Sequences and production methods for in-house-derived 12E12, 11B6, 24A3, 12B4, and CP [CP-870,893, a Pfizer Inc. agonistic antibody (4)] anti-human CD40 human IgG4 antibodies have been described (13). The HIV-5pep antigen cassette (Flex-v1-Pep-gag17-f1-gag253-f2-nef116-f3-nef66-f4-pol158) attached to H chain C-termini has been detailed (7, 14) and is contained within GenPept Sequence ID: AJD85777.1. The HPV16.E6/7 H chain adduct (9) has been described (GenBank KP684039). The Dockerin V1 (Doc) H chain adducts and Cohesin-Flu M1 protein were described elsewhere (6). Other antigen adducts are described in the sections relevant to particular protocols.

B cell proliferation assays and DC activation assays were performed as previously described (13). Briefly, B cell assays used CFSE-labeled human peripheral blood mononuclear cells (PBMCs) incubated with a dose range of the test articles with human IL-4 and human IL-21 for 5 days of culture, then cells were stained with surface and live-dead markers followed by flow cytometry analysis gating on single live cells and CFSE⁻/CD19⁺ cells. To account for donor variation, independent experiments were often collated by normalizing to the baseline and maximal responses in each experiment. Typically, baseline percentages across independent experiments using different donors were consistently in the range of ~1 to ~10% for the antibodies administered alone, and ~4 to ~22% for the antibodies administered with the sCD40L. Maximum proliferation among the experiments typically was in the range of ~46 to 97%. For DC activation, myeloid-derived human dendritic cells (MDDC), from human blood monocytes cultured with human IL-4 and human GM-CSF for 5 days, were incubated with IL-4 and GM-CSF and a dose range of the test articles. After 24 or 48 h, supernatants were analyzed for secreted cytokines by Luminex^®, and in some cases the cells were stained for cell surface activation markers.

SPR analyses were performed as previously described (13).

Cryopreserved human PBMC were thawed with 50 U/ml benzonase^® nuclease (Millipore, cat 70746), washed and rested overnight in RPMI 1640 enriched with 100X PenStrep (Gibco, 15140-122), 100X Hepes Buffer (Gibco, 15630-080), 100X Non-essential amino acids (NEAA) (Gibco, 11140-050), 100X Sodium Pyruvate (Gibco, 11360-070), 1000X 2-Mercaptoethanol (Gibco, 21985- 023), 100X Glutamax (Gibco, 35050-61) (herein called complete RPMI 1640) with 10% AB serum (GemCell, 100-512) in a 37°C 5% CO₂ incubator. The following morning cells were filtered and cultured at a concentration of 2E6 cells/ml at 37°C in 1 ml complete RPMI 1640 + 10% AB serum in a 24-well flat bottom plate. Cells were treated with different concentrations of anti-hCD40 antibodies or controls. After 48 h, 1 ml of complete RPMI 1640 with 10% AB serum and IL-2 (Proleukin, Sanofi) at a final concentration of 100 U/ml was added to each well. Half of the media was changed at day 4 and at day 6 adding fresh IL-2. At day 10, cells were harvested and washed twice in PBS with 2 mM EDTA. Cells were subsequently resuspended in complete RPMI 1640 + 10% AB serum in 50 ml tubes, counted and rested overnight at 37°C. The day after, cells were plated in a 96-well V-bottom plate in 200 µl volume per well and restimulated with 2 μM peptides or controls for 1 h at 37°C. After 1 h, Golgi Stop (BD, 51-2092KZ) and Brefeldin A (BFA) (BD, 420601) were added, and the cells were incubated for additional 4 h. Subsequently, cells were washed, and surface and intracellular staining was performed as described below in the Intracellular Staining methods section, and gating was on singlets, live cells, CD3⁺ followed by identification of TNFα⁺ and IFNγ⁺ in both CD4⁺/CD8⁻ and CD4⁻/CD8⁺ cells.

To detect processing and presentation of the melanoma gp100 Class I peptide gp100 209-217 in the context of cell surface HLA-A2, we used the TCR-like 1A9 mAb (15) H and L chain sequences from patent US20030223994A1 fused to a human IgG4 H chain with a C-terminal Dockerin V1 domain (6) and human L κ chain framework conjugated to either a Cohesin domain fused to a C-terminal CLIP tag (Sequence ID: AQS79240.1 residues 1–182 with appended GEPA) labeled with CLIP-647 according to the manufacturer’s protocol (NEB S9234), or a Cohesin-eGFP fusion protein (LDITH6 residues fused to a Cohesin domain from cellulosomal-scaffolding [Hungateiclostridium thermocellum] WP_065674352.1 residues 1044-1213 with a f1 flexible linker AVY25163.1 residues 580–608 to ABF13214.1 eGFP residues 123–361 followed by a EPEA sequence used for C-tag affinity matrix CaptureSelect™ [Thermo Fisher, 191307005]), or a Alexa Fluor^® 546 C₅ Maleimide (ThermoFisher A10258) fused to Cohesin-SH (Sequence ID: ALO70206.1 residues 1–195 with a C61A replacement and CG appended) through maleimide reaction. The specificity of this antibody conjugate was confirmed by loading gp100 peptides (gp100 G9 Short [NH2]IMDQVPFSV[COOH] or gp100 G9 Long [NH2]CSSSKRIMDQVPFSV[COOH]) at 100 μM onto the human CD40⁺ HLA class-I A2⁺ lymphoblastoid JY cell line (ECACC^® 94022533-1VL) for 24 h and comparing, via flow cytometry, binding to loaded cells versus non-loaded cells ( Supplemental Figure 1A ). To deliver the peptide via DC-targeting, anti-hCD40-IgG4 mAbs were fused via the H-chain C-terminus to two copies of the short peptide interspersed between flexible linker sequences (7). A non-DC reactive hIgG4 mAb fused to the same gp100 peptide cassette was used as a non-targeting control. The gating strategy for flow cytometry analyses is described in Supplemental Figure 1B .

Cohesin fusion proteins were expressed in CHO-S cells from expression vectors containing the HIV-1 Env gp140 sequence (16) or Env fragments (gp120 residues 94–1550; gp41 residues 1551–2064, inserted between the Cohesin domain and 6 C-terminal His codons). ELISA plates were coated overnight at 4°C with 2 μg/ml of the Cohesin fusion proteins in 0.2 M sodium carbonate-bicarbonate buffer, pH 9.4. Serial dilutions of serum starting at 1:500 in TBS blocking solution (StartingBlock T20, Pierce or Thermo Fisher Scientific) were incubated in the wells overnight at 4°C. After washing, plates were incubated with HRP-conjugated goat anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA, USA) in TBS blocking solution (Thermo Scientific, Rockford, IL, USA) for 2 h at 37°C, then washed and developed with HRP substrate (TMB, Life Technologies), stopped with equal volume of 1N HCl and read at 450 nm. The log10 transformed and normalized areas under the curves (AUC) data were calculated for each animal at each time point using GraphPad Prism 8 software. EC50 calculations were based on log10 transformed and normalized data with non-linear regression curve fit using sigmoidal dose response with variable slope constraints.

CD4⁺ T cells specific to the Eα peptide bound to class II I-A^b were adoptively transferred, as previously described (17). Briefly, six skin-draining lymph nodes (LNs), spleens, and mesenteric LNs of TEα TCR transgenic mice were disrupted through a 40 µm cell strainer, washed with sterile Hanks’ Balanced Salt Solution (HBSS), and labeled with cell trace violet (CTV; Thermo Fisher Scientific), according to the manufacturer’s instructions. The cells were resuspended in sterile PBS at a concentration of 1E6 cells/ml, and 300 µl (3E5 cells) were injected intravenously into human CD40 mice (Taconic) (13). Twenty-four hours later, mice were injected intraperitoneally with 1 µg of anti-hCD40 11B6 non-covalently linked to Eα protein or 1 µg of anti-hCD40 11B6-CD40L non-covalently linked to Eα protein, and skin-draining LNs were harvested 4 days later. Cell suspensions were stimulated in complete RPMI 1640 with PMA (Sigma) (5 ng/ml) and Ionomycin (Sigma) (500 ng/ml) for 5 h. BFA (Biolegend) was added to the cell suspension 1 h after the beginning of the incubation. After stimulation, cells were washed and stained for surface markers as described below in the Surface Staining for Flow Cytometry Analysis methods section. Intracellular cytokine staining was performed with the BD Bioscience Cytofix/Cytoperm kit (BD Biosciences, San Jose, CA, USA), according to the manufacturer’s instructions and as described below. The Eα antigen was contained within 6xHis-Cohesin-TEα-Flex-v1C3 protein, which was expressed in E. coli as 6xHis-Cohesin (Sequence ID: ALO70206.1) fused at the C-terminus to residues 52–68 (ASFEAQGALANIAVD) of the I-Eα chain (Eα peptide), followed by KAGGASCQTPTNTISCTPTNNSNPKPCPAS.

One million TNBC EO771.LMB Cyclin D1-expressing tumor cells (18) suspended in 20 μl sterile PBS were injected into the mammary fat pad of 10 to 15 weeks old female, human CD40 transgenic mice (Taconic). The mice were monitored daily by measuring tumor size with a caliper. At days 7, 17, and 24, mice were injected with 20 μg of anti-hCD40 (clone 11B6), anti-hCD40-CD40L (11B6-CD40L), IgG4 or IgG4-CD40L non-covalently linked to either Cyclin D1 constructs p1 or p2-4. The 20 μg dose was administered via injections of 10 μg vaccine, one to deliver p1 and the other to deliver p2-4. A group of mice was left unimmunized. On day 28, the mice were euthanized, and the tumors’ weight determined using an analytical scale. One Cohesin-human Cyclin D1 fusion protein was 6xHis-CthermoCohesin-Flex-v1-hCyclin D1-Peptide-1-f4-Ctag, which has a N-terminal Cyclin D1 segment (sequence ID: AAA36481.1 residues 1-48) appended onto the C-terminal Nhe I site of the His-tagged Cohesin coding sequence (6) followed by a flexible linker f4 (sequence ID: AJD85777.1 residues 697-793) terminating in EPEA, and was expressed in CHO-S cells and purified by C-tag affinity as described above. A second Cohesin-human Cyclin D1 fusion protein was 6xHis-CthermoCohesin-Flex-v1-hCyclin D1-Peptide-2-Peptide-3-Peptide-4-f4-Ctag, which has a Flex V1 sequence (Sequence ID: ALO70201.1 residues 466–496) appended onto the C-terminal Nhe/I site of the His-tagged Cohesin coding sequence (6) followed by the additional Cycln D1 segment (Sequence ID: NP_444284.1 residues 49–295) followed by a flexible linker f4 (Sequence ID: AJD85777.1 residues 697–793) terminating in EPEA, and was expressed and purified as described above.

Cells were transferred to a 96-well V-bottom plate, washed twice in PBS, and incubated for 20 min at 4°C with Live/Dead™ Fixable Aqua Dead Cell Stain Kit (Thermo Fisher Scientific, Cat. L34965) or with Live/Dead-ef780 (Thermo Fisher Scientific 65-0865-14) following the manufacturer’s instructions. Cells were washed twice with PBS and incubated for 30 min on ice with the mix of antibodies in a volume of 50 μl. Finally, cells were washed and resuspended in 200 μl BD™ stabilizing fixative (BD Biosciences, Cat. 338036) per manufacturer’s instructions. Alternatively, cells were resuspended in a volume of 100 μl of Staining Buffer (PBS + 3% Bovine Serum + 5mM EDTA) in a U-bottom plate in the presence of Live/Dead-ef780 and the mix of antibodies for 30 min on ice. Cells were subsequently washed and resuspended in Staining Buffer before being analyzed. All analysis plots were pre-gated on singlet events and live. Cells were analyzed with a FACSCanto II or an LSRFortessa (BD Biosciences). Data was analyzed with FlowJo^® Software (TreeStar; Ashland, OR, USA).

The following antibodies were used for analysis of human DC activation: hCD80-PE, clone L307.4, ref 340294 (BD), hCD83-APC, clone HB15e, ref 551073 (BD); hCD86-FITC, clone 2331 (FUN-1), ref 555657 (BD); hCD11c-PE-Cyanine7, clone 3.9, ref 25011642, (eBioscience); hHLA-DR-V450, clone G46-6, ref 561359 (BD); hCD40-PE-Cyanine5, clone 5C3, ref 555590 (BD). The following panel was used in the melanoma gp100 peptide presentation assays: hCD83-BV421, clone HB15e, ref 562630 (BD); hCD86-Pe-Cy5, clone IT2.2, ref 555666 (BD); hCD11c-APC-Cyanine7, clone Bu15, ref 337218, (Biolegend); hHLA-DR-BV711, clone G46-6, ref 563696 (BD); hCD40-PE, clone 5C3, ref 555589 (BD) (used when the panel allowed it), hCD3-PE, clone SP34-2, ref 552127 (BD) or hCD3-FITC, clone HIT3a (RUO), ref 555339 (BD), hCD19-PE, clone HIB-19, ref 561741 (BD), or hCD19-FITC, clone 4G7, ref 347543 (BD), hCD14-PE, clone 61D3, ref 12-0149-42, (Thermo Fisher), or hCD14-FITC, clone M5E2, ref 301804 (Biolegend), hCD56-PE, clone B159, ref 555516 (BD) or hCD56-FITC, clone HCD56, ref 318304 (Biolegend), hCD66b-PE, clone G10F5, ref 561650 (BD) or hCD66b-FITC, clone G10F5, ref 305104 (Biolegend). hCD19-APC, clone HIB19, ref 555415 (BD) and hCD3-PerCP, clone SK7, ref 347344 (BD) were used in the human B cell proliferation panel. The following reagents were used for the adoptive T cell transfer assay: CD90.1-PerCPCy5.5 (BioLegend 202516), IFNg-PECy7 (Tonbo Biosciences 60-7311-U100), and CD4-BV711 (BioLegend 100447). The T cell proliferation panel was as follows: hCD3-BV711, clone UCHT1, ref 563725 (BD) or hCD3-PerCP clone SK7, ref 347344 (BD), hCD4-Pe-Cy7 clone SK3, ref 348789 (BD), hCD8-PacBlue clone 3B5, ref MHCD0828 (Invitrogen), hTNFα-APC clone RUO, ref 340534 (BD), and hIFNγ-PE clone RUO, ref 340452 (BD).

Human cells were stained for surface markers as described above. After 30 min incubation on ice with the antibodies for surface staining, cells were washed in PBS twice and resuspended in BD Cytofix/Cytoperm (BD Biosciences, 554714) for 20 min at 4°C, followed by three washes in 1X BD Permwash (BD Biosciences, 554714). Cells were subsequently incubated at room temperature covered by light in 1X BD Permwash with the antibody mix for intracellular cytokines. Following the incubation time, cells were washed three times in 1X BD Permwash and resuspended in BD™ stabilizing fixative diluted 1:3. Mouse cells were stained for surface markers as described above. After the incubation time, cells were washed with Staining Buffer and resuspended overnight with the mix of antibodies. Subsequently, cells were washed in BD Permwash, resuspended in BD Permwash, and analyzed. All analysis plots were pre-gated on singlet events and live. Cells were analyzed with a FACS Canto II or an LSR Fortessa (BD Biosciences). Data were analyzed with FlowJo^® Software.

Human CD40 transgenic mice have the human CD40 region BAC inserted into a wild type C57BL/6 background [Taconic strain 12692 (13)]. CD90.1 congenic TEα Rag1^−/− CD4 TCR transgenic mice to I-Eα₅₂₋₆₈ on the C57BL/6 background were obtained from M. Jenkins (University of Minnesota). All experiments were performed with 6- to 12-week-old female and male mice. Mice were housed in micro-isolator cages and fed irradiated food and acidified water. The Baylor Scott and White Research Institute and South Dallas Veteran’s Administration Institutional Care and Use Committees approved all mouse protocols.

Data are presented as means ( ± SEM). Statistical significance was determined by Student’s t test with or without Welch’s correction. A P value < 0.05 was considered as statistically significant. GraphPad Prism^® software was used for statistical calculations. The following nomenclature was used in figures to indicate the statistical significance: * P 0.01 to 0.05; ** P 0.001 to 0.01; *** P 0.0001 to 0.001; **** is P< 0.0001.

Fusion of antigens to the C-terminus of chimeric or humanized anti-CD40 agonistic antibodies can dull or eliminate the agonistic property of the parent antibody (7). We generated a panel of agonistic anti-CD40 antibodies configured on human IgG4 isotypes, with or without a concatenated string of HIV-1 long T cell epitope-rich peptides from the Gag, Nef, and Pol gene regions grafted to their H chain C-termini (7). It is known that antibody isotype can impact CD40 agonistic potency via FcRγ interaction (2); thus, the human IgG4 isotype was selected since it has low binding to FcRγ (19), allowing a clear interpretation that the agonistic and antigen-presenting properties of our tested panel of anti-CD40-antigen fusion mAbs was due exclusively to CD40 interaction. As we described previously (13), the four selected parental mAbs, all with their variable regions configured as human IgG4 H chain and human K L chain constant regions, varied in their potency for evoking human B cell proliferation over a 100-fold range of agonist efficacies ( Figure 1A ) with rank order CP>12B4≥12E12>11B6 [CP is Pfizer CP-870,893, an agonist anti-CD40 human IgG2 antibody in clinical development, here re-configured as a human IgG4 antibody (20)]. In all cases, the agonist activity was severely decreased (>10–100-fold) when the mAb C-terminus was fused to five concatenated HIV-1 long peptide regions interspersed with glycosylated flexible linkers (HIV5pep antigen cassette) ( Figure 1A ). We also previously showed that the anti-CD40 11B6 and CP mAbs, which do not significantly block CD40L interaction with CD40, synergize strongly with soluble CD40L (13) ( Figure 1B ). The agonistic anti-CD40 11B6, 12B4, 12E12, and CP hIgG4 mAbs all became very weak agonists for eliciting B cell proliferation when the HIV5pep antigen cassette was grafted to their H chain C-termini ( Figure 1A ). Addition of suboptimal levels of monomeric sCD40L restored the B cell proliferation potency of the anti-CD40 11B6-HIV5pep and anti-CD40 CP-HIV5pep mAbs to the level characteristic of the analogous “naked” anti-CD40 11B6 and CP mAbs incubated with sCD40L, but had no significant effect on the anti-CD40 12B4-HIV5pep or anti-CD40 12E12-HIV5pep mAbs ( Figure 1B ).

Since dendritic cells (DCs) are the desired target cells for delivering antigens for potent vaccination responses, we examined these same four agonist anti-CD40-HIV5pep fusion proteins for potency of upregulation of activation markers on human myeloid-derived dendritic cells (MDDCs). As previously described (13) and as observed with the B cell proliferation assay ( Figures 1A, B ), the non-fusion antibody panel had minimal efficacy at the 10–0.1 nM concentrations tested compared to the more sensitive B cell proliferative response ( Figure 1C ). However, as previously observed (13), when assayed over this dose range together with a suboptimal dose of soluble monomeric CD40L (sCD40L), there was no effect on the dose response of the anti-CD40 12B4 and anti-CD40 12E12 antibodies, but the low 6 nM dose of sCD40L greatly increased the potency and efficacy of the anti-CD40 11B6 and CP antibodies ( Figure 1D ). As observed with B cell proliferation, suboptimal level of sCD40L also potentiated the activity on MDDCs of the anti-CD40 11B6-HIV5pep and anti-CD40 CP-HIV5pep fusion proteins, but had no effect at the tested doses on the anti-CD40 12B4-HIV5pep or anti-CD40 12E12-HIV5pep fusion proteins ( Figure 1D ). Overall, the extent of sCD40L potentiation at the sCD40L concentration tested was ~5- to 10-fold less with the anti-CD40-HIV5pep fusion proteins than that observed with the anti-CD40 11B6 and CP non-antigen fused mAbs ( Figure 1D ).

Unlike in the B cell proliferation assay, a large increase in efficacy (i.e., maximal activation marker increase achieved) between antibody alone and antibody with sCD40L was observed, indicating a greater potential for co-operation between these two CD40 agonist types on MDDCs. This effect was also seen with potentiation of cytokine secretion by non-antigen fused anti-CD40 11B6 and CP antibodies when combined with sCD40L (13).

We have shown that direct fusion of CD40L to agonistic anti-CD40 mAbs stabilizes binding to CD40 and thus potentiates their agonist activities, emulating or surpassing the synergism seen with co-administered sCD40L (13). While previous prototypes of anti-CD40-HIV5pep vaccines were based on the humanized 12E12 mAb (14, 21), we focused on the anti-CD40 11B6 mAb since it shows the strongest cooperation with soluble CD40L for agonist activity ( Figures 1B, D ) (13). Similarly to what we previously observed (13), using Surface Plasmon Resonance analyses with solid phase CD40 ectodomain to test the binding kinetic properties of our anti-CD40 constructs, direct fusion of CD40L to the anti-CD40 12E12 did not improve the affinity of this antibody, likely due to the fact that the anti-CD40 12E12 binding site interferes with CD40L binding, but in contrast, CD40L fusion greatly improves the affinity via stabilizing off-rate of the anti-CD40 11B6 antibody, which binds CD40 without blocking CD40L access ( Figure 2A ) (13). The direct fusion of the HIV5pep adduct on both the anti-CD40 12E12 and on the anti-CD40 11B6 slightly reduced the affinity of both antibodies by ~66% and ~50% respectively, likely due to changes in diffusion from increased bulk. However, direct fusion of CD40L to the anti-CD40 11B6-HIV5pep construct maintained the greatly increased affinity to CD40 observed with anti-CD40 11B6-CD40L without fused antigen, while anti-CD40 12E12-HIV5pep did not improve affinity to CD40 after direct fusion to CD40L ( Figure 2A ). These data confirm that the direct fusion of CD40L on anti-CD40 11B6, either with or without an antigen attached, improves the affinity of the antibody for CD40.

Thus, we asked if the improved binding to CD40 by anti-CD40 11B6-CD40L HIV5pep fusion protein could mirror the restoration of high agonist activity we observed by adding soluble CD40L. In B cell proliferation assays, anti-CD40 11B6-HIV5pep had minimal agonist activity; however, both anti-CD40 11B6-HIV5pep + sCD40L and anti-CD40 11B6-CD40L-HIV5pep had high agonist activity that approached that of the superagonist anti-CD40 11B6-CD40L ( Figure 2B ). This indicates that direct fusion of CD40L can render a weak agonist anti-CD40-HIV5pep fusion construct into a strong agonist. Anti-CD40 11B6-CD40L-HIV5pep constructs also had potent agonist activity on DCs ( Figures 2C–E ), exceeding the activity of the naked anti-CD40 CP agonist hIgG4 mAb or the highly active Mega sCD40 trimer (22). Thus, the observed increase in CD40 binding by the anti-CD40 11B6-CD40L-HIV5pep protein is associated with highly increased activity on B cells and DCs.

To test if CD40L fusion to anti-CD40 IgG4 L chain C-termini could also increase agonist potency while fused to other desirable antigens, we compared agonistic activities when anti-CD40 11B6 and 12E12 IgG4 mAbs were fused at their H chain C-termini to concatenated HIV-1 Gag p24 Nef Gag p17 (called GNG) or HPV16 E6/E7 (called HPV) antigens (9), with or without CD40L directly fused to the L chain. These two antigens do not significantly dull the already low potency of B cell CD40 activation of the parent 11B6 mAb, but CD40L L chain fusion potentiated the activation to levels equal to the mAbs co-administered sCD40 ( Figure 3A ). The increase on B cell proliferation activity was mirrored by robust cytokine production from MDDC elicited by anti-CD40 11B6-GNG ( Figures 3B, C ). Altogether, these data demonstrate the superagonist properties of anti-CD40 11B6-CD40L on B cells and DCs are maintained when fused to the three tested antigen modules—HIV5pep, GNG, and HPV16 E6/E7.

Peripheral Blood Mononuclear Cell (PBMC) and DC-T cell co-culture systems are useful in in vitro assays for validating DC-targeting prototype vaccine constructs, in particular for selecting the best DC receptor to target, e.g., for cellular T cell response (8), as well as confirming the efficacy of the selected fused antigen for eliciting a broad range of T cell peptide specificities for both CD4⁺ and CD8⁺ T cell recall responses across a range of HLA types (7). To investigate the potential impact of associating antigen delivery with potent CD40 activation, we first evaluated the efficacy for HIV-1 antigen-specific T cell expansion in HIV-1⁺ donor PBMC cultures of anti-CD40 mAbs fused at the C-termini to a Gag p17-Nef-Gag p24 (GNG) antigen cassette, with and without a low dose of sCD40L or with CD40L directly fused to the L chain C-terminus. PBMCs from the donor shown in Figure 4 treated with a dose ≥0.01 nM of CD40-targeting GNG constructs expanded memory CD4⁺ T cells specific to epitopes within all three Gag p17, Nef, and Gag p24 regions, while such cells were minimally represented in antigen non-stimulated or non-targeting hIgG4-GNG treated cultures ( Figure 4A ). All anti-CD40-GNG targeting constructs gave similar CD4⁺ T cell GNG-specific responses, except anti-CD40 11B6-CD40L-GNG elicited greater GNG-specific responses at the lowest tested (0.001 nM or 1 pM) dose consistently in two donors tested, although these were not statistically significant ( Figure 4A and Supplemental Figure 2A ). In this donor, all anti-CD40-GNG targeting constructs elicited CD8⁺ T cell responses specific to epitopes, especially within Nef, while such cells were minimally represented in antigen non-stimulated or non-targeting hIgG4-GNG treated cultures ( Figure 4B ). The most robust expansion of Nef-specific CD8⁺ T cells was observed in cultures treated with 1 and 0.1 nM of the highly activating anti-CD40 11B6-CD40L-GNG and anti-CD40 11B6-GNG + sCD40L constructs ( Figure 4B and Supplemental Figure 2B ).

The improved antigen-specific CD8⁺ T cell response associated with concomitant HIV-1 antigen delivery and activation via CD40 we observed above with the GNG model antigen suggested the possibility of a similar benefit for CD40 targeting of the concatenated HIV5pep antigen cassette that is well advanced towards clinical development (7, 21). HIV5pep was designed to incorporate relatively conserved CD4⁺ and CD8⁺ T cell epitopes across multiple HLA haplotypes from the HIV-1 Gag, Nef, and Pol regions within five peptides. We compared via in vitro PBMC expansion cultures various formats of anti-CD40 11B6-CD40L-HIV5pep prototype vaccines compared to the anti-CD40 12E12-HIV5pep Gen 2 vaccine (14). In the Gen 1 format, the five HIV-1 peptide regions intercalated with glycosylated flexible linkers are grafted to the mAb H chain C-termini, and this is compatible with adding CD40L to the available L chain C-termini. In the Gen 2 format, the same long peptides and linkers are used, but two are now grafted to the L chain C-termini. By configuring the same 2 + 3 peptides and linkers on separate mAb H chains using knob-in-hole (KIH) technology (23), highly activating anti-CD40 11B6-CD40L-HIV5pep KIH was also available ( Figure 2B ). Previous studies demonstrated that the Gen 1 and Gen 2 formats yield identical in vitro HIV-1 antigen-specific T cell responses (14). The Gen 1 and Gen 2 versions of the HIV5pep vaccine product were unsuitable for Good Manufacturing Practice (GMP) scale-up due to fragmentation and aggregation issues that were not evident at the research scale; the KIH Gen 3 product was developed to circumvent these problems and is now in GMP production.

We compared in vitro T cell responses to anti-CD40 12E12-HIV5pep Gen 2 to the highly activating anti-CD40 11B6-CD40L-HIV5pep Gen 1, and both vaccines elicited CD4⁺ T cell responses specific to epitopes with the Gag17, Gag253, Nef66, and Pol325 long peptide regions, as well as CD8⁺ T cell responses within the Gag17 and Gag253 regions in the illustrated donor ( Figure 5 ). However, the highly activating anti-CD40 11B6-CD40L-HIV5pep vaccine preferentially expanded the dominant Gag253-specific CD8⁺ T cell response, while the same CD4⁺ T cell responses to Gag17, Gag253, Nef66 were elicited, but to a lesser abundance ( Figure 5B ). Comparative in vitro T cell expansion studies with anti-CD40-HIV5pep versus anti-CD40-CD40L-HIV5pep constructs were replicated in three other HIV-1⁺ donors and are collated ( Figure 6 ). Overall, T cell responses were highly correlated between the vaccine pairs (Pearson’s test 0.85 for CD8⁺ and 0.7 for CD4⁺) indicating the vaccines presented a similar array of epitopes for activating populations of antigen-specific T cells. However, the HIV-1-specific CD8⁺ T cell responses were substantially (7.2X) more robust with the highly activating anti-CD40-CD40L-HIV5pep vaccines ( Figure 6A ). In contrast, the HIV-1-specific CD4⁺ T cell responses were similar or slightly depressed (0.7X) with the highly activating anti-CD40-CD40L-HIV5pep vaccines compared to anti-CD40-HIV5pep ( Figure 6B ), likely reflecting the increased dominance of antigen-specific CD8⁺ T cells in these cultures.

To extend our observations with HIV-1 donor PBMCs and to further validate the importance of our vaccine platform technology, we compared the ability of anti-CD40-CD40L to target Influenza matrix 1 (Flu M1) protein and expand Flu M1-specific memory T cells. PBMCs obtained by apheresis of Normal Donor (ND) (All Cells, CA. ID:11588) were cultured with dose ranges of Cohesin-Flu M1 alone (Flu M1) or in complex with three different CD40-targeting antibody vehicles. We have previously described a convenient method for non-covalent assembly of anti-DC antibodies and antigens using a bacterial Dockerin (Doc) domain fused to the antibody heavy chain C-terminus, and antigen such as Flu M1 fused to a Cohesin (Coh) counter-domain (6). After an expansion culture period of 10 days, cells were harvested and restimulated with three pools (clusters) of overlapping 15 mer peptides covering the entire Flu M1 protein, then analyzed by ICS for peptide-elicited production of intracellular IFNγ and TNFα. As represented by the donor shown in Figure 7 and statistically verified by four replicated experiments, compared to Flu M1 alone, anti-CD40-Flu M1 complexes efficiently expanded CD8⁺ T cells specific to cluster 1, with anti-CD40 11B6-CD40L showing a significance advantage at the 50 pM dose (p=0.018) ( Figures 7A, C ). Anti-CD40 11B6-CD40L was more potent at the low concentrations tested (10 and 50 pM) compared to the other anti-CD40 antibodies, with a significant difference at 50 pM compared to anti-CD40 11B6 (p=0.035). In these same cultures, anti-CD40 11B6 and anti-CD40 12E12 vehicles were similarly potent at expanding Flu M1-specific CD4⁺ T cells to cluster 2 and 3 peptides, but these responses were less robust with anti-CD40 11B6-CD40L targeting ( Figures 7B, C ), likely due to the dominant expansion of Flu M1-specific CD8⁺ T cells in these cultures. All the three anti-CD40 targeting agents at the higher 50 pM dose gave reduced or equal Flu M1-specific CD4⁺ T cell expansion compared to Flu M1 alone ( Figures 7B, C ), but this effect was not observed at the lower 10 pM dose or with Flu M1-specific CD8⁺ T cell expansion ( Figures 7A, C ). Repeated experiments with this donor confirmed these results ( Figure 7C ). These data, together with what we observed with GNG and HIV5pep as antigens, demonstrate the high potency of the anti-CD40 11B6-CD40L-antigen at the lower doses tested and its improved efficacy to preferentially expand CD8⁺ T cells compared to anti-CD40 12E12-antigen.

To further investigate the improved CD8⁺ responses, we used the 1A9 mAb (15) to detect processing and presentation on MDDCs of the melanoma gp100 Class I peptide gp100 209-217 in the context of cell surface HLA-A2. Indeed, the size of the CD8⁺ T cell responses is known to correlate with the duration of the antigen presentation (24). To deliver the peptide via DC-targeting, anti-human CD40 IgG4 mAbs with and without directly linked CD40L were fused via the H chain C-terminus to two copies of the peptide interspersed between flexible linker sequences (7). These constructs at 50 nM were incubated with MDDCs, and the cells were probed with the labeled 1A9 mAb after 24, 48, and 72 h and analyzed by flow cytometry. In three independent experiments, compared to anti-CD40 11B6-gp100 or anti-CD40 12E12-gp100-loaded MDDCs, anti-CD40 11B6-CD40L-gp100-loaded MDDCs had greater and more sustained specific cell surface reactivity to the 1A9 mAb, which correlates with the observed increase in expression of surface activation markers CD83 and CD86 ( Figure 8 and Supplemental Figure 1C ). These results suggest that anti-CD40 11B6-CD40L-mediated antigen delivery to DCs can prolong antigen presentation on MHC I, potentially explaining the observed increased CD8⁺ T cell responses to anti-CD40 11B6-CD40L delivery.

To test the potential of anti-CD40 11B6-CD40L-antigen constructs for increasing vaccine efficacy in the absence of co-administered adjuvant, human CD40 transgenic mice were vaccinated with anti-CD40 11B6 hIgG4 delivery vehicles, with and without fused CD40L, coupled to HIV-1 Env gp140. Vaccination with anti-CD40 11B6-CD40L directly fused to gp140 was compared to vaccination with anti-CD40 11B6-CD40L non-covalently coupled to gp140 via a Dockerin-Cohesin interaction (6), to anti-CD40 11B6 non-covalently coupled to gp140. Cohesin-gp140 was used as the non-targeted control ( Figure 9A ). Both anti-CD40 11B6-CD40L non-covalently coupled to gp140 and anti-CD40 11B6-CD40L directly fused to gp140 elicited serum anti-gp140 IgG titers that were detected as early as 13 days after a single vaccination ( Figure 9B , left panel). Responses to these anti-CD40 11B6-CD40L-based vaccines increased to similar extents when monitored 13 days after each of two subsequent vaccinations administered at 2-week intervals (day 14 and day 21) after the initial D0 vaccination. These responses were significantly increased compared to the group immunized with the non-targeted gp140 after the third vaccination (p<0.05) ( Figure 9B , right panel) where non-targeted control Cohesin-gp140 failed to elicit any detectable anti-gp140 IgG responses even after three vaccinations despite using a three-fold molar excess of gp140 compared to the CD40-targeting vaccines. Vaccination responses following the same administration and sampling schedule as above with anti-CD40 11B6 non-covalently linked to gp140 elicited serum anti-gp140 IgG titers that were detected only after a second vaccination ( Figure 9B , D27 middle panel), and increased further after the third (day 21) vaccination ( Figure 9B , D34 right panel), but the titers were substantially lower compared to the two anti-CD40 11B6-CD40L-based vaccines.

A well-established mouse model for T cell vaccine response uses adoptively transferred CD4⁺ T cells from TCR I-Eα52-68 transgenic mice (25). We compared vaccination with anti-CD40 11B6 versus anti-CD40 11B6-CD40L for targeted delivery of Eα antigen for evoking proliferation and activation of these adoptively transferred cells ( Figure 10A ). In both cases, in response to a single 1 µg dose of each construct, >80% of the adopted TEα CD4⁺ T cells proliferated over the 4 days period subsequent to vaccination, indicating efficient processing and presentation of the antigen by both constructs ( Figure 10B ). However, a significantly higher proportion of the proliferating TEα CD4⁺ T cells were primed for IFNγ production by the anti-CD40 11B6-CD40L-TEα vaccine (18± versus 5±%, p<0.0001, Figure 10C ).

To further explore the possible advantage in efficacy for raising cellular immunity by targeting antigen via anti-CD40 11B6-CD40L, we established EO771 mammary adenocarcinoma tumors [derived from a spontaneous mammary tumor in a female C57BL/6 mouse (26)] in human CD40 transgenic female mice. These tumors express Cyclin D1, a tumor-associated antigen common in breast cancers (27). At days 7, 17, and 24, mice implanted with tumor at day 0 were vaccinated with anti-CD40 11B6-CD40L-Doc, anti-CD40 11B6-Doc, IgG4-CD40L-Doc, and non-targeting IgG4-Doc non-covalently linked to human Coh-Cyclin D1 ( Figure 11A ). Only the anti-CD40 11B6-CD40L-Cyclin D1 group showed reduction in tumor growth ( Figure 11B ), indicating the low impact of non-targeted antigen, or antigen targeted to CD40 without strong CD40 activation potential, versus the significantly improved ability of the anti-CD40-CD40L vehicle to partially control tumor growth.