Lipidated brush-PEG polymers as low molecular weight pulmonary drug delivery platforms

ABSTRACT Objectives Nanomedicines are being actively developed as inhalable drug delivery systems. However, there is a distinct utility in developing smaller polymeric systems that can bind albumin in the lungs. We therefore examined the pulmonary pharmacokinetic behavior of a series of lipidated brush-PEG (5 kDa) polymers conjugated to 1C2, 1C12 lipid or 2C12 lipids. Methods The pulmonary pharmacokinetics, patterns of lung clearance and safety of polymers were examined in rats. Permeability through monolayers of primary human alveolar epithelia, small airway epithelia and lung microvascular endothelium were also investigated, along with lung mucus penetration and cell uptake. Results Polymers showed similar pulmonary pharmacokinetic behavior and patterns of lung clearance, irrespective of lipid molecular weight and albumin binding capacity, with up to 30% of the dose absorbed from the lungs over 24 h. 1C12-PEG showed the greatest safety in the lungs. Based on its larger size, 2C12-PEG also showed the lowest mucus and cell membrane permeability of the three polymers. While albumin had no significant effect on membrane transport, the cell uptake of C12-conjugated PEGs were increased in alveolar epithelial cells. Conclusion Lipidated brush-PEG polymers composed of 1C12 lipid may provide a useful and novel alternative to large nanomaterials as inhalable drug delivery systems.


Introduction
Diseases of the respiratory system pose a significant worldwide health burden [1,2].Inhaled administration of medications provides the most effective way of specifically targeting and treating respiratory diseases.However, typically rapid absorption of small molecule drugs from the lungs results in transient lung exposure and potentially high systemic adverse effects [3][4][5].This has prompted a significant worldwide interest in inhalable drug delivery systems that can better control drug exposure in the lungs over a more prolonged period of time [4,6].Nano-sized drug delivery systems, such as liposomes, have been most extensively explored as inhalable drug delivery systems [7][8][9].Nanosized materials, however, tend to reside, and release drug, entirely in the air spaces of the lungs [7].They can also stimulate local resolving inflammatory reactions that are exacerbated by extended lung residence [7,8].There is merit however, in using smaller polymeric drug delivery systems that still prolong drug exposure in the lungs, but which can better penetrate into lung endothelial cells and the interstitium [10,11].
The use of polymers based on > 5 kDa poly(ethylene glycol) (PEG), for example, can extend the residence time of conjugated drugs by up to 7 days in the lungs [10,12].On the other hand, smaller molecular weight PEGs (<2 kDa) prolong drug exposure up to 2 days in rodent models [12].Polymers based on polyethyleneimine, hyaluronic acid and chitosan have also been explored as inhalable drug delivery systems, but these tend to be larger than 5 kDa [10,11].
More recently, there has been increased interest in coupling lipid elements with hydrophilic polymer backbones to further control in vitro and in vivo cell and membrane trafficking [13][14][15][16].Lipid conjugation is thought to facilitate the integration of the polymer into natural lipid trafficking pathways -including binding to serum albumins and lipoproteins [13][14][15][16][17].This has been shown to enhance oral bioavailability, prolong systemic residence and increase lymphatic exposure, which are critical for the development of vaccines, immunotherapies and chemotherapy agents [13][14][15][16][17].This occurs as a result of the lipid element adsorbing into hydrophobic binding pockets of albumin or lipoproteins that utilize receptor-mediated active transport mechanisms such as endocytosis, transcytosis, and exocytosis to penetrate cell and tissue membranes [13][14][15]18].Access to these pathways provides opportunities for specific intracellular, tissue, and organ targeting as demonstrated for multiple systems [13][14][15]18].Further, recent work has shown that the development of inhalable drug delivery systems that bind endogenous albumin in the lungs may facilitate improved trafficking through lung epithelial membranes and access to the lung interstitium and lymphatics [19].This has the potential to better improve drug exposure and activity at interstitial sites of disease pathogenesis compared with the more common approach of relying on drug liberation in the air spaces of the lungs [19].
To this end, we recently characterized how the lipid structure on lipid-terminated brush PEG polymers (5 kDa) dictates the ability of these systems to hijack natural lipid trafficking pathways in vitro and in vivo [20].The results showed that the lipid terminal group of a brush PEG polymer dictates its pharmacokinetic and biodistribution properties after intravenous and subcutaneous administration.In this work, the lipid element (C2, C 12 , 2×C 12 , 2×C 18 or cholesterol) was found to drive the differential and specific interaction of the polymers across all natural lipid trafficking pathways in the body, including albumin and innate lipoproteins such as low-density lipoproteins (LDL), high-density lipoproteins (HDL), and chylomicrons.Notably, albumin binding increased with increasing lipid molecular weight up to 2C 12 , but showed a decrease in albumin binding when increasing molecular weight further (2C 18 ).
The present study therefore sought to describe the pulmonary pharmacokinetics, biodistribution and lung membrane trafficking of 5 kDa brush PEG polymers (Figure 1) conjugated with a single C 2 or C 12 lipid (1C 2 , 1C 12 ), or two C 12 (2C 12 ) lipids in rats.It was hypothesized that 2C 12 -PEG, and to a limited extent 1C 12 -PEG, would show increased permeability through pulmonary membranes and absorption from the lungs of rats when compared to 1C 2 -PEG by virtue of their ability to bind albumin.These lipid-modified PEG constructs were specifically investigated here since they previously show a broad range of albumin binding affinities and intravenous pharmacokinetics in rats [20].The higher molecular weight lipid (2C 18 ) was not investigated since this showed weaker albumin binding compared to 2C 12 in the initial work.An additional objective of this work was to investigate patterns of lung retention and clearance mechanisms to understand how the lipid load (and by extension, affinity for endogenous albumin) affects the in vivo behavior of these polymers in the lungs.Since the pulmonary delivery of nano-sized materials and proteins is known to commonly induce mild inflammatory changes in lungs, we also evaluated and compared the immunological response of lungs to the polymers 24 h after a single dose.To differentiate the ability of the polymers to penetrate through key membrane barriers in the lungs, their penetration kinetics through artificial lung mucus and primary in vitro models of human small airway epithelial, alveolar epithelial and microvascular endothelial membranes were examined in the presence and absence of albumin.

Article highlights
• Lipidated polymers containing 5 kDa brush-PEG showed similar pulmonary pharmacokinetic profiles and patterns of lung clearance, irrespective of the molecular weight of the attached lipids and previously established albumin binding affinities.• All polymers exhibited prolonged lung exposure after pulmonary administration, but the polymer containing a single C 12 residue showed the best safety profile in the lungs over 24 h.• The permeability of polymers through artificial lung mucus and airliquid interface monolayers of primary human small airway epithelial and alveolar epithelial cells were only subtly reduced with increased lipid MW.There were no differences in transport through monolayers of primary human lung microvascular endothelia.The presence of human serum albumin had no impact on membrane permeability.• Polymers showed increased cell uptake with increased lipid MW, with cytoplasmic appearance being both diffuse in nature and showing evidence of punctate bodies.• Cell uptake was most pronounced for the 1C 12 and 2C 12 PEG polymers in alveolar epithelial cells that are known to overexpress several albumin transporters.The presence of human serum albumin significantly enhanced the uptake of these polymers in this cell line, suggesting that cell internalization, but not penetration, was enhanced by albumin binding over the 3 h incubation.
The synthesis and characterization of the unlabeled Cy5labeled materials polymers (<7 kDa) were described previously [20].The synthesis and characterization of the 3 H-labeled materials used here for in vivo pharmacokinetic experimentation, however, are described in detail in the Supporting Information along with final characterization details for the Cy5-labeled materials.Briefly, lipidated polymers were prepared with an azlactone functional group that was used as a handle to conjugate 3 H-ethanolamine (see Figure 1).The polymers were characterized via gel permeation chromatography (GPC), proton nuclear magnetic resonance spectroscopy ( 1 H NMR), attenuated total reflectance Fourier transformed infrared spectroscopy (ATR FTIR) and scintillation counting to determine the specific radioactivity of the polymers (approximately 3 μCi/mg).

Cell culture
Primary cells were sourced and cultured as described previously [19].Briefly, Lonza human primary small airway epithelial cells (SAEC, CC-2547S) and human lung microvascular endothelial cells (HMVEC-L, CC-2527) were purchased from Capsugel (NSW, Australia).Immortalized human alveolar epithelial cells (hAELVi) were purchased from InSCREENeX GmbH (Braunschweig, Germany).Plasticware and Transwells used for the culture of hAELVi cells were coated with InSCREENeX huAEC coating solution, while Transwells used for the culture of SAEC were coated with Roche rat tail collagen (0.3 mg/ml; Sigma-Aldrich).Cells were maintained at 37°C in a humid atmosphere of 5% CO 2 .

Animals
Male Sprague Dawley rats (7-8 weeks of age at surgery, 256-285 g) were obtained from the Animal Resources Centre (WA, Australia) or the Herston Medical Research Centre (QLD, Australia).Rats were acclimatized for at least 1 week in open top cages, followed by a further acclimatization period of 1 day in individual metabolism cages prior to surgery.Rats were fed standard rodent pellets and water ad libitum, except for after surgery until 4 h after dosing, during which time food alone was withheld.Rats were maintained at 20-22°C on a 12 h light dark cycle.After surgery, rats remained in metabolism cages until the end of the study.All experiments involving rats were approved by the University of Queensland Animal Ethics Committee and were conducted in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes.

Intravenous and pulmonary pharmacokinetics and biodistribution of 3 H-labeled polymers in rats
Rats were cannulated under isoflurane anesthesia via the right carotid artery and jugular vein to facilitate blood sampling and IV dosing, respectively.Local anesthesia at incision sites and post-surgical analgesia was provided via delivery of 0.13 mg bupivacaine per incision site and 0.2 mg/kg meloxicam subcutaneously prior to surgery.Following surgical cannulation, rats were attached to a Velcro harness-swivel assembly to allow IV dosing and blood sampling from freely moving animals, before rats were returned to metabolism cages.Rats were allowed to recover in metabolism cages overnight prior to dosing.Prior to dosing, samples of blank blood (200 μl), urine and feces (collected overnight) were obtained from all rats to provide background 3 H correction for scintillation counting.Blood samples were collected into heparinized (10 IU) eppendorf tubes and centrifuged at 3,500×g for 5 min to obtain plasma.
Rats assigned to the IV dosing groups (n = 3 per group) were administered 1 mg/kg 3 H-labeled polymer in 1 ml sterile saline via the jugular vein cannula over 1.5 min, followed by a 30 s infusion of heparinized saline to flush the remaining dose into the rat.An initial (t0) blood sample was then collected as above.Additional blood samples were collected at times 0.17, 0.5, 1, 2, 4, 8 and 24 h after dosing for all polymers, and at 30 and 48 h for 2C 12 -PEG.Rats were euthanized via intracardiac injection of lethabarb under isoflurane anesthesia after the last blood sample was collected, and major organs (liver, kidneys, lungs, spleen, heart, brain) removed and frozen for later analysis of organ biodistribution.
Blank blood and urine were similarly collected from rats assigned to the pulmonary pharmacokinetics groups as described above.Rats in these groups (n = 4-5 per group) were initially administered 1 ml of a 10% charcoal solution in water via oral gavage 30 min prior to delivery of the pulmonary dose.The purpose of the oral charcoal solution was to prevent the gastrointestinal absorption of polymer cleared from the lungs (and swallowed) via the mucociliary escalator.This allowed us to determine the proportion of the dose removed from the lungs via the mucociliary escalator and allow the establishment of a plasma profile that more reliably reflected absorption of polymer from the lungs.The 30 min gap between charcoal and polymer dosing also allowed sufficient time for rats to swallow any charcoal present in their mouths and reduce the amount of mucus built up around the vocal cords to facilitate easier cannula insertion past the vocal cords. 3H-polymer (1 mg/kg) (or sterile saline control) was then dosed to the lungs of isoflurane anaesthetized rats via intratracheal instillation in 100 μl sterile saline as described previously [8,21].Rats dosed with sterile saline did not undergo microvessel cannulation or blood sampling but were administered oral charcoal and pulmonary saline vehicle.This was to provide 'blank' feces for feces biodistribution evaluation in pulmonary polymer-dosed rats and to provide 'non-polymer dosed controls' for BALF cytokine and differential immune cell evaluation.Serial blood samples (200 μl) were then collected at 0.17, 0.5, 1, 2, 4, 6, 8 and 24 h after polymer dosing.After the last (24 h) blood sample was collected, rats were anaesthetized under isoflurane, and the entire blood volume was removed via the carotid artery cannula.A 5 cm length of polyethylene tubing (1.70 mm × 1.20 mm) was then inserted into the trachea to collect BALF as previously described [22].The lungs were then flushed with 4 × 5 ml volumes of cold 1× PBS containing 0.1 mM EDTA and 0.1% BSA.The first two and the last two washes were separately pooled on ice to allow for quantification of inflammatory cytokines in the first 2 washes (to prevent excessive dilution), and analysis of the proportion of polymer remaining in the BALF plus differential immune cell counting from all 4 washes as described previously [22].All BALF samples were then centrifuged at 350×g for 5 min at 4°C to pellet cells.BALF samples (approx.1.1 ml) from each pooled fraction were collected for scintillation counting and size exclusion analysis as described below.Remaining BALF supernatant from the first 2 flushes was then frozen in aliquots for later quantification of cytokines.Cell pellets from all 4 washes were pooled, resuspended in 1.5 ml cold wash buffer and counted on a Tali cell counter before being analyzed for the total number of alveolar macrophages, T-lymphocytes and granulocytes as described below.Major organs (liver, kidneys, lung tissue, spleen, heart, brain) were also removed and frozen for later analysis of organ biodistribution as above for IV dosed rats.
Total urine (collected at times 0-8 and 8-24/48 h post dose) and feces (pooled) excreted from all IV and pulmonary dosed rats (including the pulmonary saline controls) were collected and analyzed as described below.
In addition, a separate cohort of rats (n = 4 per group) which did not undergo surgical cannulation were dosed with 1 mg/kg 3 H-polymer via the lungs as above and terminated under isoflurane anesthesia 6 h after dosing to provide information on the relative time course over which polymers are cleared from the lungs.BALF was collected from the lungs and pooled to calculate the proportion of the dose remaining in lung lining fluid before lung tissue was removed for biodistribution analysis.BALF obtained from this group of rats was not used to evaluate cytokine concentration or differential immune cell counts.

Quantification of 3 H in biological samples
Plasma (50 μl) and urine (50-100 μl) samples were mixed with 1 ml Ultima Gold in 6 ml scintillation vials and counted on a Packard Tri-carb scintillation counter.Samples of cell-free BALF (1 ml) were mixed with 10 ml Ultima Gold in 20 ml scintillation vials and similarly analyzed on a scintillation counter.
The 3 H content of major organs collected were processed and analyzed as previously described after initial dissociation in MQ water [23].Briefly, dissociated tissue samples (50-100 mg) were solubilized in Solvable overnight, followed by bleaching and addition of Ultima Gold scintillation cocktail before being analyzed via scintillation counting 4 days later.
Total feces excreted from each rat post-dose was analyzed as previously described [23], with modification.Briefly, total feces collected from each rat were solubilized in MQ water to form a homogenous slurry.Aliquots of this slurry (approx.10 ml) were dried in 20 ml scintillation vials at 60°C for several days.Samples of dry feces (20 mg) were then weighed in duplicate into 20 ml scintillation vials, resolubilized in 1 ml MQ water and then vortex mixed with 2 ml Soluene and 2 ml isopropyl alcohol.Samples were heated at 60°C overnight before being cooled to room temperature and bleached with 400 μl 30% hydrogen peroxide.IrgaSafe (12 ml) was then added to each sample, vortex mixed and samples left at room temperature in the dark for at least 4 days before being analyzed on a scintillation counter.The counting efficiency of feces samples were determined via addition of a known amount of 3 H-polymer to samples and quantifying 3 H recovery.Further, the background counts in each feces sample were subtracted by analyzing samples of 'blank' pre-dose feces or feces collected from saline dosed rats that had also been orally dosed with charcoal.Specifically, the background counts in feces samples from IV dosed rats were determined by analyzing feces collected from each animal prior to dosing.This provides the most accurate background for each animal.However, since charcoal was orally administered to pulmonary dosed rats and can interfere with scintillation counting, feces from rats dosed with saline via the lungs and given oral charcoal (as described above) were used to provide background correction for rats administered polymer via the lungs.

Size exclusion chromatography (SEC) analysis of 3 H-label in BALF and urine samples
BALF and urine (0-8 and 8-24 h) samples from all pulmonary dosed rats were analyzed via SEC on a Superdex 75 column to determine the 3 H species present in these samples and to give an indication of the relative stability of the 3 H-labeled polymers after pulmonary dosing. 3H counts in the plasma were too low for SEC analysis, but the SEC profiles in urine give a good representation of the 3 H species present in (and excreted from) plasma.The HPLC consisted of a Waters 1515 autoinjector, a 2707 isocratic pump and a Waters fraction collector (III).Mobile phase (pH 7 PBS with 0.3 M NaCl) was eluted at a rate of 0.5 ml/min.Briefly, 100 μl samples of BALF, 0-8 h urine, 8-24 h urine, polymer in mobile phase or 3 H-ethanolamine in mobile phase were injected onto the column.The eluate was collected into 0.5 ml fractions from times 10-50 min.Fractions were mixed with 1 ml Ultima Gold in 6 ml scintillation vials and analyzed on a scintillation counter.

Quantification of inflammatory cytokines and differential immune cell counts in BALF
The concentration of inflammatory cytokines TNFα, IL-1β and MCP-1 were quantified in the first 2 pooled BALF fractions from pulmonary dosed rats as described above to give an indication of the likely proinflammatory effects of the polymers in the lungs over the 24 h exposure period.Rats administered oral charcoal and pulmonary saline vehicle provided the control group.Positive controls for lung inflammation (such as lung administration of lipopolysaccharide or nickel oxide nanoparticles) were not generated since 1) we previously showed that these substances significantly increase the BALF expression of each of the 3 cytokines in Sprague Dawley rats dosed with these substances via the lungs [21] and therefore 2) this was not deemed necessary here in order to maintain compliance with animal welfare regulations.BALF samples were analyzed in duplicate using commercial ratspecific ELISA kits according to the manufacturers protocol.BALF samples were diluted to varying levels depending on the sensitivity of each kit (IL-1β [1:1], MCP-1 [1:25], TNFα [1:5]) and using the calibrator diluent provided.The concentration of each cytokine in each sample was determined by comparing optical density readings against a standard curve generated using the standards provided with each kit.Where individual samples were below the limit of quantification for the assay, cytokine concentrations were reported as the LOQ value (rather than 0).
The number of alveolar macrophages, granulocytes and T-lymphocytes in BALF from each rat was determined by multi color flow cytometry on a BD Fortessa Flow Analyzer using a slight modification to the method previously described by Haque et al. [21].Briefly, after initial quantification of the total number of cells in the BALF from each rat, aliquots containing 1 × 10 6 cells were suspended in 150 µl FACS buffer (1× PBS containing 1% BSA, 2 mM EDTA, pH 7.4).The remaining cells were used to stain control samples for each fluorescent label at the same cell density.Cells were initially incubated with 1.5 µl of anti-rat CD32 Fc receptor block (0.5 mg/ml) on ice for 10 min, followed by primary antibodies to T-lymphocytes (2 µl APC mouse anti-rat CD3; 0.2 mg/ ml), granulocytes (4 µL PE mouse anti-rat granulocytes; 0.2 mg/ml), and macrophages (5 µl rabbit monoclonal anti-F4 /80; 0.5 mg/ml) for 30 min at room temperature in the dark.All samples were then washed with 1 ml FACS buffer, centrifuged at 350 ×g for 5 min at 4°C and resuspended in 150 µl FACS buffer.Cells were then incubated with 2.5 µL goat antirabbit F(ab')2 FITC secondary antibody (1.5 mg/ml) for 30 min at room temperature in the dark and washed as described above before staining with 1.5 µl violet live/dead fixable dead cell stain for 30 min at room temperature in the dark.Cells were then washed and fixed with 150 µL formaldehyde (4% v/ v) for 30 min at 4°C.Control samples for the three fluorescent antibodies and live/dead stain were prepared using the single antibody and/or live/dead cell stain separately, with a 'dead' control sample prepared by incubating cells at 60°C for 5 min.Cells were then washed and suspended in 150 µl FACS buffer in BD trucount tubes and analyzed on a BD Fortessa Flow Analyzer on the same day.Flow cytometry data were analyzed using Flowjo v10.9.0.

Penetration of polymers through artificial lung mucus
Artificial lung mucus was prepared using a previously described method [24], and mucus properties were characterized via rheometry as described in the Supplementary Information.
The penetration of polymers through artificial mucus was investigated as described by Craparo [24].Briefly, a 0.28% w/v agarose solution was prepared in hot distilled water and sterilized by autoclaving.A 1 ml aliquot of agarose was then added to an empty 6 ml glass scintillation vial and left to sit at room temperature until solidified.A 1 ml aliquot of artificial mucus solution was then added on top of the agarose layer, followed by 100 µL of Cy5labeled polymer (1 mg/ml in PBS).Glass vials were then gently rotated on a stirring platform at room temperature for 0.5, 1, 2, 3, 4 or 6 h at which point the mucus and PBS layers were removed from the agarose (n = 3 vials per time point).Agarose layers were then melted, vortex mixed and a 100 μl aliquot added to individual wells of a black 96-well microplate.Polymer concentrations were quantified in a Tecan Spark plate reader at excitation/emission wavelengths of 646/5 and 664/5 nm against a standard curve of Cy5-polymer prepared in agarose (0.05-100 μg/ml).Data are presented as the cumulative % of polymer recovered in the agarose layer at each time point.

Permeability of Cy5-labeled polymers through lung monolayers of human small airway epithelial, alveolar epithelial, and microvascular endothelial cells
To determine the transport rates of Cy5-labeled polymers through alveolar, small airway and microvascular membranes in the lungs, we examined their apparent permeability through primary monolayers of human SAEC, HMVEC and hAELVi cells using a slight modification to a previously reported method [19].Cells were trypsinized and seeded onto the upper side of 6.5 mm Transwell permeable supports (0.4 µm pore size tissue culture treated polycarbonate membrane inserts, growth area 0.33 cm 2 ) at 50,000 cells in 0.1 ml medium.A further 0.6 ml of media was added to lower plate well.Prior to seeding cells, medium was added to Transwells to equilibrate them for 1 h in the incubator.In addition, Transwells for hAELVi cells were coated with huAEC coating solution for 2 h at 37°C, and Transwells for SAEC were coated with rat tail collagen I (0.03 mg/ml) for 45 min at 37°C.Coating solutions were then removed and Transwell membranes rinsed with PBS prior to seeding cells.Tissue culture medium in both the upper (apical) and lower (basolateral) chambers of the Transwells were changed every other day.After 3 days in culture, SAEC and hAELVi cells were air-lifted by removing media in the upper chamber of the Transwells and changing the media in the lower chamber to Lonza S-ALI differentiation medium.Cells were exposed to an air-liquid interface for a total of 7 days before being used for permeability assays.HMVECs were also cultured for 7 days prior to permeability assays.
The apical and basolateral chambers of the Transwells were prepared for permeability assays by removing the media and washing twice with prewarmed 20 mM HEPES-HBSS buffer (0.1 ml in the apical chamber and 0.6 ml in the basolateral chamber).After a 30 min equilibration period in a shaking incubator (37°C, 55 rpm), the buffer in the upper apical chambers of the Transwells were removed and replaced with 0.1 ml buffer containing 0.5 mg/ ml Cy5-labeled polymer (or FITC-dextran control), and then placed back into the shaking incubator.Samples of 0.1 ml were removed from the basolateral chamber and replaced with 0.1 ml prewarmed buffer at 15, 30, 60, 90, and 180 min.The removed samples were transferred to a black 96-well plate for analysis.Standard curves were constructed over the range of 0.1-100 µg/ml by serial dilution of Cy5-labeled polymers to quantify the levels of these compounds in the samples removed from the Transwells.Quantitation was performed using a Tecan Spark Multimode microplate reader with Ex 640/10 and Em 680/10 for detection of Cy5 and Ex 485/10 and Em 535/10 for detection of FITC-dextran.

Uptake of Cy5-labeled polymers into human small airway epithelial, alveolar epithelial, and microvascular endothelial cells
SAEC, hAELVi and HMVEC cells were seeded onto Millipore Millicell EZ 8-well slides (Merck, PEZGS0816) at 20,000 cells per well in 0.2 ml media and allowed to adhere overnight.Wells for hAELVi cells were coated with huAEC coating solution prior to seeding cells.After the overnight incubation, culture media was replaced with HEPES-HBSS containing 0.5 mg/ml Cy5-labeled polymers and cells incubated for 3 h to reflect permeability experiments.The HEPES-HBSS was removed and Hoechst 33,342 (1:2000 in HEPES-HBSS) was added to the cells for 5 min to stain the nuclei.The Hoechst solution was removed and the cells gently rinsed with HEPES-HBSS.Fluorescence microscopy was performed on live cells using an Olympus I×51inverted microscope to examine polymer cell uptake.

Cytotoxicity of PEG polymers against human small airway epithelial, alveolar epithelial, and microvascular endothelial cells
To examine whether any of the polymers adversely affected primary cell viability during the in vitro experiments described above, cell viability was examined after 24 h exposure in culture.SAEC, hAELVi and HMVEC cells were seeded at 5,000 cells per well in Greiner µCLEAR black 96-well plates (Sigma-Aldrich) and allowed to adhere overnight.Cells were then treated with serial dilutions of the different Cy5-labeled lipidated polymers at concentrations ranging from 0.06-1 mg/ml, or PBS vehicle for 24 h.Cell viability was assessed using a CyQUANT NF cell proliferation assay kit as described by the manufacturer.Cells were incubated at 37°C for 30 min with dye reagent and then fluorescence was measured in a BMG Labtech POLARstar Omega microplate reader at Ex 485/12 and Em 520.

Calculation of membrane permeability coefficients (P app )
P app values (in cm/sec) were calculated by (dC/dT x V R )/(A x C 0 ), where V R is the volume of the basal chamber, A is the monolayer surface area, C 0 is the initial concentration in the apical chamber at time 0, and dC.dT is the steady state linear rate of change in concentration in the basal chamber (in ug/sec).The rate of lipidated polymer transport was calculated by linear regression analysis.

Pharmacokinetic analysis and statistics
The concentration of polymer in plasma was determined by converting 3 H-counts to polymer mass per ml (ng/ml) using the specific activity of each polymer.Plasma pharmacokinetic parameters were therefore calculated with the caveat that the 3 H-label remains associated with the intact polymer.Elimination rate constants (k el ) were determined by regression analysis of at least 3 individual points in the elimination phases of the plasma concentration−time profiles.Plasma half-lives (t 1/2 ) were calculated as 0.693/k el .Area under the plasma concentration vs time profiles (AUC 0-t ) were calculated via the linear trapezoidal method and extrapolated to infinity to give AUC 0-∞ by dividing the last measured plasma concentration by k el .For IV data, initial volumes of distribution (V c ) were calculated by dose/initial concentration in plasma (Cp 0 ).Post-distributive volumes of distribution (V Dβ ) were determined by dividing the dose by (k el x AUC 0-∞ ).Plasma clearance was calculated by dividing dose by AUC 0-∞ .For pulmonary pharmacokinetic data, the apparent fraction absorbed from the lungs (F abs ) was determined by dividing the AUC after pulmonary administration by the AUC after IV delivery.Bioavailability (BA) was calculated as F abs x 100.Maximum plasma concentration (C max ) and time to maximum plasma concentration (T max ) were read from the individual curves.
Statistical analyses were conducted using GraphPad Prism, with statistical significance determined at a level of p < 0.05.Pharmacokinetic parameters and organ biodistribution were statistically compared between polymers via two-way ANOVA with Tukey's test for multiple comparisons.Concentrations of cytokines and differential immune cell counts in the BALF were compared between groups via one-way ANOVA with a Tukey's test for multiple comparisons.Mucus migration profiles were compared via simple two-way ANOVA.P app values between polymers in the presence and absence of HSA were statistically compared via two-way ANOVA with Tukey's multiple comparison test.

Intravenous and pulmonary pharmacokinetics and biodistribution of lipidated polymers in rats
The plasma concentration vs time profile and pharmacokinetic parameters for each polymer are shown in Figure 2 and Table 1, respectively.After IV administration, there were no discernible differences in the plasma pharmacokinetics between 1C 2 -PEG and 1C 12 -PEG, irrespective of the lipid chain length and albumin binding capacity.These polymers were cleared rapidly from plasma (Cl approx.13-16 ml/h) with a terminal half-life of approx.9 h, such that less than 1% of the dose remained in plasma after 24 h.By contrast, 2C 12 -PEG showed more prolonged plasma exposure and exhibited smaller volumes of distribution (V c 13.5 ml [p = 006 vs 1C 2 -PEG]; V Dβ 106 ml) compared to the smaller polymers.As a result, the plasma clearance of 2C 12 -PEG was significantly reduced by approx.3-fold compared to 1C 2 -PEG and 1C 12 -PEG (Cl 5.3 ml/h; p ≤ 0.0001).Approximately 50-65% of the IV dosed 3 H from each of the polymers was eliminated via the urine over the 24-48 h sampling period, suggesting relatively efficient renal elimination.By contrast, the amount of 3 H excreted in feces was below the limit of quantification.
Terminal biodistribution analysis after IV administration showed that the polymers exhibited limited distribution into the lungs, heart and brain, with less than 0.03% of the dose recovered in each organ after 24 h (1C 2 -PEG and 1C 12 -PEG) or 48 h (2C 12 -PEG) (Figure 3A and C).The polymers showed more avid biodistribution in the kidneys and liver with approximately 0.5% of the dose recovered in the kidneys and 0.4-2.2% of the dose in liver.Kidney biodistribution data exhibited a trend toward increasing retention with increasing polymer MW that was statistically significant (p < 0.0001).Interestingly, while 1C 2 -PEG and 2C 12 -PEG showed similar (not statistically different) biodistribution in the liver and spleen, with approximately 0.4 and less than 0.01% of the dose recovered in these organs respectively, 1C 12 -PEG was more avidly distributed toward these organs, exhibiting 5 to 10-fold higher uptake (p < 0.0001 in the liver; p = 0.0036 and 0.0143 compared to 1C 2 -PEG and 2C 12 -PEG respectively).
After pulmonary administration, plasma concentrations were consistently low, with C max concentrations around 1% of the Cp 0 after IV administration at the same dose (Figure 2, Table 1).In addition, with the exception of 1C 12 -PEG, the polymers did not show a clear absorption phase, but rather showed consistent plasma concentrations over the first several hours, followed by an elimination phase with a calculated halflife of between 16 and 31 h.The estimated pulmonary bioavailability based on plasma concentration data was 17, 10 and 4.5% for 1C 2 -PEG, 1C 12 -PEG and 2C 12 -PEG respectively, suggesting limited absorption form the lungs.However, the proportion of the pulmonary dose recovered in urine over the 24 h sampling period after pulmonary dosing was 21, 13 and 25% for 1C 2 -PEG, 1C 12 -PEG and 2C 12 -PEG respectively, suggesting more efficient pulmonary bioavailability compared to that calculated based on the plasma profiles.
The biodistribution of polymers after pulmonary administration are shown in Figure 3b, d and e.Consistent with the IV biodistribution data, polymers showed limited distribution into the heart, brain and spleen.Polymer distribution was most prominent in the kidneys (0.2 to 0.5% recovery) and liver (approx.0.2 to 0.3%) (Figure 3d).Interestingly, 1C 12 -PEG did not show the same prominence for uptake by the liver and spleen as after IV administration, possibly reflecting altered patterns of biodistribution for this polymer after pulmonary administration.
After 24 h, around 25% of the pulmonary dose of each polymer remained in the lungs, with approximately 20% recovered in the lung tissue and 5% in the BALF with no apparent differences in lung tissue or BALF distribution between the polymers (Figure 3e).Approximately 30% of the dose was recovered in the feces, suggesting that one third of the dose was cleared from the lungs via the mucociliary escalator.Altogether, the proportion of the pulmonary dose of all polymers accounted for in urine, feces and organs was high at 68-82%.
To give an indication of the rate of polymer clearance from the lungs, a separate group of rats were delivered a pulmonary dose of polymer and terminated 6 h later for analysis of 3 H remaining in the BALF and lung tissue.These data (Figure 4) show that while 25% of the pulmonary dose of all polymers remained in the lungs after 24 h, with the majority of the dose associated with lung tissue, 33% (1C 12 -PEG and 2C 12 -PEG) to 45% (1C 2 -PEG) of the dose remained in the lungs after 6 h.Approx.one third of the 3 H remaining in the lungs after 6 h was associated with lung tissue (approx.10-15% of the dose).This indicates that 1) the majority of the dose was cleared from the lungs within the first 6 h, 2) that the dose remaining in BALF after this time was gradually redistributed into lung tissue and 3) the polymers exhibited prolonged lung retention in rats.Further, the polymers showed relatively similar lung retention and clearance patterns, irrespective of MW or lipid component.

Speciation of 3 H in the BALF and urine of rats dosed via the lungs with lipidated polymers
The pulmonary pharmacokinetic data showed a discrepancy between the calculated bioavailability and the proportion of the dose excreted via the urine.This suggested either that the 3 H counts in plasma (which were close to the limit of quantification) were too low to accurately calculate bioavailability or that the polymers were degraded in the lungs to low MW species that exhibited different plasma pharmacokinetics.To determine what 3 H species were absorbed from the lungs and present in the lungs at termination, urine and BALF samples were analyzed via SEC (Figure 5).Plasma 3 H counts were too low to obtain an SEC profile.SEC profiles showed that for all polymers, the predominant 3 H species in terminal BALF was likely intact and protein-bound polymer (evidenced by the presence of higher MW material that eluted at around 20-23 min) (Figure 5).Likewise, the predominant species identified in urine was intact polymer, with only a small proportion of lower MW material identified in 8-24 h urine at an elution time of 34 min.The low MW material, however, was not liberated 3 H-ethanolamine which eluted at 41 min.Rather, it was likely a product of lipid hydrolysis since PEG is relatively stable in vivo.

Safety of lipidated polymers in the lungs of rats after a single pulmonary dose
Macromolecules inherently stimulate varying degrees of resolving inflammation via activation of the innate immune system in the lungs after inhaled exposure.The extent and persistence of this inflammation in the lungs is largely related to the physicochemical properties of the macromolecule and its residence time in the lungs.We investigated whether these lipidated polymers, and/or the oral administration of charcoal, stimulated an inflammatory response in the lungs after 24 h exposure by quantifying BALF concentrations of the proinflammatory cytokines TNF-α, MCP-1 and IL-1β together with differential immune cell counts.
The results show that while IL-1β concentrations in the BALF of rats administered the lipidated polymers appeared to be elevated compared to the saline control group, these differences were not significantly different due to the varied response of individual rat lungs to polymer exposure (Figure 6).Likewise, no significant differences in BALF TNF-α concentrations were found between groups, although two rats (out of four) administered 2C 12 -PEG had higher levels of cytokine compared to the other pulmonary dosed rats.This lack of a significant difference in the 2C 12 -PEG group appeared to be due to a single rat administered saline which had TNF-α concentrations that were approximately an order of magnitude higher than the rest of the control cohort, possibly as a result of aspiration of charcoal [25].No significant differences in BALF MCP-1 concentration were observed between groups.
PEG (p = 0.0135) but not 1C 12 -PEG.There were no significant differences in BALF T-lymphocyte counts between groups, but the number of macrophages in the BALF of rats administered 1C 2 -PEG were approximately 2-fold higher compared to rats in both the saline and 2C 12 -PEG groups (p = 0.0115 and 0.0008 respectively).

Permeability of lipidated polymers through artificial lung mucus and monolayer membranes of lung alveolar, small airway epithelial and microvascular cells
The pharmacokinetic data revealed that the lipidated polymers exhibit similar lung clearance kinetics and pathways, despite differences in lipid content.To better understand how the lipid content affects the membrane permeability of these polymers through key barriers in the lungs, we therefore evaluated their kinetics of permeability through artificial lung mucus and monolayers of primary human small airway epithelial, alveolar epithelial and lung microvascular endothelial cells.Unsurprisingly, the rate of 1C 2 -PEG migration through artificial lung mucus was slightly, but significantly, more rapid compared to 1C 12 -PEG and 2C 12 -PEG (Figure 7).There were no apparent differences in the rate of 1C 12 -PEG and 2C 12 -PEG migration through lung mucus.
Membrane permeability through cell monolayers were reported as P app values and are shown in Figure 8.In general, the permeability of polymers was lowest through hAELVi ALI monolayers which exhibit very tight junctions, approximately 2-fold higher through SAEC ALI monolayers and 10-fold higher through HMVEC liquid-liquid interface monolayers.In hAELVi ALI monolayers, permeability decreased with increasing lipid MW (p < 0.0001).In SAEC ALI monolayers, the permeability of 1C 2 -PEG was significantly (p = 0.0236) higher compared to 2C 12 -PEG, but strangely significantly lower (p = 0.0056) compared to 1C 12 -PEG.No significant differences in membrane permeability between polymers were observed in HMVEC monolayers.
In contrast to our expectations of enhanced membrane trafficking for 1C 12 -PEG and 1C 12 -PEG in the presence of albumin, the presence of HSA had no significant impact on the permeability of the polymers through cell membranes.
Importantly, none of the polymers were found to adversely affect primary cell viability at concentrations of up to 1 mg/ml, suggesting that the polymers did not exert any cellular toxicity during permeability experiments (see Supporting Information).

Uptake of lipidated polymers into lung alveolar, small airway epithelial and microvascular cells
Permeability data shown above suggested that the presence of albumin had no significant impact on the trafficking of these lipidated polymers through respiratory membranes.To further interrogate the cellular trafficking behavior of these polymers in the presence and absence of albumin, fluorescence microscopy was used to qualitatively examine uptake into the primary human cells used above (Figure 9).The data show that in general, polymer uptake into all cell lines increased with increasing lipid load, such that 1C 2 -PEG showed limited cell uptake, while 2C 12 -PEG exhibited extensive cell uptake.While patterns of cellular localization were largely diffuse in nature, distinct punctate bodies could also be seen in the cytoplasm.There did not appear to be any overt differences in the extent of polymer uptake between SAEC and HMVEC, but hAELVi cells clearly showed more prominent cell uptake, especially for 2C 12 -PEG.
An additional objective of this work was to examine whether these polymers could utilize endogenous albumin to enhance their trafficking through pulmonary cell membranes.For this reason, cell uptake was examined in the presence and absence of HSA.There did not appear to be any difference in the cell uptake of 1C 2 -PEG in the presence or absence of HSA in any of the cell lines examined.Images showed that the uptake of 1C 12 -PEG and 2C 12 -PEG was only slightly increased in the presence of HSA in SAECs and HMVECs.By contrast, the addition of HSA dramatically increased the uptake of 1C 12 -PEG and 2C 12 -PEG by hAELVi cells.This was most prominent for 2C 12 -PEG that was previously shown to bind albumin more avidly than 1C 12 -PEG [20].

Discussion
The use of functional materials to hijack natural lipid trafficking pathways has enormous potential for the controlled and targeted delivery of pharmaceuticals, vaccines, biological tools, and imaging agents for injected or inhaled delivery [26,27].Albumin transport processes are one of the most useful targets for enhancing the access of drugs to the lung parenchyma.This is because the lungs have well developed active transport mechanisms for albumin that are essential in maintaining fluid homeostasis and exhibit bidirectional transport across the respiratory epithelial barrier [19].However, while the lung interstitial access and systemic absorption of inhaled drug delivery systems may be improved by 'hitchhiking' albumin transporters after binding to endogenous albumin, it is not yet clear whether they can be efficiently engineered to take advantage of these pathways.Relatively low molecular weight polymeric drug delivery systems, however, may have greater potential to take advantage of endogenous albumin trafficking pathways in the lungs compared to nano-sized materials while also potentially exhibiting more favorable pulmonary pharmacokinetic behavior [19].Following on from recent work characterizing the albumin binding behavior and in vivo trafficking of lipidated brush-PEG polymers, this paper explored the potential utility of these materials as inhalable drug delivery systems both in vivo and in vitro [18,20].
To enable a complete description of the pulmonary pharmacokinetic behavior of these polymers, it was initially essential to evaluate their intravenous pharmacokinetics, biodistribution and elimination profiles.Previous work showed that Cy5-labeled 2C 12 -PEG exhibits more prolonged plasma exposure after IV administration in rats compared to 1C 2 -PEG and 1C 12 -PEG, although organ biodistribution and excretion profiles were not examined [20].Since macromolecules typically show slow and limited absorption from the lungs, Cy5 labeling was not deemed to be a sensitive enough approach for quantifying pulmonary pharmacokinetics in the present study.For this reason, tritium labeling was used to increase sensitivity and allow accurate quantification of polymer excretion profiles that cannot be obtained with fluorescent labeling.Since different labeling approaches can affect the pharmacokinetic behavior of macromolecules and the former study lacked biodistribution and excretion data, we initially re-examined the IV pharmacokinetics of the polymers.While plasma pharmacokinetic profiles were similar between the present and former study that employed Cy5-labeling, the Cy5-labeled polymers exhibited approximately 4-fold more rapid plasma clearance and higher volumes of distribution.However, fluorescent labels have the capacity to significantly enhance cellular internalization and alter distribution patterns of conjugated nanomaterials, and this may have been reflected here [28].Alternatively, the introduction of the H-label resulted in slight modifications of the polymer structure compared to the Cy5-labeled materials, including formation of amide linkages and the inclusion of hydroxy groups.Given that PEG is relatively inert, these small differences in polymer chemistry may have had an impact upon their pharmacokinetics.In general, the polymers exhibited limited distribution toward the lungs, heart, brain and spleen, but more avid retention in the kidneys and liver.Strangely, kidney retention increased with increasing lipid molecular weight.This may, however, have reflected slower renal elimination of polymers with a greater lipid load, despite overall molecular weight being well below the threshold for efficient glomerular filtration (approx.20 kDa) [29,30].Although, there was little difference in urinary elimination of the polymers (approx.50-65% excretion), 1C 12-PEG seemed to have an unusually higher affinity for uptake by the liver and spleen.This potentially suggests that this polymer more avidly targets macrophages.It is unknown, however, why increasing the lipid load by employing 2C 12 did not increase liver and spleen uptake compared to the 1C 12 polymer.Given that these are relatively small (<7kDa) hydrophilic polymers, however, they are expected to show intrinsically limited propensity for uptake by organs of the mononuclear phagocyte system.In addition, while rats administered 2C 12 -PEG intravenously were terminated 48 h after dosing, rats administered 1C 12 -PEG were terminated after 24 h.While it is possible that lipidated PEGs show initial distribution to the liver followed by elimination over time, leading to the observed difference in liver biodistribution, this is unlikely since plasma levels of 2C 12 -PEG were higher at termination compared to rats administered 1C 12 -PEG.
After pulmonary administration, all three polymers appeared to show limited systemic absorption based on plasma profiles and calculated apparent bioavailability (4.5% for 2C 12 -PEG to 17% for 1C 2 -PEG).The proportion of the pulmonary 3 H dose recovered in urine, however, was between 13% (1C 12 -PEG) and 25% (2C 12 -PEG), suggesting that bioavailability was much higher than that calculated based on the plasma profiles.The SEC profiles, however, showed that polymers appear to be absorbed intact from the lungs after pulmonary administration rather than as low molecular weight products of polymer degradation that could have explained this discrepancy.The fact that polymers were absorbed from the lungs intact and 13 to 25% of the 3 H dose was eliminated via the urine over 24 h suggests that the actual systemic bioavailability of the polymers may have been as high as approximately 30% (based on 50-65% urinary excretion over 24-48 h after IV administration).The discrepancy between calculated apparent bioavailability and the urinary excretion profiles may be due to non-linear pharmacokinetics or saturation of renal secretion processes at higher plasma concentrations.For example, plasma 3 H levels after pulmonary administration were approximately an order of magnitude lower compared to after IV administration.This may have facilitated the rapid  renal elimination of the polymers as they were absorbed from the lungs, in contrast to slower renal elimination at much higher plasma concentrations seen after IV administration.It is unlikely, however, that the discrepancy was due to saturation of plasma albumin binding.This is because albumin concentrations in plasma are very high (approx.40 mg/ml) and since saturation of albumin binding after IV administration would be expected to slow, not accelerate, renal elimination compared to after pulmonary administration.This was not, however, specifically examined in this study, but may be worth investigating further.
In general, restricted systemic access of polymers after pulmonary administration is consistent with a broad body of literature showing that the rate of absorption of large molecules from the lungs is limited [5,31,32].Endogenous macromolecules like albumin can penetrate through lung epithelial barriers via the transcellular route, with transport occurring via both cavaeolin-dependent (eg.gp60, megalin) and independent pathways [5,31].However, smaller macromolecules that cannot utilize transporters in the lungs (radius <5 nm, 4-40 kDa) have been reported to traverse the alveolar epithelial barrier mainly via paracellular pathways [33].
Lung retention data also showed that apart from initial rapid clearance from the lungs of rats during the first 6 h, likely as a result of extensive mucociliary elimination (approx.40%), the polymers exhibited relatively prolonged lung residence (approx.25% dose remaining in the lungs after 24 h).The polymers also showed few differences in patterns of lung retention or elimination.Altogether, these results suggest that pulmonary pharmacokinetics was largely driven by the 5 kDa PEG and that patterns of lung clearance were unaffected by the lipid size or load, irrespective of previously reported albumin binding affinities.This is in agreement with the results of others who have evaluated the lung retention times of PEGylated macromolecules in animals and humans [12,[34][35][36][37].
In vitro permeability data, however, showed that with the exception of HMVECs, P app values were roughly inversely correlated with overall polymer MW and cell monolayer TEER values.This suggested that the lipid MW had a small, but significant effect on cell penetration in monolayers with particularly tight intercellular junctions and high TEER values.While hAELVi and SAEC ALI monolayers exhibited a mucus layer that can hinder the access of macromolecules to underlying cells, our mucus penetration data suggested that this mucus layer was unlikely to have significantly affected permeability when compared to HMVECs.Further, lung retention data over 24 h showed that polymers were able to efficiently migrate through lung mucus in vivo and penetrate into the lung tissue.Interestingly, HSA also did not significantly increase permeability through cell monolayers over the 3 h exposure time, in contrast to our initial hypothesis.This suggests that in this model, whereby albumin was applied to both the apical and basolateral compartments, the polymers possibly trafficked preferentially through the cell membranes via the paracellular route.Mechanisms of membrane permeability and cell uptake were not explicitly interrogated in this study, however, since they were not important focus' of this study.It may be interesting, however, for future work to interrogate the precise mechanisms of cellular internalization and membrane trafficking for these polymers in order to better understand how they can be manipulated and utilized for in vivo drug delivery applications.
Interestingly, fluorescent microscopy images showed that not only are the larger 1C 12 and 2C 12 polymers avidly internalized by hAELVi cells, in drastic contrast to permeability data, but that cell uptake is increased by the presence of albumin based on qualitative fluorescent microscopy images.Importantly, alveolar epithelial cells are known to express high levels of albumin transporters and cavaeolin-dependent and independent mechanisms to aid in clearing protein from the airways and maintain fluid homeostasis in the lungs [38][39][40][41].This suggests that these polymers, which were previously shown to bind albumin, utilized albumin uptake processes in hAELVi cells to enter cells over the 3 h incubation period.Evidence to this effect was seen by the presence of punctate bodies typically seen after endocytic uptake, in addition to diffuse cytoplasmic staining, in all cell lines, but were more prominent in hAELVi cells.
This discrepancy between the permeability and cell imaging data may be explained in one of two ways.Firstly, the short time frame over which permeability was investigated (3 h) may not have been sufficient to allow albumin-bound polymer to actively traffic to a significant extent into cells and then back out into the basolateral compartment.We previously, however, showed that albumin is capable of trafficking through hAELVi ALI cell monolayers, albeit to a limited extent, over this time frame [19].Alternatively, the lack of a significant albumin concentration gradient between the apical and basolateral compartment may have restricted the driving force for albumin-bound polymer deposition on the basolateral side of the cell monolayer.In this model, albumin was applied on both sides of the membrane to avoid creating an osmotic gradient that may have artificially favored polymer penetration.In vivo, however, albumin that is translocated into the lung interstitium from the airways may be rapidly reabsorbed systemically via the lung vasculature or lymphatics, providing a greater driving force for the pulmonary interstitial access and systemic absorption of inhaled albumin-bound polymer.Regardless, these differences in cell uptake between the polymers did not translate into overt differences in pulmonary pharmacokinetics in the rat model.This is likely due to the complexity of macromolecule clearance processes in the lungs in vivo coupled with differences in the albumin binding affinity of these polymers between species [20].Some of the sources of this discrepancy, however, may be identified using in situ lung preparations, such as isolated perfused lungs, that eliminate contributing factors from the systemic circulation but maintain the relative complexity of the lung environment.To this end, this is one limitation of the monolayer permeability models used here which can usefully applies to understand trafficking through defined cellular membrane barriers, but do not recapitulate the complex cellular and biochemical environment of the lungs.
The inhaled administration of nanomaterials and macromolecules is well known to induce transient inflammatory reactions in the lungs, the extent of which are dictated in large part by lung retention time and physicochemical properties [7,8,21,22].Inflammatory effects in the lungs can also have a substantial impact on the mucociliary escalator, lung clearance mechanisms and kinetics [7,8,21,22,30,42].In addition, the cumulative administration of macromolecules over time could lead to increased lung burden and may prove a significant challenge in developing inhaled nanomedicines for chronic respiratory diseases [8,22].One objective of this work was therefore to explore the safety of these polymers in the lungs after a single 1 mg/kg dose in rats by evaluating the concentration of key inflammatory cytokines and immune cells in the BALF.In general, rats administered 1C 2 -PEG showed increases in total and immune cell counts in the BALF 24 h after pulmonary dosing and elevated, though not statistically significant, IL-1β.This was an unusual observation since this polymer shows limited cell uptake and no discernible albumin binding.Likewise, 2C 12 -PEG which shows the most significant cell uptake out of all three polymers and the most avid albumin binding affinity, also showed significantly increased levels of granulocytes and non-statistically significant elevations in IL-1β, MCP-1 and TNF-α.Elevations in these parameters, however, are commonly associated with shortlived resolving inflammatory reactions in the lungs after inhaled exposure to nanomaterials [7,8,21,22,30,42].Additionally, cytokine and immune cell counts were highly variable across most groups, including saline control rats.This was likely due to the aspiration of small amounts of charcoal in some rats, which may exacerbate inflammatory reactions of the lungs to the pulmonary delivered polymers [25].IL-1β is a pro-inflammatory cytokine produced by activated macrophages and can activate neutrophils and recruit lymphocytes.TNF-α is also released by macrophages in response to an inflammatory insult to stimulate the infiltration of immune cells.This suggests that 2C 12 -PEG appears to have the potential to activate alveolar macrophages in vivo.It is unknown, though, whether this is related to the polymer's propensity for uptake by respiratory cells.By contrast, 1C 12 -PEG showed no significant evidence of elevated cytokines or immune cell counts in the BALF.This was interesting given that this polymer also exhibited a greater propensity for uptake by the liver and spleen after IV administration, key organs of the mononuclear phagocyte system.The reason for this observation is currently unknown and further studies are required to explain the interaction of this polymer with macrophages and other immune cells.

Conclusion
The results of this study suggest that brush PEG polymers conjugated with lipids ranging in size from 1C 2 to 2C 12 exhibit prolonged lung retention after pulmonary administration, but also reasonable absorption from the lungs.The evidence suggests that 1C 12 and 2C 12 polymers may take advantage of albumin trafficking pathways in the lungs, particularly in the respiratory region which is rich in albumin transporters.Further work, however, is needed to interrogate whether differences in lung interstitial access occur for these polymers, and whether the apparent selectivity for uptake by alveolar epithelial cells may be exploited to enable specific cell targeting in the lungs.In addition, it may be useful to investigate the use of PEGs with different molecular weights and linearity/branching configurations to further interrogate the utility and flexibility of these polymers for pulmonary drug delivery applications.While rats, used here, are a good preclinical model to evaluate the pulmonary pharmacokinetics of new investigational materials, larger animal models (such as sheep) provide pulmonary pharmacokinetic data that more closely reflects inhaled delivery in humans.It would be useful, therefore, to evaluate the pulmonary pharmacokinetics of these polymers, particularly 1C 12 -PEG, in sheep which also offer the advantage of being able to investigate lung lymphatic uptake.Taken together with the in vivo toxicological data, the results of this study suggest that there is merit in further developing 1C 12 -PEG as a safe, and relatively low MW inhalable drug delivery system, but also in further exploring the flexibility and utility of lipidated PEG polymers for inhaled drug delivery applications.

Figure 1 .
Figure 1.Chemical structures of the lipidated polymers examined and location of 3 H (red) and Cy5 (blue) labels.N indicates number of repeating PEG units which, for these polymers, was between 8 and 9. Br was replaced with Cy5 in the fluorescent polymers.

Figure 2 .
Figure 2. Plasma concentration vs time profiles of lipidated polymers after IV and pulmonary dosing to rats.Plasma concentrations were normalized to a dose of 1 mg/kg.Data represent mean ± SD (n = 3-4).

Figure 3 .
Figure 3. Biodistribution of lipidated polymers 24 h after IV (panels a and c) or pulmonary (panels b, d and e) dosing to rats.Panels a and b show the mass normalized tissue biodistribution of the lipidated polymers (as % dose per gram of tissue).Panels c and d show the absolute proportion of the 3 H dose of each polymer recovered in each organ.Dose recovered in lungs after pulmonary administration represents dose recovered in lung tissue only after removal of the dose remaining in the epithelial lining fluid.Panel E shows the absolute proportion of the dose of each polymer recovered in the lung tissue, BALF, feces and urine of rats 24 h after pulmonary administration.BA represents the apparent bioavailability (F abs x 100) of polymers calculated over 24 h from the plasma concentration vs time data.Data represent mean ± SD (n = 4).Statistical analyses are only shown for mass normalized tissue biodistribution (% dose per gram of tissue) since results were comparable between mass normalized and whole organ biodistribution.*IV heart 1C 2 -PEG vs 1C 12 -PEG p = 0.0002; IV heart 1C 12 -PEG vs 2C 12 -PEG p = 0.0005; IV kidney 1C 2 -PEG vs 2C 12 -PEG p = 0.0183; IV spleen 1C 2 -PEG vs 1C 12 -PEG p = 0.0036; IV spleen 1C 12 -PEG vs 2C 12 -PEG p = 0.0134; IV liver 1C 2 -PEG vs 1C 12 -PEG, and 1C 12 -PEG vs 2C 12 -PEG p < 0.0001; pulm lung 1C 2 -PEG vs 2C 12 -PEG p = 0.012; pulm lung 1C 12 -PEG vs 2C 12 -PEG p = 0.0024; pulm heart 1C 2 -PEG vs 1C 12 -PEG p = 0.028; pulm kidney 1C 2 -PEG vs 1C 12 -PEG, and 1C 12 -PEG vs 2C 12 -PEG p < 0.0001.

Figure 4 .
Figure 4. Proportion of the dose of each lipidated polymer recovered in the lung tissue and BALF of rats 6 and 24 h after pulmonary dosing.Data represent mean ± SD (n = 4).

Figure 5 .
Figure 5. Size exclusion chromatography profiles of 3 H in the 24 h BALF and urine of rats dosed with each lipidated polymer.The top panels show the elution profile of each polymer and 3 H-ethanolamine (used to radiolabel polymers) in mobile phase.

Table 1 .
Plasma pharmacokinetic parameters of PEG polymers after IV or pulmonary administration to rats.Plasma concentrations used to calculate pharmacokinetic parameters were normalized to a dose of 1 mg/kg in all animals.Data are presented as mean ± s.d.(n = 3-4).BA: bioavailability.ND: not determined.BQ: below quantification.*Represents p < 0.05 compared to 1C2-PEG.# Represents p < 0.05 compared to 1C12-PEG.Individual p values are reported in the Supporting Information (Table