Effects of particle size on toll-like receptor 9-mediated cytokine profiles

Abstract

Biomaterials interface with toll-like receptor (TLR) 9-mediated innate immunity in a wide range of medical applications, such as tissue implants and drug delivery systems. The stimulation of TLR9 can lead to two different signaling pathways, resulting in the generation of proinflammatory cytokines (i.e. IL-6) and/or type I interferons (IFNs, i.e. IFN-α). These two categories of cytokines differentially influence both innate and adaptive immunity. Although particle size is known to be a critical parameter of biomaterials, its role in TLR9-mediated cytokine profiles is not clear. Here, we examined how the size of biomaterials impacted cytokine profiles by using polystyrene particles of defined sizes as model carriers for TLR9 agonists (CpG oligonucleotides (CpG ODNs)). CpG ODNs bound to nano- to submicro- particles stimulated the production of both IL-6 and IFN-α, while those bound to microparticles resulted in IL-6 secretions only. The differential TLR9-mediated cytokine profiles were attributed to the pH of endosomes that particles trafficked to. The magnitude of IFN-α production was highly sensitive to the change in endosomal pH in comparison to that of IL-6. Our results define two critical design variables, size and the ability to modulate endosomal pH, for the engineering of biomaterials that potentially interface with TLR9-mediated innate immunity. The fine control of these two variables will allow us to fully exploit the beneficial facets of TLR9-mediated innate immunity while minimizing undesirable side effects.

1. Introduction

Toll-like receptors (TLRs) have been well recognized for their pivotal role in the detection of pathogen associated molecular patterns (PAMPs). Twelve TLRs, each detecting a distinct set of PAMPs, have been identified in mammals [1]. TLR9 was first identified as a sensor for bacterial DNA [2], which contains unmethylated CpG motifs. Subsequently, a broad range of natural TLR9 ligands has been identified, including DNA derived from dead cells [3, 4], fungi [5, 6], viruses [7–9], and chromatin-IgG complexes [10, 11]. A recent study has suggested that TLR9 can detect any single stranded DNA [12]. Since its initial discovery, TLR9 has become one of the central targets for mediating immune responses. Many different types of synthetic TLR9 ligands, CpG motif-containing oligonuleotides (CpG ODNs), have been characterized and utilized as adjuvants for vaccines [13–16] and microbicides [17–20] against pathogen infections.

The role of TLR9 in the design of biomaterials for medical applications is implicated in three main areas: (1) particulate systems that deliver TLR9 ligands for vaccination or microbicidal activities [21–26]; (2) non-viral DNA delivery systems, which can control the interaction of DNA with TLR9 and thus minimize their immunogenicity for their application in gene therapy [27–31]; (3) tissue implants, which release degraded or eroded biomaterial fragments. These fragments are potentially deposited with DNA derived from bacteria or dead cells and endocytosed by phagocytic cells (such as macrophages). As a result, the endocytosed DNA can trigger the TLR9-based innate immune system and lead to the failure of tissue implants [32–34].

TLR9 is situated in the endoplasmic reticulum (ER) and recruited to endosomal/lysosomal compartments upon the endocytosis of pathogens or particulates [35]. The stimulation of TLR9 can lead to two distinct signaling pathways in which two transcription factors, NF-κB and IRF-7, are translocated into the nucleus, resulting in the generation of proinflammatory cytokines (such as IL-6 and IL-12) and/or type I interferons (IFNs) (such as IFN-α), respectively [1]. These two types of cytokines play independent roles in the induction and maintenance of innate and adaptive immune responses. For example, IL-6 is a key mediator of the acute phase response [36]. IL-6 is critical in the proliferation, differentiation and survival of B cells, CD4⁺ T cell differentiation to Th2 cells [37]. Type I IFNs, such as IFN-α, inhibit viral replication and induce apoptosis to clear infected cells. IFN-α promotes the proliferation and function of CD8⁺ T cells, maintains CD4⁺ Th1/Th2 cell populations. IFN-α plays a key role in the pathogenesis of the autoimmune disease, systemic lupus erythematosus (SLE) as well [38]. Additionally, proinflammatory cytokines and type I IFNs play differential roles in bone tissue engineering. IL-6 promotes osteoclast formation [39], whereas type I IFNs, such as IFN-α and IFN-β, inhibit osteoclast differentiation [40]. Due to the functional diversity of proinflammatory cytokines and type I IFNs, the ability to control the TLR9-mediated signaling pathways is essential for the design of biomaterials that interface with TLR9-based innate immunity.

The size of biomaterials is considered to be critical for the three aforementioned medical applications because it affects the level of retention in tissues [41, 42] and the efficiency of cellular uptake [43–45]. Others and we have demonstrated that the size of biomaterials also affects their intracellular trafficking [46–48]. Given the critical role of the intracellular location of TLR9 ligands which dictates the subsequent induction of signaling pathways [49–51], we hypothesized that the size of biomaterials could affect the TLR9-mediated cytokine profiles. The cytokine profiles in turn would impact the immunogenicity of biomaterials and lead to either beneficial or detrimental effects depending on the clinical applications of the designed biomaterials.

To test this hypothesis, in this study, we used polystyrene particles of defined sizes as model carriers for CpG ODNs to investigate the effect of particle size on the cytokine profiles mediated through the TLR9 signaling pathway and possible mechanisms involved.

2. Materials and Methods

2.1 Cell culture

A dendritic cell line, BC-1 (a gift from Dr. Yoshiki Yanagawa), was maintained in Isocove’s Modified Dulbecco’s Medium (IMDM) supplemented with 20% NIH-3T3 supernatant, 4 mM L-glutamine, 50 μM 2-mercaptoethanol, 5% granulocyte-macrophage colony-stimulating factor (GM-CSF)–containing supernatants, 100 units/ml penicillin and 100 μg/ml streptomycin. The fibroblast cell line, NIH-3T3, was maintained in IMDM supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin and 100 μg/ml streptomycin. Cells were maintained at 37 °C under an atmosphere with 5% CO₂. 5% GM-CSF-containing supernatants were obtained from GM-CSF conditioned medium from the GM-CSF-secreting mouse cell line J558L.

2.2 Preparation and characterization of ODN-coated PS particles (ODN-PLL-PS)

Fluorescent polystyrene (PS) particles (Polysciences, Inc., Warrington, PA) of different sizes, 100 nm, 500 nm, and 3 μm in diameter, were used as model CpG ODN carriers. The particles were first coated with poly-L-lysine (PLL; Sigma-Aldrich, Inc., Saint Louis, MO; MW = 30,000–70,000 g/mol), a cationic polymer, to facilitate the adsorption of ODNs onto particles. PLL-coated particles are designated as PLL-PS. For 100 nm particles, 250 μl of PS (2.8×10¹² particles per ml in 10 mM NaCl) was added drop-wise to an equal volume of PLL (0.8 mg/ml in 10 mM NaCl) under constant vortexing. For 500 nm and 3 μm particles, 1 ml of PS (2×10¹⁰ 500 nm particles per ml, 5×10⁸ 3 μm particles per ml) was added drop-wise to 1 ml of PLL (0.15 mg/ml (500 nm), 0.25 mg/ml (3 μm)). The final reaction volume was 500 μl for 100 nm particles, and 2 ml for 500 nm and 3 μm particles. The adsorption was conducted for 1 h at room temperature (RT) and the reaction vessel was vortexed every 15 min. PLL-PS was washed three times in 10 mM NaCl by centrifugation at 4 °C, 8000 rcf for 2 h, 300 rcf for 1 h, and 25 rcf for 30 min for 100 nm, 500 nm, and 3 μm particles, respectively. 60% of 100 and 500 nm particles were recovered, and 100% of 3 μm particles were recovered. The duration and centrifuge speed were optimized to obtain mono-dispersed particles. CpG ODNs 1826 and 2216 (TriLink BioTechnologies, San Diego, CA) were attached onto PLL-PS by using the same procedures and reaction volumes as used for the coating of PLL onto particles. The final reaction mixtures contained 4×10¹¹ particles per ml and 86 μg/ml ODN-1826, 4.5×10⁹ particles per ml and 40 μg/ml ODN-1826, and 1×10⁸ particles per ml and 33 μg/ml ODN-1826 for 100 nm, 500 nm, and 3 μm particles, respectively. The attachment of ODN 2216 onto PLL-PS was similar to that of ODN 1826. ODN-coated particles (ODN-PLL-PS are denoted as 1826-PLL-PS or 2216-PLL-PS, depending on the type of CpG ODN) were washed in the same manner as stated above. All of the washes were collected for further analysis of non-bound ODNs. The sequences of ODNs 1826 and 2216 were: 5′tcc atg acg ttc ctg acg tT3′ and 5′ggG GGA CGA TCG TCg ggg gG3′, respectively. Lowercase letters denote a phosphorothioate backone, and uppercase letters denote a phosphodiester backbone.

A Zetasizer Nano ZS (Malvern Instruments, Westborough, MA) was used to characterize the size distribution and zeta potential of the particles. 100 nm, 500 nm, and 3 μm particles were prepared at 3×10¹⁰, 5×10⁷, and 4×10⁶ particles per ml, respectively, in 10 mM NaCl with a pH of 5.9. Scanning electron microscopy (SEM) was used to confirm the particle size. SEM samples of particles were prepared in Milli-Q water, dried, sputter-coated with 12 nm of platinum using a SPI Sputter Coater (Structure Probe, Inc., West Chester, PA), and analyzed using a JEOL 7000 SEM with a beam voltage of 10.0 kV (Electron Microscopy Center, University of Washington).

2.3 Quantification of the surface density of ODNs on ODN-PLL-PS particles

The surface density of ODNs was determined by quantifying both the total amount of ODNs adsorbed and the number of fluorescent particles. The OliGreen Assay Kit for ssDNA (Invitrogen, Frederick, MD) was used to measure non-bound ODNs in the adsorption reaction solutions and washes for the indirect determination of the amount of ODNs adsorbed onto particles. The number of 100 and 500 nm particles was quantified by fluorescence spectrometry. For 100 nm particles, the particles were suspended in 10 mM NaCl and the fluorescent intensity of the particle suspension was directly measured. For 500 nm particles, particles were dried under a stream of air, dried for an additional 4 h at RT, and then dissolved in chloroform overnight. The fluorescent intensity of dissolved particles in chloroform was subsequently determined. For both sizes of particles, a series of standards with a known number of particles were used to construct standard curves. 3 μm particles were directly counted by a hemocytometer.

2.4 Regulation and measurement of endosomal pH

The pH of endosomes was controlled by exposing DCs to medium containing a weak acid, NH₄Cl [52]. The endosomal pH was measured by using a fluorescence ratiometric method that has been previously established [46]. Briefly, 100 nm amine-functionalized polystyrene particles (Polysciences, Inc., Warrington, PA) were reacted with pH-sensitive fluorescein isothiocyanate (FITC; Sigma-Aldrich, Inc., Saint Louis, MO), and pH-insensitive Alexa Fluor 647 carboxylic acid, succinimidyl ester (Invitrogen, Frederick, MD) for 4 h at RT in 0.1 M sodium carbonate buffer (pH = 8.3). The conjugation reaction was quenched with the addition of NH₄Cl, and particles were washed three times in phosphate buffered saline (PBS, pH = 7.4). To ensure the equivalent uptake of particles, DCs were seeded in a 6-well plate at a density of 3×10⁶ cells per well. DCs were pulsed with particles for 6 h. Afterwards, cells were harvested and aliquoted at 2×10⁵ cells per aliquot. One set of aliquots was incubated in different concentrations of NH₄Cl in IMDM supplemented with 0.5% FBS for 1 h at 37 °C. The other set, used for constructing a standard curve, was fixed with 4% paraformaldehyde (PFA) for 20 min at RT, permeabilized with 0.1% Triton X-100 for 2 min at RT, and incubated in 0.1 M citrate buffer solutions of known pH values for 1 h at RT. Citrate buffer solutions were prepared in Dulbecco’s phosphate buffered saline (DPBS). Cell suspensions were analyzed using a FACsCanto (Cell Analysis Facility, University of Washington). The geometric mean fluorescent intensity (gMFI) of cells was determined by FlowJo (Treestar, Inc., Ashland, OR). The ratio of gMFI of FITC to that of Alexa Fluor 647 was compared to the standard curve constructed in the citrate buffers with known pH values to obtain the endosomal pH of each sample.

2.5 Stimulation of DCs with ODN-PLL-PS

DCs were seeded at a density of 2.5×10⁵ cells per well in a 24-well plate. Soluble ODNs (0.005 mg/ml), 1826- and 2216-PLL-PS (100 nm, 500 nm, 3 μm) were added to cells in 500 μl medium in the absence or presence of designated concentrations of NH₄Cl. After 24 h of incubation at 37 °C, cell supernatants were collected and stored at −20 °C until further analysis.

2.6 Measurement of cytokine concentrations

Cell supernatants were analyzed by enzyme-linked immunosorbent assay (ELISA) for the quantification of IL-6 and IFN-α. The procedures for the ELISA for IL-6 were adapted from the manufacturer’s protocol (eBiosciences, San Diego, CA). The procedures for IFN-α were described previously [8, 53].

2.7 Statistical analysis

All experiments were repeated two to three times. The two-tailed and unpaired Student’s t test was used to analyze the differences between experimental groups as specified. A value of p < 0.05 was considered to be statistically significant.

3. Results

3.1 Characterization of ODN-PLL-PS particles

To examine the effect of particle size on the TLR9-mediated cytokine production, it was critical that ODN-PLL-PS maintained mono-dispersity and a defined size. Commercially available PS particles of defined sizes (Polydispersity Index (PDI) = 0.037 ± 0.032, 0.242 ± 0.010, 0.234 ± 0.038, for 100 nm, 500 nm, and 3 μm particles, respectively) were chosen as model ODN carriers. Because PS particles bear a negative surface charge, a cationic polymer was first adsorbed onto the surface of the particles to enhance the electrostatic adsorption of ODN onto particles. Several cationic polymers, such as poly(ethylenimine), chitosan, poly(2-dimethyl-amino)ethyl methacrylate (pDMAEMA), or poly-L-lysine (PLL), have been used to facilitate the attachment of DNA onto the carrier [54, 55]. PLL was used in this study because PLL induced minimal escape of particles from endosomal compartments and did not affect the endosomal pH [56]. PLL concentrations were optimized in order to achieve mono-dispersed particles with the maximal coverage of PLL (Supplemental Figure 1). The PLL coating shifted the surface charge of the particles from a negative to a positive zeta potential. At the maximal coverage of PLL, the zeta potential was 45, 65, and 43 mV for 100 nm, 500 nm, and 3μm particles, respectively.

Mono-dispersed PLL-PS was subsequently complexed with ODNs to yield ODN-PLL-PS. The negatively charged phosphate groups of the ODN backbone were expected to interact with the amine group side chain of lysine and result in ODN adsorption onto PS particles [57]. The size and zeta potential of 100 nm 1826-PLL-PS at different solution concentrations of CpG ODNs are shown in Figure 1. At the lowest ODN concentration examined, 0.005 mg/ml, particles tended to aggregate; the measured particle diameter and surface charge was approximately 300 nm and −10.6 mV. At the highest ODN concentration examined, 0.05 mg/ml, electrostatic stability was achieved and aggregation was minimal; the measured diameter and zeta potential was 129 nm and −19.2 mV. 500 nm and 3 μm 1826-PLL-PS exhibited a similar trend. At the optimized ODN concentrations, as stated in the Materials and Methods section, 500 nm and 3 μm ODN-PLL-PS had a surface charge of approximately −32 and −39 mV, respectively, and measured diameters of 433 nm and 3.5 μm. The measured diameter of 500 nm ODN-PLL-PS was less than 500 nm by DLS, and 500 nm PLL-PS was approximately 500 nm. Therefore, the polyelectrolyte coating may have affected the size measurement by DLS. The type of CpG ODN (1826 or 2216) did not have a significant effect on the size and zeta potential of ODN-PLL-PS (data not shown). Optimized conditions were used to prepare mono-disperse particles of uniform sizes for all experiments.

Figure 1.

Characterization of ODN-PLL-PS. (A) Characterization of the size and zeta potential of 100 nm 1826-PLL-PS. (B) A SEM image of 100 nm CpG-PLL-PS.

3.2 Differential cytokine profiles induced by CpG ODNs

Two different classes of CpG ODNs, B (i.e. CpG ODN 1826) and A (i.e. CpG ODN 2216) were adsorbed onto particles of different sizes. While both ODN types induce cytokines through the stimulation of TLR9, type B is particularly effective at inducing proinflammatory cytokines, such as IL-6, but not type I IFNs [50, 58], and type A is known to induce both proinflammatory cytokines and type I IFNs, such as IFN-α [50, 59]. For CpG ODN 1826, regardless of the size of particles, only IL-6 but not IFN-α was induced (Supplemental Figure 2), which was consistent with what was observed with soluble CpG ODN 1826.

For CpG ODN 2216, both IL-6 and IFN-α were detected when ODNs were bound to 100 and 500 nm particles, which was again consistent with what was observed with soluble ODNs (Figure 2D, E, G, H). Surprisingly, when CpG ODN 2216 was bound to 3 μm particles, IFN-α was completely diminished within the wide range of the concentrations of ODNs tested (Figure 2I), while IL-6 secretions remained at a significantly high level (Figure 2F). We note that we did not attempt to compare the efficiencies of stimulation of TLR9 by the three different sizes of particles because it is challenging to deliver equivalent amounts of ODNs due to the large difference in the surface area among the three particle sizes. Regardless of this issue, we have clearly demonstrated that ODNs bound to sub-micro and micro particles yield distinct cytokine profiles.

Figure 2.

The effect of particle size on the production of IL-6 and IFN-α. BC-1 cells were stimulated by 100 nm, 500 nm, and 3 μm (A–C) 1826- and (D–I) 2216-PLL-PS. PLL-PS and PS served as negative controls and were used at the highest particle-to-cell ratio of the corresponding particle size. Soluble ODNs (sol 1826 or sol 2216) at 0.005 mg/ml served as a positive control. The supernatants of cells were collected after 24 h and analyzed by ELISA for the (A–F) IL-6 and (G–I) IFN-α concentrations. The symbol * indicates thae no cytokine secretions were detected.

3.3 The effect of endosomal pH on cytokine profiles induced by CpG ODNs

Previous studies have suggested that pH plays a critical role in TLR9-mediated signaling. Both the recruitment of TLR9 into endosomal compartments and the interactions between TLR9 and its ligands require the acidification of the endosomes [35, 60]. Cells treated with inhibitors of acidification of endosomes, such as chloroquine or bafilomycin, exhibit impaired TLR9 stimulation [8, 60, 61]. We have demonstrated that the initial pH of endosomal compartments where 3 μm particles reside is 6.0. After 30 min, it increased to 7 and remained at 7. In contrast, submicron particles reside in endosomal compartments of pH 5.0 to 6.0 [46]. We hypothesized that the pH of endosomal compartments might affect the cytokine profiles induced by CpG ODN 2216. We tested this hypothesis by exposing both soluble CpG ODN 2216 and 100 nm 2216-PLL-PS to a broad range of endosomal pH environments.

The addition of NH₄Cl to the cell culture medium can inhibit endosomal acidification due to the rapid transport of NH₃ across the endosomal membranes and the low permeability of NH₄⁺ [52, 62]. We first confirmed the change in pH of endosomal compartments while cells were exposed to different concentrations of NH₄Cl. As shown in Figure 3, the pH of endosomal compartments increased by 1.2 while the NH₄Cl concentration increased from 0 to 16 mM. The change in endosomal pH of DCs was consistent with a previous report on the change in endosomal pH of mouse peritoneal macrophages, although only 0.16 mM NH₄Cl was needed to increase the pH by 1.6 in macrophages [62]. This discrepancy may be because the regulation of pH by NH₄Cl is cell-type dependent. By varying the concentration of NH₄Cl, we were able to obtain a distribution of pH environments that spanned from 5.21 to 6.38 in DCs.

Figure 3.

The change in the pH of endosomes by the addition of NH₄Cl. Concentrations of NH4Cl (0.063, 0.25, 1, 4, 16 mM) are presented as Log₁₀ of molar concentrations. The change in endosomal pH was calculated relative to the unmodified endosomal pH, which was 5.21.

We then studied the effect of the endosomal pH environment on the production of cytokines. A gradual increase in endosomal pH (Figure 3) corresponded to a gradual decline in IFN-α production by both soluble ODN 2216 and 2216-PLL-PS (Figure 4A). For 2216-PLL-PS, the production of IFN-α reduced by 80% when the pH increased from 5.21 to 5.82. The production of IL-6, however, remained constant (Figure 4B). At pH = 6.38, the levels of both IL-6 and IFN-α were diminished. Subsequently, we tested whether the cellular uptake of 2216-PLL-PS was affected by the presence of NH₄Cl. The addition of NH₄Cl slightly reduced the cellular uptake at concentrations up to 16 mM, which resulted the endosomal pH of 6.38. However, there was not a significant difference in the level of cellular uptake at lower NH₄Cl concentrations, except at 16 mM (Supplemental Figure 3). At pH = 6.38, the cellular uptake of 2216-PLL-PS decreased (Supplemental Figure 3), which could contribute to the reduction in the level of both IL-6 and IFN-α. In addition, the recruitment of TLR9 to endosomes at pH = 6.38 may be impaired as suggested previously [35]. Our results demonstrate that the endosomal pH affected the TLR9-mediated cytokine profiles. Therefore, the distinct cytokine profiles exhibited by micro- and submicro- particles can be attributed to the endosomal pH they are exposed to.

Figure 4.

The effect of endosomal pH on the production of cytokines. (■) 100 nm 2216- PLL-PS at 100,000 particles per cell, and (□) soluble ODN 2216 at 0.005 mg/ml, were added to BC-1 cells in the presence or absence of NH₄Cl. The cell supernatants were collected after 24 h and analyzed by ELISA for the (A) IL-6 and (B) IFN-α concentrations. The top and bottom x axes represent the Log10 of the molar concentration of NH₄Cl and the corresponding endosomal pH values, respectively. The symbols *, **, ***, +, and ++ indicate the level of the statistical significance of cytokine production concentrations: 1.0 > p > 0.1, 0.1 > p > 0.05, 0.05 > p > 0.01, 0.01 > p > 0.001, and p < 0.001, respectively. The statistical calculations were performed by using a two-tailed, unpaired Student’s t test. All p values are shown relative to the unmodified pH of endosomes, and are presented as: p value for ODN-PLL-PS / p value for soluble ODNs.

4. Discussion

By using model submicro- and micro- PS particles as TLR9-ligand (CpG ODN) carriers, we investigated the effect of the size of biomaterials on the TLR9-mediated production of cytokines. CpG ODN 2216 bound to 100 and 500 nm particles induced the production of both IL-6 and IFN-α. Surprisingly, when CpG ODN 2216 was bound to 3 μm particles, IFN-α secretions were not detected, though IL-6 secretions remained at a significantly high level. By correlating the cytokine production with the pH environment that particles of different sizes were exposed to, we further demonstrated that the endosomal pH regulated the cytokine profiles mediated through TLR9 stimulation. Our findings indicate that the size of biomaterials and the ability to regulate endosomal pH have a profound effect on the TLR9-mediated differential production of cytokines that regulate innate and adaptive immunity.

The intracellular location of TLR9 ligands has recently been shown to affect cytokine profiles. Honda et al. suggested that the retention of TLR9 agonists (CpG oligonucleotides) in early endosomal compartments led to a much higher level of type I IFNs than in late endosomal/lysosomal compartments [50]. Okuya et al. further confirmed this study and showed that the targeting of CpG ODNs to static early endosomes, which are a subpopulation of early endosomes and exhibit slow maturation kinetics and low mobility along microtubules [63], enhanced the generation of type I IFNs, but diminished the generation of proinflammatory cytokines [49]. Our previous study has shown that micrometer particles do not co-localize with LysoTracker, which was used to distinguish late endosomes and lysosomes from early endosomes, while submicrometer particles do [46]. We also confirmed that 3 μm particles were localized in early endosomes by using an early endosome-specific marker, early endosomal antigen-1 (EEA-1) (Supplemental Figure 4). The FITC-labeled 3 μm particles associated with EEA-1 (Supplemental Figure 4A). In addition, a cross-sectional image showing EEA-1 only (Supplemental Figure 4B) revealed ring-like structures that surrounded the particles, suggesting that the particles were contained within early endosomes.

Our results appeared to be inconsistent with the studies by Honda et al. [50] and Okuya et al. [49]. The inconsistent observations may arise from differences in carrier-specific intracellular trafficking and ODN release mechanisms. In this study, ODN-decorated PS particles of defined sizes were used. Our previous study showed that particles of different sizes were transported to different intracellular compartments that exhibited distinct kinetics of pH change and steady state pH. The correlation of cytokine production with the endosomal pH environment that TLR9 ligands were exposed to suggested that the endosomal pH affected TLR9-mediated cytokine profiles. In the study by Honda et al., the cationic liposome, N-[1-(2,3-Dioleoyloxy)propyl]-N,N,N-trimethylammonium methylsuflate (DOTAP), had been used as a CpG ODN carrier. It has been documented that the lipid membrane charge density and size of lipid-based DNA delivery systems determines the intracellular fate of lipoplexes [47, 64, 65]. In addition, the release of ODNs from DOTAP/ODN complexes is dependent on size and charge ratio, and is mediated by the fusogenic interaction between DOTAP and the anionic lipid membrane of intracellular compartments [66]. No detailed information in the study by Honda et al. was available to determine the size, charge ratio, and membrane charge density of the DOTAP/ODN complexes. In the study by Okuya et al., the heat-shock protein Hsp90 was used as the ODN carrier. However, the intracellular transport of heat-shock proteins remains elusive: endocytosed Hsps may transit through endosomes, into the cytoplasm, and/or into other intracellular vesicles [67, 68], which can potentially affect TLR9-mediated signaling. It is clear that endosomal markers, such as EEA-1 and LysoTracker, are not sufficient to distinguish the intracellular compartments that the three TLR9-ligand carriers, i.e. PS, DOTAP, and Hsp90, are transported to. More detailed studies are required to determine whether the carriers used in all three studies reside in the same intracellular compartment. A recent model of the bifurcation of TLR9 signaling proposed by Sasai et al. suggested that the adaptor protein 3 (AP-3), which finely controlled the trafficking of TLR9 within the endosomal system, was essential for the activation of type I IFN genes [51]. An understanding of how the carrier type and the dynamics of endosomal pH affect the recruitment of AP-3 to TLR9-residing compartments may reconcile the differences observed.

Previous studies have suggested that pH plays a critical role in TLR9-mediated signaling. Our results suggest an additional role of pH, that is, the endosomal pH affects the profiles of cytokines mediated by TLR9 activation. Carriers which possess the ability to modulate endosomal pH could potentially be used to regulate TLR9-mediated innate immunity. Collectively, our findings indicate that cytokine profiles may be manipulated by controlling the size and/or the composition of biomaterials.

5. Conclusion

In this study, we demonstrated that the size of particles affected the cytokine profiles mediated by TLR9 signaling pathways. By examining the pH of endosomes at which particles of different sizes trafficked to, we showed that type I IFNs, such as IFN-α, were much more sensitive to the change in endosomal pH compared to proinflammatory cytokines, such as IL-6. The role of TLR9 in the applications of biomaterials in medicine lies in delivery systems for vaccines, microbicides, and plasmid DNA and tissue implants. Our findings provide significant implications for two major design parameters: size and composition of biomaterials. The fine control of either or both parameters allows the exploitation of the beneficial aspects of TLR9-mediated innate immunity while minimizing unwanted side effects.