Innate Immune Activation Via TLR9 May Enhance Antibody Formation Against Factors VIII and IX in Protein and Gene Therapy

In recent years, the potential impact of innate immune responses on inhibitor formation in hemophilia has increasingly received attention, as innate immune sensing may provide a signal that leads to antibody formation, thus seriously complicating treatment of hemophilia. Similar concerns exist with the newer gene therapy approaches, where the gene transfer vector stimulates innate immunity. Curiously, we found that antibody formation against factor IX (FIX) in gene therapy for hemophilia B is independent of TLR9 sensing of the adeno-associated viral (AAV) genome. This is despite the critical role of TLR9 (toll-like receptor 9, an innate endosomal sensor of DNA) in CD8⁺ T cell responses against this vector (Blood 129:3184-3195, 2017). These results are consistent with the notion that TLRs more typically promote Th1 than Th2 immunity. However, when we co-administered an CpG-deoxyoligonucleotide (ODN-1826) agonist of TLR9 with AAV1-FIX vector (1x10¹¹ vg/mouse) to skeletal muscle of C57BL/6 mice, IgG1 formation against FIX was reproducibly observed at average titers of 3-5 ug/ml serum within 2 weeks. Control mice, which received vector only, formed <1 ug/ml to undetectable anti-FIX titers. Therefore, TLR9 activation via a different ligand (instead of the AAV genome) can induce antibody formation. Co-administration of lipopolysaccharides that are known ligands for TLR2 and TLR4 failed to induce an anti-FIX response. Although both TLR4 and TLR9 agonists upregulated co-stimulatory molecules on dendritic cells (DCs) upon intramuscular injection, only the TLR9 agonist increased the frequency of monocyte-derived DCs (moDCs), which are known to enhance T follicular helper (Tfh) cell activation, thereby enhancing germinal center and antibody formation. Indeed, we found that co-administration of ODN-1826 significantly increased frequencies if Tfh (PD-1^hiBcl-6⁺CXCR5^hiCD4⁺) cells. This occurred in the presence or absence of FoxP3⁺ Treg, although antibody formation was only increased in the presence of Treg, indication a critical role for Treg in modulating the response to the ODN. Interestingly, others had reported that ODN-1826 reduces antibody formation against FVIII in hemophilia A mice (JCI 125:3766-81, 2015), and proposed a mechanism by which Treg are induced upon TLR9 stimulation of plasmacytoid DCs, thereby suppressing anti-FVIII responses. We similarly injected FVIII (1IU, B domain deleted) intravenously once/week for 6 weeks into hemophilia A mice (F8e16-/-, C57BL6/129) with our without ODN-1826 (n=4/group). We indeed found a 4-fold lower IgG titer against FVIII in the presence of the ODN (ave. of 0.9 vs 3.6 ug/ml, P<0.05). However, inhibitor titers were substantially increased by ODN co-administration (ave. of 1675 BU vs 340 BU, P<0.001). Therefore, TLR9 activation reduced the overall antibody production but increased the formation of neutralizing antibodies. In summary, our data show that a pathway exists for TLR9-drived innate immune signaling to increase the risk of inhibitor formation in gene and protein therapy for hemophilia.