The sequences of KLH1 and of murine AM were obtained from the Universal Protein Database (UniProt). Sequences were reverse translated to the proper DNA sequence by using BCCM/GeneCorner (https://www.genecorner.ugent.be/https://www.genecorner.ugent.be/rev_trans.html) online tool.

The KLH1 protein functional sub-units (FU) a, b, c, d, e, f, g and h were analyzed looking for B cell and T cell epitope binding and immunogenicity prediction using the free epitope database and prediction resource (IEDB; http://www.iedb.org), as performed before (28, 29). Other free algorithms used to characterize KLH1 FUs were Bepipred for presence of B-linear epitopes, Kolaskar and Tongaonkar for immunogenicity prediction, Chou and Fasman for Beta Turn prediction, Karplus and Schulz for flexibility prediction, Parker for hydrophilicity prediction, and Emini for Surface Accessibility prediction. Finally, IEDB server Discotope portal was used for prediction of conformational epitopes on KLH1 protein sub-units ( Supplementary Figure S1 ).

The KLH1 FUs were further analyzed for the presence of epitopes that can bind to MHC I and II (HLA) complexes/alleles by using IEDB server MHC binding prediction portal. For MHC I, mouse alleles included in the prediction were H-2Db, H-2Dd, H-2Kb, H-2Kd, H-2Kk and H-2Ld. The prediction method used was NetMHCpan EL 4.1. and binding capacity was evaluated based on score value, where a low score correlates with higher binding capacity. MHC II mouse alleles included H-2IAb, H-2IAd and H-2IEd and the consensus method was applied for prediction. The method is based on the rank values and lower values correlate with better binders. Additionally, prediction of MHCI-KLH1 sub-unit peptide complex immunogenicity was performed using IEDB server MHCI immunogenicity prediction tool ( Supplementary Figure S2 ).

Based on the scores obtained in all analyses, the KLH1 FU with the highest scores was further evaluated. Different sequence fragments of this FU were analyzed using all methods described above in order to find the optimal fragment to be bound to mouse AM, looking for the highest scores and the minimum size of the fragment. The chosen fragment´s DNA sequence was then bound to mouse AM´s sequence to code for the fusion protein that was to constitute the vaccine´s antigen.

Besides the open reading frame (ORF) for the fusion protein, the final expression construct was complemented with the insertion of a typical 3’untranslated region (3’UTR) and a 5’untranslated region (5’UTR) that are necessary for mRNA synthesis ( Table 1 ). In addition, a standard Kozak sequence (gccaccatg) and stop codon (tag) were added. The final sequence was optimized by using Integrated DNA Technologies (IDT, eu.idtdna.com), Benchling (www.benchling.com) and GenScript (www.genscript.com) free sequence/codon optimization tools. The optimization was performed aiming to reduce rare codon presence, rebalance codon usage with emphasis on GC couples, decrease complexity, decrease secondary structure formation and uridine depletion possibilities. The sequence was built by gene synthesis and inserted into the pcDNA 3.1 plasmid by GenScript Biotech (Piscataway, NJ, USA) downstream of the T7 promoter ( Figure 1a ).

The expression vector was inserted into competent E. coli following standard heat shock protocols and selection in ampicillin plates. The vector was purified using Plasmid DNA MaxiPrep kit (Qiagen, Hilden, Germany) following manufacturer´s instructions. The DNA template was linearized using BssHII restriction enzyme (NewEngland BioLabs, Ipswich, UK) and purified by isopropanol precipitation. The linearized template was resuspended in HyClone HighPure water (Cytiva, Marlborough, MA, USA).

The therapeutic mRNA was synthesized by in vitro transcription (IVT) using T7-Scribe™ Standard RNA IVT Kit (CellScript, Madison, WI, USA) to produce standard/non-modified mRNA. Alternatively, for in vivo experiments, modified mRNA was generated by using the INCOGNITO™ T7 5mC- & Ψ-RNA Transcription Kit (CellScript), following manufacturer´s instructions. Briefly, the linearized DNA template was mixed with components of the kit and reaction incubated at 37°C for 2h. Then, DNAse I was added to remove traces of the DNA template. The reaction product was purified by standard ammonium acetate precipitation, and dissolved in RNAse-free water. The amount and purity of the RNA were analyzed on a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

Capping of modified IVT-mRNA was done post-transcriptionally using the ScriptCap™ Cap 1 Capping System (CellScript). The Cap1 structure was added by using both ScriptCap Capping Enzyme, and ScriptCap 2’-O-Methyltransferase, according to manufacturer’s instructions. The reaction was performed at 37°C for 2h. The resulting capped mRNA was further characterized by electrophoresis.

LNPs were prepared using the ionizable lipid, 4-(dimethylamino)-butanoic acid, (10Z,13Z)-1-(9Z,12Z)-9,12-octadecadien-1-yl-10,13 nonadecadien-1-yl ester (DLin-MC3-DMA, Cayman Chemical, Ann Arbor, MI, USA). In addition, three helper lipids were added, including disteraroylphosphatidylcholine (DSPC, Avanti Polar Lipids, Alabaster, AL, USA), 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol (DMG-PEG 2000, Cayman Chemical), and cholesterol (Thermo Fisher Scientific) ( Figure 1b ). To generate the particles, the lipids were dissolved in ethanol whereas the mRNA was dissolved in citrate buffer (pH 4.0). These solutions were mixed at a 1:3 volume ratio, to achieve a 40:1 weight lipid/mRNA ratio, by rapid pippeting (30). Constructs were allowed to stabilize for 15 minutes and then were transferred to Amicon Ultrafilter tubes, where they were centrifuged at 13,000 x g, at 4°C, for 30 min. The flow-through was kept for subsequent analyses, whereas the solution remaining inside the tubes was transferred to fresh tubes and centrifuged at 13,000 x g, at 4°C, for 2 min. Nanoparticles where in the pellet of the last spin. Control LNPs were prepared following the same protocol in the absence of mRNA.

Size, polydispersity (PDI) and zeta potential of LNPs encapsulating or not (control) mRNA were analyzed with Zetasizer (Malvern, Worcestershire, UK) according to manufacturer´s instructions. Briefly, samples were diluted in PBS and injected into specialized capillaries. Samples were transferred to standard cuvettes and their size and PDI measured. Six measurements were taken per sample and the mean ± SD was calculated.

To determine encapsulation efficiency, LNPs (both empty and encapsulating mRNA) were separated in two groups; LNPs in Group 1 were solubilized with 2% Triton X-100 and Tris-EDTA and heated at 37°C for 10 min while shaking. On the other hand, LNPs in Group 2 were not solubilized. Naked mRNA was used as a control. Finally, all samples were electrophoresed in 1.0% agarose gels and images recorded with a gel imaging system (BioRad, Hercules, CA, USA).

Furthermore, fluorimetric measurements were performed by using a high sensitivity kit for detection of RNA (HS RNA, Invitrogen, Carlsbad, CA, USA) and a Qubit fluorometer (Thermo Fisher Scientific) following manufacturer´s instructions. Final calculations were done based on the amount of mRNA present in solubilized samples divided by the initial mRNA used for encapsulation and multiplied by 100, with the formula:

Murine cell lines RAW 264.7 (RRID CVCL_0493, macrophages), B16-F10 (RRID CVCL_0159, melanoma) and PC12 (RRID CVCL_0481, pheochromocytoma) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). They were cultured in Roswell Park Memorial Institute (RPMI 1640, Corning, Manassas, VA, USA), for B16-F10, or in Dulbecco´s Modified Eagle´s Medium (DMEM, Corning) for the other cell lines; all of them supplemented with 10% fetal bovine serum (FBS, Thermo Fisher Scientific), at 37°C and 5% CO₂ in a humidified incubator. Cell lines were authenticated by STR profiling (IDEXX BioAnalytics, Kornwestheim, Germany).

The KLH-AM mRNA was labeled with the HighYield T7 Cy5 RNA Labeling Kit (Jena Bioscience, Jena, Germany), following manufacturer´s instructions. The Cy5-labeled mRNA, as well as unlabeled mRNA as a control, were encapsulated into LNPs. RAW 264.7 cells were seeded into 2-well chambered coverslips (Ibidi, Gräfelfing, Germany) and incubated until reaching 70-90% confluence. Then the LNPs encapsulating the labeled (or not) mRNAs were added. After 2 h, cells were washed thrice with PBS and incubated in fresh media. Cells were observed in an LSM 980 confocal microscope (Zeiss, Oberkochen, Germany) and live imaging was performed.

All experiments and animal procedures were carried out following the guidelines laid down by the European Parliament and Counsel (2010/63/UE) and Spanish Royal Decree 53/2013 on animal experimentation, and were revised and approved by the Committee for Animal Welfare of the Center for Biomedical Research of La Rioja (OEBA/CIBIR), Spain, ref. AMR-17.

Immunocompetent mice (7-week-old male C57BL/6 mice) were purchased from Inotiv (Indianapolis, IN, USA). In addition, transgenic mice that express green fluorescent protein (GFP) under the alpha smooth muscle actin (αSMA) promoter, also of C57BL/6 genetic background, were used to investigate vaccine effects on the existing vasculature. These mice (SMAA-GFP) were generated at the National Eye Institute (NEI, NIH, Bethesda, MD, USA) (31) and have been used mainly to study angiogenesis in the context of ophthalmological diseases (32). These animals were a generous gift from late Prof. John Paul Sangiovanni (NEI).

Animals were housed under specific pathogen-free conditions at the CIBIR animal facility (ES260890000992), in a temperature-controlled room with 12-h light/dark cycle and reared on standard chow and water provided ad libitum. All animals were checked by a veterinarian, and weighed, on a weekly basis, to ensure animal wellbeing and lack of vaccine toxicity.

For immunization, C57BL/6 mice were randomly divided into two experimental groups: control (n=10) and treated (n=10). The control group received empty LNPs, while the treated group received LNPs containing the KLH-AM mRNA. Animals were injected intramuscularly (i.m.) in the rear hind muscle with 50 µl of the nanovaccine containing a total of 5.0 µg of mRNA (or with empty LNPs for the control). Five immunizations were implemented at two-week intervals, with the last injection given 3 days after intravenous (i.v.) tumor challenge with B16-F10 cells. Animals were sacrificed 3 weeks after tumor challenge and blood and tissues were collected for further analysis.

Furthermore, eighth male SMAA-GFP mice were used. As with the previous group, these animals were randomly divided into two experimental groups: control (n=4) and treated (n=4) and subjected to five immunizations with either control (empty LNPs) or with LNPs containing the KLH-AM mRNA. Three weeks after the final immunization, all animals were sacrificed. After sacrifice, fresh highly vascularized tissues were photographed under UV-light (DM6000B microscope, Leica, Wetzlar, Germany) to assess potential side effects of the vaccine on existing blood vessels.

To test the antitumor effects of the vaccine, immunized animals were challenged with syngeneic B16-F10 melanoma tumor cells. For the metastasis model, mice were inoculated i.v. through the tail vein with 2 x 10⁵ B16-F10 cells/animal in order to induce lung metastases, as reported (33). Three weeks after tumor challenge, mice were sacrificed, blood was taken for antibody titration, spleens were collected for lymphocyte analysis and the number and size of metastases was assessed on the lungs.

After dissection, lungs were weighed, photographed and the number of visible metastases was recorded. Additionally, the size of the largest metastatic node was determined with a digital caliper for each animal. After that, the left pulmonary lobe was snap frozen in liquid nitrogen and stored for molecular analyses. The remaining lobes were insufflated with 10% formalin, fixed for 24h, and paraffin embedded. Tissue sections (3 µm-thick) were stained with hematoxylin and eosin and the number of metastases per section was assessed under a light microscope (DM6000B, Leica).

Additional sections were rehydrated and antigen retrieval was performed by heating in citrate buffer (pH 6.0) for 20 min at 96°C. After blocking with normal donkey serum, sections were incubated overnight with primary antibodies including anti-CD31 and anti-Ki67 ( Table 2 ). The next day, following several washes in PBS, sections were incubated with a polymer complex containing HRP (RE7200-CE, Leica, RRID AB_3674357). Immunoreactivity was developed with diaminobenzidine (DAB, Dako). Slides were lightly counterstained with hematoxylin. Pictures were taken in random fields (n=10 per mouse) and the area occupied by the blood vessels (for CD31) or the number of positive cells (Ki67) per mm² was quantified through free digital pathology QuPath software (34).

Sera was isolated from blood following standard protocols. For ELISA, COSTAR high-binding 96-well plates (Corning, New York, NY, US) were coated with 100 ng/50 µL/well of mouse AM peptide (Phoenix Pharmaceuticals, Burlingam, CA, USA), dissolved in carbonate coating buffer pH 9.6, and incubated overnight at 4°C. The next day, plates were washed with PBS and then blocked with 1.5% FBS in PBS for 1h at RT with shaking. Sera samples were serially diluted in PBS (from 1:50 to 1:1,600), added to wells and incubated at RT with shaking for 2h. Wells were washed again and incubated with anti-mouse secondary antibody bound to HRP ( Table 2 ) at a 1:10,000 dilution for 1h. After washing again, 3,3′,5,5′-Tetramethylbenzidine (TMB) substrate was added to each well. Color development was stopped with 3N HCl and plates were analyzed on a microplate reader (POLARstar Omega, BMG Labtech, Germany) at 492 nm.

In addition, to ensure that the IgGs identified by ELISA were able to recognize intrinsic cell-produced AM, PC12 cells were used as a control since they have been shown to express high levels of AM (35). Since these cells do not stick strongly to the substrate, they were collected into cytospin slides. Five thousand cells were added to each cytospin funnel and they were centrifuged at 1,500 rpm for 5 min in a Shandon Cytospin 4 (Thermo Fisher Scientific). Then, cells were fixed in 4% paraformaldehyde for 10 min, washed with PBS and permeabilized with 1.0% Triton X-100 in PBS for 10 more min. Slides were blocked with 5.0% donkey serum. Then, cells were incubated overnight at 4°C with either serum from vaccinated animals at a 1/100 dilution or serum isolated from control animals at the same dilution. In addition, other slides were exposed to PBS as a negative control. The next day, cells were washed with PBS and then incubated with secondary anti-mouse IgG AlexaFluor 647 antibody ( Table 2 ) and 1/1,000 Hoechst 33342 (Sigma Aldrich) as a nuclear stain, for 1 h at RT, in the dark. Following incubation, cells were washed several times with PBS, slides were mounted and imaging was performed using an SP5 confocal microscope (Leica).

Spleens were collected from immunized animals, minced in small pieces and passed through 70 µm cell strainers (Corning). Resulting cells were resuspended in RPMI 1640 media supplemented with 1.0% penicillin/streptomycin (Gibco). After three washes to remove cell debris, cells were exposed to Ammonium–Chloride–Potassium (ACK) red blood cell lysis buffer (Thermo Fisher Scientific) on ice. Then several washes were performed until a clear pellet was obtained. Finally, cells were resuspended in RPMI 1640 media supplemented with 10% FBS and 1.0% penicillin/streptomycin.

For cell labeling, aliquots of 1.0 x 10⁶ cells per animal were prepared. Staining was performed for surface markers CD4 and CD8 according to standard protocols. Briefly, cells were resuspended in FACS buffer (2.0% FBS plus 0.01% ammonium azide in PBS). Then, cells were incubated with Fc block (anti-CD16/32, BD, Franklin Lakes, NJ, USA) in FACS buffer for 5 min on ice. Cells were then washed, centrifuged at 1,500 rpm at 4°C, and the supernatant discarded. Cells were then exposed to an antibody mixture of CD4 APC-Cy7 and CD8 FITC ( Table 2 ) in FACS buffer and incubated for 30 min on ice, in the dark. Flow cytometry was performed on a FACSCantoII cytometer (BD).

B16-F10 cells were seeded at 1.0 x 10⁴ cells/well on 96-well plates (Corning) and incubated until reaching 70-90% confluence. Cells were then washed with PBS and splenocytes, isolated from either immunized or control animals, were added in triplicates at a ratio of 1:5 (tumor cells: splenocytes), and incubated for different times in RPMI supplemented with 10% FBS. As controls, wells with only splenocytes or only tumor cells were also used. After this time, the colorimetric CellTiter 96^® AQ_ueous One Solution Cell Proliferation Assay (MTS, Promega, Madison, WI, USA) was applied following manufacturer’s instructions. Briefly, 20 µL of solution were added to each well, cells were incubated in the dark for 4 h, and then analyzed on the microplate reader (POLARstar Omega) at 490 nm.

All data sets were tested for normalcy and homoscedasticity. Normal data were analyzed by Unpaired Student’s t-test or by 2-way ANOVA. Data that did not follow a normal distribution were compared with the Mann-Whitney´s U test. All these studies were performed with GraphPad Prism version 5.02 (GraphPad Software, Inc. La Jolla, CA). A p< 0.05 was considered statistically significant.

Mouse AM is a 50-amino acid regulatory peptide that is broadly expressed through mouse tissues and organs (36). Thus, if we want to generate an immune response and break the immune tolerance for this intrinsically generated molecule we need to add a hapten. KLH has been commonly used for this task (27) but this is a huge molecule with a molecular weight ranging from 4,500 to 13,000 kDa, depending on its glycation state and the number of subunits (37). Therefore, we decided to select a small fragment of this sea snail protein with enough immunogenicity, beta turn, flexibility, hydrophobicity and surface accessibility, with help from freely available predictive software ( Supplementary Figures S1 , S2 ). At the end, we selected the sequence that is shown in Table 1 . Following codon optimization, the final sequence containing a 5´UTR, a Kozak sequence for ribosome engagement, the ORF of the KLH-AM fusion and a 3´UTR ( Table 1 ) was built by gene synthesis and inserted into the pcDNA 3.1 plasmid through a service contract with GenScript Biotech. The plasmid provided the T7 RNA polymerase promoter and the polyA tail ( Figure 1a ).

The mRNA was synthesized in vitro in the presence of modified nucleotides and the cap was post-transcriptionally added. When the resulting RNAs were run in a gel, the apparent size was close to the expected one at about 750 nucleotides (nt) ( Figure 1c ).

The mature mRNA was encapsulated into LNPs as described in the methods section. The physicochemical characterization of empty and mRNA-containing LNPs showed very similar size, polydispersity index and zeta potential for both ( Table 3 ), indicating that the presence of the mRNA does not greatly modify the structure of the LNP.

In addition, the encapsulation efficiency for the mRNA-containing LNPs was calculated at 91.35 ± 5.25% ( Table 3 ). The encapsulation efficiency was further tested in agarose gels, demonstrating that the LNPs had a large capacity to enclose and transport the mRNA ( Figure 1d ).

To demonstrate that the LNPs were efficiently taken by cells of the innate immune system, RAW 264.7 cells were seeded in microscope-compatible slides and the LNPs, containing either unlabeled (control) or Cy5-labeled mRNA, were added. After an incubation of 2 h, cells were washed and the intracellular presence of labeled LNPs was confirmed ( Figure 2 ).

The KLH-AM vaccine was tested in C57BL/6 immunocompetent mice. During the whole process, mice were weighed weekly to detect any potential toxicity associated to the vaccine. No significant differences in weight were found between the animals receiving the empty LNPs (control) and the mRNA-containing LNPs ( Figure 3a ).

Furthermore, three weeks after the final LNP injection, blood was taken from all animals and the serum was subjected to ELISA analysis in plates coated with synthetic mouse AM. Animals subjected to the KLH-AM vaccine had significantly higher (p=0.040) titles of anti-AM IgGs than those receiving the empty LNPs ( Figure 3b ).

To further characterize the specificity of the antibodies generated in the vaccinated animals, we used the sera obtained from control and vaccinated mice to stain a cell line, PC-12, that has been shown to produce high levels of AM (35). Only the sera from KLH-AM vaccinated animals was able to produce a strong cytoplasmic signal in the pheochromocytoma cells ( Figure 4 ).

After sacrifice, the spleens of some animals were excised and the splenocytes were isolated and analyzed by flow cytometry to study the influence of the vaccine on the contents of CD4⁺ and CD8⁺ lymphocytes ( Supplementary Figure S3 ). No significant differences were found between groups for the percentage of CD4⁺ T cells ( Figure 3c ) but there was a significant increase (p = 0.033) of CD8⁺ T cells in the animals vaccinated with KLH-AM LNPs when compared to the controls ( Figure 3d ).

To demonstrate that the splenic T cells had the capability of destroying tumor cells, a co-culture experiment was set up. In each well, 50,000 CD8⁺ T cells were added on top of a culture of 10,000 B16-F10 melanoma cells. These cells were incubated for 30 min, extensively washed to remove floating lymphocytes, and the remaining number of tumor cells was estimated through an MTS assay. The number of tumor cells that were exposed to T cells isolated from KLH-AM vaccinated animals was significantly lower (p<0.001) that those exposed to control T cells ( Figure 3e ).

After sacrifice, lungs were dissected out and the number and size of the metastases were recorded. Since the B16-F10 cells express high levels of melanin, the metastases were easily visible by the naked eye ( Figures 5a, b ) and were also easily found in histological sections stained with hematoxylin and eosin ( Figures 5c, d ). Just by direct observation, it was clear that mice vaccinated with the KLH-AM LNPs had less and smaller metastases than the animals receiving empty LNPs. In addition, quantification of both the number of metastases (p = 0.028) and the size of the largest metastasis in each animal (p = 0.020) were done and both were statistically significantly lower in animals receiving the KLH-AM vaccine ( Figures 5e, f ).

Since the goal of this study was to target an angiogenic peptide, we also studied the area occupied by CD31⁺ vessels within the tumors. Within the metastases, a significant (p= 0.046) reduction in the area covered by CD31⁺ immunoreactive vessels was found in mice vaccinated with the KLH-AM LNPs when compared with the empty LNPs ( Figures 6a–c ).

Since AM has been characterized as an autocrine growth factor for tumor cells (38), we investigated whether the vaccine had an impact on the proliferation index of the tumors by using an antibody against Ki67. No significant changes were found between groups ( Figures 6d–i ).

After immunization of SMAA-GFP animals with either the empty or the KLH-AM LNPs, tissues with a high number of blood vessels were analyzed under UV light ( Figure 7 ). As can be appreciated in the images, no differences were found in the number of vessels or their distribution between both experimental groups.