HHDII-DR1 double transgenic mice expressing human α1 and α2 chains of HLA-A*0201 chimeric with α3 chain of H-2Dd allele (HHDII), also expressing HLA-DRB*0101, and knocked out for the expression of murine MHC class I (H-2b) and II (I-Ab) were provided by Dr. Lone (CNRS, Orleans, France). Animal use and care was in accordance with EU Directive 2010/63/EU and UK Home Office Code of Practice for the housing and care of animals bred, supplied, or used for scientific purposes. The studies were undertaken with UK Home Office approval. Females aged between 6-8 weeks were used for this study. Three mice per group were used to assess the immunogenicity of the vaccine in at least two independent studies, whereas a minimum of 10 mice per group were used for tumour model experiments.

The Transporter-Associated Proteins (TAP) deficient, HLA-A2+ lymphoblastoid suspension cell line, T2 (22), was a gift from Dr. J. Bartholomew (Paterson Institute, Manchester, UK). The CML derived cell line K562 was purchased from ATCC, the KCL-22 CML cell line was kindly provided by Anthony Nolan (UK, https://www.anthonynolan.org). T2 cells, KCL-22 and TCC-S cells were cultured in the Roswell Park Memorial Institute (RPMI) 1640 medium (SLS/Lonza, BE12-167F) containing 10% (v/v) foetal bovine serum (FBS, Life Technologies, 10270106) and 2mM L-Glutamine (SLS/Lonza, BE17-605E). K562 cells were cultured in Iscove’s modified Dulbecco’s medium (IMDM, SLS/Lonza, BE12-722F) containing 10% (v/v) FBS.

For the tumour implant model, the B16F1 (murine melanoma) cell line was knocked out for both murine endogenous MHC class I and II by Zinc Finger Nuclease (ZFN) technology and stably transfected with plasmids encoding both HHDII and HLA-DR1 molecules. Cells were grown in RPMI 1640 supplemented with 10% (v/v) FBS, 1% L-Glutamine, 300μg/mL Hygromycin (Insight Biotechnology, Middlesex, UK) and 500μg/mL Geneticin (Insight Biotechnology). The cells were further transfected with luciferase and HAGE constructs (23). In this study, the cells were referred to as _hB16/HAGE⁺/Luc⁺, and maintained in RPMI-1640 medium supplemented with 10% (v/v) FBS, 1% L-Glutamine, 300μg/mL hygromycin and 550μg/mL zeocin (Thermo Fisher Scientific, U.K.) to retain luciferase gene expression, and 1μg/mL puromycin (Thermo Fisher Scientific) to retain HAGE gene expression. The HEK293T human embryonic kidney cell line (Thermo Fisher Scientific) was maintained in DMEM (SLS/Lonza, Yorkshire, U.K.) containing 10% (v/v) FBS and 2mM L-Glutamine - these cells were routinely used for viral gene transduction. All cells were maintained in a 37°C incubator with 5% (v/v) CO_2.

The WT1 protein sequence was screened for peptides that have binding affinity to HLA-A2 and HLA-DR1 molecules using a web-based algorithm (www.syfpeithi.de) (24) and peptides selected according to their binding score. A 15 amino acid long peptide (VRDLNALLPAVPSLG) derived from the WT1 protein was chosen. The HAGE 30mer sequence used in this study (QTGTGKTLCYLMPGFIHLVLQPSLKGQRNR) has previously been studied by our group (23), and the same class I and class II HAGE-derived peptides were used herein. All peptides were synthesised by GenScript Ltd (Piscataway, USA), with a minimum purity of 80%. All peptides were reconstituted in 100% (v/v) dimethyl sulfoxide (DMSO) to a stock concentration of 10mg/mL and stored at -80°C in small aliquots to minimise the number of freeze/thaw cycles.

A T2 peptide-binding assay was performed to evaluate the binding affinity of the HAGE and WT1 class I peptide sequences to the HLA-A2 molecules, as predicted by SYFPEITHI software. T2 cells are an HLA-A2 human lymphoblast suspension cell line that have been genetically modified to produce mutated non-functional TAPs that are necessary to transport MHC class I restricted peptides into the endoplasmic reticulum (ER). This defect results in a failure to present internal TAP-dependant peptides and instead MHC molecules leave the ER and the Golgi compartments empty, leading to a 70–80% reduction of HLA‐A2 expression on T2 cell surface. These empty molecules are not stable and are quickly recycled. However, the empty HLA-A2 molecules on the surface of cells can be stabilised by addition of exogenous peptides for which they have an affinity (25). T2 cells are therefore frequently used to validate the binding of exogenously administered peptides. Briefly, in a sterile 96-well rounded-bottom plate, T2 cells were plated at 0.5×10⁶ cells/100µL in serum-free RPMI medium supplemented with 3µg/mL human β2m (Sigma-Aldrich, Hertfordshire, U.K.) in the presence of varying concentrations of HAGE and WT1 class I peptides or with DMSO as negative controls and then incubated at 26°C overnight. T2 cells were then harvested, washed and stained with an APC-conjugated human HLA-A2 monoclonal antibody (BioLegend, 343308, Clone: BB7.2) for 30 minutes at 4°C. Cells were then washed and stained with propidium iodide (PI, Sigma-Aldrich) to exclude dead cells, and then immediately run to assess HLA-A2 expression using a Beckman Coulter Gallios™ flow cytometer and Kaluza™ data acquisition and analysis software.

A BFA decay assay was performed to evaluate the stability of the class I peptide-HLA-A2 complex on the cell surface. For this, T2 cells were seeded in a 96-well rounded-bottom plate at a concentration of 1×10⁶ cells/100 µL in a serum-free RPMI medium containing 3µg/mL human β2m. Cells were cultured with either the candidate peptides at a final concentration of 50µg (as determined by T2 binding assay) or with DMSO as a negative control and incubated at 26°C overnight. The next day, one batch of cells was washed and incubated with 10µg/mL BFA (BioLegend, 420601) for 1 hour at 37°C (Time 0) and thereafter every two hours until 8 hours. After each incubation, cells were transferred into 12x75 mm polycarbonate tubes, washed and stained with an APC-conjugated human HLA-A2 monoclonal antibody (BioLegend, 343308, Clone: BB7.2) for 30 minutes at 4°C and analysed using the flow cytometer. Dead cells were excluded by staining with LIVE/DEAD™ viability dye (Invitrogen) or PI (Sigma-Aldrich).

The stability of peptide-HLA-A2 complex was assessed by calculating the DC₅₀ value, which is defined as the time required for a 50% reduction in the level of MFI recorded at time 0 (26), the longer the time for this to happen, the more stable the complex. The DC₅₀ was calculated according to the formula: (MFI at a given time point/MFI at time 0) X100%. T2 cells incubated in the absence of peptide, but in the presence of the same concentrations of DMSO were used as negative controls.

The CML derived cell lines, K562, KCL-22 and TCC-S were genetically engineered to either over express HAGE and/or WT1 proteins or HAGE and/or WT1 expression was silenced. K562 cells constitutively express low levels of HLA and were electroporated with the pcDNA-3.1/HHDII gene using SF cell line 4D Nucleofection™ Xkit/Amaxa according to manufacturer’s instructions. Positively transfected cells were selected using 2000µg/mL of G418. K562/HHDII⁺ cells and KCL-22 cells were transduced with HAGE using a PLenti.Puro/HAGE plasmid. An empty plasmid control PLenti.Puro/empty vector, was used as a vector control. Cells were selected with 1µg/mL puromycine and 2000µg/mL G418 in a 24-well plate at 37°C in 5% (v/v) CO₂. Gene silencing was performed using short hairpin RNA (shRNA) sequences targeting the human WT1 (MISSION shRNA/Sigma-Aldrich in TCC-S cells

(WT1.shRNA₁,TRCN0000010466,CGGATGAACTTAGGAGCCACCTTCTCGAGAAGGTGGCTCCTAAGTTCATCTTTTTG and WT1.shRNA₂, TRCN0000040064,

CCGGGCAGTGACAATTTATACCAAACTCGAGTTTGGTATAAATTGTCACTGCTTTTTG).

After have being transduced, cells were selected with 1µg/mL puromycin and assessed for either HAGE or WT1 expression using Western blot and RT-qPCR.

RT-qPCR was used to assess the expression of the genes of interest in target cells. Briefly, total RNA was isolated from cell lines using a RNeasy Mini Kit (Qiagen, Lancashire, UK), following the manufacturer’s instructions. The quantity and purity of the produced RNA was evaluated using a NanoDrop™ 8000 Spectrophotometer (ThermoFisher Scientific). 2μg of RNA was then reverse transcribed into cDNA using Promega M-MLV Reverse Transcriptase. For each qPCR reaction, 1µL of cDNA (diluted 1:1 in ddH₂O) was mixed with 6.75µL of SYBR™ Green Supermix (Bio-Rad, 172-5124), 0.5µL (5pmol) of the forward, 0.5µL (5pmol) of the reverse primer (Sigma-Aldrich) and 3.25µL of molecular grade water. Each sample was run in triplicate on a Rotor-Gene Q real-time PCR cycler (Qiagen). Table 1 below summarises primer pair sequences and annealing temperatures used in the PCR reaction. It also shows the HAGE-codon optimised pair primers which were used to detect the transduced HAGE gene in the modified cells. Expression levels of each genes were calculated using comparative CT method (27).

Immunoblotting was performed to detect HAGE and/or WT1 expression in target cells. Briefly, CML cells were washed twice with cold DPBS, lysed in bromophenol blue free-Laemmli lysis buffer containing 0.125M Tris-HCl (pH 6.8), 4% (w/v) SDS, and 20% (v/v) glycerol and 10% (v/v) 2-mercaptoethanol. Cells were then boiled for 15 minutes at 95°C to denature proteins. To determine the quantity of protein in the lysates obtained, a protein assay (Bio-Rad, 500-0116) was performed following the manufacturer’s protocol, wherein standards were prepared by serial dilutions of bovine serum albumin (BSA) diluted in bromophenol blue free-Laemmli lysis buffer. A pre-calculated volume of sample containing 30µg of protein was loaded onto on Tris/glycine SDS-polyacrylamide gels alongside 5µL of marker ladder (Precision Plus Protein™, Bio-Rad, Hertfordshire, UK). The proteins on gels were then transferred onto Amersham Hybond-P PVDF membranes (GE Healthcare, Life science, UK) for 60 minutes at 100V at 4°C. The membranes were then blocked with TBS-Tween-20 (TBST) containing 5% (w/v) Marvel milk powder under a constant agitation for 1 hour at room temperature. After being thoroughly washed, membranes were blotted with primary antibodies at a concentration recommended by the manufacturer’s instructions (rabbit anti-human WT1 antibody, Abcam, Ab89901) and (rabbit anti-human DDX43 antibody, Sigma-Aldrich, HPA031381) and kept rocking overnight at 4°C. The next day, the membrane was washed 5 times with TBST, for 5 minutes each, and incubated in horseradish peroxidase (HRP)-conjugated secondary antibodies; HRP-conjugated anti-rabbit IgG antibody (Cell Signaling Technology, 70745, 1:1000) and HRP-conjugated anti-mouse IgG (Cell Signaling Technology, 70765). At the same time, Precision Protein™ Strep-Tactin™ HRP-conjugate, (Bio-Rad, 161-0380) was also added. Membranes were kept under a constant agitation for 2 hours at room temperature. After giving an additional five washes, membranes were developed using Clarity™ Western ECL Substrate (Bio-Rad, 170-5061) and luminescence was detected by G:BOX XT4: Chemiluminescence and Fluorescence Imaging System (Syngene). For each sample, mouse anti β-actin was used as a loading control (Sigma-Aldrich, A5441).

Surface staining was performed to characterise target cells phenotypically to detect HLA-A2 expression using an (APC-conjugated anti-human HLA-A2 antibody (BioLegend, 343308, Clone: BB7.2). Phenotypic analysis was also performed to assess the success of HHDII gene transfection using an FITC-conjugated anti-human β2m antibody (Sigma-Aldrich, SAB4700012, Clone: β2M-01). Cells were incubated with PI immediately prior to running the samples. At the end of the assay, cells were re-suspended in Beckman Coulter Isoton II™ diluent and analysed immediately on the Beckman Coulter Gallios™ flow cytometer.

Expression vectors encoding HAGE- and WT1-ImmunoBody^® (28) were coated onto 1.0μm gold microcarriers, as per the manufacturer’s instructions. In brief, 200μL of 0.05M spermidine containing 16.6mg of gold was mixed with 36μg of the ImmunoBody^® DNA. After giving the sample a short sonication, 200μL of 1M CaCl₂ was then added in a dropwise manner to the mixture while vertexing, and then the DNA-gold complex was kept for 10 minutes at the room temperature. The pellet was then re-suspended in 2mL of 0.025mg/mL Polyvinylpyrrolidone after being washed twice with anhydrous ethanol. After this, and while the tube was sonicating, the sample was syringed into dried Tefzel™ tubing and kept standing in a Tubing Prep Station for five minutes. Without disturbing the gold, the ethanol was then gradually taken off using a syringe. Nitrogen gas was turned on while the tube was kept spinning for 7-10 minutes inside the station. When completely dried out, the tube was detached from the station and cut by a guillotine. Bullets were kept at 4°C until used.

HHDII/DR1 mice were immunised with DNA bullet containing 1µg of either HAGE DNA ImmunoBody^® which encodes the HAGE 30-mer sequence [QTGTGKTLCYLMPGFIHLVLQPSLKGQRNR] or WT1 ImmunoBody^® which encodes the WT1 15-mer sequence [VRDLNALLPAVPSLG] or both, administered by gene gun device on Day 1, 7 and 14. On Day 21, mice were culled and splenocytes were immediately harvested and used in an IFN-γ Enzyme-Linked ImmunoSpot (ELISpot) assay.

The immunogenicity of either HAGE or WT1 peptides was determined using splenocytes harvested from immunised mice in an ex vivo IFN-γ ELISpot assay (ELISpot kit for mouse IFN-γ/MABTECH, 3321-2A). The assay was performed as previously described (29). Briefly 0.5x10⁶ of freshly isolated splenocytes were seeded per well in triplicates into the pre-coated IFN-γ ELISpot plate along with 1μg/mL of either HLA class I or 10μg/mL of HLA class II HAGE/WT1 derived peptides, cells were then incubated at 37°C in a 5% (v/v) CO₂ incubator for 48 hours. The plates were then washed and biotinylated detection antibody against mouse IFN-γ was then added and incubated for 2 hours at room temperature. Plates were then washed followed by the addition of alkaline phosphatase (AP)-conjugated streptavidin, after which they were incubated for 60 minutes at room temperature. Plates were washed, development solution (BCIP/NBT, Bio-Rad) added left in dark at room temperature until spots could be seen. Once spots developed, the reaction was stopped by rinsing the plates under tap water. Plates were then left to dry and the spots were quantified using an ELISpot plate reader (Cellular Technology Limited). Staphylococcal enterotoxin-B (SEB), was used as a positive control, and unstimulated splenocytes (cells alone) were used as a negative control for every ELISpot assay. Animals were scored as having a positive reaction when the number of spots in the cells alone wells did not reach more than 20 spots and when the response in the peptide containing wells were at least twice that of standard deviation of the mean of the control wells.

Splenocytes derived from vaccinated mice were cultured for 7 days with mitomycin-C treated peptide pulsed LPS blast cells as previously written (29). Briefly, LPS blast cells were prepared using splenocytes extracted from naïve mice and stimulated with LPS and dextran sulphate in a T75 flask. After 48 hours, cells from LPS blasts were harvested and treated with mitomycin-C. Following the incubation, cells were washed and counted and incubated with immunogenic HLA class I peptides, as determined by the ex vivo IFN-γ ELISpot assay for 90 minutes at 37°C. The peptide-pulsed LPS cells were then washed and co-cultured with freshly isolated splenocytes from immunised mice at a ratio (1) LPS (5): splenocytes in the presence of β-mercaptoethanol and murine IL-2 (mIL-2, Sigma-Aldrich, I0523, 50U/mL) at 37°C for 6 days. After incubation, LPS blast/T cells were then harvested, washed and counted to be ready for co-culture with ⁵¹Cr-labelled target cells.

Target cells were labelled with 1.85Mbq of ⁵¹Cr in a water bath at 37°C for one hour. Cells were then washed twice and rested for another hour at 37°C in 1mL of medium. Meanwhile, in vitro stimulated T cells were harvested and counted. Effector cells and target cells were plated in a 96-well rounded-bottom plate at a final volume of 200µL at different effector: target ratios of 100:1, 50:1, 25:1 and 12.5:1. Maximum and spontaneous release were set up in 4 replicates using 1% (w/v) SDS and plain medium, respectively. After 24 hours of co-culture of the effector and target cells at 37°C in 5% (v/v) CO₂ incubator, 50μL supernatants were transferred to 96-well Luma plates to be dried on the plate. The ⁵¹Cr release was measured using a TopCount Microplate scintillation counter (beta scintillation counting) and the percentage of specific cytotoxicity was calculated using the following equation:

Fresh splenocytes isolated from vaccinated animals were stimulated with either HAGE 30-mer long peptides, WT1 15-mer long peptides or both, at a concentration of 1μg/mL in the presence of β-mercaptoethanol and 50U/mL mIL-2 for 7 days at 37°C in a 5% (v/v) CO₂ incubator. Cells were then washed and counted using a NucleoCounter™ (Chemometec).

In the prophylactic setting, mice were immunised with 1µg of gold-coated HAGE-ImmunoBody^® DNA in one flank and 1µg of gold-coated WT1-ImmunoBody^® DNA into the contralateral flank simultaneously once, on 3 consecutive weeks (Day-1, -7 and -14). On Day-21, mice received 0.75x10⁶ _hB16/HAGE⁺/Luc⁺. In the therapeutic setting, 0.75x10⁶ _hB16/HAGE⁺/Luc⁺ cells were first implanted and the next day mice received the double vaccine which was then followed by two additional injections 7- and 14 days post tumour implantation. A control group of mice received the same dose of tumour cells, but no vaccines. Animals were monitored twice weekly and tumour growth assessed by callipers until a palpable tumour was detected. Mice were then monitored twice weekly by in vivo imaging until termination. For each imaging session, luciferin was administered intraperitoneally at 150mg/kg, and anaesthesia induction began 10 minutes afterwards. All mice were anaesthetised with an appropriate concentration of isoflurane and imaged using the IVIS Lumina III system (Perkin Elmer). Mice were sacrificed when the tumour volume reached a maximum of 1.2cm² for the prophylactic group and 1.5 cm² for the therapeutic group or show clinical signs.

The immunogenic region of the HAGE antigen was previously identified in our group using combination approaches of matrix-screening method of overlapping peptides and reverse immunology. It has been found that a 30-mer peptide sequence [QTGTGKTLCYLMPGFIHLVLQPSLKGQRNR]/[position: 286-316] derived from the HAGE protein is associated with high immunogenicity, as detected using an ex-vivo IFN-γ ELISpot assay (22). From that sequence, various short peptides were predicted to bind to HLA class I (HLA-A2) and HLA class II (HLA-DR1) using the freely available software SYFPEITHI, Tables 2A, B . Only HLA-A2 and HLA-DR1 restricted peptides were selected to be used in the HLA-A2/DR1 double transgenic mice (HHDII/DR1). Similarly, a 15 amino acid HLA-DR1 binding peptide [VRDLNALLPAVPSLG] derived from the WT1 protein containing two known HLA-A2 9-mer (30) was initially assessed for its binding affinity using the SYFPEITHI database ( Tables 2C, D ).

Several factors influence the capacity of peptide-MHC complexes to stimulate T cells, of which the primary two are the affinity of peptides for the MHC molecules and the cell surface stability of the peptide-MHC complex (31, 32). Although scoring of peptide binding to HLA-A2 via the computer algorithm predictions is useful, its accuracy is less than 70–80% (33). Therefore, the strength and stability of the interaction between the respective HAGE and WT1 class I peptides and the HLA-A2 molecules were assessed experimentally using peptide-HLA-A2 T2 binding and stability assays, respectively. Overall, data in ( Figure 1 ) (A, B) demonstrate that the intensity of HLA-A2 expression as measured by Mean Fluorescence Intensity (MFI) progressively increases in a dose-dependent manner in comparison with the control (MFI produced by cells that were treated with matched doses of DMSO vehicle control), indicating a successful peptide-MHC engagement, apart from WT1/P5 which did not increase the MFI, thereby reflecting its poor binding affinity toward HLA-A2 molecules. In addition, the Fluorescence Intensity Ratio (FIR) was calculated for each peptide concentration using the MFI of HLA-A2 expressed by T2 cells incubated with peptides and the MFI of HLA-A2 expressed by T2 cells that were incubated in the absence of peptide, but with the same concentrations of DMSO as background level of HLA-A2 molecules present on the surface of T2 cells. A summary of the results is shown in Table 3 . Peptides with FIR=1 were categorised as being non-binders, 1<FIR ≤ 1.5 as weak binders, 1.5<FIR<2 as moderate binders and 2≤FIR as strong binders. All HAGE peptides demonstrated strong binding affinity towards HLA-A2 molecules, from 30μg/mL onwards ( Figure 1A and Table 3 ). WT1 peptide, P3, was found to be a strong binder even at 10μg/mL, whereas WT1/P5 exhibited the least binding affinity with a score of 1 (non-binder) at 10 µg/mL, and only weak binding capability (1.4 to 1.5) at the higher concentrations ( Figure 1B and Table 3 ).

The persistence of the peptide-HLA class I complex on the cell surface was assessed in a time course manner using a Brefeldin A (BFA) assay ( Figures 1C, D ). BFA has a trap effect by obstructing anterograde transport of vesicles from the ER to the Golgi apparatus, thereby preventing peptide-MHC proteins from being transported to the cell surface. Results in Figures 1C show that there is a steady, time-dependent reduction in the level of MFI in all complexes within the first hours of study, but that the reduction stopped after 6 hours. With regards of the DC₅₀ value, all HAGE and WT1/P3-HLA-A2 complexes on the T2 cell surface were stable as their DC₅₀ values were all more than 8 hours ( Figure 1D ). However, WT1/P5-HLA-A2 complex exhibited a poor stability, as DC₅₀ value fell before 6 hours.

To determine the immunogenicity of the HAGE (30mer) and WT1 (15mer) sequences and whether the predicted HLA-A2 and HLA-DR1 derived from these sequences were endogenously processed, HHDII/DR1 mice were immunised with either HAGE-ImmunoBody^® or WT1-ImmunoBody^® constructs on Day- 0, -7 and -14 ( Figure 2A ). On Day 21, animals were culled, and their respective spleens harvested and assessed in an ex-vivo IFNγ ELISpot assay. Results in Figure 2B confirms the immunogenicity of two class I peptides (HAGE/P5 and HAGE/P6) and one class II (HAGE/P7) in comparison to cell alone (cells that did not receive any peptide) as reflected by the high number of IFN-γ producing cells (P-value<0.0001, n=3). Whereas, HAGE/P4 and HAGE/P8 peptides did not produce any IFN-γ, indicating that these peptides were not endogenously produced and presented to T cells in vivo. WT1-specific responses induced by a novel WT1 DNA ImmunoBody ^® vaccine was also assessed and the results in Figure 2C demonstrate the immunogenicity of WT1/P1 and WT1/P3, and the poor immunogenicity of WT1/P5.

K562, KCL-22 and TCC-S CML cell line-derived targets were assessed for HLA-A2 expression. Histograms in Figure 3B shows that that only a small percentage of K562 cells naturally express HLA-A2, whereas a high percentage of KCL-22 and TCC-S cells are HLA-A2 positive ( Figures 3A, B ). Therefore, K562 cells were electroporated with PcDNA-3.1/HHDII construct encoding the chimeric HLA-A2 gene (HHDII). Figure 3C demonstrates the success of HHDII transfection of K562 cells.

The constitutive expression at mRNA and protein levels of both HAGE and WT1 was then determined in all CML targets using RT-qPCR and Western blotting ( Figures 3D, E ). All cell lines expressed WT1 mRNA and protein, whereas HAGE mRNA and protein expression was only detected in TCC-S cells. Therefore, both K562/HHDII⁺ and KCL-22 cells were then transduced with PLenti-Puro/HAGE plasmid construct, followed by antibiotic selection using puromycin to maintain HAGE expression in addition to G418, for K562/HHDII, to maintain HHDII gene expression. Post selection, the success of the transduction was demonstrated at mRNA and protein levels using RT-qPCR and Western blot ( Figures 3F, G ).

TCC-S cells were found to express HLA-A2, WT1 and HAGE proteins. It was expected that knockdown of WT1 in these cells could establish a negative control for WT1 expressing TCC-S (wild-type) upon assessing the responsiveness of WT1-specific CTLs in the cytotoxicity assay. Data in Figure 3H demonstrate the outcome of WT1 knockdown in TCC-S at protein level using two WT1.shRNA sequences in comparison with cells that were transfected with the lentiviral carrying the empty vector (PLKO.1.Puro), as a control vector. Although both sequences of shRNA were successful in significantly inhibiting WT1 protein expression, the WT1.shRNA.1 appeared to be more efficient than WT1.shRNA.2, and therefore the WT1.shRNA.1 clone was used in the cytotoxicity assay. All the modified cell types were periodically checked for the maintenance of the transduced genes.

Once all the target cells had successfully been “engineered” to express the correct HLA-A haplotype and/or antigens the ability of the vaccine induced cells to recognised and kill them was assessed. Briefly, splenocytes harvested from animals immunised with either HAGE-ImmunoBody^® alone, WT1-Immunobody^® alone or both were co-cultured in vitro with mitomycin-C treated and peptide-pulsed LPS blasts at ratio of 1:5 (LPS blast: T cell) for 6 days. Cells were then plated with ⁵¹Cr labelled K562 cells at various target: effector ratios (1:100, 1:50, 1:25, and 1:12.5) for 24 hours at 37°C. Supernatants were then transferred onto Luma plates and read using a TopCount beta scintillation counter.

Figure 4 shows that in vitro expanded cells from all vaccinated mice were able to specifically recognise and kill all the target cells expressing the antigen against which the mice were vaccinated, Figure 4 (A1, A2) (B1, B2) (C1, C2). Moreover, cells derived from the mice immunised with both HAGE- and WT1-ImmunoBody^® vaccines could consistently kill significantly more target cells expressing both antigens indicative of a synergistic effect achieved with the combined vaccines Figure 4 (A3, A4) (B3, B4) (C3, C4) (P-value<0.0001, n=3) in almost all ratios used.

In our previous studies, the B16/murine MHC knockout/HHDII+/DR1+ murine melanoma cell line was shown to constitutively express the murine WT1 protein (data not shown). Herein, these cells (with/without HAGE) were used as targets to assess the capability of HAGE- and WT1-ImmunoBody^® vaccines derived cells to provoke in vitro killing activity. Figure 4D highlights the significant difference in the lysis produced by each indicated group against these cells. Here again, a statistically significant difference was found between the killing of B16 cells expressing HAGE (almost 50%) and HAGE-negative B16 (20%) at 1:100 ratio (shown in D1), whereas no differences in killing were found against these two targets when the effector cells used were derived from mice immunised with only WT1 (shown in D2), since both express murine WT1. This indicates that the WT1 peptide used in this study was endogenously processed naturally and presented on the surface of the B16 cells and was recognised by cytotoxic T lymphocytes (CTLs). Again, as for the other targets, a significantly higher percentage of B16/HAGE⁺ cells was killed when the CTLs used were derived from mice immunised with the combined approach ( Figure 4D , at almost 60%) than when they were derived from mice immunised with either HAGE or WT1 vaccine individually (50% and 45%, respectively). This difference is shown to be a statistically significance at 1:100 (target: effector ratio) (P-value=0.0032, n=3 and P-value=0.0002, n=3) in HAGE and WT1 groups, respectively.

Overall, these findings demonstrate that the sequences contain within HAGE- and WT1-ImmunoBody^® derived vaccines either individually or in combination was endogenously processed and displayed on the surface of antigen presenting cells (APCs) in association with HLA-A2 molecules leading to the development of professional CTLs which are able of specifically recognising and killing antigen-expressing target cells. They also indicate that the combination approach improves CTL responses.

The efficacy of the combined HAGE/WT1 ImmunoBody^® vaccines was then tested in both prophylactic and therapeutic settings in female HHDII/DR1 mice (n=10/group) using the aggressive _hB16/HAGE⁺/Luc⁺ tumour. In the prophylactic setting, mice were first immunised with both HAGE- and WT1-ImmunoBody^® vaccines 3 times, a week apart, and on Day 21 mice were then challenged with subcutaneous injection of _hB16/HAGE⁺/Luc⁺ cells, as shown in Figure 5A1. In the therapeutic setting, mice were first implanted with the same dose of _hB16/HAGE⁺/Luc⁺ cells, followed by immunisation the next day, and then received two more injections of the dual HAGE/WT1 ImmunoBody^® vaccines 7 days apart. The control group received no vaccine ( Figures 5A2, A3 ).

Details of bioluminescence images and mice culling per session are shown in Supplementary Figure 2 . Mice that did not receive the vaccines had to be sacrificed as early as Day 39 post-implantation due to tumour size endpoints, whereas tumour growth was delayed in mice that combined vaccines in comparison to the controls ( Figure 5 ). Indeed, at around day-46 post-implantation, the number of surviving mice was 8/10 (80%) for the prophylactic group and 6/10 (60%) for the therapeutic group, in comparison with only 1/10 (10%) for the control group. Tumour growth, as measure by the total luminescence flux for each group was recorded and plotted in Figure 5B . Eventually, 4 mice from prophylactic group (Mouse-6, Mouse-7, Mouse-9 and Mouse-10) and one mouse from the therapeutic group (Mouse-24) were the last mice to be culled. The study was then terminated on Day 56 post-implantation before the tumours reached their end point size. This was purposely done in order to be able to perform immune characterisation of the tumour infiltrating lymphocytes extracted from the growing tumour and performed functional assay on splenocytes extracted from these mice.

Interestingly, although the total luminescence flux signals, as expected, were very similar on Day 22 (one day post-transplant), 10 days later, a statistically difference between the size of the tumours for the prophylactic group and those for the control was detected. Thereafter, the size of the tumour in the prophylactic group and for the therapeutic group remained below the 1x10⁹ threshold for the majority of the tumours. However, by Day 39, tumours in a distinct group consisting of 4 mice separated from the rest of the prophylactic group, continued to grow until they had to be culled on Day 46. In the future, it would be important to re-immunise animals showing such a trend before their tumour became too large. It is also evident from the data that for 7 out of the 10 mice that received the vaccine after tumour implantation, the tumour first regressed below the 1x10⁹ threshold, whereas the size of the tumours from the control group was, for the majority, above or very near this value (with the exception of one mouse), but then by Day-39 these increased again. This demonstrates the ability of the combined vaccines to delay the growth of the tumour, but there would likely be a requirement for further vaccination on Day-39 or additional interventions, such as checkpoint inhibitor. The survival proportion in Figure 5C clearly demonstrates that the combined HAGE/WT1 ImmunoBody^® vaccines were able to significantly delay the aggressive growth of B16 melanoma cells and increases the overall survival in the prophylactic setting in comparison to the control at **P-value<0.01. However, this level of efficacy was not shown in the therapeutic group.

Tumour volume and weight were measured as soon as animals were culled. Data in Figure 5D demonstrates that there was a significance difference in the tumour volume in the prophylactic and therapeutic group compared to the control (P-value =0.003, n=10 and P-value= 0.0271, n=10), respectively, which indicates a shrinkage of the tumour, probably due to the activation of anti-B16 CTLs.

Until Day 56 post-implantation, all mice were sacrificed due to tumour size reaching the endpoint threshold, except for Mouse-6, Mouse-7, Mouse-9 and Mouse-10 from the prophylactic group and Mouse-24 from the therapeutic group. However, it was decided to terminate the study to enable the analysis of TILs and perform functional assays, such as IFN-γ ELISpot.

The immune response generated in these mice was compared with that of the IFN-γ response assessed in a set of mice culled earlier in the study, (Mouse-11, Mouse-13, Mouse-15 and Mouse-16) from the control group and one mouse from the therapeutic group (Mouse-22). Interestingly, comparing data in Figure 6A1 versus A2 , IFN-γ production was significantly greater in the surviving mice than those which were culled early in the study due to their tumour size reaching the end point. Vaccinated mice which had to be sacrificed early in the study due to tumour size (such as mouse-22) lacked the development of a specific anti-tumour immune response and therefore could not maintain tumour growth in a similar pattern as the control. IFN-γ production from spleens of these mice was also assessed after 1week in vitro stimulation (IVS) using the IFN-γ ELISpot assay. As demonstrated in Figure 6A3 , IFN-γ secretion was significantly increased after IVS in comparison with those obtained straight ex-vivo demonstrating that these cells can be further expanded in vitro after in vivo priming.

The profile of tumour-infiltrating lymphocytes (TILs) was assessed and compared with T cells extracted from spleens from tumour bearing mice. After each culling, a small piece of tissue was immediately digested and stained with anti-mouse CD3, CD4, CD8a and CD279 PD-1 antibodies. Thereafter, samples were assessed by flow cytometry and analysed according to the gating strategy shown Figure 6B . Figures 6C, D demonstrates the expression pattern of CD3⁺, CD4⁺, CD8⁺ and PD-1 on TILs and T cells from freshly isolated splenocytes from tumour bearing mice in both treatment settings in comparison to the control group. From these data, the following points can be concluded; firstly, the majority of CD3⁺ T cells recruited in spleens were CD4⁺ cells, whereas the majority of CD3⁺ T cells in all tumours studied were CD8⁺ cells, indicating a reversal in the CD4⁺/CD8⁺ ratio of TILs in comparison with those derived from the spleens. Although this trend was observed regardless of whether the mice were vaccinated, the percentage of CD8⁺ T cells within the tumour of the prophylactic and therapeutic mice was significantly higher than the control at ***P-value= 0.0003 and *P-value= 0.0323, respectively, indicating that the combined vaccines were able to recruit more CD8⁺ cells in the vaccinated group in both the prophylactic and therapeutic settings. Secondly, more than 60% of CD4⁺ and CD8⁺ TILs expressed PD-1 compared to 5-10% in the spleens, suggesting that T cells of both types CD4⁺ and CD8⁺ in the tumour microenvironment have been jeopardised by immunosuppressive elements, PD-1, which might be induced by immune immunosuppressive cells in the hostile B16 microenvironment. Finally, the percentage of CD4⁺PD-1⁺ cells in the tumour was higher in the prophylactic group at **P-value= 0.0046 than the control.