Nanodiamonds-in-oil emulsions elicit potent immune responses for effective vaccination and therapeutics

Vaccination is an effective public health tool to contain the spread of infectious diseases worldwide [1]. Today, it plays an ever-increasing role in preventing and controlling epidemics (e.g., COVID-19) [2]. The use of vaccines is to elicit immunological responses to destroy substances containing antigens. Most currently available vaccine antigens are made of pathogen subunits or their recombinant molecules, which are generally less potent than the whole pathogens; therefore adjuvants are often used with the subunit recombinant molecules to enhance their immune responses [3]. However, when applying them to humans, safety has always been a concern [4,5]. Despite the considerable efforts which have been made in the past, only a small number of vaccine adjuvants have been licensed for human use. Alum (or aluminum salt) is the most widely used among these adjuvants. However, aluminum is a neurotoxic element and our understanding of its toxicology in the human body is still limited [6,7]. Although new types of adjuvants such as water-in-oil emulsion, squalene, liposomes and other compounds have been developed, they may have greater local and systemic reactogenicity than alum alone [8,9]. Achieving high adjuvant potency but low human toxicity remains a major challenge in vaccine development.

Complete Freund’s adjuvant (CFA) is a water-in-oil emulsion containing heat-killed mycobacteria for immunization. It is one of the strongest adjuvants known because the inactivated mycobacteria in CFA can attract macrophages and other immune cells to the injection sites, and the oil can act as an insoluble depot of antigens to achieve long-term immunostimulation [10]. Together, they produce a pronounced effect in enhancing immune responses [11,12]. However, CFA’s high reactogenicity and toxicity have prohibited its use in humans. Consequently, CFA is most commonly used for antibody production in experimental animals. A way to circumvent this limitation is to employ incomplete Freund’s adjuvant (IFA), which lacks allergenic additives (e.g., the inactivated mycobacteria), and mix it with biocompatible, nonallergenic, nontoxic nanoparticles to reduce the undesired side effects. An ideal balance of efficacy and safety is expected if synthetic nanoparticles developed for vaccine applications [13,14] are used as additives. This approach is appealing because several completed clinical trials have used IFA as a vaccine adjuvant to treat diseases like HIV infection [15]. The new formulation is thus potentially useful for human vaccination.

The nanoparticles that have been used for vaccination applications can be roughly classified into two types [13,14] organic nanoparticles, including liposomes and polymers; and inorganic nanoparticles, including aluminum hydroxides, mesoporous silica, magnetic nanoparticles, gold nanoparticles and nanodiamonds (NDs). Of particular interest are aluminum hydroxides and mesoporous silica, which have been reported to be applicable as antigen carriers and self-adjuvants for vaccine delivery [16–19]. However, the toxic levels of these nanoparticles in the human body are unclear. NDs, in contrast, are known to be chemically inert and have excellent biocompatibility and low cytotoxicity [20]. They have found a wide range of applications in biology and nanomedicine due to their high surface-area-to-volume ratios, versatile surface chemistry and ability to emit near-infrared fluorescence from color centers [21]. In addition, NDs have been demonstrated to improve the efficacy of many chemotherapeutic agents by increasing their dispersibility in water, protecting the drug from inactivation and circumventing chemoresistance [22–24]. These improvements have inspired much research on the sustained and controlled release of therapeutics such as growth factors, peptides and genes, both in vitro and in vivo, with the aid of NDs. A recent study by Zhang et al. showed that NDs could serve as an efficient delivery system for immunostimulatory CpG oligonucleotides with high potential for cancer immunotherapy applications [25]. Another study with fluorescent nanodiamonds (FNDs) by Suarez-Kelly et al. demonstrated that immune cells (including natural killer cells and monocytes) could readily take up the particles without compromising cell viability or immune-cell activities [26]. Furthermore, FNDs surface-conjugated with immunomodulatory molecules are promising agents with a unique tracking capability to stimulate and manipulate immune systems [27].

This study employed chicken egg ovalbumin (OVA) as the model antigen. The protein is a well-characterized target antigen for CD8⁺ T cells (e.g., cytotoxic T lymphocytes from OT-1 mice), which specifically recognize the OVA 257–264 peptides and thus offer an excellent opportunity to investigate antigen-specific T-cell immunity [28]. The biocompatible, nonallergenic, nontoxic nanoparticles used in this work were monocrystalline NDs of ∼100 nm diameter. The surfaces of these NDs were first oxidized in air and subsequently carboxylated by acid treatment to facilitate their conjugation with OVA through noncovalent interactions [29]. The conjugation is expected to help increase the uptake of OVA by antigen-presenting cells (APCs) through endocytosis of the noncovalently bound OVA–ND complexes, thus enhancing their antibody production. To achieve this goal, we first coated OVA on NDs by physical adsorption to enable sustained release of these surface-bound immunogens either in vitro or in vivo. We then mixed the antigen-containing buffers with IFA to yield emulsions, followed by subcutaneous injection into healthy mice and monitoring the animals’ immune responses. Furthermore, to prove that APCs could internalize the noncovalently bound complexes, we employed OVA-conjugated FNDs and acquired fluorescence images of these particles taken up by murine macrophages such as RAW264.7 cells [30]. Finally, the immunotherapeutic effects of the vaccine were demonstrated by inoculating tumor-free mice with E.G7-OVA derived from the mouse lymphoma cell line EL4 containing a single copy of the inserted gene for constitutive synthesis and secretion of chicken egg OVA in the cells [31]. The way in which the vaccination inhibited the tumor growth was examined over 1 month.

Materials & methods

Chemicals & materials

OVA, CFA, IFA, phosphate-buffered saline (PBS), oxidative acids and all other chemicals were from MilliporeSigma (Burlington, MA, USA) and used without further purification. Monocrystalline synthetic diamond powders with a nominal size of 100 nm were obtained from Element Six (Didcot, UK).

NDs & OVA-NDs

Monocrystalline diamond powders were first oxidized in air at 490°C for 2 h to remove graphitic carbon atoms on the surface, followed by microwave cleaning in concentrated H₂SO₄–HNO₃ (3:1 v/v) solution at 100°C for 3 h to remove metallic impurities and simultaneously functionalize their surfaces with –COOH groups [32]. The surface-oxidized and carboxylated NDs were then noncovalently conjugated with OVA by simple mixing of OVA solution (5 μl, 1 or 5 mg/ml) with ND suspension (30 μl, 2 or 10 mg/ml) for 1 h at room temperature.

FNDs & OVA-FNDs

NDs were converted to FNDs by radiation damage of the monocrystalline diamond powders with 10-MeV electrons and then vacuum annealing at 800°C to generate nitrogen-vacancy centers in the diamond matrix, as previously described [33]. The FNDs were then surface-oxidized, carboxylated and noncovalently conjugated with OVA, following the same procedures as for NDs, described above.

Particle characterization

Hydrodynamic sizes and ζ-potentials of bare NDs and OVA-conjugated NDs were measured with a particle size and ζ-potential analyzer (DelsaNano C; Beckman Coulter, CA, USA). Morphologies of NDs were imaged on copper grids using a transmission electron microscope (H-7650, Hitachi; Tokyo, Japan) operating at 75 kV.

Protein adsorption measurement

A UV-Vis spectrophotometer (U-3310, Hitachi; Tokyo, Japan) measured the absorption spectrum of an OVA solution (800 μg/ml) over 200–400 nm. The protein solution was then mixed with the ND suspension (1 mg/ml) at the volume ratio 1:1 and incubated at room temperature for 30 min under gentle shaking. Afterward, the OVA–ND mixture was centrifuged at 20,000 × g for 15 min to precipitate NDs. The amount of OVA remaining in the supernatant was quantified by measuring its absorbance at 280 nm.

Cells & animals

RAW264.7, EL4 and E.G7-OVA cells were obtained from Bioresource Collection and Research Center (Hsinchu, Taiwan). C57BL/6 mice (female, 6–8 weeks) were obtained from BioLASCO (Taipei City, Taiwan).

Cell cultures

RAW264.7 cells were maintained in regular DMEM (Thermo Fisher Scientific, MA, USA) supplemented with 10% fetal bovine serum (FBS) and newborn calf serum (Biological Industries; Beit-Haemek, Israel) at 37°C with 5% CO₂ in a humidified incubator. EL4 cells were grown in DMEM complemented with 10% horse serum and antibiotic–antimycotic, and E.G7-OVA cells were maintained in RPMI-1640 medium (Thermo Fisher Scientific) supplemented with 10% FBS, 10 mM HEPES, 1.0 mM sodium pyruvate and supplemented with 0.05 mM 2-mercaptoethanol and 0.4 mg/ml G418 at 37°C with 5% CO₂.

Isolation of mononucleocytes from mouse bone marrows

C57BL/6 mice were sacrificed by cervical dislocation. After disinfection with 70% ethanol, the mice were dissected to harvest their tibias and femurs, which were then soaked in RPMI-1640 medium supplemented with 10% FBS. To isolate bone marrow-derived dendritic cells (BMDCs), both ends of the bone were cut off with scissors, and the needle of a 1-ml syringe containing the medium was inserted into the bone to flush the marrow into a sterile culture dish. Afterward, the cells in each dish were collected by centrifugation at 300 × g for 5 min. The cell pellet was then resuspended in Tris-NH₄Cl red blood cell lysis buffer (Thermo Fisher Scientific) to lyse erythrocytes. Finally, the cells were collected by a second centrifugal separation and washed with PBS before use.

Induced culture of BMDCs

The murine BMDCs isolated above were suspended in RPMI-1640 medium supplemented with 10% FBS and distributed into six-well plates at a density of 1 × 10⁶ cells/ml each. Subsequently, 25 ng/ml granulocyte–macrophage colony-stimulating factor (Abcam; Cambridge, UK) was added to the medium. The cells were then cultured at 37°C in a humidified incubator containing 5% CO₂ for 12 h, after which unattached cells and cell debris were removed by replacing the medium with a fresh one supplemented with granulocyte–macrophage colony-stimulating factor. On day 7, nonadherent and loosely adherent cells were harvested by gently washing the wells with PBS, and were pooled together for use in subsequent experiments.

Antigen uptake

RAW264.7 macrophages and BMDCs were first seeded at a density of 2 × 10⁵ cells per 35-mm dish and then maintained in regular DMEM supplemented with 10% FBS and newborn calf serum at 37°C with 5% CO₂ in a humidified incubator. Afterward, the cells were incubated with OVA-conjugated FNDs in serum-free DMEM with concentrations over 0–200 μg/ml at 37°C for 4 h to facilitate cellular uptake.

Immunization

Mice were immunized by subcutaneous injection of 100 μl solutions or emulsions containing either: OVA in PBS; OVA in CFA; OVA in IFA; OVA/ND in PBS; OVA/ND in CFA; or OVA/ND in IFA on days 1, 14 and 28. The water-in-oil emulsions were prepared by mixing 35 μl of OVA/ND suspensions with 65 μl of CFA or IFA. The immunogens and the adjuvants were thoroughly emulsified by repeatedly pipetting up and down before injection.

Antitumor therapeutics

C57BL/6 mice were subcutaneously immunized with the vaccine formulation as solutions or emulsions three times at 2-week intervals. They were then challenged by E.G7-OVA lymphoma cells (5 × 10⁵ cells) or OVA-negative EL4 tumor cells (5 × 10⁵ cells) on day 7 post the final immunization. The tumor growth was monitored for over 1 month.

Confocal fluorescence microscopy

Optical images of the cells after incubation with OVA-conjugated FNDs were acquired with a laser-scanning confocal fluorescence microscope (SP-8, Leica Microsystems; Wetzlar, Germany) equipped with a supercontinuum white-light laser operating at 561 nm for the excitation of nitrogen-vacancy centers. The fluorescence emission was collected through an oil-immersion objective (63×, numerical aperture = 1.4) and detected in the 650–800 nm wavelength region. Internalization of the FND particles by cells was confirmed by 3D confocal imaging.

FND quantification

As detailed in our previous publication [34], the amounts of OVA-FNDs internalized by RAW264.7 macrophages could be selectively measured using a home-built magnetically modulated fluorescence spectrometer. Before the measurements, the FND-internalized cells were detached from the dishes by trypsinization and then sonicated in PBS for 1 h to release the internalized particles into the solution. Fluorescence intensities were measured directly for the FND suspensions in glass test tubes by laser excitation at 532 nm under magnetic modulation to achieve background-free detection. The same technique was applied to quantify the amounts of FNDs in mouse spleens. In the latter experiment, mouse spleens (0.1 g/each, wet weight) were first digested in concentrated HNO₃ (1 ml) at 120°C for several hours until the solution became clear. No isolation of the FND particles from the solution was required when conducting the fluorescence intensity measurements.

Enzyme-linked immunosorbent assay

Mouse blood was collected from the submandibular veins of vaccinated or non-vaccinated mice on various days after immunization. Antibody responses of the immunized mice were evaluated using enzyme-linked immunosorbent assay (ELISA) and an OVA-specific IgG kit (OVA sIgG; Cusabio, TX, USA) for the collected mouse sera measured in a microplate reader (GloMax, Promega, WI, USA). The same method was applied to cytokines using a mouse IL-2 ELISA kit (ab100706, Abcam) and a mouse IL-4 ELISA kit (ab100710, Abcam), following the manufacturer’s instructions.

Enzyme-linked immunosorbent spot assays

IFN-γ enzyme-linked immunosorbent spot (ELISpot) assays were undertaken following the manufacturer’s instructions (R&D Systems, MN, USA). Briefly, 1 × 10⁶ freshly isolated splenocytes were plated in triplicate into 96-well ELISpot plates and stimulated for 48 h at 37°C in an incubator with 5% CO₂. After cell removal, plates were developed overnight at 4°C in the presence of detection antibody and then added with streptavidin–alkaline phosphatase for 2 h at room temperature. Spot detection was performed following incubation for 1 h in the dark with nitro blue tetrazolium-5-bromo-4-chloro-3-indolylphosphate. IFN-γ-specific spot-forming cells were counted using an ImmunoSpot analyzer (CTL S6 Universal Analyzer, Cellular Technology, OH, USA).

Statistical analysis

Data were analyzed using the software Prism v. 8 (GraphPad; CA, USA) by the independent samples t-test and two-way or repeated measures analysis of variance, followed by the Sidak test for multiple comparisons. All the tests were two-tailed unless otherwise stated.

Results

Material characterization

NDs before and after mixing with OVA were analyzed for their sizes and ζ-potentials. Transmission electron microscopy of bare NDs after the oxidative acid washes showed that the particles were irregular in shape and varied considerably in size (Figure 1A, inset). Dynamic light scattering measurements of these acid-treated NDs in distilled deionized water (DDW) revealed a mean hydrodynamic diameter of ∼100 nm and a polydispersity index of 0.24. The average diameter of the particles after mixing with OVA in DDW increased by ∼20 nm (Figure 1A), indicating protein adsorption. The ζ-potential of the bare ND particles was -45 mV, compared with -23 mV for OVA-NDs.

Chicken OVA is a phosphorylated glycoprotein consisting of 385 amino acid residues, with a molecular weight of 42.7 kDa (or 45 kDa including the carbohydrate and phosphate portions) [35]. To find out the number of OVA molecules that could be loaded on the acid-washed NDs through noncovalent conjugation, we measured the changes in absorbance of free OVA at 280 nm before and after mixing with the nanoparticles (Figure 1B). For OVA adsorbed on 100-nm NDs, we estimated a protein-loading capacity of 0.12 g/g at saturation. Assuming a spherical shape of the adsorbent, this loading capacity suggests that each 100-nm ND (weighing ∼1.8 fg/particle) can accommodate more than 3000 OVA molecules on the surface.

Antigen uptake

RAW264.7 macrophages and BMDCs were used as the model cell lines to study the cellular uptake of the OVA antigens [28]. Specifically, we replaced NDs with FNDs to facilitate the observation of the antigen-uptake processes by fluorescence imaging [21]. Figure 2A & B show the confocal fluorescence images of RAW264.7 macrophages and BMDCs after incubation with OVA-FNDs at the 100 μg/ml particle concentration for 4 h. The images were acquired after washing the cells thoroughly with PBS to remove the excess particles and subsequent incubation in a culture medium over 3 days. As revealed by the fluorescence images, the OVA–FND conjugates could be spontaneously internalized by the cells in the culture medium. Moreover, the cells stayed healthy over 72 h of incubation, even though many FNDs were still inside the cells.

Aside from the fluorescence imaging, we quantified the number of OVA-FND particles taken up by the macrophages. This was achieved by measuring the fluorescence intensities of the cells dispersed in DDW using the magnetic modulation technique detailed in our previous publication [34]. It was found that the number of internalized FNDs scaled almost linearly with the OVA-FND concentration but gradually leveled off at concentrations above 100 μg/ml (Figure 3). For cells incubated with the nanoparticle bioconjugates at 100 μg/ml, the average number of OVA-FNDs internalized by the macrophages was 208. Given that each 100-nm FND can carry more than 3000 OVA molecules, similar to the 100-nm NDs (Figure 2A), the result indicated that the total number of OVA molecules taken up by the individual RAW264.7 macrophages could exceed 6 × 10⁵.

Immune responses

The in vivo experiments were started by mixing 5 μl OVA solution (1 mg/ml) with 30 μl ND suspension (2 mg/ml), followed by dispersing the mixtures in CFA or IFA. The corresponding control experiments consisted of 5 μg OVA only, without NDs. Before subcutaneously injecting the ND-based adjuvants into C57BL/6 mice, it is crucial to know the fraction of OVA attached to NDs, which could act as a depot of the antigens and promote their cellular uptake by APCs. UV-Vis spectrophotometric analysis of the OVA–ND mixture showed that about 40% of the OVA molecules in the mixture were attached to NDs, and the rest were free in the solution. They are in dynamic equilibrium with each other.

Depicted in Figure 4A is the timeline of immunization and blood collection in this experiment. After the immunization, the water-in-oil emulsions formed small nodules and appeared as soft capsules at the injection sites. We evaluated the OVA-specific IgG antibody responses with the sera of the immunized mice using ELISA after the second and third immunizations with OVA and OVA/ND in CFA. As shown in Figure 4B & C, the OVA/ND/CFA treatments (with 5 μg OVA each) induced a more significant amount of OVA-specific IgG antibodies in the mouse sera than OVA/CFA alone (by 3.3- and 1.4-fold, respectively), after the second and third immunizations.

Next, we investigated the dose dependence of the immune response by employing the OVA/CFA and OVA/ND/CFA emulsions containing 25 μg OVA each. The amount of NDs used in these assays increased accordingly to 300 μg. A more than twofold enhancement of the OVA-specific IgG antibody production was found associated with the OVA/CFA treatment due to this dosage increase from 5 to 25 μg OVA after the second immunization (Figure 4B). However, the response did not exceed that produced by the 5 μg treatment with OVA/ND/CFA. Notably, further increase of the OVA doses failed to boost the immune response in the OVA/ND/CFA treatment.

We investigated further whether the same level of immune response produced by OVA/ND/CFA could be maintained without allergenic components such as inactivated mycobacteria in CFA. The dose groups of 5 μg OVA were employed in this experiment. As shown in Figure 5, the OVA/ND/IFA treatment markedly boosted the levels of anti-OVA IgG. There were no significant differences in the results between the OVA/ND/IFA and OVA/ND/CFA treatments, indicating that the dead mycobacteria in CFA were not an essential additive in inducing the immune response. Compared with the OVA/CFA and OVA/IFA groups, the OVA/ND/CFA and OVA/ND/IFA treatments boosted the levels of anti-OVA IgG by 3.5- and 3.8-times, respectively.

Antitumor therapeutics

An attempt was made to apply the new formulation of NDs in oil emulsions as an antitumor therapeutic agent. We demonstrated the application using mice inoculated with the mouse lymphoma cell lines EL4 and E.G7-OVA. The latter was derived from the C57BL/6 mouse lymphoma cell line (EL4) transfected with pAc-neo-OVA plasmids [31]. They can express OVA and have been widely used in cancer immunotherapy studies. Depicted in Figure 6A is the timeline for the subcutaneous injection of OVA/ND/IFA first, followed by inoculation of EL4 and E.G7-OVA cells in C57BL/6 mice. The doses of OVA used in both groups were 5 μg. By comparison with the unvaccinated groups, we found that the treatment with OVA/ND/IFA in the EL4 model could not delay the tumor growth (Figure 6B). In contrast, the OVA/ND/IFA treatment could effectively inhibit the tumor progression in the E.G7 model over 3 weeks postinoculation of the cells (Figure 6C). Notably, half of the mice (4/7) in the E.G7 model maintained their tumor-free status for more than 15 days after the cell inoculation (Figure 6D) and survived up to 35 days post-tumor cell challenge (Figure 6E). Figure 6F shows photographs of the tumors isolated on day 24 from vaccinated and non-vaccinated mice. The difference in tumor size between these two groups (in triplicate) of mice is substantial, about ten-times in total volume.

We next verified the hypothesis that NDs played an essential role in the antitumor treatment. The verification was made by replacing NDs with FNDs in the adjuvants and searching for the particles in sacrificed mice by fluorescence detection. In this experiment, we followed the same procedures described above for the OVA/ND/IFA treatment and collected the spleen tissues of the mice inoculated with EL4 and E.G7-OVA cells on day 24 after the vaccination with OVA/FND/IFA. With the aid of magnetic modulation to achieve selective detection of FNDs in the tissue digests without any preseparation [34], we were able to identify the presence of these particles in mouse spleens by measuring the intensities of the far-red fluorescence at wavelengths longer than 750 nm. The weight of FNDs found in the spleens was 0.13 μg, obtained after subtracting the backgrounds from the signals between these two groups (Figure 6G). Given a total weight of 60 μg for the FNDs used in the vaccination, the recovery rate was estimated to be 0.2%.

Lastly, we investigated whether the addition of NDs in IFA would alter the mechanism of the immune response elicited by IFA alone, which is known to act predominantly through the Th2 pathway [36]. The question was addressed by performing ELISA for IL-2 and IL-4 in the sera of C57BL/6 mice after subcutaneous injection with OVA/ND/IFA. Only a slight difference in the IL-2 level was found between the control and treatment groups, whereas a marked elevation of the IL-4 concentration in the vaccinated group was detected (Figure 7A). Furthermore, we evaluated the T-cell response corresponding to an antigenic stimulus upon the OVA/ND/IFA treatment. As detailed above, the spleens of mice were first harvested on day 7 after subcutaneous injection with OVA/ND/IFA. Their splenocytes were then analyzed for IFN-γ (an abundant cytokine produced by Th1 cells) with ELISpot, an assay extensively used to screen immune responses in developing vaccines to prevent and treat diseases. As shown in Figure 7B, there was a similar level of IFN-γ between the OVA/IFA treatments with and without the ND additives. However, both treatments exhibited a marked increase in the IFN-γ level compared with the control.

Discussion

ND is a member of the nanocarbon family. Its surface can be readily derivatized with various functional groups, including –COOH, –COH and –NH₂ [37]. The particles are partially hydrophobic if the surface comprises graphitic carbon atoms. Extensive washing in concentrated oxidative acids is the most convenient way of carboxylating the diamond surface [32,38]. Previous studies have shown that acid-washed NDs exhibit an exceptionally high affinity for protein molecules, including cytochrome c, myoglobin, bovine serum albumin, lysozyme and luciferase [29,38,39]. More importantly, these proteins’ structure and function are not altered significantly, as evidenced by the retained enzymatic activities of lysozyme and luciferase after adsorption on NDs [38,39]. Using dynamic light scattering, this work demonstrated that NDs could also bind effectively with OVA through noncovalent interactions. The hydrodynamic size of OVA-NDs increased by ∼20 nm compared with bare NDs (Figure 1A). OVA has an isoelectric point of 4.5 [35], meaning the protein molecules are negatively charged in DDW. The change of the ζ-potential from -45 mV of NDs to -23 mV of OVA-NDs in DDW implied that forces other than electrostatic interaction (e.g., hydrogen bonding and hydrophobic forces) were involved in the protein adsorption.

This work employed FNDs to study the cellular uptake of OVA in vitro by fluorescence imaging. For FNDs and NDs, their surface properties are nearly identical because they are produced similarly, except that FNDs are pretreated with electron irradiation and vacuum annealing to create color centers in the diamond lattice before surface modifications [33]. The advantages of using of FNDs are manifold [21]. First, FNDs contain a dense ensemble (∼10 p.p.m.) of nitrogen-vacancy centers as fluorophores, which are atom-like and exceptionally photostable, well suited for quantitative analysis. Second, when excited by a green laser, FNDs produce far-red emission at ∼700 nm, where the signal level of cell autofluorescence is low. Third, the nitrogen-vacancy centers are embedded in the diamond matrix, and therefore their fluorescence properties are immune to surface modifications. Fourth, the nitrogen-vacancy centers are magneto-optical and can be detected with high sensitivity and selectivity by magnetic modulation [34]. Using this highly photostable and biocompatible additive, we showed that the OVA–FND conjugates could be readily internalized by both RAW264.7 macrophages and BMDCs in the culture medium (Figure 2). The number of internalized antigens can be more than 1 × 10⁵ molecules per cell, and they are likely to be presented on the surface of these cells and possibly other APCs as well.

Our studies align well with those of Eidi et al., who conducted a comparative experiment on the cytotoxicity of surface-modified FNDs and alum particles [40]. Cell viability assays with the NSC-34 neuron-like cell line showed that FNDs were nontoxic up to 110 μg/ml doses. In contrast, alum displayed a distinct or severely toxic effect at all doses over 1–50 μg/ml. We have previously studied the subacute toxicity of FNDs in rats by intraperitoneal injection over 5 months [41]. The measurements for water consumption, fodder consumption, body weight and organ index found no significant differences between the control and FND-treated groups (with a dose of 5 mg/kg body weight per week). The high biocompatibility, exceptional chemical stability and extraordinary photostability make it possible to use FNDs as a tag for tracking and tracing vaccine adjuvants like alum in vivo.

The in vivo experiments carried out in this work demonstrated that adding NDs in IFA as the adjuvant could elicit potent and protective immune responses against OVA in the mouse body (Figure 5). The enhancement was about fourfold; however, a saturation effect appeared to exist, where no higher levels of anti-OVA could be reached irrespective of the amounts of OVA applied. The results are in line with previous studies showing enhanced immune responses using NDs of different types as the additives or self-adjuvants, including monocrystalline NDs for recombinant HA/H7N9 in mice [42], detonation NDs for OVA in mice [43] and peptide-conjugated detonation NDs for coronaviral vaccine development [44]. An important implication of these findings is that using the ND additives in IFA can help reduce the suffering of experimental animals because no or few allergic reactions are caused. Additionally, it can decrease the consumption of experimental animals and antigens in producing the antibodies of interest, a valuable feature in the industrial production of antibodies and vaccines.

This study also examined the immunotherapeutic effects of the new formulation using tumor-free mice inoculated with OVA-negative EL4 cells or OVA-expressing E.G7-OVA tumor cells (Figure 6). Results of the in vivo experiments showed that the vaccination could effectively inhibit the tumors grown from E.G7-OVA cells over 1 month, strongly suggesting that the presently developed nanovaccines with ND/IFA as the adjuvants are promising agents for cancer immunotherapy. With the aid of FNDs as a photostable tracker, we were able to quantify the amounts of the ND-based adjuvants trapped in the spleens of the vaccinated E.G7-OVA-inoculated mice for the first time. This new combination of substances is expected to work appropriately as a trackable immune drug-delivery vehicle to promote antitumor activities with minimal systemic toxicity [26,27].

Based on all these findings, we propose a possible predominant mechanism for the initiation of the immune response by the ND/IFA-based vaccine as follows. First, the formation of nodules with loose structure in mouse tissues after subcutaneous injection of the antigen-loaded ND/IFA emulsion, in which the adjuvants act as a depot; second, active and continuous recruitment of immature immune cells to the depot; third, uptake of the antigen-loaded NDs by the immune cells through endocytosis and finally, promotion of the Th2 response, where helper T cells bind with the APCs and activate the development of B cells into antibody-producing plasma cells in the spleen. The proposed mechanism, also applicable to FND/IFA, is depicted in Figure 8.

Conclusion

We have developed a new formulation consisting of an ND suspension mixed with IFA, capable of eliciting effective and durable immune responses in mice. Without the need for allergenic additives (i.e., dead mycobacteria), the addition of highly biocompatible NDs (diameter of ∼100 nm) in the oil emulsion can not only retain the adjuvanticity of IFA but also significantly reduce the level of side effects. Compared with existing products, the ND-in-oil adjuvant has several advantages: high safety, minor side effects, low demand for antigens and broad applicability. By applying OVA as the antigen in small animals like mice, our studies indicate that ND/IFA is an active vaccine platform able to induce potent and sustained immune responses. Additionally, the adjuvants are useful as antitumor therapeutic agents, as proven by the OVA/ND/IFA treatment, which effectively inhibited the growth of OVA-expressing E.G7 tumor cells inoculated in mice. Furthermore, the agents are trackable in vivo if FNDs replace NDs. With further research, development and optimization, the carbon-based theranostic nanoplatform is expected to find real-world applications in nanomedicine.