Correlating RANK ligand/RANK binding kinetics with osteoclast formation and function

Abstract

The interaction between Receptor Activator of NF-κB Ligand (RANKL) and its receptor RANK is essential for the differentiation and bone resorbing capacity of the osteoclast. Osteoprotegerin (OPG), a soluble homodimer, acts as a decoy receptor for RANKL and thus inhibits osteoclastogenesis. An imbalance in the RANKL/RANK/OPG axis, with decreased OPG and/or increased RANKL, is associated with diseases that favor bone loss, including osteoporosis. Recently, we established a yeast surface display system and screened libraries of randomly mutated RANKL proteins to identify mutations that abolish binding to OPG while preserving recognition of RANK. These efforts yielded several RANKL variants possessing substantially higher affinity for RANK compared to their wild-type (WT) counterpart. Using recombinant RANKL mutant proteins, we find those with increased affinity for RANK produce more robust signaling in osteoclast lineage cells and have greater osteoclastogenic potential. Our results are the first to document gain of function RANKL mutations. They indicate that the physiological RANKL/RANK interaction is not optimized for maximal signaling and function, perhaps reflecting the need to maintain receptor specificity within the tumor necrosis factor superfamily (TNFSF). Instead, we find, a biphasic relationship exists between RANKL/RANK affinity and osteoclastogenic capacity. In our panel of RANKL variants, this relationship is driven entirely by manipulation of the kinetic off-rate. Our structure-based and yeast surface display-derived insights into manipulating this critical signaling axis may aid in the design of novel anti-resorptive therapies as well as provide a paradigm for design of other receptor-specific TNF superfamily ligand variants.

Keywords: BONE; KINETICS; OSTEOCLAST; RANK; RANKL; RECEPTOR; SIGNAL TRANSDUCTION; TNF-SUPERFAMILY.

Figure 1

Osteoclastogenic potential of RANKL variants. (A) The capacity of mutant RANKL proteins to generate osteoclasts from bone marrow macrophages was assessed by TRAP stain using increasing amounts of purified RANKL. (B) Titration curves of osteoclastogenesis were fit using a four-parameter dose-response curve. (C) EC₅₀ values calculated from the curve fits in (B). EC₅₀values representing the ability of each RANKL variant to generate osteoclasts in vitro is plotted against either the binding constant K_D (D) or the kinetic half-life t_1/2 (E). Assessment of osteoclastogenic markers by real-time detection of NFATc1 (F) or β3 integrin (G) mRNA levels. A RANKL concentration close to the EC₅₀ value of WT-RANKL (5 ng/mL) was used for all variants.

Figure 2

Ability of monomeric OPG to inhibit RANKL-induced osteoclast formation. (A) Varying amounts of monomeric OPG were added to cultures containing 100 ng/mL of each RANKL variants and osteoclasts were stained for TRAP activity. (B) Quantitation of TRAP activity in the presence or absence of OPG.

Figure 3

Signaling activation by RANKL variants. (A) Bone marrow macrophages were serum starved and stimulated with 100 ng/mL WT, Q236H, or KQFH RANKL. Phosphorylation of NF-kB and p38 were assessed by western blot. (B) Densitometry of western blots depicted in (A).

Figure 4

Activation of mature osteoclasts by RANKL variants. After 4 days in culture with WT RANKL, osteoclasts were plated in equal numbers on bovine bone slices and stimulated with WT, Q236H, or KQFH. The release of collagen fragments (CTx) was measured after 24 hours.