA polymers, is likely due to a dense surface coating of low MW PEG similar to that achieved with mucus-penetrating non-degradable latex beads (13, 14). It is noteworthy that particles composed of random copolymers of PSA and PEG, the subject of previous reports by our group (37, 38), were strongly immobilized in human mucus, likely due to insufficient freedom of the PEG chains to partition to the particle surface when they are flanked on both ends by PSA segments (i.e., PSA-PEG-PSA random copolymers versus the PSA-PEG diblock copolymers used here). Recently, PEGylated PLGA nanoparticles, formed by adsorbing avidin-palmitic acid to the surface of PLGA nanoparticles and subsequently attaching biotin-PEG, exhibited a 2- to 10-fold improvement in particle penetration speed in ovulatory endo-cervical mucus (OCM) compared to uncoated particles (15). In contrast to this modest increase in penetration speed in OCM, we observed a 400-fold increase in the penetration rates of our PEG-coated biodegradable particles in CVM; this difference probably results not only from differences in particle surface properties, but also differences between OCM and CVM. The fresh, undiluted CVM used in our studies has bulk rheological properties comparable to many other mucus secretions, and is approximately 1,800-fold more viscous than water under low shear. In contrast, the viscosity of OCM is as much as 100-fold lower than that of CVM (3). The increased hydration of OCM also likely increases the pore size of the mucin mesh structure compared to mucus secretions at other mucosal tissues or during non-ovulatory periods. This is consistent with the rapid migration of sperm and motile bacteria across OCM but not CVM (39, 40). Hence, OCM may not limit nanoparticle penetration to the same extent as other mucus fluids, and PEGylation can only offer limited improvement in the transport of particles that are already not slowed substantially by OCM. In agreement with this explanation, uncoated PLGA particles were slowed only 8- to 13-fold in OCM compared to their speeds in water (15), whereas we found CVM to retard the transport of uncoated PLGA particles by over 3,300-fold relative to in water. Another difference between the two works relates to the density of PEG coatings achieved. The post-formulation coating of biotin-PEG on PLGA is limited by the number of avidin molecules physically adsorbed on the nanoparticle surface, which was estimated at approximately 420 molecules of avidin per 170 nm particle (15). While it is impossible to precisely determine the number of PEG molecules on the surface of the biodegradable PSA-PEG particles used here, PSA-PEG nanoparticles penetrated CVM at speeds comparable to PEGylated latex beads (13, 14) that have in excess of 100,000 molecules of PEG per 200 nm particle. We recently found that as little as a 40% reduction in PEG density (≈60,000 PEG molecules per 200 nm latex particle) can reduce nanoparticle transport rates by nearly 1,000-fold in CVM (14). Thus, PSA-PEG nanoparticles are likely to have close to 100,000 PEG molecules per 200 nm particle, in contrast to <1,000 for particles PEGylated by attaching biotin-PEG to avidin preadsorbed to the PLGA particle surface. Our results suggest that the dense PEG coatings of PSA-PEG particles are critical for rapid nanoparticle penetration of highly viscoelastic mucus secretions, such as CVM and CF sputum.

In addition to rapid penetration of mucus fluids, important attributes of a desirable polymeric platform for mucosal drug delivery include efficient drug loading and sustained release upon exposure to water (41). We sought to address these requirements by using PSA, a member of the polyanhydride family that offers well-controlled continuous release kinetics (with little to no initial burst) for many drugs (42), as the core material of the particles. PSA-PEG also provided efficient encapsulation and sustained release of etoposide (43), gemcitabine, doxorubicin and paclitaxel, in good agreement with our previous findings that copolymers of PEG and PSA offered sustained release over several days of compounds as small as rhodamine (M_w ≈443 Da) to as large as DNA (M_w ≈5.1 × 10⁶ Da) (44).

We have previously produced nanoparticles that penetrate mucus. In 2004, Dawson et al. showed that biodegradable PLGA nanoparticles coated with DNA moved approximately 10-fold faster in reconstituted bovine gastric mucus than similar-sized latex particles (45). These particles were also shown to provide gene delivery in vitro that was at least 50-fold more efficient than naked DNA. However, such a system is likely limited to gene delivery applications since coating nanoparticles with DNA would not be generally adopted for drug delivery applications due to cost and potential immunostimulatory issues. We also showed, in 2005, that PEG coatings on various types of nanoparticles, including biodegradable particles, offered significant improvements in mucus penetration rates compared to uncoated particles (25), but the improvements were modest (<20-fold) likely due to insufficient PEG densities. Recently, we reported that a dense coating of low MW PEG on non-degradable latex particles as large as 500 nm facilitated their transport across fresh, undiluted cervicovaginal mucus at speeds up to only 4- and 6-fold lower than those of the same particles in pure water (13). This finding not only demonstrated that synthetic systems can be engineered to penetrate physiological human mucus at speeds approaching their rates in water, but also offered the prospect that large nanoparticles can be used for mucosal drug delivery. In a follow-up report, we confirmed that a high coating density and low MW of PEG are both necessary to formulate mucus-penetrating particles, which likely accounts for the previous failures to engineer such particles using PEG coatings (14). Subsequently, Cu et al. developed PEG-coated PLGA particles that exhibited a 2- to 10-fold improvement in transport compared to uncoated particles in OCM (15). In this paper, a biodegradable polymer nanoparticle system has been made that is capable of efficiently penetrating CVM and CF sputum, of which the latter is perhaps the most viscous and elastic human mucus secretion.

Conclusions

The inability for conventional therapeutic particles to penetrate the mucus barrier has precluded sustained drug delivery at mucosal surfaces for more than a few seconds to several hours (2, 36). We have developed biodegradable PSA-PEG particles that penetrate fresh, undiluted, non-ovulatory human mucus obtained from the reproductive tracts of healthy women, as well as sputum expectorated from the lungs of CF patients. In contrast to our previous reports on mucus-penetrating particles based on non-degradable latex beads, biodegradable PSA-PEG nanoparticles can be prepared using a simple one-step formulation process. The development of controlled-release PSA-PEG nanoparticles as a mucus-penetrating drug delivery platform is expected to improve drug therapies at various mucosal surfaces, including in the lung airways, sinuses, eyes, gastrointestinal tract, and female reproductive tract.

Materials and Methods

Polymer Synthesis and Particle Formulation.

Sebacic anhydride (Acyl-SA) was prepared as previously described (38). Polyethylene glycol methyl ether (M_W 5,000 Da, mPEG, Sigma) was dried under vacuum to constant weight. The diblock poly(ethylene glycol)-co-poly(sebacic acid) (PSA-PEG), and the homopolymer poly(sebacic acid) (PSA) were synthesized by melt polycondensation of Acyl-SA and mPEG or Acyl-SA alone, respectively. Briefly, predefined ratios of Acyl-SA and mPEG were melted (180 °C) in a round bottom flask before applying high vacuum for 30 min. Acetic anhydride byproduct was collected in a liquid nitrogen trap. Nitrogen gas was swept into the flask after initial melting 15 min into the polymerization to mix the melt. Polymers were cooled to ambient temperature, dissolved with chloroform, and precipitated into excess petroleum ether. Precipitate was collected by filtration and dried under vacuum to constant weight. Polymer structure was verified using ¹H NMR (Bruker Advance 400 MHz spectrometer) and polymer molecular weight was determined by gel permeation chromatography (JASCO PU-980 intelligent HPLC pump, 1560 intelligent column thermoset and RI-1530 intelligent refractive index detector). Samples were filtered and eluted in chloroform through a series of Styragel columns (guard, HR4 and HR3 Styragel columns, Waters). PSA and PSA-PEG were fluorescently labeled by dissolving the polymers in tetrahydrofuran and incubating with doxorubicin (NetQem) in DMSO (Sigma) for 30 min. To form nanoparticles, the polymer solution (5 mg/mL in THF) was added dropwise into 40 mL water where they spontaneously formed nanoparticles and stirred for 2 h to remove solvent. PLGA was purchased from Alkermes, fluorescently labeled as previously described (46), and formulated into nanoparticles using the same process. Particles were collected, washed twice, and resuspended. Size and ζ-potential were determined by dynamic light scattering and laser Doppler anemometry, respectively, using a Zetasizer Nano ZS90 (Malvern Instruments), as previously described (13). PEG coating on the particle surface was visually confirmed by preparing PSA-PEG with biotin functionalized PEG, to which rhodamine-labeled NeutrAvidin (Invitrogen) was attached. Particle images were obtained using a fluorescent microscope. Particle morphology was evaluated by scanning electron microscopy (SEM) using a cold cathode field emission SEM (JEOL JSM-6700F). Particles were lyophilized, resuspended in 100% ethanol and dried onto SEM mounts at room temperature. Particles were then sputter coated with platinum (300 s at 1500 amps, Quorum Technologies) before microscopy.

Collection of Human Cervicovaginal Mucus and Cystic Fibrosis Sputum.

Cervicovaginal mucus was collected and prepared as previously described (13, 47). Briefly, undiluted cervicovaginal secretions from women with normal vaginal flora were obtained using a self-sampling menstrual collection device following a protocol approved by the Institutional Review Board of the Johns Hopkins University. The device was inserted into the vagina for 60 s, removed, and placed into a 50 mL centrifuge tube. Samples were centrifuged at 220 × g for 2 min to collect the secretions. Cystic fibrosis sputum was collected as previously described (1, 16). Respiratory sputum was spontaneously expectorated from male and female CF patients aged 18–21 and collected following a Johns Hopkins Medicine Institutional Review Board approved protocol. Samples were transported on ice from the hospital to the laboratory for same day tracking experiments. Sample collection was performed carefully to avoid contamination with salivary enzymes.

Multiple Particle Tracking in CV Mucus and CF Sputum.

Particle trajectories were recorded as previously described (13). Briefly, particles were added to mucus or sputum samples at 3% vol/vol and incubated for 2 h at 37 °C before microscopy. Twenty second movies were captured at a temporal resolution of 66.7 ms using a silicon-intensified target camera (VE-1000, Dage-MTI) mounted on an inverted epifluorescence microscope equipped with 100× oil-immersion objective (N.A., 1.3). Trajectories for n > 150 particles per experiment were extracted using MetaMorph software (Universal Imaging). Mean-squared displacements (MSD), effective diffusivities (D_eff), and particle classifications were determined as previously described (13, 45). Three experiments were performed for each condition. A one tailed, unequal variance Student's t-test was used to evaluate significance (P < 0.05).

Modeling Particle Penetration of a Mucus Slab.

The mucus slab was modeled by a constant concentration of particles at one side (C_{x = 0,t} = 1) and no particles in the mucus slab initially (C_{x = 0,t = 0} = 1, C_{0<x≤ΔL,t = 0} = 0). The concentration at the other end of the slab (x = ΔL) was determined by numerical integration (Matlab) of Fick's second law using the effective diffusivity for each particle:

The effective diffusivity was obtained from experimental data for each particle. Briefly, the time scale independent parameters for each individual particle (α and D₀) were obtained by fitting MSD to Eq. 2.

Using α and D₀, the time (τ) required for a particle to penetrate a given mucus thickness was calculated. The effective diffusivity for each particle was then calculated from Eq. 3:

The concentration profiles represent an arithmetic mean of the individual profiles for each particle type.