Vaccination against neoantigens induced in cross-priming cDC1 in vivo

All animal work was conducted under the approval of the University of Miami Institutional Animal Care and Use Committee (IACUC) in accordance with federal, state, and local guidelines. Female, 10–12-week-old C57Bl/6 mice were used for all studies and purchased from The Jackson Laboratory.

Antibody conjugations were carried out using a modified protocol from Vector Laboratories for protein-oligonucleotide conjugation. All reagents, unless otherwise indicated, were purchased from Vector Laboratories. The antibody-oligonucleotide calculator (https://​vectorlabs.​com/​wp-content/​uploads/​2022/​11/​VL_​A-9002-001.​UserGuide.​LBL-01990.​pdf) available from Vector Laboratories was used throughout the protocol. Conjugation was carried out in two steps. First, a 5′ amino-modified ODN was labeled with S-4 formylbenzamide (4-FB). Next, the 4-FB modified ODN was covalently linked to a S-Hynic labeled antibody molecule using an analine catalyst to generate a stable antibody-ODN conjugate. To label the ODN with 4-FB, the 5′ amino-modified ODN was resuspended in 1× Modification Buffer at a concentration of 0.5 OD/µL, diluted 1:200 in H₂O and the concentration (OD₂₆₀/µL) determined using a NanoDrop. This value, along with the total volume, was entered into the calculator. The ODN, the calculated volume of andrydrous DMF, and the resuspended S-4-FB were combined and incubated for 2 h at RT on a rotator. The 4-FB ODN was purified and desalted twice into 1 × Conjugation Buffer using a 7 K MWCO Zeba Column (Thermo Scientific) according to the manufacturer’s instructions. Following purification, the ODN concentration was determined, and it was aliquoted and stored at − 80 °C. Aiming at conjugating 8–10 ODN molecules per antibody, the antibody (InVivoMAb anti-mouse Clec9A (CD370) Clone 7H11, or rat IgG1 isotype control clone HRPN; BioXCell) was desalted into 1 × Modification Buffer using Amicon Ultra 30 K MWCO centrifugal filters (Millipore). Three buffer changes were done according to the manufacturer’s instructions. The antibody was collected, and the protein concentration was measured using a Nanodrop and adjusted to be not more than 2 mg/ml. S-Hynic was resuspended in anhydrous DMF to label the antibody with a 50-molar excess of S-Hynic (µg IgG = 6.6 pmol) and incubated for 2.5 h at RT on a rotator. Excess S-Hynic was removed by buffer exchange into 1X Conjugation Buffer using Amicon filters as described above. The S-Hynic modified antibody was incubated with 10-molar excess 4-FB modified ODN in 10X Catalyst Buffer. Following incubation, the antibody-ODN was desalted into PBS (no Ca⁺⁺ or Mg⁺⁺, Gibco) using the Amicon filters as described above. The antibody concentration was measured using the Pierce BCA protein assay (Thermo Scientific) and stored at 4 °C for not more than 1 month. For each experiment, complementary ODN (cODN) modified siRNA and/or CpG ODNs were hybridized to the antibody conjugated ODNs at a molar ratio of 1:1 at 41 °C for 15 min in PBS with divalent cations to yield 8–10 ODN/antibody molecule.

To determine that ODNs were conjugated to the antibody and that excess free ODNs were removed, 1 mg of antibody-ODN conjugate was subjected to 2% agarose gel electrophoresis in TBE buffer, stained with ethidium bromide and visualized with UV. To determine whether ODN conjugation has affected antibody binding, the antibody-ODN conjugates were incubated with Clec9a-Fc-beads and checked for fluorescence saturation by flow cytometry. To prepare the beads, 10 µl Protein A DynaBeads (Thermo Scientific) were washed three times with bead buffer (20 mM Hepes, pH 7.4, 10 mM NaCl, 2 mM CaCl₂, 2 mM MgCl₂, 0.02% azide, 0.01% Tween-20). The washed beads were incubated with 10 µg of murine rClec-9A-Fc chimeric protein (R&D) in 100 µL bead buffer for 1 h at RT with rotation and washed as before followed by incubation with 10 µg IgG isotype antibody. Finally, the beads were washed as before and stored in 1 mL of bead buffer at 4 °C until use. To estimate the number of ODNs conjugated per antibody, 1 µg of Clec9A-ODN was incubated with increasing molar excess of cODN-AlexaFluor-647. Following hybridization, the reaction mixture was incubated with 10 µL of Clec9a DynaBeads for 10 min at RT. Beads were washed in mL FACS Buffer (PBS + 2% FBS) and analyzed by flow cytometry as detailed below. An unrelated antibody (mPD-1 clone RMP1-14; BioXCell)-ODN was used as a negative control for specificity of the Clec9a-ODN antibody to the bead.

The following ODN sequences were used in the study as indicated in figure legends. All ODNs were either purchased from Integrated DNA Technologies or Trilink BioTechnologies, Inc. Where indicated, 2’ O-methylated (m) pyrimidines or phosphorothioate (*) modified oligos were used to increase stability. Duplexed Tap 2 siRNA (Sense: 5′-GmCmUGmCAmCAmCGGmUmUmCAGAAmU; Antisense: 5′-AUUCUGAACCGUGUGCAGCmUmU) or Scrambled Control (Sense: 5′-mUAAAGAAmCmCAmUGGmCmUAAmCmC; Antisense: 5′ GGUUAGCCAUGGUUCUUUAmUmU) were delivered to target cells by conjugation to Nucleolin aptamer, CpG, or antibody. The Nucleolin aptamer (5′-GGTGGTGGTGGTTGTGGTGGTGGTGG) extended at the 3′ end with a 12-carbon spacer followed by a linker (5′-GmUmAmCAmUmUmCmUAGAmUAGmCmC) was annealed to the 5′ end of Tap 2 or control siRNA using a 3′ complementary linker sequence separated from the sense sequence by a nine-carbon spacer and was used to deliver Tap 2 siRNA to RMA cells. CpG 1668 (5′-T*C*C*A*T*G*A*C*G*T*T*C*C*T*G*A*T*G*C*T*) extended at the 3′ end with a 12-carbon spacer followed by the linker: 5′-mCGAGGmCmUAmUmCmUAGAAmUGmUAmC was annealed to a complementary sequence on the Tap 2 siRNA as above and used to deliver Tap 2 siRNA to TLR9 expressing cells. Antibodies were used to deliver Tap 2, control scrambled siRNA, CpG, or control CpG (5′-T*C*C*A*T*G*A*G*C*T*T*C*C*T*G*A*G*C*T*T*) ODNs by annealing to the ODN (5′-/5AmMC6/rAmUrAmGmUrAmCrAmUmUmCmUrArGrAmUrArGmCmC) labeled antibody with a 3′ complementary linker separated from the sequence by a nine-carbon spacer and was used to deliver Tap 2 siRNA or control sequence to Clec9A-expressing cells. Annealing was carried out at equimolar ratio of the complementary sequences in PBS with Ca²⁺ Mg²⁺ for 5 min at 85 °C in a heat block and allowed to cool to room temperature and stored at − 80 °C until use.

Mice were euthanized, and spleens were removed at timepoints indicated in figure legends. A single cell suspension of the spleen was prepared and filtered through a 70 µm nylon filter. Red blood cells were lysed using ACK Buffer (ThermoFisher), and cells were resuspended in FACS Buffer. FcR was blocked (Fc Block; BD Biosciences), and splenocytes were stained with antibody cocktail diluted in Brilliant Stain Buffer (BD Biosciences) for 30 min at 4 °C and washed with PBS. Cells were incubated with Live-Dead Blue (ThermoFisher Scientific) for 15 min at RT, washed twice with PBS and fixed with BD Cytofix/Cytoperm according to the manufacturer’s instructions. Intracellular staining was performed in two steps. First, cells were stained with anti-Tap2 antibody for 30 min at RT, washed with Perm Wash Buffer (BD), and then stained with anti-rabbit AlexaFluor647 for 30 min at RT. Cells were washed a final time and then collected using a Cytek Aurora spectral analyzer flow cytometer (Flow Cytometry Shared Resource; UMiami Sylvester Cancer Center). Data was rendered and analyzed using FlowJo software (TreeStar). Multicolor flow cytometry was performed with the following antibodies: CD3e-AlexaFluor488 (145-2C11), CD49b-AlexaFluor488 (DX5), Ly6G-AlexaFluor488 (1A8), MHCII(I-A/I-E)-APC-Fire750 (M5/114.15.2), XCR1-PE (ZET), CD11b-PE-Cy5 (M1/70), Ly6C-BV605 (HK1.4), CD172a (SIRPa)-PE-Cy7 (P84), and CD317-BV711 (927) from BioLegend; F4/80-APC-R700 (T45-2342), Siglec-H-BV421 (440c), CD19-BUV805 (1D3), and CD11c-BB700 (N418) from BD Biosciences; and rabbit polyclonal anti-Tap2 (PA5-37,414) and anti-rabbit AlexaFluor647 (A21443) from ThermoFisher Scientific.

RMA (Tap2 sufficient) and RMA-S (Tap2 deficient) cell lines were used for all tumor studies and have been described [13, 14]. Cells were cultured in IMDM supplemented with 10% FBS (Cytiva), Pen/Strep, sodium pyruvate, non-essential amino acids, and 2-mercaptoethanol (all from Gibco). For tumor cell injection, cells were collected and washed twice in PBS.

When variables studied were normally distributed, statistical analysis of multiple comparisons was performed using one-way ANOVA with Tuckey or Dunnet post-test. To compare the mean differences between groups that have been split into two independent variables (treatment/number of doses), analyses were performed using two-way ANOVA. Nonparametrical methods were applied for not normally distributed variables. For these statistical analyses, multiple comparisons were performed using Kruskal–Wallis with Dunn post-test, and comparisons between just two groups were performed using Mann–Whitney U-test. Significance of overall survival was determined via Kaplan–Meier analysis with log-rank (Mantel–Cox) analysis. All statistical analyses were performed with Graphpad Prism 6 and 7 (GraphPad). Error bars show standard error of the mean (SEM) and p < 0.05 was considered statistically significant.

To downregulate TAP in resident cDC1 in mice with a TAP specific siRNA, we used a modular multivalent antibody targeting platform (Fig. 1a) whereby 8–10 copies of a 19 nt long 2’-O-methyl (2’OMe)-modified ODN are chemically conjugated to a Clec9a antibody (Step 1). A fluorophore-labeled complementary ODN was used to determine the average number of ODNs conjugated per antibody (Fig. 1b). In Step 2, the cargo (Fig. 1a, X) is conjugated to the ODN-modified antibody by hybridization via an attached complementary sequence (cODN) resulting in a multivalent configuration whereby the antibody delivers multiple cargo molecules to the targeted cell. 8–10 copies of a TAP or control siRNA were conjugated to a Clec9a antibody via complementary sequences engineered at the 3′ end of the siRNA sense strand (Fig. 1c). Hybridization of the siRNA to the ODN-modified Clec9a antibody was monitored by agarose gel analysis (Fig. 1d).

Downregulation of TAP was evaluated by flow cytometry in cDC1, cDC2, B cells and macrophages isolated from the spleens of mice treated with Clec9A antibody or CpG ODN targeted TAP siRNA. Figure 1a shows the results from a representative mouse. For example, 72 and 96 h post-treatment with Clec9a targeted TAP siRNA a significant downregulation of TAP in cDC1 cells can be seen when compared to treatment with Clec9A antibody targeting a control siRNA, whereas treatment with CpG ODN targeted TAP siRNA downregulation was less pronounced and more transient, seen at 72 h but not at 96 h post-treatment. TAP downregulation averaging three mice per group is shown in Fig. 2b. CpG ODN targeted TAP siRNA led to transient TAP downregulation in all subsets analyzed, consistent with the broad distribution of TLR9 in DC, B cells, and monocyte/macrophages [15‐17]. In contrast, Clec9A antibody targeted TAP siRNA led to TAP downregulation in cDC1, but not in cDC2, B cells, or macrophages, consistent with the selective expression of Clec9a in cDC1. Clec9a mediated TAP downregulation was also more sustained than CpG ODN mediated TAP downregulation, conceivably reflecting the longer half-life of antibodies in the circulation. Isotype antibody conjugated TAP siRNA failed to downregulate TAP in cDC1 cells (data not shown). Taken together this experiment shows that Clec9a Ab targets its attached TAP siRNA cargo to cDC1 in vivo that results in TAP downregulation.

We next tested whether treatment of mice with TAP siRNA targeted to cDC1 can inhibit tumor growth. Figure 3 shows that treatment of mice bearing the TAP-deficient RMA-S tumor that naturally present TAP TEIPP with Clec9a antibody targeted TAP siRNA inhibited tumor growth, reducing the rate of tumor growth (Fig. 3a) and extending the survival of the treated animal (Fig. 3b). Figure 3 also shows that Clec9a targeting of the TAP siRNA was more effective than CpG ODN targeting.

To test whether Clec9a-TAP siRNA can also inhibit TAP in TAP sufficient RMA tumor cells, expression of TAP was partially and transiently reduced in the tumor bearing mice by treatment with a TAP siRNA targeted to the RMA tumor cells by conjugation to a nucleolin binding DNA aptamer (Nucl-TAP siRNA) as previously described [13]. Figure 4 shows that treatment of the TAP sufficient RMA cells with Clec9a antibody targeted TAP siRNA inhibited tumor growth provided mice were also treated with Nucl-TAP siRNA. Clec9a targeting was more effective than CpG ODN targeting, but the difference was less pronounced than seen in RMA-S tumor bearing mice (Fig. 3). Taken together, these experiments show that Cle9a antibody targeted delivery of TAP siRNA to cDC1 cells in mice inhibits tumor growth and is more effective than CpG ODN targeting. This difference could reflect the sustained TAP downregulation in cDC1 cells (Fig. 2b) and/or the restricted expression pattern of Clec9a largely confined to cDC1 (Fig. 2) precluding the presentation of the induced TAP TEIPP by immune suppressive myeloid cells like MDSC, M2 macrophages or incompletely matured DC.

While Clec9a targeting provides more specificity than CpG ODN targeting (Fig. 2), unlike engagement of TLR9 by CpG ODNs cross-linking of TAP siRNA with the Cle9a antibody does not induce DC activation/maturation which is required to elicit an effective immune response against the antigens presented by the cDC1. Activation/maturation stimuli were, therefore, provided separately by co-treatment with an agonistic CD40 antibody [18, 19]. In patients, treatment with systemically administered CD40 antibody was associated with immune related adverse effects [20]. To dispense with the need of using the agonistic CD40 antibody and reduce the risk of toxicity, we conjugated both the TAP siRNA and a CpG ODN to the Clec9a antibody by hybridization via common complementary sequences as shown in Fig. 5, thereby limiting activation to cDC1 that present the induced antigens.

Figure 6a and b shows that treatment of mice bearing subcutaneously implanted RMA-S tumors with Clec9a antibody co-targeted TAP siRNA and CpG ODN (Clec-CpG/Tap) was more effective than treatment with Clec9a targeted TAP siRNA and CD40 antibody (Clec-Tap + CD40) and was dependent on co-delivery of the TAP siRNA and CpG ODN to the same cell because treatment with a mixture of Cle9a targeted TAP siRNA and Cle9a targeted CpG ODN (Clec-Tap siRNA + Clec-CpG) did not inhibit tumor growth. The ability of CpG ODNs to target DC (Figs. 3 and 4) did not contribute to tumor inhibition because an isotype antibody conjugated with TAP siRNA and CpG ODN (Isotype-CpG/Tap) was ineffective. A likely reason for the reduced antitumor activity of Cle9a-TAP siRNA + CD40 antibody seen in this experiment (Fig. 6b) compared to the experiment shown in Fig. 3b was that we used half the dose of TAP siRNA, 4 versus 8 siRNAs per antibody, respectively, in order to accommodate 4 control siRNAs on the Clec9a antibody.

To further examine the mechanism of action of the Clec9a Ab conjugated CpG ODN, we measured the downregulation of TAP isolated from mice as described in Fig. 2. Figure 6c shows that Clec9a targeted TAP siRNA (Clec-Tap/Ctrl + CD40 and Clec-CpG/Tap) led to robust and sustained downregulation of TAP in cDC1 but not B cells or macrophages. A small and transient downregulation of TAP was also seen in cDC2 that was not seen in the experiment shown in Fig. 2, perhaps reflecting low levels of Clec9a expression in this preparation. Notably, no TAP downregulation was seen in cDC1 when mice were treated with isotype antibody conjugated to both CpG ODN and TAP siRNA (Isotype-CpG/Tap). This is consistent with the immunotherapy experiment (Fig. 6a and b) showing that the antibody immobilized CpG ODN was not responsible for targeting the Clec9a bound TAP siRNA to DC. Taken together, Fig. 6 shows that co-targeting TAP siRNA and CpG ODNs to cDC1 leads to a superior antitumor response, presumably by limiting DC activation to cells also presenting the induced antigens.