Biodegradable polymeric PLGA-PEI NPs were prepared by the solvent extraction-evaporation of a double emulsion (w/o/w) method. DLS was used to analyse the Z-average diameter and PdI of the NPs, and the Z-potential was determined through Laser Doppler micro-electrophoresis. Both of them measured by a Malvern® Zetasizer NanoZS Model ZEN3600 (Malvern Instruments Ltd., UK). NPs exhibited homogeneous size and low PdI (below 0.2), which confirms narrow size distribution. Positive Z-potential values were also obtained, which confirms efficient PEI coating. These results are consistent with other PLGA NPs functionalised with PEI [50, 51], and meet the requirements for cellular uptake as nanometric particles have demonstrated better uptake than bigger particles [52]. Moreover, a similar encapsulation efficiency (EE%) and protein loading for both peptides were observed (Table 2), which fall within the usual range of antigen encapsulation in PLGA NPs [53, 54].

Poly(I:C) is a negatively charged molecule that can interact with the positively charged surface of NPs. Due to the ability of a single poly(I:C) molecule to bind with multiple NPs simultaneously, establishing the correct NP:poly(I:C) ratio is crucial to avoid NP aggregation and ensure full surface coverage (Fig. 1). To achieve this, we first determined the optimal NP concentration by using a fixed concentration of poly(I:C) and increasing concentrations of NPs. After identifying the optimal NP concentration, we then determined the best concentration of poly(I:C) by testing a range of concentrations. Size analysis was used as a means of detecting potential NP aggregation.

w/o/w PLGA-PEI nanoparticles with poly(I:C) covering and charge interaction diagram. A Positively charged PLGA-PEI NP covered with negatively charged poly(I:C). B Suboptimal coverage of the poly(I:C) causes the different charges of the NPs to interact with each other causing them to aggregate. C Complete poly(I:C) covering of the NPs. Created with BioRender.com

An important factor to bear in mind when covering the NPs is that the interaction among NPs affects the detected size. Positively charged particles covered with negatively charged poly(I:C), can interact with poly(I:C) attached to other particles. This charge interaction affects NP movement in the fluid – or Brownian motion –, and therefore, greater sizes were detected.

To stablish the optimal NP:poly(I:C) ratio, firstly, 150 µg/ml of poly(I:C) was used at different NP concentrations, ranging from 0.9–15 mg/ml. We observed a greater tendency for aggregation at 15 mg/ml NP concentration, which indicated high NP:poly(I:C) ratio. However, the smaller ratios significantly reduced NP aggregation, resulting in an adequate NP size (Fig. 2A). For the following experiments, the concentration of 7.5 mg/ml NPs was selected as it allowed an efficient coating of the NPs with the smallest possible volume, thus optimising the experimental conditions.

NP characterization. A Size of NPs with a specific concentration of poly(I:C) (150 µg/ml) (***p < 0.001 with respect to 15 mg/ml NP). B Size of NPs (7,5 mg/ml) coated with increasing concentrations of poly(I:C) (*p < 0.05 and *p < 0.005 with respect to 50 µg/ml poly(I:C)). C Z-potential of NPs (7,5 mg/ml) covered with different poly(I:C) concentrations. Increasing poly(I:C) concentrations lowered Z-potential values until no statistical differences were observed form 150 µg/ml on (*p < 0.05 and ***p < 0.001 in regard to 100 µg/ml of poly(I:C)). D Percentage of incorporated poly(I:C) that had been efficiently loaded onto the NPs (*p < 0.05 and ***p < 0.001 in regard to 50 µg/ml of poly(I:C)). E Poly(I:C) quantity loaded per 1 mg of NP. F Size determination of NP covered with poly(I:C) in NTA. G, H NP morphology with H or without G 250 µg/ml poly(I:C), in SEM

Next, poly(I:C) concentration was stablished. The NP concentration was maintained at 7.5 mg/ml and the poly(I:C) concentration ranged from 50 to 350 µg/ml. Following previous observations, low concentration was related to higher aggregation tendency due to its capacity to bind more than one NP. On the contrary, high poly(I:C) concentration allowed complete NP covering, less NP interaction, and finally, less aggregation tendency. For nano size, at least 100 µg/ml of poly(I:C) is needed (Fig. 2B). Same experiment was carried out in more acidic environments (pH 4.6) to study the best conditions for the NPs, however, we observed that an acidic pH did not allow the desired studies as it favoured the precipitation of the NPs (Supplementary Table 1). Although the pH of melanoma is slightly acidic, the intention of this vaccine is phagocytosis by APCs, with special interest in DCs, so its administration is intended to be far from the tumour to be treated, and therefore far from its acidic environment. As a result, the acidic pH was discarded for the following experiments, which is interesting as a more neutral pH would allow for subcutaneous administration.

Negatively charged poly(I:C) addition to positively charged NP decreased surface Z-potential. As in size, lower immunostimulant concentrations showed lower NP covering, leading to positive Z-potential values. While increasing poly(I:C) concentration, Z-potential decreased until around -20mV values were achieved (Fig. 2C). Results showed good covering efficiency and higher attachment in a dose dependent manner, with no statistical differences from 150 µg/ml on.

Next we wanted to determine the binding efficiency of the poly(I:C) to the NPs, as well as poly(I:C) quantity loaded per mg of NP. Concentrations below 50 µg/ml showed complete poly(I:C) attachment, while higher ones led to lower covering efficiency (Fig. 2D). Meanwhile, Fig. 2E shows an increased attachment at higher concentrations, although no statistical differences were observed. This data indicates that NP surface was completely covered with concentrations above 100 µg/ml of poly(I:C). Moreover, results were in accordance with aggregation and surface potential values (Fig. 2B, C).

Taken together, the results showed that complete covering of NPs is achieved with 7.5 mg/ml of NPs and with poly(I:C) concentrations at least of 150 µg/ml. However, more reproducible values with smaller deviation were detected with higher poly(I:C) concentrations, so following experiments were carried out with 250 µg/ml poly(I:C).

NP size alterations were confirmed with NTA (Fig. 2F) and SEM images (Fig. 2G, H), in which poly(I:C) covered NPs showed same sizes as non-covered ones. This confirms NP size was not altered in the covering process and that the size differences observed with Zetasizer equipment were due to the charge interaction between NPs.

Overall, we can conclude that poly(I:C) covering was successfully achieved by surface functionalisation of PLGA-PEI NPs.

Phagocytosis and cytotoxicity studies were performed to ensure NPs uptake by cells and assess the potential NP toxicity.

In phagocytosis studies, RAW 264.7 macrophage cell line was cultured with NPs dyed with DiD. After 24 h, cells were fixed in microscopy covers and stained with phalloidine and DAPI and observed under microscopy. In Fig. 3A we see the NPs in red, the cytoplasm in green and the nucleus in blue. The image shows efficient NP uptake and localization within the cytoplasm, surrounding the nucleus.

NP internalization. A NP internalizations observed in microscopy. RAW 264.7 cell membranes are stained in green with Alexa488-phalloidine. NPs are stained in red with DiD. Blue nucleus in DAPI. B Quantification of viable cells by CCK8 to determine cytotoxicity of NPs. C Phagocytosis of NPs by DCs was analysed by flow cytometry, in which an increase in complexation (SSC) demonstrates uptake of NPs

To assess NP cytocompatibility, they were incubated with RAW 264.7 cells. Cells were cultured with two different concentrations of poly(I:C) or NP poly(I:C) (1 and 5 µg/ml), and another group was included with uncoated NPs in an equivalent amount. After 24h, cell viability was measured by CCK-8 assay. The results illustrated in Fig. 3B show that none of the concentrations tested were cytotoxic (> 70%. cell viability).

Finally, cytometry analysis carried out in DCs confirmed NP uptake, in which DCs treated with NPs exhibit more complexity (SSC signal increase) (Fig. 3C). In particular, iDC and mDC groups have about 4% of the population in the complexity gating, while the NP-treated group has more than 40% of the population.

Adequate DC activation is crucial for T-cell-targeted cancer vaccines, as it allows for proper antigen presentation, and thus effective activation of the immune response [55]. Therefore, to select the most appropriate NP poly(I:C) dose, the following experiments were carried out on the target cells, i.e. dendritic cells (DCs). For this purpose, we examined whether NPs promote the up-regulation of maturation markers (HLA-DR, CD80, CD83 and CD86) and cytokine release (TNF-α) by human DCs.

Following previous experiments in RAW 264.7 cells, in which the two doses tested were non-toxic, we treated monocyte-derived iDCs with 1 and 5 µg/ml of poly(I:C), NP poly(I:C) and uncoated NP (NP). In this case, we compared the dose of 1 µg/ml poly(I:C) stimulating for 24 h, and 5 µg/ml poly(I:C) in culture for 6 h. As a maturation control, mDCs were prepared using a proinflammatory cytokine cocktail.

Results revealed that the dose of 5 µg/ml NPpoly(I:C) for 6 hours was too stimulating (Supplementary Fig. 1). A reduction in phagocytic capacity compared to uncoated cells was observed, since a decrease in SSC was apparent. In addition, the trend of the CD14 marker changed drastically because the increase in complexity was observed only in CD14− cells, which were also negative for the maturation markers (Supplementary Fig. 2), and therefore, were not DCs. Furthermore, no maturation differences were detected among DCs of stimulated groups. Hence, even reducing the incubation time to 6h, a dose of NP poly(I:C) equivalent to 5 µg/ml of poly(I:C) was still too strong a stimulus for this cell type. Therefore, we concluded that 5 µg/ml NP poly(I:C) for 6 hours was not appropriate for DC stimulation.

In contrast, the 1 µg/ml dose of poly(I:C) provided optimal maturation of DCs (Fig. 4A). On the one hand, the cells efficiently took up the NPs and increased their complexity accordingly. On the other hand, DCs increased maturation markers after stimulation, mainly in the poly(I:C) containing groups. In particular, compared to iDCs, non-covered NPs were able to stimulate the activation of DCs for all three markers, probably because the structure of the NPs resemble virus particles. However, with the addition of poly(I:C), the maturation capacity was enhanced. Indeed, both NPpoly(I:C) and free poly(I:C) show similar maturation to control mDCs in HLA+CD80+ and HLA+CD86+ graphs. Therefore, the addition a dose of NPpoly(I:C) equivalent to 1 µg/ml of poly(I:C) seemed to help in its purpose of maturing DCs.

DC maturation with 1 µg/ml poly(I:C). A Representative flow cytometry plots from DCs after maturation with 1 µg/ml poly(I:C), NPpoly(I:C) and non-covered NPs. B DC maturation markers are represented in cell percentage (cell%). C TNF-α cytokine secretion was measured by ELISA. (*p < 0.05, **p < 0.005 and ***p < 0.001 in regard to NPpoly(I:C))

Afterwards, TNF-α secretion of the 1 µg/ml poly(I:C) group was determined by ELISA. As shown in Fig. 4C, TNF-α secretion correlated with flow cytometry results since the poly(I:C)-containing groups showed the highest levels of the detected cytokine. Although the quantified amount of TNF-α in both groups were lower than the positive maturation control (mDC), the levels of TNF-α secreted appear to be sufficient to induce an adequate immune response.

Taking into account the ELISA results together with the observations made in flow cytometry, the results showed a better maturation profile with the poly(I:C)-containing groups compared to the uncovered NPs. Overall, the main conclusion is that the incorporation of poly(I:C) on NPs improves DC activation. However, it should be highlighted that although the level of DC maturation obtained with NPpoly(I:C) is similar to that achieved with free poly(I:C), it offers several advantages. On the one hand, the inclusion of the adjuvant in a nanoparticulate system manages to combine its effects and direct its immunostimulatory capacity to the same phagocytic cell, thus reducing off-target effects. On the other hand, the formulation developed allows the incorporation of another compound to direct the immune activation to the target of interest, which would allow a specific response to be achieved.

After testing the maturation capacity of the NPs, the lymphocyte activating capacity was tested. In particular, the specific response generated by the NPs.

For this purpose, we carried out an activation-induced marker (AIM) assay in PBMCs of healthy donors: this approach allows for the identification of T cells recently activated by antigen, measured as the upregulation of CD154 and CD69 in CD4+ T cells; the quantification of CD69 upregulation also allows for the interrogation of the activation status of CD8+ T cells [56,57,58]. CD69 is a membrane receptor used as an early marker of activation. Its expression is low in resting lymphocytes but increases rapidly after cell activation [59]. Meanwhile, CD154 —also known as CD40L— is a transmembrane molecule that is temporarily expressed on activated CD4+ T cells following T-cell receptor (TCR) stimulation, making its expression antigen-dependent. CD40L interacts with CD40 expressed on antigen-presenting B cells and DCs to induce antibody formation and cellular immune responses [60].

PBMCs were incubated overnight with 1 µM of each peptide or the equivalent amount of poly(I:C) or NPs. After, cells were harvested and stained for flow cytometry analysis (Supplementary Fig. 3). Cytometry results of the activated cells were donor-dependent and varied significantly among individuals. Consequently, the results were interpreted as a ratio between the groups of interest, and the statistics were obtained by pairing the responses of each individual for both groups compared. In particular, there were three questions that needed to be addressed in this section.

First of all, does encapsulating the Ag enhance the lymphocyte response? The first question was assessed comparing free Ag and encapsulated Ag cytometry results (NP(Ag) vs Ag). As shown in Fig. 5A, B, after peptide encapsulation both types of T cells increased the population expressing the activation markers analysed. In particular, all donors reacted more strongly to NP(Ag) than to free Ag, probably because NPs favour antigen uptake and presentation by APCs.

Specific response studies of T cells performed in PBMCs. Ag, NP(Ag), NPpoly(I:C) and NP(Ag)poly(I:C) effect comparison on A CD4+ T cells and B CD4− T cells. C IFN-\(\gamma\) cytokine release. (*p < 0.05, **p < 0.005 and ***p < 0.001)

Secondly, does the poly(I:C) in the formulation improve the ability to activate lymphocytes? The effect of the poly(I:C) on the activation capacity of the NPs was stablished by comparing the results of Ag containing NPs with poly(I:C) (NP(Ag)poly(I:C)) with the results of NP without poly(I:C) (NP(Ag)). As shown in Fig. 5A, B, all donors increased the percentage of activated CD4− cells with NP(Ag)poly(I:C), however, there was no significant improvement on activated CD4+ cells. Evidence suggests that poly(I:C) acts primarily on cytotoxic lymphocytes as their activation is greatly enhanced by the addition of the adjuvant to the particle, which is of interest in cancer immunotherapy [61].

And third, is the effect enhanced by adding the Ag and the poly(I:C)? For that aim, two NPs with poly(I:C) were compared (NPpoly(I:C) vs NP(Ag)poly(I:C)), and the activation gap among them was related to the presence of the Ag in the formulation, thereby detecting the specific response. Figure 5A, B shows how the addition of the Ag to the formulation increased the percentage of activated cells, once again mainly in the CD4− group with 90% of donors responding. Surprisingly, in the CD4+ group there was no statistical difference between the group with or without Ag in the formulation, but a modest response can be appreciated with a 60% of responding donors. However, it should be noted that the selected Ag does not come from any of the volunteers tested, so it could be expected that by using neoantigens from the patients themselves, these responses would be improved [62, 63].

Cytometry data was supplemented with an IFN-\(\gamma\) ELISA. IFN-\(\gamma\) is a cytokine secreted mainly by activated T-lymphocytes and its functions include: regulating the immune response, stimulating antigen presentation on MHC-I and MHC-II molecules, coordinating leukocyte-endothelium interaction, and controlling cell proliferation and apoptosis [64]. ELISA results showed that IFN-\(\gamma\) secretion was highly influenced by the presence of poly(I:C). The poly(I:C) groups secreted significantly more IFN-\(\gamma\) than those without adjuvant, and yet, although the presence of the Ag hints at an upward secretion trend, the groups did not differ significantly (Fig. 5C). This implies that the presence of poly(I:C) promotes the activation of lymphocytes and produces a large release of IFN-\(\gamma\), and therefore favors the execution of lymphocyte functions.

In summary, nanoparticulating the neoantigen peptides and adding poly(I:C) improves the immune response by increasing the number of active cells and their degree of activation, but without affecting antigen-specific recognition, enabling personalised and effective activation of the immune system. On the other hand, the addition of poly(I:C) to the formulation enhances the lymphocyte response by CD4− cells and promotes IFN-\(\gamma\) secretion. This is of great importance for a complete and competent cytotoxic response [65, 66]. Compared to previously published work on poly(I:C) in PLGA particles for immunotherapy, our vaccine is the first nano-sized, PEI-functionalised, poly(I:C)-coated particles that encapsulates melanoma patient-derived neoantigens [37, 67,68,69]. In addition, it is the only one that tests NPs in human PBMC to analyse the specific response generated. All of this makes the developed vaccine a novelty in various aspects of particle development and experimental trial design. Nevertheless, it should be noted that the selection of the personalised neoantigen was based on the expression on the patient's tumour cells, and on the affinity with their HLA. The results obtained, however, come from PBMC from healthy volunteers, where the frequency of lymphocytes responding to these patient-derived antigens is very low. Thus, although this is a clear limitation of the assay, it is likely that in the case of patients themselves this response would be favoured [49]. In summary, the incorporation of poly(I:C) onto the peptide-based vaccine promotes and amplifies the activation of the immune response, which is expected to be greater in patients.