Membrane fusion FerA domains enhance adeno-associated virus vector transduction

Highlights

• Diverse FerA domains augment AAV transduction in AAV serotypes-independent manner.

• FerA domains directly interact with AAV virion.

• FerA domains fuel AAV binding, trafficking, and traversing ability.

• FerA domains slow down the blood clearance of AAV.

• ScAAV8/hFIX-FerA injection improves hemostasis in FIX-/- hemophilia B mice.

Abstract

The recombinant adeno-associated virus (rAAV) vector has been successfully employed in clinical trials for patients with blindness and bleeding diseases as well as neuromuscular disorders. To date, it remains a major challenge to achieve higher transduction efficiency with a lower dose of rAAV vector. Our previous studies have demonstrated that serum proteins are able to directly interact with AAV virions for transduction enhancement. Herein, we explored the effect of the FerA domains, which are derived from ferlin proteins and possess membrane-fusion activity, on AAV transduction. Our results show that FerA domains from dysferlin, myoferlin, and otoferlin proteins are able to directly interact with AAV vectors and enhance AAV transduction in vitro and in mice through either intravenous or intramuscular injections. The enhanced AAV transduction induced by human/mouse FerA domains is achieved in various cell lines and in mice regardless of AAV serotypes. Mechanism studies demonstrated that the FerA domains could effectively enhance the ability of AAV vectors to bind to target cells and cross the vascular barrier. Additionally, FerA domains slow down the blood clearance of AAV. Systemic administration of AAV8/hFIX-FerA complex induced approximate 4-fold more human coagulation factor IX expression and improved hemostasis in hemophilia B mice than that of AAV8/hFIX. Collectively, we show, for the first time, that multiple FerA domains could be tethered on the AAV capsid and enhance widespread tissue distribution in an AAV serotypes-independent manner. This approach therefore holds a promise for future clinical application.

Introduction

Recombinant adeno-associated virus (rAAV) vector has been a leading vehicle for clinical application of gene therapies in view of its safety, broad tissue tropism, and sustained effectiveness [1,2]. Currently, 13 serotypes and numerous AAV variants and mutants have been isolated and studied as gene delivery vehicles [3]. Several AAV serotypes, such as AAV2, AAV8, and AAV9, have been extensively employed in clinical trials and achieved therapeutic effects [[4], [5], [6]]. Nevertheless, emerging concerns about the rAAV limited transduction efficacy and the high vector dose requirement remain crucial barriers for ongoing AAV-based gene therapy in preclinical and clinical settings [7,8]. Data from AAV-human factor IX (hFIX)-related clinical trials in hemophilia B patients have shown that high vector dose directly correlates with the cellular immune responses to vector capsid [9,10]. Meanwhile, several studies have demonstrated that AAV capsid or transgene-specific cytotoxic T lymphocytes (CTLs) response and late innate immune activation after long-term AAV transduction hinder the AAV vector transduction efficiency [[10], [11], [12]], which could be strengthened under the condition of high dose of vehicles. In this context, tremendous efforts have been directed to improve the AAV vector transduction efficacy while simultaneously decreasing the vector dose. Several approaches, including but not limited to direct evolution, capsid/genome engineering and polyploid AAV capsids modification, have been undertaken in an effort to produce and impel AAV variants into desired tissues or cell types and achieve therapeutic efficacy [[13], [14], [15], [16], [17]]. Meanwhile, numerous studies with cell surface targeting of AAV vectors by genetic modification of the capsid proteins VP1, VP2, or VP3 have been extensively explored to improve safety and efficacy in preclinical/clinical application [[18], [19], [20]]. Additionally, covalent coupling of targeting ligands to intact AAV particles and magnetically guided AAV delivery system have been also exploited to achieve selective gene transfer in distinct cell types, even in non-permissive cell types [21,22]. However, the achieved results being implemented in specific mouse strains and other small-animal models cannot always be extrapolated to that of other species such as non-human primates (NHPs) [[23], [24], [25], [26]]. One study has shown that the therapeutic efficacy for liver transduction in large-animal models compared with mouse models is approximately 50-100-fold less efficient [27]. Notably, engineered modification of the AAV capsid may lead to unpredicted structural and tropism changes [28]. Although those risks could be mitigated or eliminated by direct evolution approaches [17], the most mutants developed from direct evolution were isolated from cell lines in vitro or in animal models and have unknown human tropism. Therefore, further exploration of safer and more effective strategies to improve AAV transduction at a lower vector dose, especially in a manner that induces consistent transgene expression without limitation of species and alteration of tissue tropism, is still of critical clinical importance.

It has been demonstrated by multiple studies that serum proteins effectively bind to AAV and affect transduction efficacy via different mechanisms in a serotypes- and species-specific manner [[29], [30], [31], [32], [33], [34], [35]]. For instance, the mouse-derived but not human-derived C-reactive protein can bind to the AAV6 virions [29]. However, both of them did not appreciably interact with either AAV8 or AAV9 [35]. Our recent studies have demonstrated that several human serum proteins are able to directly interact with the specific AAV capsid and increase its transduction [[32], [33], [34]]. Pei et al. found that the low-density lipoprotein and transferrin could be appropriated by AAV8 virions to enhance the liver transduction by increasing AAV's ability to bind to target cells, although the enhancement is only limited to hepatocytes after systemic administration [33]. Chai et al. has demonstrated that serum proteins could directly interact with AAV9 and further enhance AAV9 vascular permeability for global transduction [34]. Nevertheless, it is worthy to note that the effect of serum proteins may have limited application only to specific AAV serotypes [29,30,[32], [33], [34], [35]]. In light of the described events, it will undoubtedly require the exploration of novel proteins with the characterization of generally working with multiple AAV serotypes to achieve broader AAV preclinical/clinical application.

Ferlins are a family of type II transmembrane proteins and are characterized by the multiple tandem C2 domains (Ca2+-regulated, phospholipid-binding domains), a centrally positioned FerA domain, and anchored by a C-terminal transmembrane domain [36,37]. Ferlins are implicated in membrane repair and membrane fusion processes. To date, there are six Fer-1-like genes that form the ferlin family: dysferlin (Fer1L1), otoferlin (Fer1L2), and myoferlin (Fer1L3), as well as three additional yet not well-characterized members Fer1L4, Fer1L5, and Fer1L6 in humans and most mammals [38]. A recent study found that the α-helical ~15-kDa FerA domain possesses Ca²⁺-dependent membrane-associating activity, and may contribute to the overall membrane fusion activity of ferlin proteins [39]. If the isolated FerA domain can help mediate vesicle fusion, can it then be used to improve the AAV transduction process?

In the present study, we demonstrated for the first time that the isolated FerA domains from the dysferlin, myoferlin, and otoferlin proteins could generally enhance the AAV transduction in vitro and in vivo in a species- and AAV serotypes-independent manner, thus overcoming some of the limitations of the previous AAV adjuvants. Underlying mechanism studies suggest that FerA domains directly bind to AAV, promote the AAV binding activity, delay the blood clearance, and enhance the transcytosis ability of AAV. We applied this approach for treating FIX-deficient hemophilic mice. After systemic administration of self-complementary (sc) AAV8/hFIX vector preincubated with FerA domain, enhanced transgene hFIX expression and hemostasis improvement were achieved.