Cell targeting and immunostimulatory properties of a novel Fcγ-receptor independent agonistic anti-CD40 antibody in rhesus macaques

Targeting CD40 by agonistic antibodies used as vaccine adjuvants or for cancer immunotherapy is a strategy to stimulate immune responses. The majority of studied agonistic anti-human CD40 antibodies require crosslinking of their Fc region to inhibitory FcγRIIb to induce immune stimulation although this has been associated with toxicity in previous studies. Here we introduce an agonistic anti-human CD40 monoclonal IgG1 antibody (MAB273) unique in its specificity to the CD40L binding site of CD40 but devoid of Fcγ-receptor binding, we demonstrate rapid binding of MAB273 to B cells and dendritic cells resulting in strong activation in vitro on human cells and in vivo in rhesus macaques. Dissemination of fluorescently labeled MAB273 after subcutaneous administration was found predominantly at the site of injection and specific draining lymph nodes. Phenotypic cell differentiation and upregulation of genes associated with immune activation were found in the targeted tissues. Antigen-specific T cell responses were enhanced by MAB273 when given in a prime-boost regimen and for boosting low preexisting responses. MAB273 may therefore be a promising immunostimulatory adjuvant that warrants future testing for therapeutic and prophylactic vaccination strategies.


Introduction 1
proliferation was clear with both MAB273 and CP-870,893, we did not find detectable 1 cytokines such as IL-12 p40, IL-6, IFN-γ and TNF in the cell culture supernatants after 2 stimulation (data not shown). 3 4 CD40 binding and activation capacities remain after removing the Fc region of MAB273 5 In order to evaluate if MAB273 is FcγR-independent and the role of avidity, we generated 6 F(ab')2 fragments by pepsin digestion to cleave off the Fc region but maintain the hinge 7 region as well as Fab fragments by papain digestion to remove both the Fc region and hinge 8 region ( Figure 2A). We confirmed that the Fab and F(ab')2 fragments of MAB273 still bound 9 to CD40 but were not detected by an anti-Fc antibody ( Figure 2B). Human B cells and MDCs 10 exposed to MAB273 or the F(ab')2 fragment showed similar ability to block the CD40 11 staining antibody in a dose dependent manner, while the Fab fragment showed weaker CD40 12 blocking capacity ( Figure 2C). In addition, MAB273 and the F(ab')2 exposure resulted in 13 similar upregulation of CD80, CD70 and CCR7 on B cells ( Figure 2D) and MDCs ( Figure  14 2E), while the Fab fragment retained the activation capacity on MDCs ( Figure 2E) but was 15 weaker for B cells ( Figure 2D) suggesting that B cells may require higher avidity for 16 activation ( Figure S1A). In addition, B cell proliferation induced by MAB273 or F(ab')2 was 17 similar but Fab showed weaker induction ( Figure 2F). This demonstrates that the F(ab')2 18 fragment of MAB273 has retained immunostimulatory capacities in vitro after removing the 19 Fc region, but the Fab fragment was less potent for B stimulation. Nevertheless, we conclude 20 that MAB273-induced activation is not dependent on FcγR crosslinking. 21

MAB273 binds CD40 and activates rhesus macaque PBMCs in vitro 23
With the further aim of utilizing a physiological in vivo animal model, we next tested the 24 ability of MAB273 to bind and stimulate rhesus macaque PBMCs in vitro by repeating a large subset of the above in vitro experiments. The immune cells expressed CD40 as 1 expected where B cells and MDCs had the highest expression and T cells and neutrophils low 2 expression ( Figure S1B). Phenotypic differentiation after stimulation with MAB273, CP-3 870,893 or the isotype control antibody (IgG1-LALA) were analyzed. Again, the signal of 4 the CD40 staining antibody was blocked when the cells had been exposed to MAB273 or CP-5 870,893 but not to the isotype control ( Figures 3A and 3B). In addition, MAB273 6 upregulated CD80 on B cells ( Figure 3A) and MDCs ( Figure 3B). Less upregulation was 7 found by CP-870,893 and no upregulation was found by the isotype control antibody. B cell 8 proliferation was also induced by MAB273 exposure but not by the isotype control antibody 9 ( Figure 3C). MAB273 induced low but detectable levels of IL-12 p40, IL-6 and TNF 10 secretion ( Figure 3D). IFN-γ was not detected (data not shown). This shows that MAB273 11 can bind to rhesus macaque CD40 and activate immune cells. 12 13

MAB273 induces innate immune activity in vivo in rhesus macaques 14
Six rhesus macaques were thereafter divided into three groups to receive MAB273 15 administration at different doses and routes ( Figure 4A). Standard clinical chemistry analyses, 16 including a series of liver and kidney function and complete blood count (CBC) 17 measurements were performed in addition to immunological analyses. The first group that 18 received the highest dose of 1 mg/kg by intravenous (i.v.) administration showed that several 19 clinical chemistry parameters including alkaline phosphatase (ALP), alanine transaminase 20 (ALT), gamma-glutamyl transferase (GGT), bile acid (BA), total bilirubin (TBIL) and blood 21 urea nitrogen (BUN) were elevated above the normal reference range ( Figure S2A). This 22 group also showed side effects characterized by loss of appetite and reduced activity behavior 23 for up to 5 days. In contrast, the lower dose of 0.1 mg/kg given i.v. did not induce any 24 detectable side effects and most of the clinical chemistry parameters remained within the healthy range ( Figure S2A). The low dose was thereafter tested with subcutaneous (s.c.) 1 administration which neither induced side effects. As visualized by CBC, there were 2 fluctuations of cell numbers following administration of MAB273 found with both doses and 3 routes. A rapid decline in platelets was observed already at 0.5-4 hours accompanied by a 4 rapid increase in white blood cell counts, especially granulocytes, was found after MAB273 5 administration in all groups ( Figure S2B). Frequencies of specific cell subsets identified by 6 flow cytometry and normalized to the CBC data confirmed a rapid increase in neutrophils 7 while there was a transient decline in both B cells and MDCs. The cell fluctuations were 8 dose-dependent and with a notably more dramatic effect in the 1 mg/kg i.v. group (Figure  9 S2C). The transient fluctuation of immune cells after MAB273 administration may stem from 10 redistribution of activated cells leaving the circulation to migrate to tissues followed by a 11 replenishment of cells from the bone marrow as has been proposed earlier [11,22]. Body 12 weight remained stable during the entire study period in all groups ( Figure S2D). 13 14 Analysis of the pharmacokinetics (PK) of MAB273 in plasma showed that the levels were 15 readily detectable after 0.5 hour of administration in both of the i.v. groups ( Figure 4B). In 16 the high dose group, the levels of MAB273 peaked around 0.5-4 hours and then declined 17 gradually until it was undetectable after 2 weeks. In the low dose group, the highest level was 18 detected at 0.5 hour, then continually decreased and was undetectable after 1 week. In the s.c. 19 group, MAB273 was detectable at 0.5-4 hours but at 2-4 log lower levels compared to the i.v. 20 groups. However, the level of MAB273 in the s.c. group was sustained for a week and 21 undetectable at 2 weeks. This suggests that s.c. administration results in a depot effect and 22 slower release of MAB273 into the circulation. 23 Binding of MAB273 to CD40 in vivo was evaluated by quantifying the loss of detection 1 signal from the staining CD40 antibody as performed in the in vitro experiments. Rapidly 2 (0.5 hour) after administration of MAB273, detection of CD40 was blocked on B cells and 3 MDCs ( Figure 4C). Lack of CD40 signal was sustained for 72 hours in the high dose i.v. 4 group while this was found for a shorter period for the low dose i.v. group. The s.c. group 5 also showed reduced signal for CD40, but this was noticed later (at 4 hours) and sustained for 6 2 weeks in line with the observed pharmacokinetics of MAB273 in plasma. As mentioned 7 above, the return of detectable CD40 expression may be due to replenishment of new cells 8 into the circulation as well as the half-life of MAB273. Accompanied with MAB273 binding 9 to CD40 on immune cells, a rapid increase in CD80 and CCR7 expression was observed 10 especially in i.v. groups ( Figure 4D). The expression gradually returned to baseline levels or 11 even below which may be explained by that newly recruited cells exhibit a more immature 12 phenotype. The upregulation of CD80 and CCR7 on MDCs was less noticeable than on B 13 cells ( Figure 4E). Secretion of IL-12 p40, IL-6 and IFN-γ was detected in one of the animals 14 receiving the high dose while most animals did not show detectable levels ( Figure 4F). TNF 15 was not detected (data not shown). In conclusion, MAB273 induces strong innate immune 16 activation with regards to cell recruitment and activation while being well-tolerated at the 17 dose of 0.1 mg/kg given either i.v. or s.c. in rhesus macaques. Since s.c. administration 18 demonstrated clear immune stimulation while potentially offering a depot effect of MAB273 19 for slower release and better tolerability, this route may be more attractive for clinical 20 development and hence this was used in our subsequent studies. 21

MAB273 targets and activates immune cells at the site of injection and draining lymph 23
nodes 24 To understand the biodistribution of MAB273 in different tissues after administration, the 1 antibody was labeled with AlexaFluor 680 fluorochrome to enable tracking in vivo. The 2 fluorescent signal and unaltered CD40 binding and activation capacities of the labeled 3 MAB273 were validated in vitro before administered in vivo ( Figures S3A-S3C). Three 4 animals were immunized s.c. and biopsies were collected after 24 (n=1) or 48 hours (n=2) 5 from the sites of injection, lymph nodes (LNs) and other selected tissues ( Figure 5A). To further understand the immune activation profile induced by MAB273 administration in 6 vivo, we performed RNA sequencing analyses on the draining LNs and skin from the site of 7 injection as well as the blood ( Figure 6A). This revealed a significant number of differentially 8 expressed genes (DEGs) in the MAB273 targeted skin and LNs compared to the donor-9 matched saline control sites ( Figures 6B and 6C). In addition, blood taken before MAB273 10 administration compared to 24-48 hours after showed significant gene modulation (  The results indicated activation and recruitment of T cells, B cells, NK cells, monocytes, and 20 DCs to the site of injection ( Figure 6B). MAB273-draining LNs also showed that there were 21 genes upregulated compared to the saline-draining LNs. These genes were fewer and were 22 distinct from those observed in the skin. The upregulated genes in the LNs were mainly 23 linked to antigen presentation (IRAG2) and interferon (IRF6), and a few downregulated genes 24 were linked to RNA processing (U2, U3, U4, RNaseP). The enrichment analysis indicated an upregulation of modules related to cell proliferation (mitotic cell division and cell cycle 1 modules) ( Figure 6C). Furthermore, in the blood, the genes differentially expressed 24-48 2 hours after MAB273 administration were mainly associated with interferon signatures, such 3 as ISG15, RSAD2, and SKIV2L as well as genes associated with monocytes and DC 4 activation ( Figure 6D). MAB273 therefore induces significant innate immune activation 5 characterized by monocyte and DC activation in the blood, recruitment of immune cells to 6 the site of injection while cell proliferation and antigen presentation processes were more 7 dominant in the draining LNs. 8 9

MAB273 exhibits adjuvant effects for induction of antigen-specific CD4 and CD8 T cells 10
Finally, we evaluated the effect of MAB273 to act as an adjuvant both for therapeutic 11 vaccination where low degree of immunity already exists and also to enhance primary 12 immune responses as in prophylactic vaccination. Three animals therefore first received 13 seven well characterized HIV-1 envelope glycoprotein (Env) peptides[24] as model antigen 14 alone two times to establish low levels of immunity before receiving boost immunizations 15 with MAB273 co-administered s.c. to mimic a therapeutic vaccination ( Figure 7A). In a 16 separate group, three animals received MAB273 together with the Env peptides in a prime-17 boost schedule of four immunizations to mimic prophylactic vaccination. The final 18 immunization was performed with an additional recombinant trimer Env protein ( Figure 7A). 19 20 Low frequencies of Env-specific T cell responses were induced by Env peptide immunization 21 alone ( Figure 7B). The responses were enhanced in two out of three animals when they 22 received a boost with Env peptides and MAB273. This effect was evident for both systemic 23 Env-specific CD4 and CD8 T cells in blood ( Figure 7B) as well as in bronchoalveolar lavage 24 (BAL) ( Figure 7B). Two out of the three animals immunized with Env peptides in combination with MAB273 already at prime immunization induced higher levels of Env-1 specific CD4 and CD8 T cell responses compared to the animals receiving Env peptides only 2 ( Figures 7B and 7C). Although the subsequent boost immunizations re-activated T cell 3 responses, they did not reach the peak levels found after the prime immunization ( Figure 7C). 4 This was observed both in blood and BAL and may be a consequence of the low dose of Env 5 peptides (0.1 mg/kg) and the induction of antibodies against the humanized MAB273 in 6 rhesus macaques ( Figures S4A and S4B). 7 8 The activation profile of MAB273 based on the RNA sequencing and blood transcriptome 9 analysis comparing the activation at pre-immunization compared to the second boost showed 10 that the differences were negligible indicating that recurrent MAB273 administration may 11 result in lower innate immune activation ( Figure S4C). Nevertheless, reactivation of memory 12 T cell responses to peak levels occurred after the fourth immunization of MAB273 when 13 using trimer Env protein in combination with Env peptides to provide more antigen ( Figure  14 7C). No detectable IgG to Env peptides was found ( Figure S4D) but IgG to Env protein was 15 detected ( Figure S4E). Taken    and blood (D) compared to their respective control samples, color gradient is based on 23 normalized gene set enrichment scores. All statistical comparisons were adjusted by the Benjamini-Hochberg procedure, adjusted p-values < 0.05 were considered significant. "D0" is 1 the day before prime immunization, "D2" is 48h after prime. See also Figure S3. were administered 0.1mg/kg s.c. of Env peptides for the first two immunizations then 1 mg/kg 6 Env peptides plus 0.1 mg/kg MAB273 s.c. for the last immunization; in the prophylactic 7 vaccination group, rhesus macaques (n=3) were co-injected with 0.1 mg/kg Env peptides plus 8 0.1 mg/kg MAB273 s.c. for the first two immunizations then 1 mg/kg Env peptides plus 0.1 9 mg/kg MAB273 s.c. for the third immunization, followed 1 mg/kg Env peptides, 0.1 mg/kg 10 MAB273 plus 100