Preparation and Characterization of Block Copolymer Micelles

P(PEGMA-co-PDSM)-b-(LMA-co-MAA) micelles were obtained by dissolving 10 mg of polymer in 200 μL of methanol:dimethyl sulfoxide (MeOH:DMSO; 80:20 v/v) followed by the slow addition of 0.1 M HEPES buffer (pH 7.4, Gibco) yielding a final concentration of 5 mg/mL. For in vivo studies MeOH:DMSO was removed by buffer exchange into 0.1 M HEPES via 4 cycles of centrifugal dialysis (Amicon, MWCO 3000, Millipore). Prior to use, micelle solutions were filter sterilized (0.22 μm, Pall Corporation) and copolymer concentrations were determined via UV-Vis spectroscopy (Synergy H1 Multi-Mode Reader, BioTek). The critical micelle concentration (CMC), size, surface charge, and morphology of each block copolymer were characterized through fluorescence spectroscopy (Synergy H1 Multi-Mode Reader, BioTek), dynamic light scattering (DLS), ζ-potential, and transmission electron microscopy (TEM, FEI Tecnai Osiris), respectively. Polymer CMC values were determined using fluorescence and particle stability studies. Nile red (NR, Sigma Aldrich), a hydrophobic dye that exhibits strong fluorescence within intact micelles,⁵⁰ was utilized as the fluorescence probe. Polymer solutions were prepared at varying concentrations from 0.0001 to 1 mg/mL and were incubated with NR stock solution (1 mg/mL in MeOH:DMSO (80:20)) at a 100:1 (v/v) ratio for 24 h under continuous stirring in the dark at RT. Samples were filtered (0.45 μm PTFE, Thermo Scientific) and the excitation spectra of NR was measured in 96-well plates (UV-Star® Microplates, Greiner Bio-One) at an excitation wavelength of 550 nm and the emission was monitored from 580 to 680 nm. The stability, hydrodynamic size, and ζ-potential of polymer micelles were analyzed by DLS using a Malvern Instruments Zetasizer Nano ZS Instrument (Malvern, USA) equipped with a 4 mV He-Ne laser operating at λ = 633 nm, an avalanche photodiode detector with high quantum efficiency, and an ALV/LSE-5003 multiple tau digital correlator electronics system. To determine the concentration at which the micelles exhibited morphological changes, particle sizes were measured over 10-fold range of serial dilutions with HEPES from 1 mg/mL to 0.001 mg/mL. TEM samples were prepared by placing 10 μL of micelle solution (1 mg/mL) onto copper grids (formvar stabilized with carbon 400 mesh, Ted Pella Inc.). After 2 min of incubation, the excess solution was removed with filter paper and the grids were left to dry overnight before being imaged by TEM.

Drug Encapsulation Efficiency and Release Profile

A polymer micelle solution of 5 mg/mL was incubated with IMQ (Santa Cruz Biotechnology®, Inc) at a ratio of 3.3:1 (w/w) under continuous shaking (1100 rpm) for 48 h at RT. Non-encapsulated and insoluble IMQ was removed by centrifugation (4000 × g for 5 min) followed by sterile filtration (0.22 μm). IMQ concentration was quantified in solvent-resistant UV-Star® microplates via diluting 10 μL of IMQ-loaded micelle solution in 90 μL of DMSO and measuring fluorescence at excitation/emission wavelengths of 325/365 nm. A calibration curve of fluorescence intensity vs IMQ concentration (ranging from 6.25 to 200 μM with an R² value of 0.99) prepared in a similar fashion (DMSO:HEPES 9:1, v/v) was used for calculating the amount of IMQ encapsulated in the micelles. Drug loading capacity (DLC) was determined through the ratio of encapsulated IMQ to total micelles (mg/g).⁵¹

To assess IMQ release kinetics, 100 μL of IMQ loaded micelles were placed in a dialysis device (Slide-A-Lyzer™ MINI Dialysis Device, MWCO 3500, 0.5 mL, ThermoFisher Scientific) and dialyzed against 5 mL of buffer solutions (0.3 M sodium acetate (pH 5.0), 0.1 M phosphate buffer (pH 7.4) or phosphate-buffered saline (PBS) supplemented with 10% Heat Inactivated-Fetal Bovine Serum (HI-FBS, Gibco)) under continues stirring. The buffer was collected and replaced with 5 mL of fresh buffer at predetermined time points. The collected dialysates were lyophilized, reconstituted in buffer-dependent solvent mixtures (DMSO:H₂O (1:1 v/v) for pH 5.0, DMSO:TFA:HCl (4:6:1 v/v) for pH 7.4 and DMSO:H₂O:HCl (2:4:1 v/v) for PBS supplemented with 10% FBS buffered samples) and the concentration of IMQ was determined via fluorescence spectroscopy using a calibration curve comprising standard solutions of IMQ in their respective solvent mixtures.

Assessment of In vitro Cytotoxicity

Antigen presenting cells [bone marrow-derived dendritic cells (BMDC) and RAW264.7 (macrophages)] and tumor cells [B16-F10 murine melanoma cell line and A549 human lung adenocarcinoma epithelial cell line] were used to assess the toxicity of both empty and drug loaded micelle solutions at various concentrations. RAW264.7 and B16-F10 lines were grown in Dulbecco's Modified Eagle Medium w/L-Glutamine (DMEM, Gibco) containing 4.5 g/L glucose whilst A549 cells were cultivated in Kaighn's Modification of Ham's F-12 Medium (F-12K, ATCC) containing 1.5 g/L sodium bicarbonate; both growth mediums were supplemented with HI-FBS. BM cells were isolated from both the femurs and tibias of 8-12 week old male wild-type C57BL/6J mice. After muscle tissue removal and ethanol sterilization of the bones, bone marrow was flushed out over a strainer with Dulbecco's phosphate-buffered saline (DPBS; Corning). Cells were then rinsed with DPBS, lysed using ACK Lysis buffer, (Gibco) and resuspended in complete growth medium GM-RPMI (Roswell Park Memorial Institute medium w/L-Glutamine (RPMI-1640, Gibco) supplemented with 10% FBS, 1% Penicillin Streptomycin (Gibco), 10 mM HEPES, 1× non-essential amino acids (Sigma-Aldrich), 1 mM sodium pyruvate (Gibco), 55 μM 2-mercaptoethanol (Sigma-Aldrich), 50 μg/mL Gentamycin (Life Technologies), 2.5 μg/mL Amphotericin B (Corning) and 10 ng/mL Granulocyte-macrophage colony-stimulating factor (GM-CSF, Peprotech)). BMDC emerging from this were cultured and supplemented with additional GM-RPMI on d 3 and 7 were employed in experiments on d 8 to 12 after harvest.

BMDC, RAW264.7, B16-F10, and A549 were seeded into 96-well plates (appropriate number of cells needed for each well was determined from growth curves: 5 × 10³ cells/well for B16-F10 and 10 × 10³ cells/well for BMDC, RAW264.7, and A549) and incubated at 37 °C in 5% CO₂ and 95% humidity for 24 h to ensure adherence. Upon surface attachment, cells were incubated with both empty and IMQ-loaded micelles at varying concentrations (5 - 500 μg/mL). After 24 h incubation, 20 μL of AlamarBlue dye was added to each well, and incubated for an additional 5 h.⁵² Absorbance was measured at 570 nm and percent viability was expressed with respect to untreated cells (i.e., 100% viable). The treatments were done with four replicates and the assay was independently repeated three times.

In vitro Evaluation of the TLR7 Pathway Activation in RAW-Blue Macrophages

RAW-Blue™ reporter cells (InvivoGen), a RAW264.7 macrophages derivative that contains an NF-κB and AP-1 inducible secreted embryonic alkaline phosphatase (SEAP) construct, were employed to investigate the effect of micellar IMQ delivery on TLR7 activation. Cells were seeded in 96-well plates at 15 × 10⁴ cells/well concentration (180 μL/well) and cultivated with empty micelles, free and encapsulated-IMQ treatments (20 μL/well) at various concentrations for 24 h. Owing to poor solubility in 0.1 M HEPES, free IMQ treatments were prepared from a stock solution in DMSO (1 mg/mL) which was then diluted into cell culture medium to desired concentrations. To measure SEAP levels, cell culture supernatants (50 μL/well) were incubated with 150 μL/well of the SEAP substrate QUANTI-Blue™ (InvivoGen) in a 96-well plate for 3 h at 37 °C. SEAP levels were determined by UV-Vis measurements at 650 nm.

In vitro BMDC Activation and Maturation

BMDC's were seeded at 2 × 10⁵ cells/well (in 24-well plates) and 1 × 10⁶ cells/well (in 6-well plates) 24 h prior to treatment. Cells were incubated with empty micelles, free, and encapsulated-IMQ treatments at various concentrations for 24 h. BMDCs treated with only 0.1 M HEPES and only lipopolysaccharide (LPS, 5 μg/mL, eBioscience) in RPMI medium were used as negative and positive controls, respectively. Cell supernatants in the 24-well plates were collected and analyzed for different cytokines (IL-6, TNFα, IL-1β, and IL-12) by ELISA (ELISA Ready-SET-Go!® eBioscience) following the manufacturer's protocol. Cell surface immunostaining was employed for investigating BMDC maturation. Accordingly, cells cultivated in 6-well plates were harvested, pre-incubated with TruStain fcX™ (BioLegend) for 10 min on ice to block the Fc receptors and subsequently stained with fluorescently labeled antibodies anti-CD40-PE, anti-CD80-APC, anti-CD86-PE-Cy7, and anti-MHC-II-APC-Cy7 (BioLegend) for 30 min at 4 °C. LPS stimulated BMDCs were stained with relative isotype controls (BioLegend) and used in background subtraction. Ten thousand events were acquired by flow cytometry (BD LSRII) and the emergent data was analyzed via FlowJo software.

Preparation of Micelle-Antigen Conjugates

Polymer micelles were decorated with ovalbumin protein [OVA; Sigma Aldrich (for in vitro studies) or EndoFit™ ovalbumin, InvivoGen (for in vivo studies)] via thiol-disulfide exchange. Following manufacturer's instructions, OVA was first labeled with either fluorescein isothiocyanate isomer (FITC, Thermo Scientific) or AlexaFluor488-TFP (Invitrogen) with a degree labeling of ∼1 dye/OVA. OVA was then thiolated using 25 molar excess of 2-iminothiolane (Traut's reagent, Thermo Scientific) and buffer exchanged into 1× PBS using a Zeba desalting column (0.5 mL, MWCO 7000, Thermo Scientific) as previously described.³³ After sterile filtration (0.2 μm) the number of thiols per protein (4-5 thiols/OVA) was determined with Ellman's reagent (Thermo Scientific). Polymer micelle solutions were reacted with thiolated-OVA at various molar ratios in 0.1 M HEPES (pH 7.4) overnight in the dark at RT under continuous shaking. Antigen conjugation and subsequent disulfide reduction with 20 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP, Sigma-Aldrich) was verified via gel electrophoresis using 4-20% Mini-Protean® TGX™ Precast Protein Gel (Bio-Rad). The free micelle lanes (LMA50 and LMA50-IMQ) were used for subtracting background fluorescence from their respective conjugates (OVA-LMA50 and OVA-LMA50-IMQ). Gels were imaged (Gel Doc EZ, Bio-Rad) and conjugation was quantified with Image Lab (Bio-Rad) software.

In vitro Antigen and IMQ Uptake by APCs

Intracellular uptake of vaccine formulations was evaluated in a dendritic cell line (DC2.4) using AlexaFluor488-labeled OVA (OVA488), a model fluorescent hydrophobic drug (NR), and IMQ. Cells were plated at 7.5 × 10³ cells/well in 24-well plates and allowed to adhere overnight. DCs were then incubated with treatments groups containing 10 μg OVA488 with or without 2.5 μg IMQ and/or 0.12 mg polymer for 24 h. The formulations were as follows: free OVA488, OVA488 conjugated to polymer (OVA488-LMA50), IMQ encapsulated by polymer-OVA488 conjugates (OVA488-LMA50-IMQ), OVA488 mixed with IMQ (OVA488+IMQ), free NR, NR encapsulated by polymer-OVA488 conjugates (OVA488-LMA50-NR) and free OVA488 mixed with NR (OVA488+NR). To provide an accurate comparison amongst treatment groups, OVA controls used for in vitro uptake (OVA488) and MHC-I presentation (OVA) studies were thiolated proteins. Consistent with the encapsulation protocol used in CMC investigations, polymer solutions (5 mg/mL) were incubated with NR (1 mg/mL in MeOH:DMSO (80:20)) for 24 h at RT. Excess dye was removed by sterile filtration, the encapsulated dye content was quantified and matched to the dye content of all NR formulations. After 24 h, cells were collected, rinsed twice with PBS, and resuspended in flow cytometry staining (FACS) buffer (PBS containing 2% HI-FBS). Flow cytometry (Guava® easyCyte™, EMD Millipore) was used to monitor the uptake of OVA and NR.

In vitro MHC-I Antigen Presentation

DC2.4 were employed as model antigen presenting cells to analyze the effect of IMQ and polymer carriers on MHC class I (MHC-I) presentation. DC2.4 (H-2K^b-positive) cells were kindly provided by K. Rock (University of Massachusetts Medical School) and cultured in RPMI supplemented with 10% HI-FBS, 2 mM l-glutamine, 1% Penicillin Streptomycin, 55 μM 2-mercaptoethanol, 1× nonessential amino acids and 10 mM HEPES. This study utilized a 25-D1.16-PE monoclonal antibody (eBioscience) which recognizes the immunodominant OVA_257-264 epitope SIINFEKL bound to H-2K^b, a murine MHC-I molecule.⁵³ DC2.4 cells were seeded in 12-well plates at 2 × 10⁵ cells/well and incubated for 24 h to ensure adherence. Cells were incubated with a variety of formulations (free OVA, OVA-LMA50, OVA-LMA50-IMQ, and OVA+IMQ) featuring 40 μg OVA with or without 10 μg IMQ and/or 0.5 mg polymer for 24 h. As a positive control, untreated cells were pulsed with 20 μM OVA_257-264 (InvivoGen) peptide for 2 h prior to staining. After incubation, cells were rinsed with PBS, detached and resuspended in fresh FACS buffer. After TruStain fcX treatment, cells were stained with 25-D1.16-PE for 30 min at 4 °C, washed twice with FACS buffer and analyzed via flow cytometry (Guava® easyCyte™, EMD Millipore).

Animals

Male C57BL/6 mice (6-8 weeks) were purchased from The Jackson Laboratory (Bar Harbor, ME), maintained at the animal facilities of Vanderbilt University under conventional conditions, and experimented upon in accordance with the regulations and guidelines of Vanderbilt University Institutional Animal Care and Use Committee (IACUC).

Intranasal Immunization

Low endotoxin grade OVA (<1 EU/mg, EndoFit™) and sterile buffer solutions (1× PBS or 0.1 M HEPES at pH 7.4) were used for vaccine formulations. Endotoxin levels of the copolymer vaccines were tested using a ToxinSensor™ Chromogenic LAL endotoxin assay kit (GenScript) and found to be ≤5 EU/mL as recommended by the United States Pharmacopoeia.⁵⁴ The two experimental groups were IMQ encapsulated by polymer-OVA conjugates (OVA-LMA50-IMQ) and OVA mixed with soluble IMQ (OVA+IMQ). Formulations were prepared 3 d prior to vaccination and were stored sterile at 4 °C. Polymer solutions buffer exchanged into 0.1 M HEPES were incubated with IMQ for 48 h. Following the removal of free IMQ via centrifugation and sterile filtration as described above, IMQ-loaded polymers were conjugated with freshly thiolated endo-free OVA (4 thiols/OVA in PBS) overnight. The control formulation, OVA+IMQ, was prepared by mixing endo-free OVA and soluble IMQ (5 mg/mL in 0.1 M lactic acid, Sigma-Aldrich) overnight.

Mice were anesthetized with ketamine/xylazine (10 mg/mL ketamine hydrochloride injection, Vedco; 1 mg/mL xylazine hydrochloride injection; Vanderbilt Pharmacy) by intraperitoneal (IP) injections (100 μL anesthesia/18 g mice). Groups of anesthetized mice (n=5) were immunized intranasally (IN) on d 0 with formulations containing 40 μg OVA and 10 μg IMQ with or without 0.5 mg polymer. Vaccine formulations were delivered through the nostrils into the lungs at 100 μL/mouse. Animals were monitored every day from d 0 to d 12 for weight loss and signs of distress.

Analysis of OVA-specific CD8⁺ T cell responses

On d 12 after vaccinations, mice were anesthetized and administered 200 μL of anti-CD45-APC antibody (BD Biosciences) intravenously (i.v.). This procedure stains leukocytes in the vasculature. Five minutes after injection, mice were euthanized with CO₂, and broncheoalveolar lavage (BAL), lungs, and spleen were collected from each mouse. Organs were harvested and processed as previously described.⁵⁵ In brief, organs were passed through a 70 μm cell strainer and were lysed with ACK lysing buffer (Gibco) forming single cell suspensions. Prior to processing, lung samples were minced with a scalpel and incubated for 1 h at 37 °C in complete RPMI medium supplemented with 2 mg/mL collagenase (Sigma-Aldrich) and 50 nM dasatinib (LC Laboratories). Single cell suspensions from the BAL fluid, lungs, and spleens were stained on ice with anti-B220-FITC (BD Biosciences), anti-CD4-FITC (BD Biosciences), anti-CD11b-FITC (Tonbo Biosciences), anti-CD11c-FITC (Tonbo Biosciences), anti-CD8α-PacificBlue (BD Biosciences), and 1.5 μg/mL of PE-labeled OVA_257-264 (SIINFEKL)-H-2K^b tetramer prepared according to a previously reported procedure.⁵⁶ After 1 h, cells were rinsed and stained with propidium iodide (BD Biosciences) to discriminate live from dead cells. Flow cytometry (BD LSRII) determined the frequency of OVA-specific CD8⁺ T cells (tetramer-positive, CD8-positive population) and all data were analyzed using FlowJo software.

Measurement of Antigen-Specific IgG Production in Sera of Immunized Mice

Retro-orbital blood draws were performed using plain capillary tubes (Fisher Scientific) two days prior to sacrifice (d 10), serum was isolated and stored at -20 °C until further analysis. ELISA microtiter plates (Nunc MaxiSorp™) were coated with 10 μg/mL OVA in 1× PBS and incubated overnight at 4 °C and blocked with PBS/0.01% Tween 20 (PBST; Sigma-Aldrich) for 1 h at RT. Starting from a 1:25 dilution, serum was added to the wells in 3-fold serial dilutions in PBST. In parallel, control-serum from naïve mice was used to determine the background. Following a 2 h incubation period, plates were rinsed with PBST (4X) and incubated with secondary antibody (goat anti-mouse IgG conjugated to horseradish peroxidase (HRP), Southern Biotech; diluted 1:5000) for 1 h at RT. After washes, antibody binding was detected with 100 μL of 1-step ultra TMB-ELISA substrate (ThermoFisher); after stopping HRP reaction with 2 N H₂SO₄ (Sigma-Aldrich), absorbance was measured at 450 nm, endpoint titers were determined from the reciprocal of serum dilution corresponding to 2.5× higher absorbance value than that of the control-serum.³⁰

Diblock Copolymers Self-Assemble into Micellar Nanoparticles with a Fatty Acid-Mimetic Core

To generate nanoparticles with a fatty-acid mimetic core for the encapsulation of IMQ, polymer micelles were formed by the self-assembly of P(PEGMA-co-PDSM)-b-(LMA-co-MAA) chains in aqueous solution (Figure 1A-B). Solubility trials were conducted using a range of water-miscible, organic co-solvents, and a mixture of MeOH:DMSO (80:20 v/v) was ultimately found to solubilize all polymers in the series with varying oleophilic content (i.e., LMA25, LMA50, and LMA75). Micelles were prepared by dissolving the copolymers in the organic co-solvent at 50 mg/mL followed by diluting the solutions 10-fold into HEPES buffer (pH 7.4). The gradual dilution process allowed the polymers to spontaneously self-assemble into micelles with P(PEGMA-co-PDSM) forming an outer shell around the P(LMA-co-MAA) core. The size, surface charge, and morphology of LMA75, LMA50, and LMA25 copolymers was characterized in comparison to control-block copolymers, LMA100 and LMA0 ((PEGMA)-b-(MAA); M_n,theo=30000 g/mol), which were synthesized to represent both ends of the copolymer composition spectrum.

DLS measurements indicated average hydrodynamic diameters of 51.6, 30.1, and 21.4 nm for LMA75, LMA50, and LMA25 respectively (Figures 1C and S6A-B, Supporting Information). The hydrodynamic diameter of micelles (Table 2) displays a strong correlation between increased LMA content and the micelle size, a likely result of the sterically-demanding LMA units within the core increasing chain persistence length and occupying a higher volume.^{65, 67} TEM imaging (Figure S6C, Supporting Information) of LMA50 confirmed particle size and indicated a spherical particle morphology. Also consistent with micellar assembly, the negative charges, originating from MAA moieties, were successfully shielded by the PEGMA corona in LMA75 and LMA50 copolymers resulting in ζ-potential measurements of 0.025 and 0.018 mV, respectively. A decrease in the ζ-potential is observed for LMA25 (-8.45 mV) due to the higher mole ratio of MAA and lower feed molar ratio of steric LMA groups, which are proficient at hindering charge,⁶⁸ within the copolymer structure.

Two complimentary methods, a fluorescence probe technique and DLS stability assay, were used to determine the critical micelle concentration (CMC) of micelles of variable core LMA content.^69-70 The fluorescence intensity of NR, a well-established lipophilic florescent probe,^71-72 was measured as a function of polymer concentration. At concentrations above the CMC, NR exhibits dramatically higher fluorescence emission, indicating that the dye is encapsulated in a hydrophobic micellar core (Figure S7A-B, Supporting Information). The CMC values of LMA75, LMA50, and LMA25 copolymers were estimated to be 0.093, 0.088, and 0.165 mg/mL, respectively, which is similar to other block copolymer micelles used for hydrophobic drug delivery reported in the literature (Table 1).^{51, 73-74} To validate CMC measurements, a DLS-based dilution method was used in which hydrodynamic diameters of copolymers were measured at concentrations ranging from 1 mg/mL to 0.001 mg/mL (Figure S7C and Table S1, Supporting Information). Particle size remained stable between 0.1 - 0.01 mg/mL (for LMA75 and LMA50) and 0.5 - 0.1 mg/mL (for LMA25), consistent with the CMC range determined with NR encapsulation.

Compared to the other copolymers, LMA25 demonstrated a nearly two-fold increase in its CMC value as result of its higher negative charge and reduced core hydrophobicity. As expected, LMA0 was not able to form micelles, likely existing as unimers in aqueous solution due to the absence of hydrophobic micellar driving forces in the second block. The small size, neutral zeta potential, and brush-like PEGMA shell render fatty-acid mimetic carriers well-suited for increasing drug circulation time, enhancing lymphatic drainage from injection sites,³⁹ and facilitating passive accumulation at tumor sites via the enhanced permeation and retention (EPR) effect.^{40, 75} Though micelles were specifically designed to accommodate the unique solubility requirements of IMQ, they have potential utility as carriers for hydrophobic drugs or cationic macromolecules.

IMQ Encapsulation and Release from Fatty Acid-Mimetic Micelles

Based on the high solubility of IMQ in fatty acids, but not other organic solvents, we postulated that a micelle core comprising a statistical copolymer of long alkyl chain and carboxylic acid monomers would provide a favorable environment for IMQ loading. To evaluate this and to elucidate how the molar ratio of alkyl chains and acidic groups influences IMQ solubility, the DLC of IMQ in each micelle formulation was quantified. Pre-formed micelles were incubated with IMQ for 24 and 48 h under continuous stirring. Under the same conditions, a control sample without micelles was prepared to determine IMQ solubility in 0.1 M HEPES (4.5 and 6.0 μg/mL for 24 and 48 h, respectively). As additional controls, structurally analogous diblock copolymers lacking acidic groups (LMA100) or without a micellar morphology (LMA0) were prepared. Free IMQ was removed by centrifugation and filtration, followed by a 10-fold dilution into DMSO, facilitating the release of IMQ which was then quantified via fluorescence spectroscopy. In order to determine the maximum DLC, the weight ratio of polymer to IMQ was varied between 10:1 - 1.25:1 at a constant polymer concentration, resulting in a plateau (maximum DLC) at 3.3:1 ratio (Figure S8, Supporting Information), which was used for all subsequent encapsulation studies.

The loading efficiency of LMA100 micelles was largely negligible (0.16 w/w%), with a similar solubility to IMQ in HEPES buffer, whereas LMA0 increased solubility nearly 3.5-fold, likely due to IMQ's preference for acidic groups and ability to form hydrogen bonds with MAA.⁷⁶ These results demonstrate the importance of acidic groups for solubilization of IMQ and motivate the design of micelles comprising both LMA and MAA in the core. Indeed, the loading efficiency of LMA75, LMA50, and LMA25 were revealed to be 0.51, 2.23, and 1.84 w/w%, respectively (Table 2). The hydrodynamic diameter of the loaded micelles revealed a slight increase in size (Table 2) as a result of chain expansion in the hydrophobic core upon encapsulation. As expected, LMA75, the copolymer with the least acidic content, demonstrated the lowest loading of the three P(PEGMA-co-PDSM)-b-(LMA-co-MAA) block copolymer micelles, though comparable to the non-micellar LMA0 formulation, suggesting an interplay between IMQ's interaction with acidic moieties and a hydrophobic environment. Interestingly, the highest IMQ loading was obtained with LMA50, consistent with mimicking the 1:1 molar ratio of alkyl chains to acidic groups characteristic of fatty acids. Notably, on a per mass basis, IMQ solubility in LMA50 micelles (2.2 w/w%) is comparable to reported solubility in isostearic acid (1.9 w/w%)²¹ and approved topical formulations of Aldara® (3.75 and 5 w/w%).²⁰

The protonation of aromatic amines in IMQ under acidic conditions led us to postulate that the reduced pH encountered within tumor microenvironments or endo/lysosomes would enhance the release of encapsulated IMQ.⁷⁷ Specifically, IMQ binds TLR7 localized within acidic endosomal vesicles, and, therefore, must be released from micelles in response to decreases in pH.^{29, 78} To explore this, the release of IMQ encapsulated in micelles of varying LMA composition was assessed via dialysis against buffer solutions mimicking the cytosolic (pH 7.4) and late endosomal (pH 5.0) environments (Figure 2). At pH 7.4, 50% of IMQ was released from LMA75, LMA50, and LMA25 micelles after 12.1 ± 0.6, 25.7 ± 0.3, and 32.7 ± 0.4 h, respectively. Consistent with the interactions between IMQ and MAA, the copolymer with the lowest acidic content, LMA75, demonstrated the fastest drug release profile, and increasing the acidic character of the micelle core increased the retention time of IMQ by 2- and 3- fold for LMA50 and LMA25, respectively. Slightly faster release was observed with micelles in the presence of serum proteins (Figure S9, Supporting Information). The 50% release time points for LMA75, LMA50, and LMA25 polymer micelles were revealed to be 4.6 ± 2.1, 16.2 ± 1.6, and 22.1 ± 1.1 h in PBS buffer supplemented with 10% FBS, respectively. This may reflect a destabilizing effect of serum proteins on the micelle core and reduced the retention time of the drug within the micelle, resulting in 1.5 to 2.5 times faster release at similar pH. At pH 5.0, the release kinetics of IMQ was significantly increased, displaying a burst-like effect due to IMQ's high solubility under acidic conditions with 50% IMQ released at 2.6, 5.1, and 8.4 h time points for LMA75, LMA50, and LMA25, respectively. As a further demonstration of pH-responsive release kinetics, IMQ release from LMA50 was studied at the early endosomal pH of 6.2. As expected, the release profile laid between pH 5.0 and pH 7.4, exhibiting 50% drug release at 17.7 ± 0.2 h. We note that relative to a recently described polylactide-based IMQ delivery system, to our knowledge the only previously described micellar delivery system, fatty acid-mimetic micelles enable more stable IMQ encapsulation at physiological pH (i.e., 50% release at >10 h vs. 1 h) and a more dramatic pH-triggered release profile.⁷⁴ Therefore, by rationally designing block copolymer micelles with an optimized fatty acid-mimetic core, we have developed a novel water-soluble, nanoparticle-based IMQ formulation with a smaller size and higher encapsulation efficiency than other sub-100 nm IMQ delivery vehicles reported to date.^{74, 78-79}

Open in a separate window

Figure 2

The core composition of fatty acid-mimetic micelles influences the release kinetics of IMQ at physiological, early, and late endosomal pH values. IMQ was encapsulated in (A) LMA75, (B) LMA50, and (C) LMA25 polymer micelles, dialyzed against buffers at pH 7.4, 6.2, and 5.0 and cumulative release was quantified over 48 h. 50% IMQ release values were determined by using exponential one-phase decay nonlinear fit analysis with R² values greater than 0.98. The 50% release time points for LMA75, LMA50, and LMA25 polymer micelles were revealed to be 12.1 ± 0.6, 25.7 ± 0.3, 32.7 ± 0.4 h (at pH 7.4) and 2.6 ± 0.7, 5.1 ± 0.6, 8.4 ± 0.4 h (at pH 5.0), respectively. Graphs represent the means and standard deviation from four individual experiments.

Fatty Acid-Mimetic Micelles are Highly Cytocompatible

Micelles were designed with a PEGMA-rich corona to impart high colloidal stability and minimal cytotoxicity. Prior to studying the immunomodulatory properties of IMQ-loaded micelles, the cytotoxicity of empty and IMQ-loaded micelles was investigated in antigen presenting cells (RAW264.7 and BMDCs) and cancer cell lines (B16-F10 and A549), owing to the exploration of IMQ as both a vaccine adjuvant and a cancer immunotherapeutic. Cells were treated with empty micelles at varying concentrations (5 - 500 μg/mL) for 24 h to determine the biocompatibility of the carriers in vitro. Consistent with the neutral/anionic nature of the copolymers, no significant toxicity was observed in any of the cells tested (Figures 1D and S10, Supporting Information).

Subsequently, polymer carriers were loaded with IMQ (0.2, 1, 2, and 10 μg/mL drug encapsulated in 10, 50, 100, and 500 μg/mL micelle solutions, respectively) and cytotoxicity evaluated in the cells described above (Figure S11, Supporting Information). Consistent with the empty micelles, IMQ-loaded particles displayed no cytotoxicity in the cell lines tested, though a dose-dependent increase in cell viability was observed for BMDCs (Figure S10B). This likely reflects IMQ's ability to increase BMDC proliferation and improve DC survival, which has been leveraged to enhance the efficacy of DC-based therapies.^{32, 80} The high cytocompatibility of fatty acid-mimetic micelles bodes favorably for in vivo applications, potentially enabling IMQ delivery via multiple routes.

Micellar Delivery Enhances IMQ Activity

To exert its immunomodulatory effects, IMQ must be delivered to endosomal vesicles where TLR7 is stationed. Based on the propensity of macrophages and DCs to endocytose nanoparticles and the rapid pH-responsive release of IMQ from fatty-acid mimetic micelles (Figure 1B), we postulated that micellar delivery of IMQ would enhance immunostimulatory activity. Binding of IMQ to TLR7 stimulates innate immunity through the adapter molecule MyD88, resulting in the activation of the transcription factor NF-κB and the production of proinflammatory cytokines.¹⁶ RAW-Blue™ cells that report on NF-κB induction via production of SEAP were utilized to evaluate the immunostimulatory activity of free and encapsulated-IMQ formulations. Free and encapsulated-IMQ treatments varied from 0.1 to 5 μg/mL, while empty micelle concentrations were equivalent to their IMQ-loaded counterparts. Both free and encapsulated-IMQ formulations stimulated NF-κB -dependent SEAP in a dose-dependent manner (Figure 3A), with all encapsulated formulations demonstrating increased activity over free IMQ at concentrations above 1 μg/mL. Empty micelles did not trigger NF-κB (Figure S12A, Supporting Information), supporting the non-inflammatory nature of the carriers and confirming IMQ-dependent NF-κB activation. Delivery of IMQ with LMA50 resulted in the most significant enhanced activity, with a ∼3.5-fold increase of SEAP levels at concentrations >1 μg/mL. While micelle-mediated enhancements in IMQ activity are likely due in part to increased bioavailability owing to enhanced aqueous solubility, differences observed between carriers suggest a relationship between core composition and IMQ delivery. For example, the superior activity of LMA50 relative to LMA25 and LMA75 may reflect a requirement for balance between micelle stability, high drug loading, and rapid endosomal release. Elucidating the mechanisms by which fatty acid-mimetic micelles enhance IMQ activity will require further investigation. Nonetheless, to our knowledge this represents the first investigation into how material properties can be tuned to optimize IMQ solubility and activity.

Open in a separate window

Figure 3

Micellar IMQ delivery enhances NF-κB activity resulting in increased cytokine production. (A) In vitro evaluation of the NF-κB activation in RAW-Blue cells stimulated with 0.1 – 5 μg/mL free and encapsulated IMQ for 24 h. Statistical differences were observed between the loaded polymer micelles and soluble IMQ at higher doses and are represented as (#) LMA75-IMQ vs IMQ; (*) LMA50-IMQ vs IMQ; and (ˆ) LMA25-IMQ vs IMQ. BMDCs were stimulated for 24 h with free and encapsulated IMQ at varying doses (0.2, 1, and 2 μg/mL). A higher IMQ dose, 10 μg/mL, was included in the studies with LMA50-IMQ and LMA25-IMQ being the only two formulations able to achieve this dose. Levels of proinflammatory cytokines (B) TNFα, (C) IL-6, and (D) IL-1β were measured by ELISA. The lower detection limit of each cytokine is represented by a dotted line. Statistical analysis revealed no significant difference between treatment groups at 0.2 and 10 μg/mL for all cytokines. Statistical analysis illustrated by (*) represents significance for each loaded micelle vs free IMQ, while significance amongst polymer micelles are indicated with tracing lines.

Micellar IMQ Delivery Increases Dendritic Cell Activation and Maturation

Activation of TLR7 signaling induces DC maturation and stimulates the expression of proinflammatory cytokines, motivating exploration of IMQ as a vaccine adjuvant, cancer immunotherapeutic, and anti-viral agent.^{16, 81} Hence, BMDCs were incubated with empty micelles, free, and encapsulated-IMQ treatments over a range of IMQ concentrations (0.2, 1, 2 and 10 μg/mL) and secreted TLR7-driven cytokines TNFα, IL-6, IL-1β, and IL-12p70 were monitored. Overall, micellar formulations tended to enhance cytokine levels over free IMQ, although some compositional dependence was observed (Figure 3B-D). All formulations increased IL-1β production at 1 and 2 μg/mL compared to free IMQ 9- and 3-fold, respectively. Micellar IMQ encapsulation also enhanced TNFα production at 2 μg/mL, with LMA75 and LMA50 carriers mediating 2-fold higher TNFα production over LMA25 and 5 times greater than free IMQ. IL-6 expression was most strongly enhanced by a 3-fold increase in IL-6 production at both 1 and 2 μg/mL LMA50-IMQ over free IMQ. Likewise, LMA50-IMQ enhanced production of IL-12p70, at 2 μg/mL, although the levels were low and near the lower detection limit of the assay (Figure S13, Supporting Information). Additionally, the high loading efficiency of LMA50 and LMA25 micelles enabled dosing at 10 μg/mL IMQ concentrations that was beyond its solubility limit, resulting in further enhancement in cytokine production. Consistent with NF-κB studies, empty micelles (without IMQ) corresponding to the same concentrations (10, 50, 100, and 500 μg/mL) did not elicit detectable cytokine production (Figure S12B-D, Supporting Information).

The effect of micellar IMQ delivery on BMDC maturation was assessed via flow cytometric detection of the maturation markers CD80, CD86, CD40, and MHC-II (Figure 4). Treatment with IMQ-loaded micelles increased the frequency of marker-specific BMDCs with LMA50-IMQ and LMA25-IMQ mediating the most significant enhancements compared to free IMQ. Similarly, delivery with LMA50 and LMA25 increased expression levels (as determined by MFI of marker positive cells) of the costimulatory molecules CD40 and CD86 relative to free IMQ (Figure S14, Supporting Information). In accord with BMDC activation studies, empty micelles (without IMQ) had no effect on BMDC maturation (Figure S15, Supporting Information), verifying that the enhanced immunostimulatory effects results from the combination of fatty acid-mimetic micelles and IMQ. Collectively, these results demonstrate that fatty acid-mimetic micelles facilitate the delivery of IMQ in a highly water soluble formulation with increased immunostimulatory activity, and, hence, have the potential to enhance the efficacy and expand utility of IMQ for in vivo applications. LMA50 micelles, which facilitated the highest IMQ loading, were particularly effective at enhancing TLR7 signaling, proinflammatory cytokine production, and DC maturation. Therefore, LMA50 was selected as the lead candidate for a novel subunit vaccine formulation based on nanoparticle-mediated co-delivery of encapsulated IMQ and covalently conjugated protein antigen.

Open in a separate window

Figure 4

Micellar IMQ delivery enhances TLR7-driven BMDC activation and maturation. Flow cytometry was used to measure surface expression of BMDC maturation markers (A) MHC-II, (B) CD86, (C) CD40, and (D) CD80 induced by free and encapsulated IMQ. BMDCs were treated with 2 μg/mL IMQ for 24 h and were evaluated for their relative marker surface expressions (top) and percentage of marker positive cells (bottom). Representative histograms of untreated (control), IMQ, and LMA50-IMQ are shown for each marker. (*) represents the significant difference between micellar-IMQ vs free IMQ, while connected lines illustrate significance amongst different polymer micelles.

Fatty Acid-Mimetic Micelles Enable Dual-Delivery of Antigen and IMQ

A number of studies have demonstrated improved T cell and antibody responses with subunit vaccines that co-deliver protein antigens and TLR7 agonists with the same carrier.^{18, 30, 36-37} To exploit this phenomenon, micelles were designed with a hydrophilic first block comprising PEGMA and thiol-reactive PDSM groups for the conjugation of thiolated-protein antigens. The reversible disulfide bond would liberate antigens upon negotiating the reducing environment of the endosome or cytosol, resulting in improved antigen processing and enhanced immunogenicity.⁴⁵ Hence, the model protein antigen OVA was fluorescently tagged, thiolated and reacted with LMA50 micelles. SDS-PAGE was used to evaluate antigen conjugation efficiency, by monitoring the band shift and disappearance of FITC-labeled OVA (Figure 5A).^{33, 43-44} Using lane densitometry, approximately 88% conjugation was attained at a 1:20 molar ratio of protein:polymer (OVA:LMA50). As anticipated, coupling was not observed with the physical mixture of native OVA and polymer (OVA+LMA50 1:20 ratio) owing to the lack of free cysteine residues in the protein.^{33, 82} Antigen conjugation to IMQ-loaded micelles at a 1:20 ratio was slightly less efficient, potentially due to reduced accessibility of PDSM groups after IMQ encapsulation, though OVA conjugation remained >80%. Protein release and disulfide bond reversibility, two properties proven to enhance conjugate-based vaccine delivery,⁴⁵ was demonstrated through treatment with the reducing agent TCEP. Conjugation results in distributing OVA over a broad range of molecular weights, indicating the variation in the number of OVA conjugated per polymer chain.⁴³ The hydrodynamic diameter of conjugates were analyzed by DLS, revealing no significant change in size (35.5 ± 0.1 nm OVA-LMA50 and 34.0 ± 0.3 nm OVA-LMA50-IMQ), which signifies negligible particle cross-linking or aggregation.

Open in a separate window

Figure 5

Conjugation of OVA to PDSM groups on micelle coronas enables co-loading of antigen and IMQ, resulting in enhanced intracellular uptake and antigen cross-presentation. (A) FITC-labeled OVA was thiolated (4-5 thiols/OVA) (represented as I in the top gel and V in the bottom gel) and reacted with empty (III: OVA-LMA50) or IMQ-loaded micelles (VI: OVA-LMA50-IMQ) at 1:20 (OVA:polymer) ratio and SDS-PAGE was used to assess antigen conjugation and release of OVA upon incubation with TCEP (VII: OVA-LMA50-IMQ + TCEP). A physical mixture of non-thiolated OVA and LMA50 (II: OVA+LMA50) as well as formulations lacking OVA (IV: LMA50 and VIII: LMA50-IMQ) were run as controls. Flow cytometry was used to assess the effect of micellar delivery on the uptake of OVA488 and a model hydrophobic, fluorescent drug, NR by DC2.4 cells after 24 h incubation at 37 °C. (B) Uptake of OVA488: Representative histograms (top) and MFI values (bottom) of DC2.4 cells incubated with OVA488 (red), OVA488-LMA50 (orange), and dual-delivery carriers (OVA488-LMA50-IMQ; green) treatment groups for 24 h. (C) Uptake of NR: DC2.4s treated with NR (red), OVA488-LMA50 (orange), a physical mixture of OVA488 and NR (OVA488+NR) (blue), and dual-delivery carriers (OVA488-LMA50--NR; green) are illustrated in representative histograms (top) and MFI graphs (bottom). (D) The effect of IMQ and/or micellar delivery on MHC-I presentation by DC2.4 cells was assessed via flow cytometry using an antibody that recognizes MHC-I (H-2K^b)-bound SIINFEKL. Antigen cross presentation was illustrated by representative histograms (top) and MFI graphs (bottom) of DC2.4 cells incubated with OVA (red), OVA-LMA50 (orange), OVA+IMQ (blue), and dual-delivery carriers (OVA-LMA50-IMQ; green) for 24 h. Statistical analysis illustrated by (*) represents significance for each ova conjugated-loaded micelle vs control (free antigen or dye) and significance between different groups are shown with connected lines.

To initiate adaptive cellular immunity, antigens must be first be internalized by professional APCs and subsequently processed and presented as peptide epitopes on MHC for recognition by T cells. Based on the well-established proclivity of macrophages and DCs for endocytosis of micro- and nanoparticles,⁸³ we postulated that conjugation of OVA to the micelle corona would enhance antigen internalization. To evaluate this, OVA488 was conjugated to LMA50 micelles, with and without IMQ, and flow cytometry was used to measure uptake relative to free OVA488 (Figure 5B-C). Conjugation of OVA488 to empty LMA50 micelles resulted in a significant, but only modest increase in OVA488 uptake, a likely consequence of the highly PEGylated, stealth-like micelle corona. Strikingly, OVA488 with IMQ-loaded micelles resulted in a 4-fold increase in OVA488 uptake relative to free antigen (Figure 5B). We speculate that the increased OVA488 uptake was a consequence of the enhanced immunostimulatory activity of micellar IMQ formulations, as TLR activation acutely stimulates macropinocytosis by DCs,⁸⁴ a common mechanism of nanoparticle uptake. Even though DC maturation typically down-regulates endocytic capacity, this effect may be TLR-specific, as DC phagocytic activity is maintained after TLR7/8 activation.⁸⁵ To evaluate the ability of micelles to enhance delivery of a hydrophobic drug, a model hydrophobic molecule, NR, was used as a fluorescent analogue of IMQ to monitor internalization by flow cytometry. Similarly, OVA488-LMA50-NR revealed 12-fold higher levels of cell-associated NR relative to incubation with the free dye (Figure 5C). Taken together, these results underscore the importance of particle-based carriers to enhance the intracellular delivery of both antigens and water insoluble, hydrophobic compounds.

Dual-Delivery of Antigen and IMQ Enhances MHC-I Presentation

The predominant fate of endocytosed soluble antigen is lysosomal degradation and presentation by MHC-II molecules and minimal presentation on MHC-I (i.e., cross-presentation), resulting in negligible CD8⁺ T cell responses. That notwithstanding, TLR agonists, including IMQ,^86-87 as well as a number of particle-based carriers⁸³ have been shown to enhance antigen cross-presentation. DC2.4 cells, a dendritic cell line derived from BMDCs, are capable of cross-presenting exogenous antigen⁸⁸ and can also be activated with free or encapsulated IMQ (Figure S16, Supporting Information), and therefore, were used to investigate the synergistic effects of dual-delivery of antigen and adjuvant. To determine the effect of micellar antigen delivery on antigen cross-presentation, DC2.4 cells were incubated with free OVA, OVA-LMA50 conjugates with or without IMQ, and a physical mixture of OVA and IMQ. In vitro presentation of the immunodominant MHC-I restricted OVA epitope (SIINFEKL) was assessed by flow cytometry using an antibody that recognizes MHC-I (H-2K^b)-bound SIINFEKL. Conjugation of OVA to micelles modestly enhanced cross-presentation (Figure 5D), consistent with enhancement in uptake via a carrier that lacks an active mechanism to enhance cytosolic delivery to the classical class I pathway.⁴³ Significantly, dual-delivery of OVA and IMQ using LMA50 micelles enhanced SIINFEKL presentation by class I molecules by 3-fold over free antigen, 2.5-fold over OVA-LMA50 conjugates, and 2-fold over OVA+IMQ mixture (Figure 5D). This is consistent with the observed micelle-mediated enhancements in IMQ activity as well as increased antigen uptake, indicating a potential synergism via delivery of both antigen and IMQ using fatty acid-mimetic micelles.

We gratefully acknowledge the laboratory of Prof. Craig Duvall for the use of GPC chromotography as well as the core facilities of the Vanderbilt Institute of Nanoscale Sciences and Engineering (VINSE) and for the use of both the dynamic light scattering and TEM instruments. We also acknoledge the VMC Flow Cytometry Shared Resources, supported by the Vanderbilt Ingram Cancer Center (P30 CA68485) and the Vanderbilt Digestive Disease Research Center (DK058404), for the usage of 3-laser BD LSRII flow cytometery. This research was supported by NSF CAREER Award CBET 1554623 (JTW), NIH 1R21AI121626-01A1 (JTW), the Vanderbilt University Discovery Grant Program (JTW), and the Vanderbilt University School of Engineering (JTW). SJ would like to acknowledge the US Department of Veterans Affairs for the Merit Award (BX001444).