Recent success in AAV liver gene therapy for hemophilia A (HA) and hemophilia B (HB) has demonstrated that the lower the vector dose, the better the safety profile. Strategies to enhance vector performance currently being investigated in clinical trials include the use of gain of function clotting factor variants such as Factor IX (FIX) Padua (R338L), a hyperactive variant with an activity to antigen ratio ~8 fold higher than FIX wild type (WT). However, early studies in HB mice suggested that immune tolerance induction following AAV liver gene therapy required FIX antigen levels above a certain threshold (Mingozzi et al . JCI 2003). These results raise the concern that lower antigen levels with gain of function FIX variants could increase the risk of an anti-FIX immune response, including inhibitor development. Inhibitor development is associated with a substantial increase in mortality and morbidity and remains a major safety concern for any new HB therapeutic.

To evaluate this potential risk in a large animal model, we have undertaken a study in adult dogs with severe HB. Two canine (c) models with different underlying F9 gene mutations (missense (MS) and null) and distinct risks of inhibitor development received AAV liver- or skeletal muscle-mediated expression of cFIX Padua. A naïve MS HB dog was treated with a sub-therapeutic dose of AAV-cFIX-Padua for skeletal muscle gene therapy to investigate the role of antigen level in transgene immune tolerance. This strategy resulted in cFIX activity level of only ~1% and undetectable cFIX antigen levels (assay sensitivity ~0.3%, 15 ng/mL). However, even this minute antigen level was sufficient to maintain immune tolerance despite several treatments with cFIX-WT protein for bleeding. In addition, we have previously treated 3 MS HB dogs from the same colony with higher vector doses and achieved cFIX activity levels of 3.5-8% and antigen levels of 0.4-1.5%. No FIX antibody was detected in these animals, even after further challenges with cFIX-WT protein.

We have reported that liver-directed AAV-cFIX-Padua gene therapy can eradicate a pre-existing anti-cFIX immune response and induce immune tolerance in an inhibitor-prone null HB dog (Wiley) (Crudele et al . Blood 2015). This immune tolerance has been maintained despite multiple challenges with recombinant cFIX protein. Here we report eradication of a pre-existing anti-cFIX immune response in a second inhibitor-prone HB dog (Otis) with similar gene therapy. Both dogs had detectable cFIX antigen levels within 1 week of vector delivery (1-1.5%), but low activity to antigen ratios for cFIX Padua (<1), consistent with circulating immune complexes of FIX and antibody, as is observed in HB inhibitor patients (Nilsson et al. PNAS 1986). Wiley's activity to antigen ratio only normalized after inhibitor eradication. Interestingly, Otis was challenged with recombinant cFIX-WT protein after his inhibitor titer was ≤0.6 BU and his FIX-specific IgG1 and IgG2 were baseline, but his cFIX activity to antigen ratio was still <1. This resulted in an anamnestic response with substantial increases in the cFIX-IgG1, but not IgG2 (equivalent to human IgG4). Importantly, continued expression of cFIX Padua ultimately resulted in the disappearance of the cFIX-IgG1, as well as the normalization of his cFIX Padua activity to antigen ratio. Neither Wiley nor Otis has bled since gene therapy. The experience of Otis suggests that decreased activity to antigen ratio is a marker for a continued anti-FIX immune response, even in the absence of other laboratory findings, likely due to immune complexes. After eradication, these dogs expressed cFIX-Padua antigen at ~10% with the expected activity to antigen ratio of ~8 for ≥3.5 years.

Combined these results support the concept that the continuous uninterrupted transgene expression after gene therapy, even at undetectable antigen levels, can prevent and/or eradicate inhibitors. These data are in agreement with the results of the International Immune Tolerance Induction trial for HA patients with inhibitors, where the rates of inhibitor eradication at a 3 year period were similar between high- and low-dose cohorts (Hay et al . Blood 2012). Overall, this work supports the emerging understanding that gene therapy utilizing gain of function variants for inherited protein deficiencies can simultaneously promote immune tolerance and provide therapeutic activity levels.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.