Inhibitor-Positive Hemophilia a Patients-Derived Humanized Mouse Models Revealed the Connection of PD-1/PD-L1 Interactions to Successful Immune Tolerance Induction (ITI) Therapy

Background: Patients with hemophilia A (PWHA) are at risk of developing FVIII inhibitors. To date, immune tolerance induction (ITI) therapy is still the only way to induce tolerance with FVIII. Developing mouse models for ITI testing offers a promising approach to benefit PWHA. However, humanizing immunodeficient mice with patients' immune cells poses challenges such as graft-versus-host disease (GvHD) and poor survival of human B cells, limiting their utility for ITI studies. Furthermore, while studies suggest associations between FVIII immune responses and PD-L1/PD-1 expression in Treg/plasma cells and TFH cell activation, conclusive mechanisms underlying successful ITI induction remain elusive.

Aims: We aim to establish an ITI testing platform using PWHA’s peripheral blood nuclear cells (PBMCs) to humanizing a novel immunodeficient mouse model that minimizes GvHD and enhances B cell survival. We also investigate PD-1/PD-L1 expression in Treg/plasma cells and TFH cell activation to elucidate the mechanisms of ITI therapy.

Methods:

Using CRISPR/Cas9 technology, we generated NSG-KDIA-HA mice lacking mouse major histocompatibility (H2) genes and the F8 gene, which were further transduced with AAV-hBLyS for long-term expression of human B lymphocyte stimulator (hBLyS). Genotyping and phenotyping were by DNA sequencing, flow cytometry and ELISA and the one-stage clotting assays. The hBLyS-NSG-KDIA-HA mice were engrafted with peripheral blood mononuclear cells (PBMCs) from 5 FVIII inhibitor-positive PWHA (FVIII-PWHA), and from hemophilic C57BL/6 mice challenged to carry high titer FVIII inhibitors (FVIII-B6-HA). Mice underwent either an ITI protocol (FVIII 200 IU/kg, BIW) mimicking patient conditions or received regular FVIII infusions (maintain groups, FVIII 50 IU/kg, BIW) as controls. Samples were collected during and at experimental endpoints for analysis.

Results: We generated the hBLyS-NSG-KDIA-HA mice and confirmed their loss of MHC molecules, their residual FVIII activity of 3.0 ± 1.7 %, and their expressing hBLyS in plasma (35~110 ng/mL). The historic FVIII inhibitor titers in the 5 FVIII-PWHA were 18.3~147.2 BU, and 215.1 ± 53.0 BU in FVIII-B6-HA mice (n=32). The FVIII-B6-HA’s PBMCs-engrafted BLyS-NSG-KDIA-HA mice (n=12) exhibited 3.2 % B cells (H2K^b+mCD19+) at endpoint. Their FVIII inhibitors were 2.7 ± 2.3 BU on week 8 and decreased to < 0.6 BU (0.5 ± 0.3 BU) in the ITI groups (n=4), in contrast to the 1.2 ± 0.7 BU in the maintain groups (n=4). In the FVIII-PWHA humanized hBLyS-NSG-KDIA-HA mice, we detected 28.0 ± 20.3% patients’ cells (hCD45+, n=20) at week 6 and gradually decreased to 9.0 ± 6.6% at endpoint (n=10). The FVIII inhibitor titers peaked to 1.4 ± 0.9 BU at week 8 (n=18). Of the 5 patients, successful ITI were confirmed in 4, showing titer < 0.6 BU at endpoint, whereas 1 failed in ITI, exhibiting 1.1 ± 0.3 BU/1.8 ± 0.2 BU at week 8/week 20 (endpoint). At endpoint analysis, the FVIII-PWHA-humanized mice carried different B cell subsets, as 56.7% of B cells were mature B cells (hCD27-hCD38+) in the spleen and 51.8% of B cells were plasma cells (hCD27+hCD38+) in the bone marrow.

To investigate the mechanism for ITI, we engrafted the hBLyS-NSG-KDIA-HA mice with FVIII-B6-HA splenocytes. The engraftment rate was 24.9 ± 13.0 % donor cells (H2K^b+) in circulation at week 2 which expanded to 66.6 ± 5.0 % at week 20 (n=7). In the ITI groups, the FVIII inhibitor titers were 16.0 ± 6.2 BU (week 8) decreased to 6.5 ± 2.0 BU (n=3) at endpoint, while in the maintain group, the titers were consistently > 20 BU (n=4). Germinal center-like structure was identified by H&E staining and by immunofluorescence staining of recipient’s spleen, revealing a “white pulp-like” structure with donor T cells recruited at the central arteriole surrounded by B cells. Analysis of the splenocytes showed upregulated PD-L1 in the Treg cells and upregulated PD-1 in the B cells (mCD19+mCD138+) in the ITI group as compared to respective counterparts in the maintain group (n=7, p<0.05).

Conclusion: We have developed a novel humanized ITI testing mouse model that significantly shortens testing duration for FVIII-PWHA. Our findings highlight the potential involvement of the PD-L1/PD-1 axis in ITI mechanisms.