Background
In recent years, peptides derived from tumor-associated antigens (TAA) have been identified for a variety of human cancers [
1]. Thus far, however, effective peptide vaccination of patients with cancer has been limited to very few trials [
2]. The relative paucity of responsiveness after conventional peptide vaccination is mostly due to the high levels of protein degradation, limiting antigen delivery. Polymeric nanoparticles (NP) may allow encapsulation of peptides inside a polymeric matrix, protecting them against enzymatic and hydrolytic degradation. In addition, the nanoparticle vaccine approach offers the possibility of developing tailor-made vaccines containing specific targets or molecules that may improve their function [
3].
Dendritic cells (DC) are the most potent professional antigen-presenting cells (APC), having the ability to initiate primary immune responses [
4]. Therefore, immunotherapy utilizing DC has become a promising therapeutic modality in recent years [
5‐
7]. However, the lack of efficient and long-lasting antigen presentation by DC
in vivo has been a major difficulty in the development of effective vaccines. These obstacles could be circumvented through the development of nanoparticles, which can efficiently deliver the antigenic peptides into the APC.
Our recent studies have characterized distinct subsets of regulatory CD4+FOXP3- and CD8αα+TCRαβ+ T cells that target autoaggressive Vβ8.2+ T cell responses for down-regulation and protection against autoimmune disease [
8‐
13]. Several non-classical MHC class I, Qa-1a-restricted CD8αα+TCRαβ+ T cell and MHC class II-restricted CD4+ T cell clones and lines have been characterized [
8,
9,
11]. The CD4+ T cells recognize a TCR-derived peptide in the context of a class II MHC molecule, I-Au [
11]. However, CD8αα+TCRαβ+ T cells are cytotoxic and recognize another TCR-derived peptide bound to a class Ib molecule, Qa-1 [
8,
9]. These cells are physiologically primed and operate in unison to assist in recovery from T cell-mediated experimental autoimmune encephalomyelitis in H-2
u mice [
10,
13,
14]. Interestingly, recent data suggest that class Ib MHC-restricted cytotoxic CD8+ T cells also play an important role in anti-tumor immunity [
15,
16]. Therefore, it is important to examine whether CTL can be effectively primed using nanoparticle containing class Ib-binding peptides.
To engineer our nanoparticle-based vaccines, we utilized the biodegradable and biocompatible polymer PLGA, which is already approved by the US FDA [
17,
18]. PLGA is easy to formulate into different devices and has generated immense interest because of its favorable properties, which include good biocompatibility, biodegradability, and mechanical strength [
17,
18].
Here we demonstrate for the first time efficient nanoparticle-facilitated loading of class I-restricted, clinically relevant TAA-derived peptides to human DC, and the development of nanoparticles incorporating MHC class II and non-classical MHC class I, Qa-1-binding peptides that are able to stimulate helper CD4+ and cytotoxic CD8αα+TCRαβ+ T cells. Importantly, we confirmed that the enhanced peptide presentation by the NP-loaded DC is associated with more efficient generation of antitumor CTL.
Methods
Antibodies and reagents
Antibodies: Anti-human IFN-γ (mAb 1-D1K) - Mabtech Inc. (Mariemont, OH), anti-HLA-A2-FITC, anti-HLA-DR-PE, anti-CD83-PE, anti-CD80-FITC, anti-CD86-FITC, mouse-IgG1-FITC, mouse-IgG1-PE, and mouse-IgG2a-PE all from BD (San Diego, CA). Cytokines: GM-CSF - BERLEX (Richmond, CA), Interleukin 4 (IL-4) - PeproTech (Rocky Hill, NJ), IL-2 - Chiron (Emeryville, CA). Peptides: MART-127-35, gp100209-217 and mSTEAP326-335 - GenScript Corp. (Piscataway, NJ), murine TCR Vβ8.2 chain peptides: B5 (76-101) - Caltech (Pasadena, CA), and p42 (42-50) - Synthetic Biomolecules (San Diego, CA). PLGA - Birmingham Polymers (Birmingham, AL), bovine serum albumin (BSA) and Poly (vinyl alcohol) (PVA) - Sigma-Aldrich (St. Louis, MO), coumarin 6 - Polyscience (Warrington, PA).
Cell lines
T2 cell line [
19] was purchased from ATCC (Manassas, VA). Human melanoma cell lines 624 and 1351, as well as human tumor-infiltrating lymphocytes (TIL) cell lines TIL1235 and TIL1520 were kindly provided by Dr. John R Wunderlich (NIH/NCI, Bethesda, MD). CD4+TCRαβ+ (B5.1) and CD8αα+TCRαβ+ (XT14), [
8] T cell lines were generated in H-2u background and were specific for the Vβ8.2TCR-derived peptides, B5 and p42-50, respectively. The CD4+TCRαβ+ (B5.1) cell line was generated from the draining lymph nodes of a H-2u mouse immunized i.p. with 20 μg of TCR peptide B5.
Nanoparticle formulation
Peptide-containing NP were formulated using a double emulsion-solvent evaporation technique as we described previously [
20]. For optimizing the peptide dose entrapped in the NP, 300 μg, 600 μg and 1 mg of each peptide was formulated into the PLGA polymer in each NP batch. Some NP were formulated with a fluorescent dye (coumarin 6) by adding 100 μg of coumarin 6 to the polymer solution prior to emulsification.
Nanoparticle characterization
To characterize the surface morphology of the NP we utilized a scanning electron microscope (SEM). Particle size analysis and zeta potential determination was carried out using a Zetasizer (Malvern, Worcestershire, UK). The peptide content of the peptide-loaded NP was determined by HPLC using a C18 column (Waters, Milford, MA). Specifically, the peptides and nanoparticles were separated and identified using ultraviolet (UV) detection and known standards, at a wavelength of 280nm (attenuation 0.002 AU). An aliquot (50 μl) was injected onto the column and eluted with a mobile phase containing a gradient mixture of reagent A, 0.1% trifluoroacetic acid (TFA) in water (Sigma Aldrich St Louis, MO, USA), and reagent B, 0.1% TFA in Acetonitrile. The gradient times were as follows: 0-23 minutes, 75% A and 25% B; 23 -25 minutes, 0% A and 100% B. Total run time was 25 minutes at a flow rate of 0.8 ml per minute.
Generation of Human DC
Peripheral blood mononuclear cells (PBMC) isolated from buffy coats of healthy donors were allowed to adhere in 6-well plates for 1 hour. The adherent cells were cultured with 1000 U/ml GM-CSF and 300 U/ml IL-4, with cytokines added on days 2, 4, and 6. Lipopolysaccharide (LPS) was added to the culture medium on day 7, and two days later, the mature DC (mDC) were harvested and characterized by FACS using antibodies against HLA-DR, CD80, CD83, and CD86.
Generation of murine bone marrow-derived dendritic cells (BMDC)
Murine DC were derived from tibias and femurs by flushing out the bone marrow with RPMI medium as described [
21] and cultured with 10 ng/ml IL-4 and 25 ng/ml GM-CSF for 5 days, with cytokines added on day 3.
Nanoparticle uptake imaging studies
Human immature DC (imDC) were seeded overnight on sterile cover slips and incubated with NP containing coumarin-6 for 1-hr at 37°C. Then, the cover slips were washed and observed with a fluorescent microscope. For confocal microscopy, imDC were incubated in 4-well chamber slides for 1 hour at 37°C with 100 μg/ml Coumarin 6-containing NP, washed and fixed with paraformaldehyde. After washing and staining with DAPI, imDC were mounted on glass slides. Confocal images were obtained with a Leica TCS SP2 UV confocal microscope.
FACS analysis of NP-loaded DC
NP-loaded and non-loaded DC were stained with the following antibodies for 30 min at 4°C: PE-anti-human HLA-DR, PE-anti-human CD83, FITC-anti-human CD80, and FITC-anti-human CD86. All data was analyzed using the Cell Quest software.
Antigen presentation by the NP-loaded DC
Human imDC were collected and pulsed with the peptides MART-127-35, or gp100209-217, or incubated for 1 hour with 100 μg/ml nanoparticles formulated with the same peptides at 300 μg, 600 μg or 1 mg peptide per batch. LPS (100 ng/ml) was subsequently added. Two days later, the mDC were harvested, tested by FACS, and co-cultured for 20 hours with TIL1235 (recognizing MART-127-35) or TIL1520 (recognizing gp100209-217), and the efficiency of the antigen presentation was evaluated in an IFN-γ ELISPOT assay.
Induction of CTL with the NP-loaded mDC
The mDC were mixed with HLA-A2+/CD8+ T cells at a ratio of 1:10 in complete medium, and incubated at 37°C. Four days later, 20 U/ml of IL-2 and 30 U/ml of IL-7 were added. On days 7 and 14, the cultures were re-stimulated with peptide-pulsed adherent autologous CD8− cells in complete medium. Specifically, irradiated CD8− cells were incubated for 2 hours with β-2 microglobulin (at 5 μg/ml) and peptide (at 5 μg/ml), washed once and used as stimulators of the CTL. Seven days later, the CTL were tested by IFN-γ ELISPOT assays and CytoTox96 cytotoxicity assays.
Discussion
Epitope-based peptide vaccines can be designed to include multiple epitopes from one or several antigens, and can be easily analyzed for purity and produced economically on a large scale. Currently however, there are no human peptide-based cancer vaccines on the market, mostly resulting from difficulties associated with their poor immunogenicity, stability and delivery. We [
22,
23] and others [
24,
25] have described strategies to enhance the peptides' immunogenicity and stability. Direct peptide delivery to dendritic cells using particulate delivery systems is a promising new approach. In addition to having a depot effect on the peptide antigens, inherent properties of the particles themselves engender immunogenicity of the peptides, and allow uptake of an immunogenic package of peptides and other molecules [
26]. This approach is exemplified with the use of liposomes and immunostimulatory complexes [
27], as well as virosomes [
28] and exosomes [
29].
Cultured DC are very suitable for the delivery of peptide vaccines, as their direct loading bypasses the processing requirements and allows for precise delivery of peptide antigens to the immune system [
30]. However, the potency of DC-based vaccines is significantly reduced by the short persistence of the MHC/peptide/β2-microglobulin complexes on the DC surface, especially when the antigen-derived peptides are bound from the outside and not processed [
31].
It is therefore essential that the vaccine-carrier systems are capable of delivering the vaccines inside the APCs, in order to facilitate a potent and prolonged antigen presentation. Liposome-based systems are commonly used, but their delivery efficiency is sub-optimal, and the duration of their effects is relatively short [
32]. In addition, the liposome is not a thermodynamically stable system and therefore has multiple physical stability issues such as rapid drug leakage, merging of vesicles and low loading efficiency. The viral vector delivery is limited by a difficult large-scale production and potential for toxicity [
33], immune and inflammatory responses [
34], as well as insertional mutagenesis and oncogenic effects [
35].
Nanoparticle-based vaccine delivery systems offer significant advantages due to their safety profile, ease of manufacture and storage, and most importantly, their versatility in designing customized products for specific targeting applications [
36]. We developed previously PLGA nanoparticles designed for sustained release of drugs or DNA to human umbilical vein endothelial cells [
37], and prostate cancer cells [
20]. In the current study, we used PLGA-based nanoparticles as a delivery system to load clinically-relevant, tumor antigen-derived peptides into human DC, and found that 100% of imDC internalized NP after just 1-hour co-incubation with the nanoparticles (Figure
2). Using our new PLGA nanoparticles containing MHC class I peptides as a vaccine delivery system offers distinct advantages over the administration of the corresponding soluble peptides. These NP are composed of solid PLGA polymers, therefore avoiding the drug-leakage problems involved in liposome formulations, thus preventing the proteolytic degradation of the antigen. These polymeric NP can also be easily lyophilized for long-term storage with better stability than liposome and other liquid carrier systems.
We also considered the influence of PLGA uptake on the properties of the DC, and its potential negative effect on some important parameters for the DC function. The hydrolysis of PLGA leads to the liberation of lactic and glycolic acids, and therefore we expected that the resulting acidification could negatively affect the cellular functions of the PLGA-loaded DC. It this study we did not observe any negative effects and/or reduced viability of the PLGA-loaded DC. We also found that our NP formulations did not have an effect on the maturation of human or murine DC. Our findings are in agreement with a similar study, which reported that immature DC loaded with PLGA particles exhibited a similar DC phenotype to those without any loading [
38]. In contrast, another study showed that a maturation process has been induced by polystyrene nanospheres, as the maturation markers HLA-DR and CD86 were upregulated [
39]. A similar result was observed using PLGA nanoparticle formulations in cord blood derived DC [
40], as well as murine bone marrow derived DC [
41]. These discrepancies are most likely due to the different culture conditions and differences in the nanoparticle preparations in these studies. We conclude that the intake of our PLGA NP does not adversely affect important DC functions required for their use as vaccines in the clinic.
In the present study, the efficiency of the antigen presentation by human DC was significantly enhanced after just 1-hour incubation with our NP containing class I-restricted peptides (Figure
4A and
4b). The level of antigen presentation was related to the amount of peptide incorporated inside the NP. Importantly, in these studies we used patient-derived TIL lines and peptides used in many clinical trials, which suggests a speedy utilization of these data in clinical trial designs.
For the presentation of MHC class I restricted T cell epitopes from PLGA-encapsulated peptides, the involved antigen presenting cells must be able to "cross present" the exogenous peptides onto MHC class I molecules by either the classical proteasome and TAP-dependent pathway [
42], or by an alternative TAP-independent pathway [
22,
23] of antigen presentation. DC, macrophages and some endothelial cells have been shown to be able to cross present [
43]. Cross presentation of soluble proteins by DC can occur, but it is extremely inefficient, as it usually requires the incubation with high concentrations of protein antigens. Remarkably, in this study the amount of peptides encapsulated inside our NP was significantly lower than the amount of peptides used to pulse the DC externally. Still, we observed a greatly enhanced antigen presentation by the NP-loaded DC. We suggest that this enhancement of antigen presentation is due to the slow hydrolysis of the PLGA NP in the endosomes the DC, which provides a continuous supply of peptide ligands for newly synthesized MHC class I and II molecules. We also demonstrated that the NP-loaded DC were able to induce more potent CTL than the DC pulsed with the same peptides externally (Figure
4C). The resulting CTL were able to recognize and kill efficiently not only peptide-pulsed target cells, but also HLA-matched melanoma cells expressing the corresponding antigens. Similarly, murine DC could present antigens more efficiently as a result of the antigen encapsulation inside our nanoparticles (Figure
5). These results provide further confirmation of the usefulness of this approach for induction of potent and specific anti-tumor responses, and its potential for clinical application.
Our nanoparticles were also effective in stimulating class II-restricted CD4+ T cells as well as non-classical class I-restricted CD8αα+TCRαβ+ cytotoxic T cells. Notably, these CD4+ and CD8αα+TCRαβ+ T cells are involved in the negative feed back regulation of autoimmunity [
8,
9,
11‐
13]. Our earlier studies have shown that priming of these regulatory T cells following peptide or DNA immunization results in immune regulation [
8,
11,
12,
44]. Data presented in this paper further indicate that nanoparticles containing appropriate peptides can be used to generate effective vaccines not only against tumors but also in the intervention of autoimmune diseases. Class Ib MHC-restricted cytotoxic T cells also play an important part in the anti-tumor responses [
15,
16]. It is clear from our data that nanoparticles containing class Ib MHC (Qa-1 or HLA-E in humans) binding tumor-associated antigens can also be designed.
Conclusions
The development of nanoparticle-based vaccines derived from clinically relevant tumor antigens holds great promise. Encapsulating antigens in PLGA nanoparticles offers unique advantages such as higher efficiency of antigen loading, prolonged presentation of the antigens, prevention of peptide degradation, specific targeting of antigens to APC, improved shelf life of the antigens, and easy scale up for pharmaceutical production. In addition, a variety of targeting strategies may be readily utilized, including ligand-receptor mediated targeting, antibody-antigen interaction, lectin-carbohydrate interaction, etc. Recent advances in polymer chemistry also allow for many variations of the nanoparticle design, including simultaneous delivery of a combination of vaccines, immunomodulators, drugs or other compounds, creating a potent multivalent therapeutic strategy. This paper is therefore highly significant to the development of optimized clinical grade vaccines, and the induction of CTL for adoptive immunotherapy of cancer.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
WM carried out and participated in all of the studies, including nanoparticle preparation and characterization, DC isolation and loading, CTL induction in vitro, data analysis and manuscript preparation. TS carried out murine DC isolation and loading, analysis of presentation of murine MHC class II-restricted peptides and murine non-classical MHC class I, Qa-1-restricted peptides encapsulated into nanoparticles, and data analysis. VB participated in the design of the study, helped with the statistical analysis and manuscript preparation. YZ participated in nanoparticle characterization, DC loading and imaging and data analysis. CO YZ participated in nanoparticle characterization, DC imaging and data analysis. MO participated in nanoparticle characterization, DC loading and imaging and data analysis. MH participated in DC loading and CTL induction. SS participated in DC loading and CTL induction. EC participated in the design of the study, helped with the statistical analysis and manuscript preparation. DM participated in nanoparticle characterization, DC loading and data analysis. VK participated in data analysis and supervised studies related to murine class II and Qa-1-restricted T cell presentation. BM designed, supervised and coordinated the study, performed the statistical analysis and drafted the manuscript. All authors read and approved the final manuscript.