Designed trimer-mimetic TNF superfamily ligands on self-assembling nanocages

Abstract

Presentation of an endogenous bioactive ligand in its native form is a key factor in controlling and determining its bioactivity, stability, and therapeutic efficacy. In this study, we developed a novel strategy for presenting trimeric ligands on nanocages by designing, optimizing and testing based on the rational design, high-resolution structural analysis and agonistic activity assays in vitro and in vivo. We successfully designed a nanocage that presents the TNF superfamily member, TRAIL (TNF-related apoptosis-inducing ligand) in its native-like trimeric structure. The native structure of TRAIL complexes was mimicked on the resulting trimeric TRAIL-presenting nanocages (TTPNs) by inserting sufficient spacing, determined from three-dimensional structural models, to provide optimal access to the corresponding receptors. The efficacy of TTPNs as an anti-tumor agent was confirmed in preclinical studies, which revealed up to 330-fold increased affinity, 62.5-fold enhanced apoptotic activity, and improved pharmacokinetic characteristics and stability compared with the monomeric form of TRAIL (mTRAIL). In this latter context, TTPNs exhibited greater than 90% stability over 1 mo, whereas ∼50% of mTRAIL aggregated within 2 d. Consistent with their enhanced stability and ultra-high affinity for the TRAIL receptor, TTPNs effectively induced apoptosis of tumor cells in vivo, leading to effective inhibition of tumor growth. Although TRAIL was used here as a proof-of-concept, all members of the TNF superfamily share the TNF homology domain (THD) and have similar distances between ecto-domain C-termini. Thus, other TNF superfamily ligands could be genetically substituted for the TRAIL ligand on the surface of this biomimetic delivery platform.

Introduction

Members of the tumor necrosis factor (TNF) ligand and receptor superfamily play important roles in hematopoiesis, morphogenesis, and the regulation of immune responses, and there is growing interest in the development of TNF-targeted therapeutics [1,2]. The TNF superfamily is composed of 27 ligands, all of which share a structural feature, an extracellular TNF homology domain (THD), which triggers the formation of non-covalent homotrimers [3]. The endogenous TNF ligand that binds and activates its cognate TNF receptor (TNFR) to induce downstream signaling is a homotrimer; thus, the stability and biological function of TNF ligands is critically dependent on formation of a trimeric structure [3,4].

TRAIL (TNF-related apoptosis-inducing ligand), a member of the TNF superfamily, binds to five members of the TNFR superfamily: TRAIL R1 (death receptor 4 [DR4]), TRAIL R2 (death receptor 5 [DR5]), TRAIL R3 (decoy receptor 1 [DcR1]), TRAIL R4 (decoy receptor 2 [DcR2]), and osteoprotegerin [[5], [6], [7]]. Of these receptors, DR4 and DR5 contain a cytoplasmic ‘death domain’ (DD) and induce cellular apoptosis. DcR1 and DcR2, which have lack or a truncated DD, act as antagonist decoy receptors, blocking the apoptosis signal. Normal cells can evade TRAIL-induced apoptosis by upregulation of antagonist decoy receptors [[8], [9], [10]]. Unlike other apoptosis-inducing ligands (e.g., TNF-ɑ and Fas-ligand), TRAIL has proven to be more effective in selectively inducing apoptosis of tumor cells [[11], [12], [13], [14]]. Preclinical studies have demonstrated that TRAIL agonists show prominent antitumor activity in a variety of tumor types, but little or no activity towards normal cells [15,16]. Because of its tumor-specific apoptotic activity, TRAIL is considered a promising anticancer agent.

Like other members of the TNF superfamily, endogenous TRAIL exists as a homotrimeric complex, which is crucial for its stability, solubility, and bioactivity. Recent studies have reported a number of strategies for preparing recombinant TRAIL trimeric formulations, including FLAG and His tag-mediated crosslinking; linkage to the Fc portion of IgG; fusion of trimerization domains, such as a leucine zipper or isoleucine zipper; conjugation to various nanoparticles (e.g., lipid, polymeric, magnetic, gold, and human serum albumin nanoparticles); and stabilization of trimers with cations. All are designed to improve biological properties such as stability, delivery, and cytotoxic activity and selectivity towards tumors [[17], [18], [19], [20], [21], [22], [23], [24], [25]]. Despite the encouraging results obtained in vitro and in vivo, none of these approaches has yielded an effective anticancer therapy for cancer patients [6,18,[26], [27], [28]]. A number of considerable challenges remain, including differences in the sensitivity of different tumor types, inability to form a stable trimeric structure, hepatotoxicity, poor pharmacokinetic characteristics, insufficient agonistic activity, and low stability in physiologic environments [13,[17], [18], [19],22,23,[29], [30], [31], [32]]. Importantly, many TRAIL-based therapeutic agents in development have been reported to aggregate at high concentrations, consequently showing dose-limiting toxicity in clinical studies [29,30,33,34]. Several TRAIL-targeted agents, particularly His- or Flag-tagged TRAIL, have been shown to aggregate uncontrollably and induce severe apoptosis in hepatocytes [35,36].

Nanocages derived from naturally occurring products have attracted considerable research attention because they can be manipulated for various applications owing to their perfect and complex symmetry, uniform size and shape distribution, biocompatibility, and biodegradability [[37], [38], [39], [40], [41], [42], [43]]. Among such nanocages, human ferritin heavy chains self-assemble to form a constant 24-subunit structure with a spherical cage-like architecture. As described in recent reviews [[44], [45], [46]], these nanocages possess desirable physical characteristics, such as structural stability under non-native conditions, high production levels in Escherichia coli, reversible control of assembly/disassembly in a pH-dependent manner, and the ability to transfer metal cations in and out of the nanocage through eight hydrophilic channels. Importantly, their surface can also be engineered to acquire specificity by incorporation of active proteins or small molecules through simple genetic and chemical modifications. The potential applications of ferritin nanocages in drug and vaccine delivery, diagnosis, and as biomineralization scaffolds and more, have been extensively evaluated over the last 20 years [[44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54]].

Here, we developed a biomimetic delivery platform using ferritin protein nanocages to present a highly stable homotrimer of TRAIL. TRAIL complexes formed trimer-like structures on the surface of ferritin nanocages, as predicted based on the threefold axis of symmetry characteristic of the surface structures of nanocages. This biomimetic delivery platform provides a scaffold for presenting native-like trimeric TRAIL with enhanced affinity, stability, pharmacokinetic characteristics, and excellent apoptotic activity in vitro and in vivo.