General Updates | Ongoing challenges associated with using AAV-based gene therapy in hemophilia

I think the most important point is that hemophilia A has proven to be a more difficult problem for AAV gene therapy than hemophilia B. With factor IX, we have seen relatively stable expression for many years. With factor VIII, expression is variable, tends to decline over time, and liver enzyme elevations are much more common than in hemophilia B. So although there are broader questions about AAV vectors, in hemophilia A, the payload itself may be one of the main limitations. Factor VIII is a very challenging protein for a cell to manufacture. It is large, difficult to fold, and inefficiently secreted. Removing the B domain made AAV delivery possible for factor VIII, but even these factor VIII cassettes strain the vector’s packaging capacity, and the protein still has a major secretion bottleneck. There is also a biological mismatch. Current liver-directed gene therapies ask hepatocytes to produce factor VIII, whereas its principal natural cellular source is liver sinusoidal endothelial cells. If hepatocytes are pushed to produce large amounts of this difficult protein, factor VIII can accumulate in the endoplasmic reticulum and activate cellular stress and inflammatory pathways. Preclinical data suggests that this may contribute to declining expression or translational shutdown. Clinically, higher early factor VIII levels have often been followed by a steeper decline, while lower vector doses have generally been associated with less liver toxicity. That has led to an important design principle. For hepatocyte-directed factor VIII gene therapy, lower may be better. The most immediate strategy is therefore not simply to give more vector. We learned that all too well. It is to build a better factor VIII molecule so that each vector genome and each molecule of protein does more useful work, basically.

And one example is the V3 factor VIII design, which restores several native glycosylation sites to improve folding and secretion. Early clinical experience suggests improved durability, although it has not eliminated transaminitis. That implies that declining expression and liver enzyme elevations may have partly distinct mechanisms. And another interesting gain-of-function factor VIII was recently developed through comparative biology. The investigators introduced 36 amino acid substitutions found across factor VIII proteins from dogs, pigs, cattle, and sheep, which happened to perform much better hemostatically. And the resulting molecule was secreted more efficiently, accumulated less inside cells, and produced less evidence of an unfolded protein response than conventional B-domain-deleted factor VIII. It also had substantially greater functional activity relative to the amount of protein present. In non-human primates up to approximately 15-fold, depending on the assay.

Many more such gain-of-function variants are in development, applying similar principles, and the clinical hope is that we could normalize hemostasis with much less vector and much less factor VIII production, thereby reducing the burden on hepatocytes. However, protein engineering introduces its own safety questions. If we make factor VIII too active or too resistant to normal inactivation, which is one of the approaches, we may increase thrombotic risk. If we introduce too many non-human sequences, we may breach immune tolerance and cause factor VIII inhibitors despite the otherwise tolerogenic environment that the liver provides. We have already seen a warning signal. A heavily engineered human-porcine, human-pig factor VIII program was stopped after the first two treated individuals developed high-titer inhibitors toward factor VIII, so more engineering is not automatically better; there is a fine line there, and every candidate must be evaluated carefully for immunogenicity, thrombogenicity, cellular stress assay behavior, and long-term durability. Ultimately, the goal is not to make hepatocytes produce the maximum possible amount of factor VIII. That’s the bottom line, really. It is to achieve reliable physiologic hemostasis with the lowest feasible vector dose and the smallest possible protein production burden without liver injury, inhibitors, or thrombosis. Improving the factory transgene itself may be the key to unlocking the next generation of gene therapy.

This transcript is AI-generated. While we strive for accuracy, please verify this copy with the video.