|Home | About | Journals | Submit | Contact Us | Français|
Current cancer immunotherapy strategies target tumors by modulating the host immune responses against tumor antigens [1, 2]. Antigen-based cancer immunotherapy relies on the uptake of antigen by dendritic cells (DCs) in the presence of immuno-stimulatory molecules (adjuvants), followed by the DC localization into draining lymphoid tissues and antigen cross-presentation to CD8+ T cells [3, 4]. Effector T cells can then target and kill tumor cells, while memory T cells can control tumor relapse [5, 6]. Many of these cancer immunotherapy strategies have shown promise in the clinic, but there are a number of limitations inherent to these approaches. First, dendritic cell-based adoptive immunotherapies are expensive and have shown modest increases in patient survival , which could be attributed to the apoptosis of reinfused cells post-transplantation and failure of cells in retaining their effector functions . Second, most systemically administered cancer vaccines are not taken up by the specific immune cells and have a short retention time in the body. Finally, current antigen-based cancer vaccines are soluble biomolecules that have poor retention in tissues, along with sub-optimal antigen processing and presentation by DCs. The low retention in the circulation and tissues imposes a significant limitation because multiple booster doses are often needed, leading to dose limiting systemic toxicity , and therefore safer immunotherapy approaches are needed. Several current immunomodulation approaches have been reviewed elsewhere by us [10, 11] and others [1, 3, 6, 12-14].
Materials-based vaccines may overcome the limitations of current cancer immunotherapies by targeting dendritic cell populations and deliver antigen/adjuvant payloads in a sustained manner, therefore potentially reducing high dose related systemic toxicity. Recent works by Mooney and colleagues [15, 16] have resulted in the first biomaterials-based implantable cancer vaccine, which is currently in human clinical trial against melanoma and works through subcutaneous implantation. Yet, there is a strong need to develop a simple but versatile approach that could be applied to other cancers, including non-cutaneous and systemic tumors, such as lymphomas, carcinomas, and myelomas. These cancers express tumor specific antigens, which however are often poorly immunogenic [10, 17, 18] and require administration of additional immuno-stimulatory molecules [14, 19]. The immuno-modulatory agent should preferably be delivered simultaneously to the same DC that phagocytoses the antigen . At the same time, it is important for these cues to be presented to a substantial number of DCs for a prolonged duration [15, 21]. Pathogen-associated molecular patterns (PAMPs) and Toll Like Receptor agonists can provide danger signals to activate the immune system [14, 22], but many of these immune-stimulators are insoluble hydrophobic molecules (e.g. monophosphoryl lipid A), which imposes delivery limitations with current aqueous vaccine formulations . As an alternative, nanomaterials-based immunotherapy has the potential to overcome several of the current limitations. First, nanometer sized particles can deliver immune-modulatory cargo to the lymphoid tissues in a size-dependent manner . Second, the enhanced retention of therapeutics at the injection site or the lymphoid tissues can reduce systemic toxicity, therefore enabling safer immunotherapy approaches . Finally, nanoparticle formulations can often be functionalized with ligands to target specific immune cells and could be formed using immune-modulatory biomaterials that function as adjuvants or immune-potentiators [13, 25]. However, many existing polymers (e.g. PLGA ) used for fabricating nanoparticles have poor protein encapsulation, sub-optimal dosing characteristics, limited stability (liposomes), and importantly, can only form nanoparticles with a subset of proteins. There is a need for a versatile protein antigen delivery system that can readily form nanoparticles with any protein antigen of interest and allow for a precise control over particle size within nanometer range. We recently developed self-assembling nanogels containing poly(hydroxyethyl methacrylate) (pHEMA) backbone with a functionalized hydrophobic side chain of pyridine . The pHEMA-pyridine polymer self-assembled with purified bovine serum albumin (BSA) or therapeutic adhesion protein, fibronectin. The protein-polymer ratio enabled control over nanogel size in a wide 200–900 nm range and demonstrated significantly longer nanogel-protein retention (> 14 days) in the rat stifle joints through single intra-articular injections as compared to bolus protein doses . Here we report the development of self-assembly nanogels vaccines using a modified catalyst reaction for pHEMA-pyridine that provided a fine control of nanogel sizes within a narrow 145–160 nm range. The resulting immune-stimulatory nanogels delivered protein antigen to DCs and resulted in efficient priming of OVA-specific CD8+ T cells.
pHEMA offers a unique advantage of being biocompatible and availability of its hydroxyl end groups for further modifications. Previously, we have reported the synthesis of pHEMA-pyridine using pyridine in THF . However, pyridine is known to be toxic and relatively difficult to remove; therefore, purification of the final product is often challenging. Here we report a modified protocol where pHEMA-pyridine synthesis was carried out through esterification of the hydroxyl group on pHEMA using DMAP catalyst that allowed pHEMA to readily dissolve in THF, and separated out the excess chloride salt on nicotinoyl chloride (Figure 1A, B). The resulting white powder was recovered following purification (see methods) and the esterification was confirmed by 1H NMR with the following peaks: 7.2-9.0 for pyridine-H (four peaks), 4.0-4.5 for –OCH2 (two peaks), 1.5-2.0 for –CH2, and ~0.5-1.0 for –CH3 (Figure 1C). The hydroxyl group of pHEMA disappeared in the NMR spectrum post-reaction, indicating successful esterification of pHEMA. These peaks were similar to those previously reported by us using pyridine instead of DMAP , however with 94% substitution as compared to near 90% observed earlier (see supplementary methods). To further confirm substitution of pyridine using DMAP, we performed diffusion slope decay analysis on the NMR results and found identical slopes between the pHEMA linkage and pyridine ring (Figure 1D, Supplementary Figure S1). These results indicated that the pyridine protons had the same diffusion coefficient as the pHEMA, therefore confirming that pyridine was functionalized to pHEMA. The 13C NMR and H/13C 2D HSQC spectrum of pHEMA-pyridine in DMSO-d6 are indicated in Supplementary Figure S1.
The modification of pHEMA with pyridine groups decreased the solubility of pHEMA in water, which in turn allowed for aqueous protein solutions to form self-assembled nanogels. The protein self-assembly was promoted by pyridine while the carbonyl group present between pHEMA backbone and functionalized pyridine groups served as an electron-withdrawing group to lower the pKa of the protonated pyridine. This hypothesis was supported by the NMR results where the protons on peak 5 and 6 (CH2-CH2) appeared to be de-shielded by approximately 0.5 ppm, indicating a presence of an electron withdrawing ester group in the vicinity of the protons.
Following the synthesis of pHEMA-pyridine using DMAP as the reaction catalyst, we assessed the ability of pHEMA-pyridine to form protein nanogels using FITC-conjugated bovine serum albumin (FITC-BSA) as a model protein and ovalbumin as a model vaccine antigen to test against immune cells from OVA-specific mouse. The self-assembly of protein with the pHEMA-pyridine was accomplished by incubating the modified polymer in aqueous buffer solution containing the protein (Figure 1E). The protein-polymer complexation occurred rapidly and the resulting nanogels were collected through overnight lyophilization and reconstitution in cell culture media. Notably, presence of benzene-containing FITC enabled successful centrifigation and reconstistitution of nanogels, whereas unconjugated BSA or ovalbumin resulted in a pellet that required lyophilization and resuspension. As suggested earlier by us , we anticipated the nanogel assembly process to be primarily entropically driven, with electrostatic interaction and hydrogen bonding acting as patches that stabilized the final structure. The displacement of proteins between the aqueous water and insoluble pHEMA–pyridine at pH 7.4 minimized the interfacial energy that resulted in nanogels self-assembly. It is important to note that pHEMA-pyridine can only form a nanogel in the presence of a protein (any type of protein, e.g. BSA, fibronectin, ovlabumin (OVA) etc.) In the absence of a protein, pHEMA-pyridine remains as a un-complexed molecule.
We observed that by keeping the amount of pHEMA-pyridine constant, increasing amount of FITC-BSA resulted in the smaller diameter of nanogels, however within 100-200 nm range. As indicated in Figure 1E, 60 μg/mL of protein or 0.75:1 protein-to-polymer ratio resulted in nanogels with average diameter of 160 ± 0.5 nm. Increasing the ratio to 1.5:1 by using 120 μg/ml of protein and 3:1 ratio with 240 μg/ml of protein produced smaller particle size with 150 ± 0.4 nm and 145 ± 0.7 nm in an average diameter, respectively. A further increase in protein concentration beyond 240 μg/mL did not further decrease the size of the particles (Supplementary Figure S2A). The resulting hydrodynamic diameter for the nanogels formed with 60 μg/ml of protein showed low polydispersity as indicated by the narrow nanogel size distribution (Supplementary Figure S2B). While we have previously demonstrated the use of pHEMA-pyridine (using pyridine as catalyst) in forming nanogels with size ranging from 200 to 900 nm, our results indicate that the use of pHEMA-pyridine synthesized using DMAP as a catalyst enabled a control of nanogel size within 145 – 160 nm range even with an increase in protein amount to three-fold.
Analysis on zeta potential for nanogels formed using 60, 120, and 240 μg/mL of protein resulted in similar zeta potential values of −20.2 ± 0.5, −20.0 ± 0.6, and −19.5 ± 0.57 mV (Supplementary Figure S2C). Electrophoretic mobility for the three categories of nanogels were −1.57 ± 0.04, −1.56 ± 0.05, and −1.52 ± 0.05 μmcm/Vs (Supplementary Figure S2D). The nanogel characterization data indicates that while changing protein to polymer ratio resulted in tunable particle sizes, both particle stability and movement in an electric field were not affected. These findings highlight that pHEMA-pyridine nanogels retain characteristic properties across a range of hydrodynamic diameter, making them a suitable protein delivery platform for targeting many biological locations through precise size control without any unexpected change in the physical characteristics. Protein loading studies indicated that 30 μg protein was loaded per 400 μg pHEMA-pyridine polymer. In addition, use of both 60 and 120 μg/ml BSA protein stock solutions resulted in 30 μg protein loading while further increase in the protein solution 4 times to 240 μg/mL resulted in 3 times as much protein encapsulation (Supplementary Figure S2E). The resulting particles were biodegradable and incubation in 5 mL PBS over 48 hr resulted in 25% reduction in the number of nanogels (Supplementary Figure S3). We determined the optimal nanogel formulation by developing a two-dimensional particle selection matrix that took into account particle number and mean fluorescence. As indicated in Figure 1F, nanogels formed with 60 μg/mL protein solution resulted in the maximum particle count and protein loading, and were used for future studies.
To further examine the biological characteristics of pHEMA-pyridine nanogels, we determined the nanogel cytotoxicity and DC phagocytosis. Material cytotoxicity was evaluated by incubating JAWSII DC cell line with an increasing dose of nanogels (33.75 – 150 μg/mL). The FITC-BSA nanogels demonstrated no cytotoxicity when incubated with DCs over the period of 24 hours and 120 hours (Figure 2A, Supplementary Figure S4). We performed confocal microscopy that indicated near uniform uptake and localization of fluorescent particles inside the cells in the cytoplasm (Figure 2B). The uptake of fluorescent nanogels was next analyzed by determining the mean fluorescent intensity (MIF) of the FITC-BSA nanogels in the DC population using flow cytometry. The addition of nanogels at a concentration of 75,000 nanogels/mL (22 μg/mL protein) resulted in 25 ± 0.8 fold increase in MIF value as compared to an untreated control group, while doubling the nanogel dose to 150,000 nanogels/mL (45 μg/mL protein) doubled the MIF by 50 ± 1.4 fold (p < 0.05; Figure 2C). Further increase in the nanogel concentration to 300,000 nanogels/mL (90 μg/mL protein) and 450,000 nanogels/mL (135 μg/mL protein) resulted in MIF values increase by 84 ± 5.6 and 113 ± 2.3 fold, respectively, as compared to the untreated control group (Figure 2C). These results indicate that increase in the nanogel dose resulted in an increase in fluorescence intensity, however the fold change in MIF was proportionate to the fold change of nanogel concentration at lower doses only, perhaps indicating a threshold for particle uptake by the DCs. We also observed an increase in cell granularity that indicate a growth in internal cell complexity, which is supported by the confocal studies that indicate successful uptake of nanogels by the DCs.
Previous studies have demonstrated that hydrophobic material could potentially serve as an adjuvant [18, 28]. A recent study demonstrated that, if exposed, the hydrophobic component of any material can serve as damage-associated molecular pattern, and a high concentration of such materials could activate the corresponding TLR receptor . Since pHEMA-pyridine is functionalized with hydrophobic side groups, we hypothesized that the nanogel could potentially immune-modulate DCs by upregulating the surface expression of co-stimulatory molecules, such as CD80 or CD86, and antigen-presenting major histocompatibility complex (MHC)-II molecules. We observed an increase in CD 80 and MHC-II surface activation markers with increasing dose of nanogels (Figure 2D, E). At the low protein nanogel dose of 75,000 nanogels/mL (22 μg/mL protein), the CD80 and MHCII expression level were found to be 2.4 ± 0.05 and 5.2 ± 0.9 fold, respectively, as compared to untreated controls, respectively, and were higher than lipopolysaccharide (LPS). As we increased the nanogel dose to 450,000 nanogels/mL (135 μg/mL protein), CD80 and MHC-II expression further increased to 6.8 ± 0.1 fold and 19.2 ± 0.4, respectively. These findings supported our hypothesis that high doses of nanogels would enhance the activation of DCs in the absence of immune-stimulatory antigens.
Since TLR immune-modulators like CpG oligonucleotides and poly I:C are often used to facilitate stronger immunological reactions, we determined the effect of soluble TLR agonist on DCs as compared to nanogel. DCs were incubated with FITC-BSA nanogels or soluble FITC-BSA with soluble poly I:C that functions as a TLR3 agonist and helps in antigen cross-presentation to T cells. We observed a significant reduction in MHC-II and CD80 as compared to soluble poly I:C (Figure 2F).
The ability of DCs to process antigen and presentation to CD8+ T cells is the hallmark of cancer immunotherapy. To determine the cross presentation of antigen from nanogels to support cytotoxic T cell stimulation, we engineered OVA-nanogels and incubated with JAWSII DCs for 48 hr. The activated DCs were then co-incubated with CFSE-labeled CD8+ T cells isolated from OT-I OVA mice at 4:1 T:DC cell ratio for 72 hr. Flow cytometry analysis of CFSE indicated that stimulation of DCs with OVA-nanogel at 135 μg/mL OVA concentration resulted in 35,500 ± 1,500 OT-I CD8+ T cells which was significantly higher than soluble OVA group (22,800 ± 2,500 OT-I CD8+ T cells) and LPS (p < 0.05; Figure 3A). Increasing the OVA dose to 270 μg/mL in nanogels resulted in similar level of T cell proliferation as observed with 135 μg/mL OVA-nanogel (p > 0.05). Interestingly, with 90 μg/mL OVA-nanogel concentration, the T cell proliferation was significantly higher than a higher dose of soluble OVA (135 μg/mL), clearly indicating that using nanogels, a lower dose of antigen results in superior immune-modulation and antigen-specific T cell proliferation. We next evaluated the influence of poly I:C on cross-presentation and T cell stimulation and compared soluble OVA + polyI:C to OVA-nanogels. As indicated in Figure 3B, OVA-nanogels alone induced comparable response to that observed with exogenous addition of poly I:C to soluble OVA antigens. We further confirmed that the observed effect was specific to the OVA-nanogel and not any bacterial endotoxins. Observations from LAL assay indicated no detectable endotoxin units (EU)/mL for OVA-nanogels at 45-135 μg/mL dose.
Our results indicated that the proteins could be released from the nanogels intracellularly, processed by DCs, and cross-presented to cytotoxic T cells. Activation of CD8+ T cells can only occur when they recognize small peptides from larger proteins presented to them by MHC class I molecules. These small peptides are generated by processing of antigen, which can occur by 2 well-described pathways, an endogenous pathway and an exogenous/cross presentation pathway. The endogenous pathway processes protein antigens that are synthesized by the cell, and the resulting peptides get access to MHC class I molecules that then can activate CD8+ T cells. In the exogenous/cross presentation pathway DCs pick up antigen from the outside, and process them into smaller peptides that can enter the cross presentation pathway, and get access to MHC class I molecules which can activate CD8+ T cells (see Ref. ). The ability of DCs exposed to the nanogel to activate CD8+ T cells is a direct effect of the antigen getting access to the cross presentation pathway, since the OVA protein needs to be processed by the pathway before CD8+ T cells are able to recognize it and become activated. The evidence for this is: (1) the presence of fluorescent particle within DCs as indicated by the confocal imaging showing that the nanogel has gotten inside the DC, (2) the significantly higher number of antigen-specific T cells following co-culture with DCs that received the whole OVA antigen indicating that these CD8+ T cells have been activated and have divided. These CD8+ T cells from OT-I mice carry a T-cell Receptor that is specific for a particular peptide that is generated only when the OVA protein has been processed (“SIINFEKL”) and presented by H-2Kb of MHC molecules. Taken together, our results demonstrated that pHEMA-pyridine antigen interaction results in nanogel vaccines. Importantly, nanogels-delivered antigens induced robust DC activation and T cell response at lower antigen doses than soluble antigens, therefore making the nanogels a safer immunotherapy approach. The nanogel approach makes it a versatile system to simply use any protein antigen (tumor or infection or autoimmune diseases) and engineer nanogels by a self-assembly process without exposing to solvents, emulsification process, or sonication. At the same time, the protein-nanogel approach is relevant to a wide range of therapeutic protein delivery applications for regenerative medicine. Future studies will determine mechanistic understanding of nanogel and immune pathway (TLR) interactions and establish the immunomodulatory nanogels as prophylactic and/or therapeutic cancer vaccines in animal models.
pHEMA-pyridine synthesis was carried out through esterification of the hydroxyl group on pHEMA. Briefly, pHEMA (150 mg, 1.2 mmol OH residues) was dispersed in anhydrous tetrahydrofuran (THF, 5 mL) at room temperature. DMAP (451 mg, 3.6 mmol) was added, and the mixture was allowed to stir at room temperature until the solution became homogeneous. Nicotinoyl chloride hydrochloride (411 mg, 2.3 mmol) was dispersed in THF and added dropwise to the pHEMA solution mixture under inert atmosphere (N2). The esterification was allowed to proceed at room temperature for 48 h. A precipitate appeared within 24 h and was separated from the liquid by centrifugation at the end of the reaction. The precipitate was further washed with THF, and the supernatant was combined together and precipitated in cold diethyl ether. The final solid precipitate was rinsed thoroughly with ether three times to remove excess DMAP. Finally, the precipitate was collected and dried under vacuum for 3 days at room temperature.
Nanogels were formed by dropwise addition of 0.2 mL of pHEMA-pyridine (2 mg/mL in Dimethylformamide (DMF)) into a 5 mL solution of Ovalbumin protein as a model antigen (or BSA as a control) in PBS under stirring at 900 rpm at room temperature. The resulting product was separated from unreacted components by centrifugation, washed, was then lyophilized for 24 hours using LabConco FreeZone Freeze Dry System. Particle characterization of size, zeta potential, and electrophoretic mobility was performed using Zetasizer Nano ZS (Malvern Instruments). Analysis on particle number and mean fluorescence intensity of FITC-BSA-nanogels were performed using a BD C6 Accuri Flow Cytometer. Reported protein concentrations represent the original concentration of protein solution mixed with the polymer prior to nanogel formation. Nanogel counts were performed using flow cytometry. We determined the protein loading efficiency using FITC-BSA nanogels and a Biotek Synergy™ H1 plate reader (488 nm excitation and 525 nm emission). Protein amount was quantified using a standard curve (0-60 μg/ml of FITC-BSA). FITC-BSA nanogels were prepared using 60, 120, and 240 μg/mL stock protein solution concentrations with 0.2 mL of 2 mg/mL pHEMA-pyridine polymer. Samples were centrifuged 3 times, thoroughly washed after the removal of supernatant, followed by resuspension in fresh PBS.
T cell studies were performed by incubating JAWSII DCs with different concentration of OVA-nanogels and/or 40 μg/ml TLR agonist for 48 hours, followed by co-incubation with CFSE-labelled OT-I CD8+ T cells for 72 hours. The co-culture was initiated by adding T cells (100,000 cells in 100 μL of RPMI media) into primed DCs (25,000 cells in 100 μL of DC media) cultured in 96 well plates. The primed DCs were washed twice with 200 μL of PBS following priming period prior to the co-culture. Viable T-cell proliferation was evaluated using flow cytometry as reported earlier by us .
Analysis of variance (ANOVA) statistical analyses was performed using GraphPad Prism software with Tukey's test (1-way ANOVA) or the Bonferroni correction (2-way ANOVA). A p-value of less than 0.05 was considered significant. All studies were performed in triplicates unless otherwise noted. All values are reported as Mean ± S.E.M. For more details on experiments, see supplementary methods.
The authors acknowledge partial financial support from the National Institutes of Health (1R21CA185236-01, A.S.), the National Science Foundation (DMR-1554275, A.S.), the Sibley School of Mechanical and Aerospace Engineering at Cornell University, the Howard Hughes Medical Institute (HHMI 56006761, A.P.), and Cornell University's Dean's excellence fellowship (K.R). The authors acknowledge assistance from Dr. Ivan Keresztes at Cornell NMR facility for the data acquisition and analysis. We acknowledge the Cornell University Biotechnology Resource Center (BRC) Imaging Facility and NIH 1S10RR025502 for data collected on the Zeiss LSM 710 Confocal. We acknowledge Cornell University Nanobiotechnology Center (NBTC) for access to shared facilities. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.
Alberto Purwada, Meinig School of Biomedical Engineering, College of Engineering, Cornell University, Ithaca, NY 14853, USA.
Ye F. Tian, Sibley School of Mechanical and Aerospace Engineering, College of Engineering, Cornell University, Ithaca, NY 14853, USA.
Weishan Huang, Department of Microbiology and Immunology, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853, USA.
Kathleen M. Rohrbach, Department of Material Science and Engineering, College of Engineering, Cornell University, Ithaca, NY 14853, USA.
Simrita Deol, Department of Microbiology and Immunology, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853, USA.
Prof. Avery August, Department of Microbiology and Immunology, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853, USA.
Prof. Ankur Singh, Sibley School of Mechanical and Aerospace Engineering, College of Engineering, Cornell University, Ithaca, NY 14853, USA.