A self-aggregating peptide forms nanoparticles useful in the design of thermostable vaccines

Abstract Background: The use of biomaterials has been expanded to improve the characteristics of vaccines. Recently we have identified that the peptide PH (1-110) from polyhedrin self-aggregates and incorporate foreign proteins to form particles. We have proposed this peptide can be used as an antigen carrying system for vaccines. However, the immune response generated by the antigen fused to the peptide has not been characterized. In addition, the adjuvant effect and thermostability of the particles has not been evaluated. Results: In the present study we demonstrate the use of a system developed to generate nanoparticles carrying as a fusion protein peptides or proteins of interest to use as vaccines. These nanoparticles are purified easily by centrifugation. Immunization of mice BALB/c with the nanoparticles in the absence of adjuvant result in a robust and long-lasting immune response. Proteins contained inside the nanoparticles are maintained for over one year at ambient temperature, preserving their immunological properties. Conclusion: The rapid and efficient production of the nanoparticles in addition to the robust immune response they generate position this system as an excellent method for the rapid response against emerging diseases. The thermostability conferred by the nanoparticle system facilitates the distribution of the nanovaccines in developing countries or areas with no electricity.

maintained for over one year at ambient temperature, preserving their immunological properties. Conclusion: The rapid and efficient production of the nanoparticles in addition to the robust immune response they generate position this system as an excellent method for the rapid response against emerging diseases. The thermostability conferred by the nanoparticle system facilitates the distribution of the nanovaccines in developing countries or areas with no electricity. Keywords : vaccines, nanoparticles, self-assembling, thermostable, immunology Background Vaccines are considered one of the most important medical advances in the history of humanity, preventing and eradicating diseases [1,2]. The World Health Organization (WHO) estimates that vaccines save around 2-3 million lives a year [3]. Traditional vaccines are based on two main methodologies: live-attenuated and inactivated/killed 3 pathogens [4]. Even though vaccines produced with these methods are immunologically effective, they still show some disadvantages, such as the need for a cold-chain, reduced shelf life and the time-consuming processes involved in the production and purification [5][6][7]. On the other hand, other methodologies have been used such as subunit and recombinant vaccines that weakly stimulate the immune system and their immunological effect is of short durability, so they require the use of adjuvant to potentiate their effect.
Currently available adjuvants may lead to unwanted effects such as the generation of granulomas, allergies and neurotoxicity due to the different components used [5,8,9].
However, even with the evolution of vaccines, vaccination continues to represent a high cost mainly for developing countries, due to the fact that they have the highest number of people vulnerable to infectious diseases [10][11][12]. An effective, low cost technology to produce thermostable vaccines would represent a major advancement in the fight against infectious diseases worldwide, and may significantly reduce the risk of pandemics [13].
For this reason, in recent years new technologies have been developed to advance in the production of more efficient and safer vaccines [10,14,15]. Lowering the cost of vaccines is an essential step to facilitate massive vaccination especially in isolated areas where the cold-chain cannot be maintained easily [14,16]. This last point is especially important since the cold chain represents about 80% of the cost of vaccines [17,18].
The use of biomaterials are a central part of novel strategies to develop next-generation vaccines [19,20], delivery systems [21,22] and improved thermostability [23]. Some insect viruses have developed a remarkable strategy to maintain virus viability for years at ambient temperature. The strategy is based on the generation of crystal structures known as polyhedra, where the virus is occluded and protected from the environment for several years. Most interestingly, a single protein (known as polyhedrin) forms the polyhedra crystal during the infection of insect cells. Polyhedrin self-aggregates 4 inside the nucleus and during aggregation viruses get occluded inside the crystal. Thus polyhedra is a natural preservative of proteins, whose function is to maintain the virus viable for many years at ambient temperature [24,25].
One of the most studied insect viruses that forms polyhedra is the Autographa californica multiple nucleopolyhedrovirus (AcMNPV) [26,27].
We have recently identified an amino acid sequence in the polyhedrin protein from AcMNPV, which maintain the self-aggregating properties of the full-length protein. This sequence includes the first 110 amino acids from polyhedrin (PH  ). We have shown also that PH  self-aggregates even when other proteins or peptides are fused to its sequence.
In the present study we fused the green fluorescent protein (GFP) to the PH  sequence to produce a fusion recombinant protein that self-aggregates. The use of GFP facilitated the characterization of the nanoparticles using confocal microscopy. We used this fusion protein to characterize the formation of nanostructures and to explore its thermostability for several months as well as their capacity to generate antibodies when immunized in mice.
The results obtained show that the nanostructures formed by PH  preserve the function of the protein contained within for at least one year at ambient temperature. The nanostructures formed by PH  generate a robust immune response raising antibodies that recognize GFP. The nanostructures showed adjuvant properties, since no adjuvant was required to generate a robust immune response. The PH  nanostructures are easily purified by centrifugation since they form water insoluble nanoparticles, reducing significantly the cost of purification. All these results position PH  as a novel platform for the production of thermostable vaccines contained inside nanoparticles (which we 5 named here as nanovaccines).

Design of recombinant baculoviruses
For the generation of recombinant baculoviruses was used the expression vector pFastbac TM 1 of Bac-to-Bac ® baculovirus expression system (Thermo Fisher, USA, cat. no.10359-016). Under the promoter of polyhedrin (polh), the genetic sequence of the first 330 bp of the N-terminal region of the polyhedrin was cloned and the genetic sequence of the GFP was ligated into its C-terminal in an open reading frame to generate a fusion protein called PH  GFP [28]. The polh promoter and the polyhedrin sequence were taken from Autographa californica multiple nucleopolyhedrovirus virus (AcMNPV). For the generation of the PH-WT-GFP chimeric polyhedra, the pFastbac TM Dual expression vector (Thermo Fisher, USA, cat.No.10712024) was used, the WT polyhedrin was cloned under the p10 promoter and the PH  GFP under polh promoter. The baculoviruses were amplified, purified and titrated by following the recommendations and protocols provided by the supplier (Thermo Fisher, USA).  [28].

Production and purification of PH (1-110) GFP NPs
SF9 cells (2x10 6 cel/ml) were infected using a multiplicity of infection (moi) of 10 with the recombinant baculoviruses, the cells were maintained at 27 °C under agitation at 100 RPM, 72 hours post infection (hpi) the cultures were centrifuged at 4200 g for 15 min to recover the viruses and obtain the cell pellet. The pellets were resuspended in phosphate buffered saline (PBS, 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 , pH 7.4) and were sonicated with 5 cycles of 20 seconds per pulse with 30% amplitude (Qsonica 700, USA). Between each cycle were maintained on ice for 5 minutes. After the last cycle, the PH  GFP NPs were washed 5 times with PBS, between each wash the samples were centrifuged at 14,000 g. Finally, they were resuspended in PBS. In addition, chimeric polyhedra were generated by infecting SF9 cells with baculovirus with the WT polyhedrin and recombinant polyhedrin PH  GFP.

Separation of PH (1-110) GFP NPs by sucrose gradients
The PH  GFP NPs were separated in a discontinuous gradient of sucrose. To form the discontinuous gradient, 3 different sucrose concentrations were used, 40%, 50% and 60% (w/v) in distilled water, ultracentrifugation was performed at 17,738 g (SW 40 ti rotor, Beckman Coulter, USA) for 10 min at 4 °C. The PH  GFP NPs of the different gradients were recovered and 3 washes were carried out with PBS, centrifuging the samples at 14,000 g after each wash.

7
The total protein of the lysates and the nanoparticles of the different gradients was determined using the Pierce TM BCA Protein Assay Kit (Thermo Fisher, USA, cat. no. 23225) based on bicinchoninic acid (BCA) for colorimetric detection.

Cell confocal microscopy
SF9 cells infected with recombinant baculoviruses, 72 hpi were washed with PBS and incubated for 5 minutes with DAPI (4', 6-diamino-2-phenylindole) to mark the nucleus (Thermo Fisher, USA, cat. no. D3571) at a 1:1000 dilution and fixed in slide glass (76x26 mm) with DAKO Fluorescent Mounting Medium (Agilent, USA, cat. no. S3023) [28]. The GFP of the PH (1-110) GFP NPs was excited at 473 and DAPI was excited at 405 nm. Fluorescence emission was collected at 510 nm for GFP and 420 nm for DAPI. All images were taken with a Fluoview FV10i confocal microscope (Olympus ® , Japan), using the 60 × NA 1.35 oil immersion objective (UPLSAPO60XO). The images were analyzed with the software, FV10ASW.

PH (1-110) GFP NPs confocal microscopy and 3D reconstruction
The purified PH  GFP NPs were fixed with DAKO Fluorescent Mounting Medium in glass slides (76x26 mm). To obtain the images, we used a wide-field inverted IX81 Olympus ® microscope with 60 × 1.42 NA oil immersion objective, to MT-20 illumination system and EMCCD camera iXon-897 (Andor Technology South Windsor, CT, USA). The used excitation and emission filters were 470 and 520 nm/40 bandpass respectively. The images were analyzed using ImageJ software. Imaris software was used for the 3D recostruction of confocal images (Additional file 1).  GFP were centrifuged, the pellet was washed with cacodylate buffer (0.08 M, pH 7.4) and fixed with 0.6% glutaraldehyde and 0.4% paraformaldehyde in cacodylate buffer for 10 min. Post-fixation was made with 1% osmium tetroxide in cacodylate buffer. The cells were included in an epoxy resin and cuts of 90 nm thickness were made. Then the samples were contrasted with uranyl acetate 1% for 10 min and with lead citrate for 2.5 min. The JEOL JEM 12,000 EXII microscope at 80 kV (Jeol USA, USA) was used to observe the samples.

Scanning electron microscopy (SEM)
Briefly, the particles were purified and fixed with 2.5% glutaraldehyde in phosphate buffer  nm. Laser intensity settings of 1% were sufficient to illuminate the fluorescent label 9 without causing significant photobleaching. The images were analyzed with ZEN 2012 software (blue edition, Carl Zeiss, Germany) and the final images were edited with ImageJ 1.52n (NIH, USA). We compared the recovery of fluorescence between the PH  GFP NPs (n = 14) and the chimeric particles PH-WT-GFP (n = 10) at different times. For this experiment, the PH  GFP NPs and the PH-WT-GFP particles were prepared in the same way as for confocal microscopy.

Western blot
For western blot analysis, proteins contained in the SDS-PAGE were transferred to a nitrocellulose membrane (Merck Millipore, USA, cat. no HATF00010) at 100 V for 1 h in wet chamber using transfer buffer (48 mM Tris base, 39 mM glycine, 0.037% SDS, 20% methanol). Membrane was blocked with 5% fat-free milk in Tris-buffered saline (TBS, 50 mM Tris-Cl, pH 7.6, 150 mM NaCl) over night (ON). The anti-GFP antibody utilized in these studies was produced in mice in our laboratory. The antibody was used at a 1:2000 dilution in TBS-T (0.05% Tween) and 0.5% fat-free milk. Membranes were incubated with anti-GFP antibody for 3 hours with agitation at room temperature (RT). The secondary antibody was horseradish peroxidase-coupled (HRP) anti-mouse IgG (Sigma-Aldrich, USA, cat. no. A9044) was used at dilution 1:5000 in TBS-T and 0.5% fat-free milk. The secondary antibody was incubated 1 h in agitation at RT. The membranes were analyzed with a C-Digit Blot scanner (LI-COR, USA) and the signal generated by the SuperSignal ® West Femto substrate (Thermo Fisher, USA, cat. no. 34095) was taken using the Image Studio software.

Purity and conservation of PH (1-110) GFP NPs
The purity of fresh PH  GFP NPs (Additional file 2) was evaluated by run electrophoretic assays using the Agilent Bioanalyzer 2100 (Agilent Technologies, USA) equipped with the Protein 230 assay kit according to the manufacturer's recommended protocol. The electropherograms and gel-like images results were analyzed with Agilent 2100 expert software (Agilent technologies, USA).

Nanoparticle tracking analysis (NTA)
The NanoSight instrument (Malvern Panalytical, UK) was used to determine the size of the polyhedrin nanoparticles produced by the recombinant baculoviruses. The PH  GFP NPs resuspended in sterile water were injected in a volume of 1 ml into the sample chamber. Five readings were made for each sample processed to obtain the average particle sizes. The NanoSight software (Malvern Panalytical, UK) tracked the Brownian motion in real-time to determine the center of the PH  GFP NPs and determine the diffusion coefficient of each particle. Finally, the software based on the Stokes-Einstein equation calculated the size of the particles [29,30].

Animal studies
For the care, feeding and management of the animals, we followed the guidelines

Immunization studies
The animals used were obtained from the bioterium of the Institute of Cellular Physiology and were kept in a pathogen-free environment. Female BALB/c mice 6-8 weeks of age and 20-25 grams in weight were used for all experiments. All groups were randomly formed from approximately 3 litters. The immunization route was intramuscular (i.m.). The treatments were suspended in PBS and all treatment groups received the dose of antigen on days 0 and 14, this was decided after the dose response assay (Additional file 3). The control groups were treated first, then the groups without adjuvant and finally the groups with adjuvant. Blood samples were collected from day 0 until the end of each study at 2week intervals. Samples were centrifuged, and the sera were stored at -70 °C until analysis by Enzyme Linked Immunosorbent Assay (ELISA). For the proliferation assay, sevoflurane was used as an inhaled anesthetic and cervical dislocation as a method of euthanasia for mice. Details about experiments are described in the next sections.

PH (1-110) GFP NPs vs aluminum hydroxide (Alum)
From the dose response assay the treatment with PH

Immunological memory evaluation
In week 24 th week of the experiment PH  GFP NPs vs aluminum hydroxide (Alum), all groups received a boost with 5 μg of free GFP without adjuvant. Serum samples were obtained at day 4, 7, 14 and 21 post-immunization (Additional file 5).

Thermostability evaluation
Stocks of PH   condition was taken and 6 mice group (n = 5) including a control group (PBS) were immunized. The RTD NPs were dehydrated using a vacufuge TM concentrator 5301 (Eppendorf, Germany, cat. no. 5301) at a centrifugal force of 240 g at 30 °C for 30 min and were resuspended in PBS before being injected. In this experiment, no adjuvant was used. Blood sampling was performed for two months at 2-week intervals. incomplete Freund's adjuvant (IFA) (Sigma-Aldrich, USA, cat. no. F5506) were used. We decided to use the CFA and IFA in this experiment because, unlike Alum, broader stimulation of the cellular response has been previously observed [31,32]. Blood samples were taken for 6 weeks at 14-day intervals (Additional file 6A).

Flow cytometry analysis
Cell proliferation was evaluated using standard flow cytometry protocols. [33,34]

Enzyme-linked immunosorbent assay (ELISA)
To determine the presence of GFP-specific antibodies in immunized mouse sera, samples were analyzed by ELISA. ELISA analysis was carried out using microtiter plates (Corning,

Statistical analysis
All statistical analyses were performing using GraphPad Prism 7 software (GraphPad software, USA). Results were expressed as the means ± SD. All experiments were repeated at least once with comparable results. Data were analyzed by two-way ANOVA with a Tukey or Dunnett post-tests to correct for multiple comparison test. In the FRAP experiment to obtain the percentage of fluorescence recovery, the initial post-bleaching value (10 min) was subtracted from the last value obtained (140 min). To calculate FRAP differences was used in an unpaired, two-tailed Student's t-test. To determine the cutoff in the titration of antibodies, the previously described methodology was used [35]. In the lymphoproliferation assay to obtain the absolute percentage of proliferation the PBS+Alum group value was subtracted from the other groups. A p-value <0.05 was considered statistically significant. * p <0.05; ** p <0.01; *** p <0.001 and ns = not significant.

PH (1-110) peptide fused to GFP form nanoparticles (NPs).
We developed a universal system to produce fusion proteins using as template the first 110 amino acids from AcMNPV polyhedrin protein. A transfer plasmid containing the strong polyhedrin promoter drives the expression of the PH  followed by a poly-linker [36], which allows the insertion of any sequence to generate the fusion protein ( Figure 1A). In this particular case we introduced the sequence from the Green Fluorescent Protein (GFP) to produce the fusion protein PH  GFP [28]. This plasmid was utilized to produce recombinant baculovirus expressing the fusion protein in Sf9 insect cells. Sf9 insect cells infected with our recombinant baculovirus carrying the gene to expressed the fusion protein PH  GFP were sonicated to release the nanoparticles. Nanoparticles were centrifuged at low speed and the protein purified in this manner was subjected to SDS-PAGE analysis ( Figure 1B). As control we utilized a pure soluble form of GFP. As illustrated in the figure, a simple centrifugation results in highly pure PH  GFP protein, showing that the major component is the expected protein. Figure 1C illustrates an example of a Sf9 insect cells expressing PH  GFP visualized by confocal microscopy (for a 3D reconstruction of the nanoparticles please refer to Video, Additional file 1). Notice that all nanoparticles were contained within the nucleus (labeled with DAPI in blue). Electron microscopy (transmission electron microscopy in panel D and scanning electron microscopy in E) shows that the PH  GFP NPs are polydisperse. A capillary electroforesis analysis indicates that over 80% of the protein content is PH  GFP (see the Additional file 2). PH (1-110) nanoparticles. To determine the rigidity of the PH  GFP NPs, we conducted fluorescence recovery after photobleaching (FRAP) studies using confocal microscopy. The idea behind this study was that in a crystal structure the PH  GFP protein should have no diffusion, since a crystal lattice is rigid, whereas in a less rigid structure some diffusion should be observed. The speed of diffusion should be related to the laxity of the structure [37]. To conduct this experiment a region of interest (ROI) in each of the nanoparticles was photobleached to eliminate the fluorescence of GFP in the ROI. Fluorescence recovery inside the ROI was monitored for several hours.

Fusion protein have slow diffusion inside
Experiments were conducted with wild type polyhedrin fused to GFP (PH-WT-GFP) and particles formed by PH  GFP (Figure 2A). Fluorescence recovery was followed for 140 minutes. As expected, the particles formed by PH-WT-GFP showed no recovery after photobleaching, indicating the lack of mobility of the GFP inside the crystal. Most interestingly, the PH  GFP showed a partial recovery of fluorescence after 140 minutes. The time course of recovery after photobleaching is illustrated in figure 2B-C and quantification of recovery in figure 2D. PH  GFP NPs showed around 5% recovery after FRAP within the first 140 minutes, indicating a very slow diffusion of several hours.
PH  nanoparticles are purified with a single centrifugation. The use of a sucrose gradient aims in the separation of different nanoparticle sizes. Indeed the PH  GFP NPs are polydisperse. A Coomassie Blue Staining from the SDS-PAGE shows a main protein component, corresponding to the molecular weight expected for PH  GFP ( Figure 3A). A sucrose gradient from 40-60% facilitates the separation of PH  GFP NPs of different sizes, which are evident when subjected to confocal microscopy imaging ( Figure 3B). Because some of the nanoparticles are smaller than the light diffraction limit of light microscopy, we conducted an additional analysis using nanoparticle tracking analysis (NTA, Methods). This analysis method clearly evidenced the multiple sizes in nanoparticles with the most abundant nanoparticles at around 115 nm ( Figure 3C).

PH (1-110) nanoparticles produce a robust immune response in mice without
adjuvant. Using the PH  GFP NPs we immunized mice to evaluate the antibody response generated by our nanoparticles. Initially we explored if one or two immunization would make a difference in the immune response and found no significant differences (see the Additional file 3). The immunization protocol included 2 vaccinations a week apart ( Figure 4A). Blood samples were taken in two-week intervals for 24 weeks to assay for antibodies against GFP used as model antigen ( Figure 4A). GFP is poorly immunogenic and adjuvants are required in order to obtain antibodies when using soluble GFP as antigen. Most interestingly, ELISA assays show the generation of anti-GFP antibodies in all animals vaccinated with the PH  GFP NPs obtained from the sucrose gradients shown in Figure 3. As indicated by the data, no significant differences in the antibody generation were observed with any of the 40-60% sucrose gradient nanoparticles ( Figure 4B). When the different nanoparticle sizes where compared with the combined mixture (particle mix), no statistically significant differences in antibody levels were observed. Thus nanoparticle size does not appear to influence in the production of immunoglobulin IgG. Therefore, in the subsequent experiments we use the mixture of particles. The presence of anti-GFP antibodies was observed even after 24 weeks, indicating the induction of a long lasting immune response by the PH  GFP NPs ( Figure 4C). The antibody titers obtained with our PH  GFP NPs were high (12800 dilution) without the use of any adjuvant. Using the gold standard adjuvant aluminum hydroxide (Alum) resulted in higher antibody titers (51200, Figure 4D). High antibody titers were maintained for at least to 24 weeks postvaccination (see the Additional file 4). Most noticeable, when GFP was used in the absence of any adjuvant, no antibodies were produced (Fig. 4C-D, green triangles). These results show that even though the use of adjuvant improves the immune response, the PH  GFP NPs can induce a robust, long lasting immune response comparable to that obtained with the use of an adjuvant ( Figure 4C-D). Thus, our results indicate that nanoparticles have an adjuvant effect, since using free GFP (without the nanoparticles) requires adjuvant in order to induce a measurable immune response ( Figure 4C-D).
Furthermore, we evaluated if the vaccination with PH  GFP NPs may induce immunity memory. Using a protocol to determine if subsequent exposure to the antigen may reactivate the immune response clearly showed that the original vaccination with PH  GFP NPs induce long lasting immune memory (see the Additional file 5).

The immune response induced by PH (1-110) nanoparticles is both cellular and
humoral. Comparing pre-immune sera with sera obtained from mice after 8 weeks post vaccination we observed high levels of immunoglobulin IgG2a ( Figure 5A) and IgG2b ( Figure 5B) as well as IgG1 ( Figure 5C). The IgG2a/ IgG1 ratio obtained suggest a Th1 and Th2 mixed response, strongly suggesting that the PH  GFP NPs may induce the production of antibodies and moderate phagocytic activity ( Figure 5D). To further confirm that the vaccination with PH  GFP NPs induce also cellular immunity, we conducted a 20 cell proliferation experiment (see the Additional file 6) comparing the PH  GFP NPs against Freund's adjuvant that generates a strong cellular response [31]. As illustrated in the figure, the PH  GFP NPs induced a moderate cellular proliferation, characteristic of cellular immunity.
These results indicate that the PH  GFP NPs induce both humoral (antibodies) and cellular immunity.
PH  nanoparticles are stable for over one year at ambient temperature.
Vaccines and drugs are usually stored at 4 °C or frozen, for this reason we designed a protocol to maintain PH  GFP NPs for up to a year under the following conditions: 1) stored at -70°C, 2) at -20°C, 3) at 4°C, 4) at room temperature in solution and 5) at room temperature as dry powder. Every month during a year an aliquot of PH  GFP NPs stored at the different conditions specified above were used to immunize mice. The immune response was evaluated during the entire year by ELISA analysis of anti-GFP antibodies present in the sera from immunized animals. As illustrated in Figure 6, storing the PH  GFP NPs at 4°C impacted negatively in its ability to generate antibodies after six months and more evidently after 1 year of storage. Most interestingly, keeping the PH  GFP NPs at room temperature as dry powder (R.T.D.) maintained the efficacy of the nanoparticles to induce a robust immune response ( Figure 6A-D). Similar results were obtained with PH  GFP NPs stored at -20°C and -70°C. These results indicate that storing the PH  GFP NPs as dry powder is equivalent to storing the nanoparticles at -70°C, since its ability to induce a robust immune response is not compromised after a year of storage.

Discussion 21
The use of polyhedrin protein as a biotechnological tool has increased mainly due to the intrinsic capacity of self-aggregation [27]. In addition, we have shown that using only the first 110 amino acids of polyhedrin (PH  ) retains the self-aggregation property [28].
The fusion protein spontaneously forms nanoparticles ranging in size from 100 nanometers to 1 micron. Unlike the particles formed by wild type polyhedrin which show a uniform size of around 1-2 microns [28], nanoparticles formed by the peptide PH  are polydisperse ( Figure 3). Also the geometry of the wild type particles are polyhedral (hence the name polyhedra) while the nanoparticles produced by the peptide PH  are irregular.
Recently, this property has been used for biomedical purposes to incorporate antigens into nanoparticles for vaccine development [36]. However, the thermostability of the nanoparticles formed has not been characterized until now. In the present study we use GFP (a poorly immunogenic protein) to characterize by confocal microscopy the nanoparticles formed by the fusion protein PH  GFP.
We perform a microscopic characterization of the particles ( Figure 1C, 1D and 1E). Our experiments provide data on the physical nature of the nanoparticles formed. The wild type polyhedrin forms crystals of polyhedra, as demonstrated elsewhere [38]. Recently the crystal structure of wild type AcMNPV polyhedra has been elucidated using X-ray crystallography with a 3 Å resolution [39]. However, there are no studies aimed to determine the nature of the nanoparticles formed by PH  . This is important because it could help determine in later studies the degradation time of the PH  particles in vivo. We observed a slow recovery after FRAP in PH  GFP NPs (Figure 2) suggesting that the particles may form an dense structure similar to a liquid crystal [40]. Diffusion coefficients of proteins in agarose gel have been experimentally determined in dozens of 22 hours [37].
One of the main problems in modern vaccines is the costly and lengthily purification process [10]. For this reason, the first step was to obtain a highly pure antigen easily and quickly. Because the PH  GFP NPs are insoluble in aqueous solutions, their purification is simple involving a one-step centrifugation process at low speed [28,36]. These results show that the recombinant baculovirus system can produce large amounts of our fusion protein and that a simple centrifugation step results in highly pure PH  GFP NPs (Figure 3 and S1).
By analyzing the type of immunoglobulins produced during an immune response, one can estimate if the response is T helper type 1 (Th1) or type 2 (Th2). Th1 lymphocytes stimulate Th1 immunity, which is characterized by intense phagocytic activity; while Th2 cells stimulate type 2 immunity, characterized by high antibody levels [41,42]. A mixed response is desired in vaccines, which gives greater effectiveness in preventing diseases [41]. In the sera of mice immunized with our PH  GFP NPs we observed the stimulation of both the Th1 and Th2 response, in addition, high antibody and long duration titers were observed. The immuniglobulin profile clearly shows the mixed response, which is consistent with the lymphoproliferation analysis. However, this study was not aimed to an exhaustive analysis of the immune response, which can include the evaluation of subpopulations of T lymphocytes as well as cytokines induced by the nanoparticles.
On the other hand, an important finding was the adjuvant effect of the PH  peptide.
Adjuvants are primarily designed to improve the presentation of antigens, increase the immune response, as well as reduce doses [4]. In PH  GFP NPs a robust immune response was observed without the need for adjuvant. The use of only PH  GFP NPs without adjuvant achieved the same response as that achieved with aluminum hydroxide 23 ( Figure 4C). The antibody titres remained high for more than 14 weeks ( Figure S3) and a robust immunological memory was achieved ( Figure S4).
It has been observed that the size of some particles affects the immune response [43,44].
PH  GFP particles of different sizes were evaluated without finding a difference in the response of immunoglubulin G.
Finally, the most relevant finding is the thermostability of PH   GFP NPs. An ideal vaccine should be also thermostable to avoid the so-called "cold chain", which represents approximately 80% of the price of modern vaccines [17,18]. There are biomaterials that have shown thermostability, however, few have shown stability after one year at room temperature [23]. Our result strongly suggest that the PH  fragment retains the capacity to preserve proteins found in the wild type polyhedra [24,25]. Interestingly, storing nanoparticles at 4°C for more than six months compromised their ability to generate a robust immune response, suggesting that the tridimensional structure o integrity of the antigen was compromised during storage at this particular temperature.
This was not observed with nanoparticles stored at room temperature or frozen (-20°C or -70°C). The reason behind this result remains to be explored.

Conclusions
We have developed a universal system to generate nanoparticles using peptides and proteins of interest as antigens. We coupled our method to the baculovirus expression system in order to generate large amounts of our fusion protein. Nanoparticles are purified by a single centrifugation step, showing purity higher than 80%. The nanoparticles are stable for at least one year at room temperature, preserving the antigenicity of the proteins of interest. This finding opens the possibility to significantly reduce the costs of conservation and distribution of vaccines.
Immunization with nanoparticles results in a robust humoral and cellular immunity.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.     The antibody response to GFP induced by the polyehedrin nanoparticles suggests a Th1 and Th2 profile. After 8 weeks when the immune response was established immunoglobulin G subtypes were evaluated by ELISA; A. IgG2a, B. IgG2b and C.
IgG1 showing no significant difference between the groups: GFP+Alum, PH(1-110)GFP and PH(1-110)GFP+Alum, but these groups had at least a p-value <0.05 against the PBS+Alum and GFP groups. D. The IgG2a/IgG1 ratio was obtained to determine the type of predominant profile (Th1 or Th2) of the immune response generated by the PH(1-110)GFP NPs. Error bars indicate the means ± SD (n = 5). *** p <0.001; ns = not significant. (Two-way ANOVA with Tukey post-tests).