A Bioluminescence Resonance Energy Transfer-Based Approach for Determining Antibody-Receptor Occupancy In Vivo

摘要

class="head no_bottom_margin" id="sec1title">IntroductionMonoclonal antibodies (mAbs) are often regarded as “magic bullets” (), which have been applied toward the treatment of an array of human diseases (). These therapeutic mAbs are engineered to specifically bind their cognate antigens with high affinities and have been deployed for neutralizing pathologic factors, blocking cellular signaling, and stimulating immune functions (). Therapeutic mAbs have shown great promise in cancer treatments given their therapeutically desirable characteristics of long plasma half-lives, high selectivity, and limited off-target toxicity (). To date, over 30 mAbs (and rising) have been approved for treatment of various types of cancers, including hematologic malignancies and solid tumors (, , , , ).Like other targeted therapies, mAbs can only elicit their desired pharmacological effects when directly bound to their cognate targets. Therefore elucidating the target engagement of a given mAb is a crucial step toward characterizing its therapeutic potential and in determining its pharmacological dynamics, which helps define the optimal dosing regimens to achieve maximal therapeutic efficacy. Target engagement, or receptor occupancy (RO), is the ratio of occupied receptors of interest over total receptors of interest present on the targeted cells. Establishing the RO profile of any therapeutic mAb via preclinical or clinical studies is critical toward projecting the first-in-human dosages, to ensure minimal anticipated biological effect level and minimize potential dose-limiting toxicity (, ; ). Antibody RO is often a valuable intermediate measurement for establishing dose (or exposure)-response relationships, especially at early stages of mAb development when defined biomarkers for an mAb's pharmacological effects are not available (, , ). Although many other factors should be considered when interpreting RO, such as receptor epitope properties (href="#bib31" rid="bib31" class=" bibr popnode">Lipman et al., 2005, href="#bib46" rid="bib46" class=" bibr popnode">Rook et al., 2015), antibody-receptor binding is the first step required to elicit a pharmacological effect, and the binding kinetics of a given mAb to its targets within the tumor microenvironment dictates its general therapeutic potential.Tremendous efforts have been expended toward creating a reliable and cost-effective method to quantify antibody RO. Flow cytometry (FCM), owing to its ease of operation, is routinely used to determine RO (href="#bib58" rid="bib58" class=" bibr popnode">Topalian et al., 2012, href="#bib30" rid="bib30" class=" bibr popnode">Liang et al., 2016); however, FCM is only ideally suited to antibodies that have targets present on circulating blood cells. Moreover, the constraints on sampling accessibility and high spatial heterogeneity often hinder the use of FCM toward antibodies targeting peripheral tissues. Large disparities have been observed between antibody concentrations in circulating plasma and in solid tumors (href="#bib55" rid="bib55" class=" bibr popnode">Suh et al., 2016, href="#bib3" rid="bib3" class=" bibr popnode">Bartelink et al., 2018). Owing to the large sizes, high binding affinities, and high target specificities (href="#bib63" rid="bib63" class=" bibr popnode">Weinstein et al., 1987), the distribution of antibodies in dense interstitial matrix is often limited to the perivascular area, resulting in fractional accessibility of targets to mAbs. In solid tumors, antibody-target binding kinetics and the resultant RO are subject to complex biological variables, including tumor-blood perfusion, antibody extravasation across the tumor vasculature, tumor extracellular matrix densities, and the expression levels and accessibility of antigens on tumor cells that are recognized by mAbs. All these factors complicate reproducibly quantifying antibody-target binding kinetics and the resultant RO in solid tumors.One approach to quantify antibody RO in solid tumors is to perform immunohistochemistry staining on tumor biopsies. However, this approach lacks temporal resolution and often fails to incorporate dynamic factors present in in vivo situations that could greatly influence mAb-target interactions (href="#bib21" rid="bib21" class=" bibr popnode">Gebhart et al., 2016). Another approach to assess antibody RO within solid tumors is to perform radiotracer replacement studies, which usually require two steps: first, giving subjects a small dose of radiolabeled antibody, and then giving increasing doses of unlabeled antibody. Owing to competitive binding, the radioactivity levels in the tumors decrease as doses of unlabeled antibody increase, indicating an increased RO, until a plateau is gradually achieved. Determining mAb RO using this approach is often complicated by the rapid endocytosis of the radiotracers by tumors (href="#bib4" rid="bib4" class=" bibr popnode">Boswell et al., 2010). The estimation of RO is further biased by the unstable radioactivity in the control group, which should have relatively constant radioactivity without competitive replacement by unlabeled antibodies (href="#bib12" rid="bib12" class=" bibr popnode">Cunningham et al., 2005).Other radiolabeling methods, including positron emission tomography/single-photon emission computed tomography, are often applied to quantify mAb pharmacokinetics (PK), tissue distribution, and tissue-specific RO. These approaches raise safety concerns when determining the mAb RO due to elevated radiation accumulation (href="#bib6" rid="bib6" class=" bibr popnode">Burvenich et al., 2018). Fluorescence imaging has also been explored for both preclinical and clinical applications (href="#bib47" rid="bib47" class=" bibr popnode">Rosenthal et al., 2015, href="#bib61" rid="bib61" class=" bibr popnode">Warram et al., 2015, href="#bib48" rid="bib48" class=" bibr popnode">Saccomano et al., 2016, href="#bib28" rid="bib28" class=" bibr popnode">Lamberts et al., 2017, href="#bib17" rid="bib17" class=" bibr popnode">Fornasier et al., 2018, href="#bib35" rid="bib35" class=" bibr popnode">Miller et al., 2018). However, fluorescent imaging suffers from fluorescence quenching that is caused by external excitation light, and poor signal:noise ratios due to the high autofluorescence of biological tissues. Aside from these intrinsic disadvantages, most current non-invasive in vivo imaging methods have a common drawback to RO quantification, namely, they are unable to distinguish signals arising due to specific target engagement versus non-specific background signals. At the tissue level, it is difficult to distinguish the signals of bound mAbs from those of free mAbs present in blood circulating within tissues. Probes or tracers can exhibit non-specific binding and residualization in tumors, which greatly bias RO quantification (href="#bib12" rid="bib12" class=" bibr popnode">Cunningham et al., 2005, href="#bib40" rid="bib40" class=" bibr popnode">Patel and Gibson, 2008, href="#bib39" rid="bib39" class=" bibr popnode">Ogawa et al., 2009). Therefore a non-invasive imaging technology that exclusively enables the visualization of antibody-target interactions in vivo is greatly desired.In the present study, we developed a bioluminescence resonance energy transfer (BRET)-based system to non-invasively quantify antibody RO in live animals. BRET detection schemes are based on Förster resonance energy transfer, in which resonance energy is transmitted from a luciferase molecule (donor) during substrate catalysis to a fluorescent molecule (acceptor), which then re-emits the light according to its own emission spectra (href="#bib13" rid="bib13" class=" bibr popnode">Dragulescu-Andrasi et al., 2011). In BRET-based protein-protein interaction studies, the donor (luciferases) and acceptor (fluorophores) molecules are tagged onto two distinct proteins of interest. Interaction between the proteins of interest, upon appropriate stimuli, brings the luciferase and fluorophore into close proximity, enabling the luciferase to efficiently transmit energy to the fluorophore resulting in BRET (href="#bib34" rid="bib34" class=" bibr popnode">Mandic et al., 2014, href="#bib10" rid="bib10" class=" bibr popnode">Ciruela and Fernandez-Duenas, 2015, href="#bib33" rid="bib33" class=" bibr popnode">Machleidt et al., 2015, href="#bib11" rid="bib11" class=" bibr popnode">Coriano et al., 2016, href="#bib22" rid="bib22" class=" bibr popnode">Goyet et al., 2016, href="#bib2" rid="bib2" class=" bibr popnode">Alcobia et al., 2018, href="#bib41" rid="bib41" class=" bibr popnode">Rathod et al., 2018). BRET efficiency is governed by both the distance and orientation of the donor and acceptor molecules relative to each other. Given the stringent requirements of distance separation (∼10 nm) between donor-acceptor molecules for efficient BRET, it offers a large signal:noise ratio and high sensitivity at physiologically relevant temporal resolutions and therefore has found wide utility in ligand-target interaction studies (href="#bib33" rid="bib33" class=" bibr popnode">Machleidt et al., 2015, href="#bib36" rid="bib36" class=" bibr popnode">Mo and Fu, 2016). Recently, BRET imaging was applied to visualize a propranolol-dye conjugate (acceptor-ligand) binding to an N-terminal NanoLuc (NLuc)-tagged human G-protein-coupled receptor β2-adrenoreceptor (donor-receptor) in real time (href="#bib2" rid="bib2" class=" bibr popnode">Alcobia et al., 2018), demonstrating the noteworthy performance of BRET imaging for monitoring ligand-receptor binding in vivo. In the present study, we extended the BRET approach to clinically attractive mAb therapies. Cetuximab (CTX), a therapeutic mAb currently deployed in many clinical trials for solid tumors, and its cognate target receptor, epidermal growth factor receptor (EGFR), were selected as a model system. EGFR is one of the most well-studied receptors; mediates key growth factor response pathways driving cell survival, proliferation, and growth; and has been implicated (overexpression/mutations) in numerous human malignancies (href="#bib25" rid="bib25" class=" bibr popnode">Han and Lo, 2012, href="#bib51" rid="bib51" class=" bibr popnode">Seshacharyulu et al., 2012, href="#bib9" rid="bib9" class=" bibr popnode">Ceresa and Peterson, 2014, href="#bib54" rid="bib54" class=" bibr popnode">Song et al., 2016, href="#bib32" rid="bib32" class=" bibr popnode">Liu et al., 2017, href="#bib53" rid="bib53" class=" bibr popnode">Sigismund et al., 2018). Herein, we present a BRET-based imaging approach to directly monitor the temporal profiles of antibody-target RO in live animals using CTX binding to EGFR.