Cowpea Mosaic Virus Nanoparticles and Empty Virus-Like Particles Show Distinct but Overlapping Immunostimulatory Properties

Samples of purified wild-type CPMV, eCPMV produced in agroinfiltrated plants (eCPMV/p), and eCPMV produced in insect cells (eCPMV/i) were denatured and separated by SDS-PAGE (Fig. 1A). The L and S subunits of wild-type CPMV presented as single bands of 42 and 24 kDa, respectively, and both eCPMV preparations yielded the same L and S bands. Nondenaturing agarose gel electrophoresis, through which the nanoparticles travel intact, confirmed the lack of nucleic acids in both eCPMV particles, as indicated by the absence of nucleic acid staining (Fig. 1B). Two bands were observed after electrophoretic separation, and these are due to the presence of two electrophoretic forms for intact eCPMV/CPMV particles, the slow and fast forms; the fast electrophoretic form of CPMV results from a loss of 24 amino acids at the C terminus of the small coat protein, a process that occurs by proteolysis in plants or insect cells (5, 13, 14). Wild-type CPMV traveled farther through the gel than the eCPMV particles, probably reflecting the lack of internal RNA and its strong impact on electrophoretic mobility, which is consistent with our previous observations (15). Interestingly, the eCPMV/i particles were less mobile than the eCPMV/p particles, even though the coat protein composition was identical in both cases (see below). The major difference between the eCPMV/i and eCPMV/p particles was their degree of purity. Insect cell expression was less efficient, and the low yields resulted in a greater quantity of impurities than in the plant-based system, as determined by size exclusion chromatography (SEC) (Fig. 1C). We speculate that residual contaminants in the eCPMV/i formulation led to the formation of aggregates and/or surface-bound molecules that inhibited electrophoretic mobility.

Consistent with the native gels, UV-visible spectroscopy revealed a low A₂₆₀/A₂₈₀ ratio (0.7) for the eCPMV particles, indicating the absence of nucleic acids, compared to the typical A₂₆₀/A₂₈₀ ratio of 1.8 for wild-type CPMV particles. SEC was carried out to determine the purity and structural integrity of the particles, revealing no significant difference between wild-type CPMV and eCPMV/p particles with the same elution volume of 9.7 ml (Fig. 1C). Although the same elution volume was observed for the eCPMV/i particles, a series of peaks between 25 and 40 ml indicated impurities from the insect cell expression system. Due to low yields, we were unable to purify the samples any further.

Finally, dynamic light scattering (DLS) and transmission electron microscopy (TEM) confirmed the size and icosahedral shape of the CPMV and eCPMV particles (Fig. 1D and E). In agreement with the SEC data, icosahedral particles were found in all samples, but smaller proteins and/or fragmented particles were found in the eCPMV/i preparation. Generally, these data indicated that there were no significant differences between CPMV and eCPMV from different expression systems, except for the lower purity of the eCPMV/i preparation.

Reversed-phase liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis was carried out next to identify and quantify the peptide compositions of the 42- and 24-kDa bands revealed by SDS-PAGE. By comparing the data from each sample to the CPMV sequences, 18 target peptide sequences were selected for the L subunit, and 10 were selected for the S subunit (Fig. 2). For all three samples, the intensities of most commonly identified peptides were similar, confirming that the coat protein sequence of both eCPMV formulations is indistinguishable from that of wild-type CPMV.

Endotoxins are lipopolysaccharides (LPSs) that act as pathogen-associated molecular patterns (PAMPs), and they are found in the cell walls of most Gram-negative bacteria, including Agrobacterium tumefaciens, which was used to generate eCPMV in plants. LPS initiates a strong innate immune response when bacteria infect humans (16, 17). The presence of endotoxins can cause adverse reactions, including shock when endotoxin-contaminated reagents are used clinically (18). Therefore, it is essential to remove LPS from recombinant proteins, drugs, and other biological and pharmaceutical products to avoid adverse reactions. Various LPS removal procedures have been reported (19, 20), and we selected a method based on Triton X-114 to extract the LPS from our samples (18, 21).

We also considered the possibility that LPS could be encapsulated within VLPs. Therefore, we heated the eCPMV/p samples to denature the capsids and release any contents. This step was applied before or after the removal of LPS using the Triton X-114 method. We found that endotoxin levels remain unchanged regardless of whether the particles were intact or disassembled (Fig. 3), which indicates that Triton X-114 is sufficient to remove any LPS from the formulation, whether on the particle surface or encapsulated. The data may also suggest that no significant quantities of LPS are enclosed within the capsid.

Importantly, LPS contamination was apparent only in eCPMV samples produced by agroinfiltration of plants. Wild-type CPMV particles were produced by mechanical inoculation using infectious particles and black-eyed pea plants; these preparations avoid the use of Gram-negative bacteria, such as A. tumefaciens, which are the source of LPS contaminants, therefore highlighting the advantages of the mechanical inoculation method. Obtaining pure particles without LPS contaminants streamlines manufacture and improves yields; any additional postharvest purification steps ultimately reduce yields and require additional steps for characterization, driving up the production costs. For translational applications to avoid the release of a plant pathogen, wild-type CPMV particles could be treated with UV light or chemicals to render them noninfectious toward plants (22) while maintaining the presence of RNA and keeping the manufacture LPS free.

In situ vaccines trigger a local innate immune response that leads to the systemic attack of cancer cells carrying the same tumor antigens as a primary tumor. Although ovarian cancer is a devastating disease, metastases are frequently restricted to the peritoneal cavity (where the tumor microenvironment is directly accessible), so local immunotherapy can harness this effect more efficiently (23). To determine whether in situ vaccination with CPMV or eCPMV can increase the survival of tumor-bearing mice and reduce the tumor burden, we inoculated mice intraperitoneally (i.p.) with the luciferase-labeled ID8-Defb29/Vegf-A cell line on day 0. ID8-Defb29/Vegf-A cells were stably transfected with luciferase using a retrovirus as described previously (24) and in Materials and Methods.

Treatments started on day 7 after tumor challenge; treatments (100 μg/200 μl CPMV, eCPMV/p, or eCPMV/i versus the phosphate-buffered saline [PBS] control) were given six times in weekly intervals via i.p. injection (n = 5 per group). Endotoxin levels of all samples were determined using an endotoxin detection assay, and a level of <50 endotoxin units (EU)/mg protein was considered negligible. If necessary, LPS was removed prior to immunization. Tumor growth was monitored by measuring body weight, abdominal circumference, and total bioluminescence, all of which showed a close correlation (Fig. 4A to E). Mice were euthanized when their weight reached 35 g or when moribund. Both eCPMV/p and eCPMV/i reduced the tumor burden significantly in terms of total bioluminescence (P < 0.0001) and prolonged survival (P < 0.01) compared to the control group (Fig. 4D). CPMV prolonged survival even more than the eCPMV formulations, although the differences were not statistically significant. These data confirm that in situ vaccination with either CPMV or eCPMV can be highly efficacious against an aggressive ovarian tumor in mice.

Next, we compared the effects of CPMV and eCPMV ex vivo and in vivo stimulation on the tumor microenvironment, to determine whether the immunomodulatory activity of eCPMV (i.e., its ability to induce cytokine secretion and the activation of immune cells) is influenced by the lack of virus RNA. We therefore harvested nonadherent peritoneal cavity cells on day 35 from untreated ID8-Defb29/Vegf-A tumor-bearing mice, stimulated these cells ex vivo with 10 μg CPMV or eCPMV, and collected the supernatant 24 h later for cytokine analysis (Fig. 5A). The concentration of the proinflammatory cytokine interleukin-6 (IL-6) increased sharply in cells stimulated by CPMV (P < 0.0005) but not in those stimulated by eCPMV, in each case compared to the unstimulated control (Fig. 5A). However, when we increased the stimulation dose (50 μg), a significant increase of IL-6 was observed in cells stimulated by eCPMV/p, although the level of IL-6 was still 1.8 times lower than the IL-6 level measured after CPMV stimulation (data not shown). The levels of other proinflammatory cytokines (IL-1β, IL-12, interferon beta [IFN-β], and granulocyte-macrophage colony-stimulating factor [GM-CSF]) increased slightly in response to eCPMV but not CPMV (Fig. 5A). Both eCPMV/p and eCPMV/i inhibited the secretion of transforming growth factor β (TGF-β) (P < 0.0001) and IL-10 (P < 0.05 for eCPMV/p and P < 0.01 for eCPMV/i), whereas CPMV inhibited the secretion of only TGF-β (P < 0.0001). All three formulations induced the secretion of IFN-γ (P < 0.05 for CPMV and eCPMV/p and P < 0.01 for eCPMV/i). To further explore the mechanistic differences and similarities between CPMV and eCPMV, cytokine secretion after in vivo stimulation was measured. On day 35 after tumor inoculation, CPMV or eCPMV/p was i.p. administered. After 24 h, peritoneal cavity washes were harvested, and the cytokine levels in the supernatant were measured (Fig. 5B). Different from the ex vivo stimulation results, the level of IL-6 secretion in the CPMV-treated group was similar to the levels in the eCPMV/p- and PBS-treated groups. Besides IL-12, the levels of secretion of other proinflammatory cytokines (IL-1β, TNF-α, IFN-β, and GM-CSF) were increased by both CPMV and eCPMV/p. Moreover, in situ treatment of CPMV as well as eCPMV/p led to reduced TGF-β (P < 0.05 for eCPMV/p and P = 0.0738 for CPMV) and IL-10 (P < 0.05 for eCPMV/p and P < 0.01 for CPMV) secretion. A significant increase in IFN-γ secretion was observed only in CPMV-treated mice (P < 0.001). These data not only highlight the differences between CPMV and eCPMV but also underline the differences between the ex vivo and in vivo assays, the most striking difference being that in vivo, only CPMV stimulated IFN-γ.

We next characterized the phenotype of peritoneal leukocytes after 24 h of stimulation with each formulation (Fig. 6). The gating strategy is shown in Fig. 7. There was no significant difference in monocytic myeloid-derived suppressive cell (M-MDSC) populations between the stimulated groups and the unstimulated control group, while the abundance of granulocytic myeloid-derived suppressive cells (G-MDSCs) was significantly increased by all three formulations (P < 0.0001 for CPMV, P < 0.001 for eCPMV/p, and P < 0.01 for eCPMV/i). The number of tumor-infiltrating neutrophils (TINs) or type 1 neutrophils (N1) was significantly enhanced by CPMV (P < 0.01). Tumor-associated macrophages (TAMs), type 1 TAM (M1) populations (P < 0.01 for CPMV and no significant difference for eCPMV), and type 2 TAM (M2) populations (P < 0.0001 for CPMV and eCPMV/p and P < 0.0005 for eCPMV/i) were increased by all three formulations. The population of NK cells was increased following treatment with eCPMV/p (P < 0.05) and CPMV (no significant difference) stimulation. Furthermore, we observed a significant increase in the population of dendritic cells (DCs) (P < 0.0001) following stimulation with CPMV (but no enhancement of the activated DC subtype). The immune cell profile was also evaluated after in vivo stimulation. Either CPMV or eCPMV/p was i.p. injected into tumor-bearing mice (35 days postinoculation [dpi]), and after 24 h, peritoneal cavity cells were analyzed (Fig. 6B). Consistent with the ex vivo study results, the population of TINs (P < 0.0001) was dramatically enhanced by CPMV in situ treatment; M1 (P < 0.0001 for CPMV and P < 0.05 for eCPMV/p) and NK cell (P < 0.0001) abundances were increased by both CPMV and eCPMV/p treatments. However, different from the ex vivo results, we observed a potent decrease in the G-MDSC population (P < 0.0001) and an increase of M-MDSCs (P < 0.0001) by CPMV and eCPMV/p. CPMV significantly increased the levels of total TAMs and M2, while eCPMV/p reduced the M2 percentages in CD45⁺ cells. Notably, more activated DCs were found in the tumor microenvironment after CPMV treatment (P < 0.0001).

We therefore conclude that both CPMV and eCPMV can enhance antitumor leukocytes (M1 and NK cell) populations and regulate important cytokines that contribute to an immunostimulatory tumor microenvironment but that the presence of RNA in the CPMV particles has a specific impact, particularly on potent antigen-presenting cells (APCs), such as TINs and activated DCs.

CPMV nanoparticles were produced in plants as previously described (10). LPS was quantified using the Endozyme II recombinant factor C endotoxin detection assay (Biovendor), and a level of <50 endotoxin units (EU)/mg protein was considered negligible.

N. benthamiana plants were grown in soil for 6 weeks with a 16-h photoperiod (10,000 lux) at 24°C and 60% humidity before transient expression of the CPMV coat proteins using the agroinfiltration system based on A. tumefaciens (52). A preculture of 10 ml lysogeny broth (LB) supplemented with streptomycin and kanamycin was inoculated with 100 μl of an A. tumefaciens LBA4404 glycerol stock carrying the plasmid pEAQ-VP60-24K encoding the L and S coat protein subunits. After overnight cultivation at 28°C and 160 rpm, the preculture was used to inoculate a large culture for vacuum infiltration, which was grown under the same conditions for 48 h. For infiltration, the bacterial suspension was diluted to an optical density (OD) of 1.0 with 2× infiltration medium (100 g/liter sucrose, 4 g/liter glucose, 1 g/liter Jack’s professional 20-10-20 Peat Lite water-soluble fertilizer). To activate the vir genes, we added 200 μM acetosyringone, and the suspension was incubated at room temperature for 1 h.

Prior to infiltration, all plants were sprayed with tap water to facilitate infiltration. Plants were submerged into 2 liters of the bacterial suspension in a bucket in a desiccator. The desiccator lid was replaced, and a vacuum was applied. After pressure equalization, all noninfiltrated leaves were removed, and the plants were incubated under the conditions described above. After 5 days, all leaves were harvested and either directly processed or stored at −80°C.

The harvested plant material was homogenized in a blender with 3 volumes of 0.1 M potassium phosphate (KP) buffer (pH 7.0) and supplemented with 2% (wt/vol) polyvinylpolypyrrolidone (PVPP). After stirring, the material was filtered through two layers of Miracloth and centrifuged at 14,000 × g for 20 min at 4°C before adding 0.25 volumes of polyethylene glycol (PEG) precipitation buffer (1 M NaCl, 20% PEG 6000) to the supernatant and stirring overnight at 4°C. After centrifugation at 14,000 × g for 20 min at 4°C, the supernatant was kept, and the pellet was resuspended in 10 ml 0.01 M KP buffer and centrifuged again at 27,000 × g for 20 min at 4°C. The supernatants from both steps were combined for ultracentrifugation over a 10-ml 30% sucrose cushion at 133,000 × g for 3 h at 4°C. The sucrose fraction was collected and dialyzed against 0.01 M KP buffer for 48 h, with frequent buffer exchange.

The eCPMV/i formulation was produced in Sf9 insect cells by Creative Biolabs. The sequences of the CPMV L and S coat protein subunits as well as the 24K protease were codon optimized for insects and introduced into the pFastBac dual-expression vector (Thermo Fisher Scientific) to generate plasmid pFBD-VP60/24K.

CPMV and eCPMV particles were denatured at 100°C for 5 min in 4× lithium dodecyl sulfate (LDS) loading dye (Thermo Fisher Scientific) to obtain a final volume of 20 μl containing 20 μg of particles. The L and S subunits were separated for 40 min at 200 V and 120 mA in 4 to 12% NuPAGE precast gels in 1× morpholinepropanesulfonic acid (MOPS) buffer (Thermo Fisher Scientific). Gels were photographed before and after staining with 0.25% (wt/vol) Coomassie brilliant blue using the AlphaImager imaging system (Protein Simple) under white light.

CPMV and eCPMV samples (10 μg in 6× loading dye) were fractionated by 0.8% (wt/vol) agarose gel electrophoresis at 100 V for 30 min in Tris-borate-EDTA (TBE) buffer. Gels were visualized under UV light and after staining with 0.25% (wt/vol) Coomassie brilliant blue.

Samples were analyzed at 25°C in plastic disposable cuvettes with a path length of 10 mm using a Wyatt DynaPro NanoStar DLS instrument. The diameter was calculated as the mean of the intensity distribution, and the polydispersity index was determined from the intensity autocorrelation function.

TEM was performed using a Zeiss Libra 200EF microscope. Samples were mounted on 400-mesh hexagonal copper grids bearing Formvar support film, stained with 2% uranyl acetate, and allowed to dry for 12 h.

One hundred microliters of each sample (1 mg/ml) was injected into a Superose 6 column on an Äkta Explorer chromatography system (GE Healthcare) at a flow rate of 0.5 ml/min in 10 mM KP buffer (pH 7.4), and the absorbance at 260 and 280 nm was recorded.

L and S gel bands from each sample (20 μg) were excised, cleaned, and digested with lysyl endopeptidase (Wako Chemicals) at an enzyme/substrate ratio of 1:20 for 2 h at 37°C, followed by overnight trypsin digestion using sequencing-grade trypsin (Promega) at an enzyme/substrate ratio of 1:20 at 37°C (53). The digested peptides were analyzed by LC-MS/MS using an LTQ-Orbitrap Velos mass spectrometer (Thermo Fisher Scientific) equipped with a nanoAcquity ultra-high-pressure liquid chromatography system (Waters). Full-scan MS spectra (m/z 380 to 1,800) were acquired at a resolution of 60,000, followed by 20 data-dependent MS/MS scans. MS/MS spectra were generated by collision-induced dissociation of the peptide ions (normalized collision energy of 35%, activation Q of 0.250, and activation time of 20 ms) to generate a series of b and y ions as major fragments. LC-MS/MS raw data were acquired using Xcalibur v2.2 SP1 (Thermo Fisher Scientific). Raw files (one for each sample) were imported into Rosetta Elucidator v3.3.0.1.SP.25 (Rosetta Bio-software) for processing (54). The peak list (.dta) files were searched against the UniProt virus database (550,552 sequences) using Mascot v2.1 (Matrix Science) with the following settings: trypsin enzyme specificity, mass accuracy window for precursor ions of 10 ppm, mass accuracy window for fragment ions of 0.8 Da, carbamidomethylation of cysteines as a fixed modification, oxidation of methionine as a variable modification, and one missed cleavage. The peptide identification criteria were a mass accuracy of ≤10 ppm, expectation P value of <0.05, and estimated false discovery rate (FDR) of <2%. The results were imported into Rosetta Elucidator for the automated differential quantification of peptides (54). Retention time (RT) alignment, feature (RT × m/z surface) identification and peptide ion extraction (peak), and background subtraction and smoothing across both RT and m/z dimensions were performed using PeakTeller. Normalization of signal intensities across samples was achieved using the average signal intensities obtained in each sample. For each peptide, the average intensity for all samples within a group was calculated.

Endotoxins were removed by treating all particles (CPMV, eCPMV/p, and eCPMV/i) three times with 1% (vol/vol) Triton X-114 (18, 21). After the addition of Triton X-114, samples were incubated on a rolling platform at 4°C for 20 min and then at 37°C for 10 min before centrifugation at 20,000 × g for 10 min at room temperature. The clear layer from the phase separation was collected and transferred to a new tube. After the last step, the samples were centrifuged at 150,000 × g for 1 h at 4°C. The supernatants were discarded, and the pellets were resuspended in 0.1 M potassium phosphate buffer (pH 7.2). Subsequently, Triton was removed using Pierce detergent removal spin columns (Thermo Fisher) with 95% removal efficacy according to manufacturer’s instructions.

We heated 10 μg of CPMV or eCPMV/p particles with or without LPS for 10 min at 90°C, followed by brief centrifugation. To test for endotoxins, 10 μg of CPMV or eCPMV/p with and without LPS was diluted 1:10 in endotoxin-free water and assessed using the Endozyme II recombinant factor C endotoxin detection assay (Biovendor) according to the manufacturer’s instructions.

Female C57BL/6J mice (6 to 8 weeks old) were purchased from The Jackson Laboratory. All mouse studies were approved by the Institutional Animal Care and Use Committee of Case Western Reserve University. ID8-Defb29/Vegf-A cells were transfected with luciferase as previously described (24). Briefly, 10 μg of plasmids vesicular stomatitis virus G and pBabe-puro Luc was added to the GP2-293 packaging cell line following a 30-min incubation with 60 μl of TransIT (Mirus). GP2-293 cells and plasmids were a generous gift from William P. Schiemann, Case Western Reserve University. Forty-eight hours following the addition of plasmid, medium was collected, filtered to remove cellular debris, and added to ID8-Defb29/Vegf-A cells in the presence of Polybrene (8 μg/ml; Santa Cruz Biotechnology) at a ratio of 50:50 with modified RPMI 1640 medium typically used with this cell line. ID8-Defb29/Vegf-A cells were treated with the retroviral particles for 24 h, allowed to recover in fresh medium for 24 h, and selected for transduction with puromycin for 5 days. We implanted 2 × 10⁶ live cells/200 μl PBS orthotopically into mice by i.p. injection (55). One hundred micrograms of CPMV or eCPMV was administered once a week by i.p. injection in 200 μl PBS for a total of 6 treatments. The mice were monitored weekly for signs of tumor progression, including abdominal distension, weight, circumference, and other morbidity indicators. Tumor growth was also monitored twice weekly by total bioluminescence imaging, based on the i.p. injection of 150 mg/kg of body weight luciferin (Thermo Fisher Scientific), followed by analysis using an Ivis spectrum imaging system (PerkinElmer). Total bioluminescence was determined using Living Image software (PerkinElmer). Regions of interest were quantified as average radiance (photons per second). Mice were euthanized when their weight reached 35 g or when moribund.

Peritoneal cavity washes and cell culture supernatants were tested by an enzyme-linked immunosorbent assay (ELISA) to detect interleukin-1β (IL-1β), IL-6, IL-10, IL-12, tumor necrosis factor alpha (TNF-α), transforming growth factor β (TGF-β), interferon beta (IFN-β), granulocyte-macrophage colony-stimulating factor (GM-CSF), and IFN-γ (BioLegend) according to the manufacturer’s instructions. For ex vivo stimulation, 10⁵ peritoneal cavity cells from tumor-bearing mice (35 days postinoculation [dpi]) were collected, plated in 96-well plates, and then stimulated with 10 μg CPMV or eCPMV. After 24 h, the supernatant was collected for cytokine quantification; nonadherent cells were washed away with PBS, and the profile of surviving cells was evaluated using flow cytometry. For in vivo stimulation, 100 μg of CPMV or eCPMV/p was i.p. administered to tumor-bearing mice (35 dpi). After 24 h, the peritoneal cavity wash was collected. Cells were spun down by centrifugation (500 × g for 5 min), the supernatant was used for cytokine quantification, and the cells were resuspended and analyzed by flow cytometry.

Cells were washed in cold PBS containing 1 mM EDTA and resuspended in staining buffer (PBS containing 2% fetal bovine serum, 1 mM EDTA, and 0.1% sodium azide). Fc receptors were blocked using anti-mouse CD16/CD32 (BioLegend) for 15 min and then tested with the following fluorescence-labeled antibodies (BioLegend) for 30 min at 4°C: CD45 (30-F11), CD11b (M1/70), CD86 (GL-1), major histocompatibility complex class II (MHCII) (M5/114.15.2), Ly6G (1A8), CD11c (N418 A), F4/80 (BM8), Ly6C (HK1.4), NK1.1 (PK136), and isotype controls. Cells were washed twice and resuspended in staining buffer for data acquisition. Flow cytometry was carried out using a BD LSRII cytometer (BD Biosciences), and the data were analyzed using FlowJo software (TreeStar). OneComp eBeads (eBioscience) were used as compensation controls.