TY - JOUR
AU - Lilly, Michael, B
AB - Abstract Castration-resistant prostate cancer cells exhibit continued androgen receptor signaling in spite of low levels of ligand. Current therapies to block androgen receptor signaling act by inhibiting ligand production or binding. We developed bispecific antibodies capable of penetrating cells and binding androgen receptor outside of the ligand-binding domain. Half of the bispecific antibody molecule consists of a single-chain variable fragment of 3E10, an anti-DNA antibody that enters cells. The other half is a single-chain variable fragment version of AR441, an anti-AR antibody. The resulting 3E10-AR441 bispecific antibody enters human LNCaP prostate cells and accumulates in the nucleus. The antibody binds to wild-type, mutant and splice variant androgen receptor. Binding affinity of 3E10-AR441 to androgen receptor (284 nM) was lower than that of the parental AR441 mAb (4.6 nM), but could be improved (45 nM) through alternative placement of the affinity tags, and ordering of the VH and VK domains. The 3E10-AR441 bispecific antibody blocked genomic signaling by wild-type or splice variant androgen receptor in LNCaP cells. It also blocked non-genomic signaling by the wild-type receptor. Furthermore, bispecific antibody inhibited the growth of C4-2 prostate cancer cells under androgen-stimulated conditions. The 3E10-AR441 biAb can enter prostate cancer cells and inhibits androgen receptor function in a ligand-independent manner. It may be an attractive prototype agent for prostate cancer therapy. Introduction Signaling through the androgen receptor (AR) plays a critical role in the progression of advanced prostate cancer (PCa) (Claessens et al., 2008; Karantanos et al., 2013; Hu et al., 2014). Methods to suppress AR signaling are based on blocking the production of ligand (inhibitors of testosterone or dihydrotestosterone (DHT) synthesis) or prevention of ligand binding to the AR (anti-androgens) (Liu et al., 2014; Tan et al., 2015). However the effects of these therapies are transient and eventually cancer growth resumes as castration-resistant prostate cancer (CRPC). Multiple molecular lesions have been identified to account for the maintenance of AR signaling in the absence of exogenous testosterone, typical of CRPC. These include amplification of AR, point mutations in AR that lead to altered response phenotypes to anti-androgens and alternative ligands, expression of autonomously active splice variants or mutants that lack ligand-binding domains (LBDs), and molecular interactions with kinases or other binding partners that lead to ligand-independent receptor activation (Godbole and Njar, 2011; Sadar, 2011; Eisermann et al., 2013; Drake et al., 2014). Some of these variant AR functions are mediated through receptor domains other than the C-terminal LBD. Molecular ligands that target other AR regions such as the N-terminal association domain or the DNA-binding domain (DBD), might offer novel therapeutic approaches to CRPC. Currently, several classes of therapeutic agents that target AR signaling in a ligand-independent manner are under development. Recently, small-molecule inhibitors of the N-terminal domain (NTD) or DBD have been described, some of which have anti-tumor activity in rodent models (Quayle et al., 2007; Myung et al., 2013; Banuelos et al., 2014). Their clinical activity is under study. Global inhibition of AR expression through shRNAs has been the subject of extensive preclinical studies (Sun et al., 2010; Wang et al., 2017), and one such agent is now in clinical trials. However, the transfer of genetic material into cells presents challenges due to poor efficiency, inflammatory reactions and cytotoxic off-target effects. Monoclonal antibodies (mAbs) have been used as therapeutic agents since they have low target-independent toxicity and long half-lives. Moreover, antibodies can be used to deliver a variety of payloads (drugs, toxins, siRNA or radioisotopes) to tumor cells (Jiang et al., 2013; Blakkisrud et al., 2017; Dotan et al., 2017; Wolska-Washer et al., 2017). However, mAbs have traditionally been used to target cell surfaces or extracellular targets since they cannot penetrate the cell membrane. Attempts have been made to deliver antibodies into cells through covalent or non-covalent fusion to polybasic peptides (Lim et al., 2013; Estanqueiro et al., 2015) or lipids (Parhi and Sahoo, 2015; Zalba et al., 2015). Gene transfer techniques have also been used to introduce antibody cDNAs into target cells, which then synthesize ‘intrabodies’ (Maleki et al., 2013; Rinaldi et al., 2013; Cetin et al., 2017). All of these methods have been effective in cultured cells. However, none have been reliably used to deliver antibodies intracellularly following systemic administration to rodents or in clinical settings. Recently, several antibodies have been found to spontaneously enter cells in intact animals, and to target intracellular antigens, but such antibodies remain rare (Lee et al., 2010; Li et al., 2012). We have recently developed a new platform for the intracellular delivery of antibodies and other therapeutic payloads based on the anti-DNA mAb 3E10 (Weisbart et al., 2003, 2004). This antibody was isolated originally as an anti-DNA autoantibody produced by mice with lupus nephritis. In addition to its DNA-binding activity, mAb 3E10 was found to penetrate mammalian cells rapidly (Weisbart et al., 2003). Furthermore, protein payloads coupled to mAb 3E10 were also delivered intracellularly (Weisbart et al., 2003; Hansen et al., 2007a; Zhan et al., 2010). The uptake of mAb 3E10 by mammalian cells was found to be dependent on the nucleotide salvage receptor ENT2, (Weisbart et al., 1990; Zack et al., 1996; Hansen et al., 2007b) and uptake could be completely destroyed by a single amino acid mutation in the kappa light chain domain (Zack et al., 1996; Hansen et al., 2012; Weisbart et al., 2012). These properties are shared by both the parental 3E10 mAb and its scFv equivalent. The 3E10 mAb and its scFv form have been used to deliver protein payloads in rodents, resulting in the expected biologic phenotypes (Zack et al., 1996). We have extended this technology to the design of bispecific antibodies (biAbs) by joining the 3E10 scFv through a linker to another scFv that binds to an intracellular target. In a recent study, we observed that a 3E10-3G5 biAb that targets HDM2 entered cultured melanoma cells to produce the expected biochemical and biologic phenotypes, and inhibited the growth of melanoma xenografts (Weisbart et al., 2012). To develop a cell-penetrating biAb that will inhibit AR function, we have used the 3E10 platform to join scFv 3E10 with another scFv based on the anti-AR mAb AR441 (Libertini et al., 2007; Yang et al., 2007). The scFv version of AR441 is an attractive component of the bispecific molecule for two reasons: First, its epitope (aa 299-315) is located in the NTD of AR. This epitope is not lost from any of the constitutively active AR splice variants, and it is not commonly involved in cancer-associated point mutations. Thus the resulting biAb is likely to be active against many PCa cells. Furthermore, AR441 has been shown to inhibit the growth of hormone-sensitive and -resistant LNCap cells when the antibody is introduced through microinjection (Zegarra-Moro et al., 2002). Here we describe the production and preclinical characterization of the 3E10-AR441 cell-penetrating biAb in PCa models. Results Production of bispecific antibodies The biAbs were produced as recombinant single-chain peptides in yeast, with Myc and His tags either at the C-terminus (3E10-AR441) or between the scFv domains (3E10-MycHis-AR441; Fig. 1a). Purified biAbs were visualized on a Coomassie Blue-stained gel as prominent bands at ~60 KDa while the 3E10scFv domain was observed as band at 30 KDa (Fig. 1b and c). The 3E10-AR441 bispecific antibody was identified in an immunoblot by an anti-Myc antibody. The 3E10-AR441 biAb was used for the majority of the following studies, while the 3E10-MycHis-AR441 biAb has had limited use. Fig. 1 Open in new tabDownload slide Cell-penetrating bispecific antibody design and production. (a) schematic design of biAbs 3E10-AR441 (top) and 3E10-Myc-His-AR441 (bottom); (b) Coomassie Blue stain of 3E10scFv (1) and 3E10-AR441biAb (2); an identical lane of sample 2 was used for an immunoblot probed with anti-Myc antibody, and visualized with enhanced chemiluminescence reagent; (c) Coomassie Blue stain of the biAb 3E10-AR441 (lane 1) and biAb 3E10-MycHis-AR441 (lane 2). Fig. 1 Open in new tabDownload slide Cell-penetrating bispecific antibody design and production. (a) schematic design of biAbs 3E10-AR441 (top) and 3E10-Myc-His-AR441 (bottom); (b) Coomassie Blue stain of 3E10scFv (1) and 3E10-AR441biAb (2); an identical lane of sample 2 was used for an immunoblot probed with anti-Myc antibody, and visualized with enhanced chemiluminescence reagent; (c) Coomassie Blue stain of the biAb 3E10-AR441 (lane 1) and biAb 3E10-MycHis-AR441 (lane 2). BiAb 3E10-AR441 binds to AR in cell lysates Antigen binding by 3E10-AR441 was tested through an immunoprecipitation assay (Fig. 2). Wild-type AR and its splice variants present in LNCaP (Zhu et al., 2003), VCaP (Guo et al., 2009) and 22Rv1 (Guo et al., 2009) were studied, in addition to a CRPC-associated truncated AR(Q640X) (Lapouge et al., 2007) stably expressed in HEK 293 cells. Freshly prepared cell lysates were incubated overnight with 3E10-AR441, control 3E10 scFv, AR441 mAb or media in the presence of protein L-agarose coated beads. Protein L has high-binding affinity for kappa light chains, including domains present in most scFv species. Cell lysates were also run on the gels as positive controls. The AR441 parental mAb and the 3E10-AR441 biAb were both able to bind to wild-type AR (110kd) in LNCap cell lysate while 3E10scFv did not (Fig. 2; far left panel). The AR441 mAb bound to both wild-type AR and AR splice variants (75kd) in VCap cells, while 3E10-AR441 bound strongly to wild-type AR and less strongly to variant AR (Fig. 2, near right panel). The original mAb AR441 and the 3E10-AR441 biAb bound to both wild-type AR and variant AR isoforms (75, 77kd) from 22Rv1 cells as well, although again, the biAb bound more weakly to the splice variant proteins (Fig. 2; near left panel). The binding specificity of mAb AR441 or 3E10-AR441 towards a truncated mutant AR protein, AR(Q640X) (70kd), was generally similar to that seen with the splice variants (Fig. 2; far left panel). 3E10 scFv did not bind to wild-type AR or to variant or mutant AR proteins. Fig. 2 Open in new tabDownload slide Binding of biAbs to AR. Immunoblot of immunoprecipitated AR. Cell lysates containing wild type, variant or mutant AR were immunoprecipitated with AR441 mAb (lane 1; 2 μg/0.5 mg lysate protein), 3E10-AR441 biAb (lane 2; 5 μg/0.5 mg lysate protein) or 3E10scFv control (lane 3; 5 μg/0.5 mg lysate protein). The bound AR proteins (arrows) were compared to those in a corresponding cell lysate (lane 4; no IP). LNCaP lysates were used as source of wild-type AR while 22Rv1 and VCaP lysates were source of wild-type and splice variant AR. HEK/ARQ cell lysates provided the AR(Q640X) mutant found in CRPC. Precipitated proteins were resolved by PAGE and immunoblotted with an anti-AR antibody recognizing the N-terminus. Each immunoprecipitation was repeated at least twice. Fig. 2 Open in new tabDownload slide Binding of biAbs to AR. Immunoblot of immunoprecipitated AR. Cell lysates containing wild type, variant or mutant AR were immunoprecipitated with AR441 mAb (lane 1; 2 μg/0.5 mg lysate protein), 3E10-AR441 biAb (lane 2; 5 μg/0.5 mg lysate protein) or 3E10scFv control (lane 3; 5 μg/0.5 mg lysate protein). The bound AR proteins (arrows) were compared to those in a corresponding cell lysate (lane 4; no IP). LNCaP lysates were used as source of wild-type AR while 22Rv1 and VCaP lysates were source of wild-type and splice variant AR. HEK/ARQ cell lysates provided the AR(Q640X) mutant found in CRPC. Precipitated proteins were resolved by PAGE and immunoblotted with an anti-AR antibody recognizing the N-terminus. Each immunoprecipitation was repeated at least twice. Uptake and nuclear localization of biAb 3E10-AR441 Bispecific antibody uptake and nuclear localization were visualized by confocal microscopy (Fig. 3). LNCaP cells were incubated with 3E10-AR441, 3E10scFv control, or medium only overnight and then examined by immunocytochemistry for the presence of Myc-tagged proteins. Antibodies 3E10-AR441 and 3E10scFv were internalized in LNCap cells and localized in the nucleus. Cells treated with medium only did not show any fluorescence in the FITC channel. These results show that the 3E10-AR441 biAb retains the cell-penetrating activity of 3E10 scFv. We characterized the kinetics of biAb ingress and egress in LNCaP cells by measuring cell-associated biAb by ELISA (Fig. 4). Maximum uptake of biAb or 3E10scFv occurred within 6 h of antibody addition to the medium, with most of the uptake occurring in the first hour. After antibody removal from the medium, much of the cell-associated antibody was rapidly lost. However, a minority of antibody remained stably associated with the cells. The same ingress and egress pattern was detected when 3E10-Myc-His-AR441 biAb was measured in LNCaP cells by ELISA (not shown). The intracellular level of 3E10scFv was higher than that of the biAb. Fig. 3 Open in new tabDownload slide 3E10-AR441 internalization and nuclear localization. LNCap cells grown on polylysine-coated coverslips were incubated with medium only, 3E10scFv (35 μM), or 3E10-AR441 (35 μM) overnight at 37°C. Cells were then fixed and permeabilized for staining with FITC-labeled anti-Myc Ab. Visualization with phase contrast and confocal microscopy. The white line is 30 µ. Fig. 3 Open in new tabDownload slide 3E10-AR441 internalization and nuclear localization. LNCap cells grown on polylysine-coated coverslips were incubated with medium only, 3E10scFv (35 μM), or 3E10-AR441 (35 μM) overnight at 37°C. Cells were then fixed and permeabilized for staining with FITC-labeled anti-Myc Ab. Visualization with phase contrast and confocal microscopy. The white line is 30 µ. Fig. 4 Open in new tabDownload slide 3E10-AR441 uptake and efflux in LNCap cells. Cells were incubated with 3E10-AR441 biAb or 3E10 scFv (6 μM) × 24 h. Antibodies were then removed and fresh (antibody-free) medium was added for an additional 24 h. At intervals, aliquots of cells were lysed and antibodies in cell lysates were measured by ELISA. Data present data from one of two similar experiments. Each point is the mean (±SD) of triplicate determinations. Fig. 4 Open in new tabDownload slide 3E10-AR441 uptake and efflux in LNCap cells. Cells were incubated with 3E10-AR441 biAb or 3E10 scFv (6 μM) × 24 h. Antibodies were then removed and fresh (antibody-free) medium was added for an additional 24 h. At intervals, aliquots of cells were lysed and antibodies in cell lysates were measured by ELISA. Data present data from one of two similar experiments. Each point is the mean (±SD) of triplicate determinations. Binding affinity of BiAb 3E10-AR441 to AR The differences in immunoprecipitation efficiency between AR441 and 3E10-AR441 prompted us to examine the binding affinity of biAbs 3E10-AR441 and 3E10-MycHis-AR441 for wild-type and mutant/variant AR. We employed a competitive ELISA format to estimate the relative binding affinity. The measured KD values are shown in Table I (mean ± standard deviation of triplicate determinations). Examples of competitive ELISA measurements are provided in the supplementary information. Calculated KD values showed that the AR441 parental mAb binds to wild-type AR with very low-nanomolar affinity, typical of many mAbs. The affinity of the parental antibody for wild-type and variant ARs is similar (≤4-fold difference among values). The 3E10-AR441 and 3E10-Myc-His-AR441 biAbs have lower, and more variable binding affinities towards wild-type and variant AR, though in all cases the 3E10-Myc-His-AR441 molecule shows higher affinities. Binding affinity of either biAb to the AR(Q640X) mutant molecule may be higher than for wild-type AR expressed in HEK cells, suggesting that it may be possible to create biAbs specific for mutant target molecules. Table I. Binding affinities of bispecific antibodies for AR proteins AR protein (as cell lysate) . KD (nanomolar) . mAb AR441 . 3E10-AR441 . 3E10-MycHis- AR441 . LNCaP (WT) 5.2 ± 3.3 284.5 ± 76.2 45 ± 4 HEK/AR (WT) 17 ± 16.9 352 ± 117.4 100.3 ± 2.5 HEK/AR(Q649X) 23 ± 7.8 158.5 ± 55.9 42 ± 12.8 HEK/ARv7 19.3 ± 2.9 327 ± 45.4 91 ± 40.7 AR protein (as cell lysate) . KD (nanomolar) . mAb AR441 . 3E10-AR441 . 3E10-MycHis- AR441 . LNCaP (WT) 5.2 ± 3.3 284.5 ± 76.2 45 ± 4 HEK/AR (WT) 17 ± 16.9 352 ± 117.4 100.3 ± 2.5 HEK/AR(Q649X) 23 ± 7.8 158.5 ± 55.9 42 ± 12.8 HEK/ARv7 19.3 ± 2.9 327 ± 45.4 91 ± 40.7 Table I. Binding affinities of bispecific antibodies for AR proteins AR protein (as cell lysate) . KD (nanomolar) . mAb AR441 . 3E10-AR441 . 3E10-MycHis- AR441 . LNCaP (WT) 5.2 ± 3.3 284.5 ± 76.2 45 ± 4 HEK/AR (WT) 17 ± 16.9 352 ± 117.4 100.3 ± 2.5 HEK/AR(Q649X) 23 ± 7.8 158.5 ± 55.9 42 ± 12.8 HEK/ARv7 19.3 ± 2.9 327 ± 45.4 91 ± 40.7 AR protein (as cell lysate) . KD (nanomolar) . mAb AR441 . 3E10-AR441 . 3E10-MycHis- AR441 . LNCaP (WT) 5.2 ± 3.3 284.5 ± 76.2 45 ± 4 HEK/AR (WT) 17 ± 16.9 352 ± 117.4 100.3 ± 2.5 HEK/AR(Q649X) 23 ± 7.8 158.5 ± 55.9 42 ± 12.8 HEK/ARv7 19.3 ± 2.9 327 ± 45.4 91 ± 40.7 BiAB 3E10-AR441 inhibits genomic signaling of AR LNCap/ARELuc cells were used to detect DHT-activated transcription from a consensus androgen response element (ARE)-regulated firefly luciferase gene (Fig. 5a). Cells were pretreated overnight with androgen-deficient media. After 1 h incubation with 3E10-AR441, DHT was added and cells were incubated with both agents for an additional 7 h. Luciferase activity in the cells was then measured. DHT alone increased firefly luciferase activity by 6-fold over that of unstimulated cells. Preincubation of LNCaP/ARELuc cells with the 3E10-AR441 biAb resulted in dose-related inhibition of DHT-induced luciferase activity, with near complete abrogation at 15 μM. The inhibition was significantly greater than that seen with the 3E10scFv antibody at all dose levels, indicating a specific effect on DHT-regulated transcription. Additional cells were treated with enzalutamide as a positive control. This agent inhibited DHT-dependent AR signaling completely at <1 μM concentration (data not shown). Fig. 5 Open in new tabDownload slide Bispecific antibody blocks genomic AR signaling. (a) Artificial AR-dependent promoter. LNCaP/ARELuc cells were pretreated with low-androgen medium overnight (phenol red-free RPMI 1640 with 5% charcoal-stripped fetal calf serum). Test antibodies were added, followed in 1 h by DHT (10 nM). Incubation continued for 7 h more, before measuring luciferase activity. Luciferase activity per well was expressed as the mean ± SEM of triplicates of two independent experiments. Statistically significant values (P < 0.05) between 3E10scFv and 3E10-AR441 are denoted with *. (b) Endogenous promoter. LNCaP cells were pretreated with low-androgen medium overnight. Test antibodies (2.5 μM or 10 μM) were added, followed in 1 h by DHT (10 nM), and incubation was continued for five more hours. PSA (=KLK3) mRNA expression was quantified using real-time PCR, with respect to GAPDH. Gene expression values were expressed as mean ± SEM of triplicates of three independent experiments. Significant values (P < 0.05) are denoted with *. Fig. 5 Open in new tabDownload slide Bispecific antibody blocks genomic AR signaling. (a) Artificial AR-dependent promoter. LNCaP/ARELuc cells were pretreated with low-androgen medium overnight (phenol red-free RPMI 1640 with 5% charcoal-stripped fetal calf serum). Test antibodies were added, followed in 1 h by DHT (10 nM). Incubation continued for 7 h more, before measuring luciferase activity. Luciferase activity per well was expressed as the mean ± SEM of triplicates of two independent experiments. Statistically significant values (P < 0.05) between 3E10scFv and 3E10-AR441 are denoted with *. (b) Endogenous promoter. LNCaP cells were pretreated with low-androgen medium overnight. Test antibodies (2.5 μM or 10 μM) were added, followed in 1 h by DHT (10 nM), and incubation was continued for five more hours. PSA (=KLK3) mRNA expression was quantified using real-time PCR, with respect to GAPDH. Gene expression values were expressed as mean ± SEM of triplicates of three independent experiments. Significant values (P < 0.05) are denoted with *. The inhibitory activity of 3E10-AR441 biAb on genomic AR signaling through an endogenous promoter was tested by quantifying the expression of mRNA for prostate-specific antigen (PSA; gene symbol = KLK3) after DHT treatment of LNCap cells (Fig. 5b). LNCap cells were pretreated overnight with androgen-deficient media. After 1 h incubation with 3E10-AR441, DHT was added and cells were incubated with both agents for an additional 5 h. Cells were then lysed and RNA extracted. Prostate-specific antigen mRNA expression was quantified by real-time PCR. Cells that were treated with 3E10-AR441 biAb (at 2.5 μM and 10 μM) did not show increased PSA mRNA expression after DHT stimulation. However, DHT induced a 20-fold and 15-fold increment in PSA mRNA expression in LNCaP cells treated with 2.5 μM and 10 μM 3E10scFv, respectively. Cells treated only with 3E10scFv control without DHT stimulation had similar PSA mRNA expression levels as the untreated unstimulated control. Because the 3E10-AR441 biAb bound to the autonomously active ARv7 splice variant with a similar affinity as to wild-type AR, we examined its effects on genomic signaling by this receptor. LNCaP/ARELuc cells were transiently transfected with an ARv7 mammalian expression plasmid, then plated in microtiter plates to recover for 24 h. The cells were documented to express the ARv7 protein (Fig. 6a). Following an overnight starvation in low-androgen medium, the 3E10-AR441 biAb or the 3E10scFv negative control antibodies were added to the cells for 24 h. Cells transfected with ARv7 plasmid constitutively increased luciferase expression, from 6-fold to 12-fold higher than that of cells transfected with an empty plasmid (Fig. 6b). The 3E10-AR441 biAb inhibited ARv7-dependent luciferase expression in a dose-dependent manner by up to 62%, whereas the 3E10scFv produced no more than 25% inhibition (Fig. 6c). As expected, enzalutamide did not inhibit ARv7-directed transcription (data not shown). Fig. 6 Open in new tabDownload slide Bispecific antibody blocks genomic signaling by ARv7 splice variant. LNCaP/ARELuc cells were transiently transfected with a mammalian expression plasmid encoding ARv7, or empty plasmid. After 24 h, they were placed into low-androgen medium for 48 h and inhibitors or antibodies were added for the final 24 h. (a) Immunoblot of LNCaP lysate probed sequentially with antibodies to AR and beta-actin (ACTB). Expression of the ARv7 protein is specifically associated with transfection of the ARv7-encoding plasmid. (b) Luciferase activity in LNCaP/ARELuc cells. ARv7 plasmid transfection increases firefly luciferase activity compared with that of cells transfected with empty plasmid. Luciferase activity is the mean (±SEM) of triplicate determinations from two separate experiments. (c) 3E10-AR441 biAb inhibits ARv7-dependent luciferase activity to a significantly greater amount than does 3E10scFv. Luciferase activity is the mean (±SEM) of triplicate determinations from three separate experiments, each performed in duplicate. Significant values (P < 0.05) are denoted with *. Fig. 6 Open in new tabDownload slide Bispecific antibody blocks genomic signaling by ARv7 splice variant. LNCaP/ARELuc cells were transiently transfected with a mammalian expression plasmid encoding ARv7, or empty plasmid. After 24 h, they were placed into low-androgen medium for 48 h and inhibitors or antibodies were added for the final 24 h. (a) Immunoblot of LNCaP lysate probed sequentially with antibodies to AR and beta-actin (ACTB). Expression of the ARv7 protein is specifically associated with transfection of the ARv7-encoding plasmid. (b) Luciferase activity in LNCaP/ARELuc cells. ARv7 plasmid transfection increases firefly luciferase activity compared with that of cells transfected with empty plasmid. Luciferase activity is the mean (±SEM) of triplicate determinations from two separate experiments. (c) 3E10-AR441 biAb inhibits ARv7-dependent luciferase activity to a significantly greater amount than does 3E10scFv. Luciferase activity is the mean (±SEM) of triplicate determinations from three separate experiments, each performed in duplicate. Significant values (P < 0.05) are denoted with *. BiAb 3E10-AR441 inhibits non-genomic signaling by AR Binding of DHT to membrane or cytoplasmic AR activates several parallel signaling pathways, including the release of intracellular calcium (Peterziel et al., 1999; Heinlein and Chang, 2002). The blocking effect of 3E10-AR441 biAb for non-genomic AR signaling pathway in LNCap cells was demonstrated by measuring the amount of intracellular calcium released upon stimulation with DHT (Fig. 7). LNCaP cells treated with the calcium ionophore A23187 showed a rapid increase in calcium5 fluorescence, indicating mobilization of calcium. Treatment of the cells with DHT also triggered an immediate, though less intense fluorescence, indicating non-genomic signaling through the AR. Pretreatment of the cells with biAb 3E10-AR441 produced a dose-related blockade of DHT-induced calcium release, with complete blockade at 4 μM and no effect at 0.5 μM. Pretreatment with 3E10scFv did not block the calcium release, resulting in similar fluorescence intensities as the cells stimulated only with DHT. Fig. 7 Open in new tabDownload slide Bispecific antibody blocks non-genomic AR signaling. LNCap cells were treated with 3E10scFv or 3E10-AR441 (up to 4 μM), plus Calcium5 dye, for 2.5 h at 37°C. Intracellular calcium ion release was triggered by the activation of AR upon addition of DHT (100 nM), by A23187 (100 μM; positive control), or by DMEM (negative control). Calcium5 fluorescence intensity is directly proportional to the amount of intracellular calcium Ca2+ released. Data are from one of three similar experiments. Fig. 7 Open in new tabDownload slide Bispecific antibody blocks non-genomic AR signaling. LNCap cells were treated with 3E10scFv or 3E10-AR441 (up to 4 μM), plus Calcium5 dye, for 2.5 h at 37°C. Intracellular calcium ion release was triggered by the activation of AR upon addition of DHT (100 nM), by A23187 (100 μM; positive control), or by DMEM (negative control). Calcium5 fluorescence intensity is directly proportional to the amount of intracellular calcium Ca2+ released. Data are from one of three similar experiments. BiAb 3E10-AR441 inhibits androgen-dependent growth of C4-2 PCa cells The LNCaP subline C4-2 demonstrates androgen-stimulated growth in response to DHT treatment. C4-2 cells were grown in 24-well plates at low density under androgen-depleted conditions, then treated with DHT plus other growth regulators (enzalutamide, BiAb 3E10-AR441 and 3E10scFv) for 14 days (Fig. 8). Viable cells were quantified by the MTT assay (Gerlier and Thomasset, 1986). DHT alone stimulated growth of C4-2 cells, resulting in higher cell numbers. This effect was substantially blocked by the anti-androgen enzalutamide. The 3E10-AR441 biAb also inhibited DHT-stimulated C4-2 cell growth to a similar degree, which was significant greater than that seen with the 3E10 scFv co-treatment. Fig. 8 Open in new tabDownload slide The 3E10-AR441 biAb inhibits growth of androgen-sensitive C4-2 cells. Cells were plated in 24-well plates at 400 cells/well, in phenol red-free RPMI1640 medium with 5% charcoal-stripped serum. Twenty-four hours later DHT and growth regulators were added at the indicated concentrations. After 7 days, the medium and growth regulators were completely renewed, and incubation was continued. After a total of 14 days treatment time, viable cells were enumerated by the MTT assay. Data from one of two similar experiments. *P < 0.05, **P < 0.01. Fig. 8 Open in new tabDownload slide The 3E10-AR441 biAb inhibits growth of androgen-sensitive C4-2 cells. Cells were plated in 24-well plates at 400 cells/well, in phenol red-free RPMI1640 medium with 5% charcoal-stripped serum. Twenty-four hours later DHT and growth regulators were added at the indicated concentrations. After 7 days, the medium and growth regulators were completely renewed, and incubation was continued. After a total of 14 days treatment time, viable cells were enumerated by the MTT assay. Data from one of two similar experiments. *P < 0.05, **P < 0.01. Discussion Existing therapeutic antibodies have been targeted to extracellular or cell membrane targets, since these large molecules have no facile means to enter cells. We have used a bispecific antibody platform to develop prototype antibodies that will target intracellular molecules for cancer treatment. In the present report, our studies of 3E10-AR441 confirm that the biAb exhibits the hypothesized features of cell penetration and target engagement, resulting in therapeutically useful biochemical and biologic phenotypes. By inhibiting transcriptional activity of the ARv7 splice variant, 3E10-AR441 becomes one of a very small number of compounds that bind to the NTD of the AR. Only one class of agents has been shown to actually inhibit ARv7 transcriptional activity in a ligand-independent manner. The most advanced of these compounds is EPI-001 and its analogs, one of which is now in clinical trials (Myung et al., 2013; Brand et al., 2015). This agent appears to also bind to the N-terminus of the AR, and to block the transcriptional activity of splice variants lacking the LBD. The 3E10-AR441 biAb appears to bind to the wild-type and splice variant AR at nanomolar concentrations, which is achievable in cells. There are no comparable quantitative data on binding of EPI-001 and its analogs to the AR; however, EPI-001 appears to enter cells, and to produce inhibition of AR signaling at extracellular concentrations similar to those where 3E10-AR441 is active. Off-target effects of EPI-001 are now being reported (Brand et al., 2015). It remains to be determined whether the 3E10-AR441 biAb will penetrate cancer cells in vivo. In principle the 3E10 vector motif can enter most cells. We have recently published data that demonstrate a requirement for extracellular DNA for uptake of the 3E10 scFv in cells (Weisbart et al., 2015). This suggests that uptake will be most efficient in tumors that are hypoxic, with spontaneous or treatment-induced cell death. This hypothesis is consistent with our previous data showing uptake of a 3E10-vectored heat shock protein in a rat stroke model was largely confined to the hypoxic tissue (Lee et al., 2010). Thus there may be little loss of the biAb in the normal tissues. Our binding affinity studies suggest that 3E10-AR441 biAb binds to both wild-type and splice variant ARs with similar high affinity. However, the immunoprecipitation experiments consistently showed less binding of the splice variants compared to that of the wild-type protein. These seemingly contradictory observations may reflect conformational changes in the vicinity of the binding site. The direct binding affinity studies utilized isolated native AR protein, or recombinant AR proteins produced in HEK cells, as the binding targets. In contrast the immunoprecipitation studies utilized lysates of PCa cells VCaP or 22Rv1 that continually produce both wild-type and splice variant ARs. The differences may reflect the actions of AR-binding cofactors that could affect AR conformation in the vicinity of the binding site. We note, however, that 3E10-AR441 inhibition of genomic signaling through the wild-type AR is more efficient than through the ARv7 variant receptor. Possibly these alternative receptors form different molecular complexes in LNCaP cells. The antibody ingress/egress studies confirmed quantitatively the uptake of both biAb 3E10-AR441 and 3E10scFv by the target cells. In spite of some rapid loss with antibody washout, a pool of antibody is retained in the cell, and appears stable for at least 24 h. This likely represents the binding of antibody to DNA through the 3E10scFv component, rather than binding to AR through the AR441scFv motif, since persistent levels of 3E10scFv are actually higher than for the biAb. This prolonged binding likely accounts for the biologic features attributable to 3E10scFv. In our studies, there is a slight, but reproducible, inhibition of genomic signaling through either wild-type AR or AR/v7 in 3E10 scFv-treated cells. These findings suggest that DNA-bound 3E10 affects gene transcription. Previous reports using 3E10scFv have indicated that this antibody inhibits both single- and double-stranded DNA repair, resulting in enhanced toxicity to BRCA2-deficient cells or to cells treated concurrently with DNA-damaging agents (Noble et al., 2014). Thus the 3E10scFv may not be a completely ‘negative’ control. Nevertheless, the effects of 3E10-AR441 far exceed those of the monospecific fragment in all assays, confirming that the AR itself is being targeted to cause the response phenotype. While binding affinity of the 3E10-AR441biAb to AR is in the high-nanomolar range, further improvement is necessary. Because of the large molecular weight of the 3E10-AR441 molecule, we may need milligrams of antibody per mouse dose. Improvements in production parameters and binding affinity may result in a reduction in the amount of antibody needed to inhibit AR function. Introducing the AR441 scFv protein into a bispecific scaffold improved binding affinity immensely. Further improvement was achieved by varying the arrangement of the VH and VL regions, and the placement of the affinity tags (Libyh et al., 1997; Cao and Lam, 2003). Further improvement in affinity may result from selection of tight-binding proteins through yeast surface display technology. Another theoretical method to improve binding affinity would be through producing antibodies with multiple copies of the antigen-binding domain. However, this would cause the molecule to be larger (~30 kd for one scFv motif), leading to impaired tissue penetration. Empiric studies will be needed to refine the biAb sequence. Materials and methods Bispecific antibody production, purification and characterization A hybridoma producing mAb AR441 was a gift from Dr Dean Edwards, (Baylor University). Sequences corresponding to AR441 Vk and VH were cloned by reverse transcription-PCR from cDNA with a panel of degenerate primers designed to identify mouse immunoglobulin heavy and light chain variable region genes. The AR441 scFv construct was then produced by overlap extension PCR as previously described (Weisbart et al., 2003, 2004, 2012) and termed as AR441scFv. This cDNA construct was then inserted into plasmid pPICZaA-FV between the EcoR1 and Xba1 sites, placing the AR441 scFv downstream and in frame with the 3E10 scFv sequences, and in frame with C-terminal Myc and His tags (=biAb 3E10-AR441; Fig. 1a). In this construct, the linker between the scFv domains is a 15-amino acid, glycine-rich sequence (GGGGS)3. A variant was also produced (3E10-MycHis-AR441) in which the Myc and His tags were placed between the two scFv domains and the order of the light and heavy chain fragments in AR441scFv were switched to VH and Vk. The linker between the scFv domains in this case is a 19-amino acid motif consisting of CH1 sequences and a swivel sequence as described (Weisbart et al., 2004). The Pichia expression plasmid encoding either biAb was transfected by electroporation into Pichia pastoris X-33 cells. Colonies were selected with Zeocin and identified with anti-his6 antibodies, and a higher-secreting clone was identified as described previously (Weisbart et al., 2004). Stably transfected clonal yeast cells were grown in BMGY, and protein secretion was induced for 72 h at 28°C in 2 l shaker flasks each containing 500 ml BMMY with 0.5% methanol. Secreted protein was concentrated with Centricon Plus 70 centrifugal concentrators (Millipore, Billerica, MA, USA). The recombinant protein was purified by metal chelation chromatography with Ni-NTA-Agarose (Qiagen,), eluted with 500 mM imidazole, and dialyzed against serum-free DMEM tissue culture medium. The purity of recombinant proteins was assessed by PAGE and by immunoblots probed with an anti-Myc (clone 9E10) antibody. Fractions containing bispecific antibody were concentrated and buffer exchanged into DMEM. The 3E10-AR441 or 3E10-Myc-His-AR441 antibodies were then filter-sterilized and the protein concentrations were measured with the BCA assay (ThermoScientific). Antibodies were stored at 4°C for up to 2 months. Cell culture and transfections The LNCap, VCap and 22Rv1 PCa cell lines, and the HEK293 human epithelial kidney cell line, were obtained from the American Type Culture Collection. PCa cells were cultured in RPMI 1640 medium (Cellgro) supplemented with 10% fetal calf serum (Atlanta Biologicals) and 1% penicillin/ streptomycin (Cellgro). The HEK293 cells were grown in Dulbecco’s Modified Eagle’s Medium (Cellgro) with 10% fetal calf serum and 1% penicillin/streptomycin. Wild-type or mutant (ARQ640X) human AR cDNAs cloned in the pCMhAR mammalian expression plasmid were provided by Dr. Kent Nastiuk (University of California, Irvine). HEK293 cells were transfected using Lipofectamine, and stable clones (HEK/AR and HEK/ARQ640X) were selected in medium containing 3 μg/ml G418. Mammalian expression plasmids encoding the AR splice variant ARv7 was kindly provided by Dr. Scott Dehm (University of Minnesota) (Chan et al., 2012). HEK293 were transiently transfected with these plasmids using Lipofectamine. Cell populations with stably or transiently expressed AR variants were characterized by immunoblotting, using an anti-AR antibody (clone PG21; Millipore). Immunoprecipitation Cell lysates of LNCap, VCap, 22Rv1 and HEK/ARQ640X were prepared in RIPA buffer and clarified by high-speed centrifugation and preincubation with mouse IgG-agarose. Five hundred micrograms of total cellular protein from each cell line was immunoprecipitated with mAb AR441 (Santa Cruz Biotechnologies), biAb 3E10-AR441, and 3E10scFv control. We used 2 μg of AR441 or 5 μg of the recombinant antibodies for each immunoprecipitation, to have approximately the same number of antigen-binding sites for each sample. Protein L-agarose beads (Santa Cruz Biotechnologies) were used as the solid phase. Precipitation proceeded overnight at 4°C. The washed agarose pellets were then resuspended in 40 μl of sample buffer under reducing conditions and boiled for 5 min before loading them into a polyacrylamide gel and subjecting them to electrophoresis. After the proteins were separated by electrophoresis, they were transferred onto polyvinylidene difluoride membranes, blocked with milk, and probed with a rabbit polyclonal antibody that targeted the N-terminus of AR (PG21, Millipore). Immunofluorescence Nuclear localization of the 3E10-AR441 biAb was visualized by immunocytochemistry using anti-Myc tag antibody, since the scFv scaffold has a Myc tag. LNCap cells were grown on a polylysine-coated coverslip and incubated with control media, 3E10scFv (35 μM), or 3E10-AR441 (35 μM) for 24 h at 37°C. Cells were gently washed with PBS, then fixed and permeabilized with cold methanol at −20°C for 30 min. The cells were then blocked with 2% BSA and stained with a FITC-labeled anti-Myc tag antibody (clone 71D10) (Cell Signaling) overnight at 4°C. Uptake was visualized using a differential interference contrast confocal microscope equipped with a digital camera (Olympus FV10i laser scanning confocal). 3E10scFv ELISA For quantitation of monospecific or bispecific antibodies containing scFv 3E10 in biologic fluids, we developed a sandwich ELISA assay. Rabbit polyclonal antibodies against purified scFv 3E10 were raised by a commercial service (Pacific Immunology). Polyclonal IgG was purified from post-immunization serum by sequential fractionation with ammonium sulfate precipitation and protein A agarose chromatography (Santa Cruz Biotechnology). An aliquot of the purified immunoglobulin was biotinylated by EZ-link sulfo-NHS-Biotin (Thermo Fisher Scientific) according to the manufacturer’s protocol. The ELISA was formatted by coating wells of a polystyrene 96-well plate with unlabeled anti-scFv3E10 immunoglobulin (10 μg/ml) overnight. The next morning the wells were washed with PBS containing 0.1% Tween 20 and blocked with 1× ELISA blocking buffer (eBiosciences) for 2 h at room temperature. Samples were then added for 1.5 h at 37°C, followed by repeated washing. Biotinylated anti-scFv 3E10 (2 μg/ml) was then added for 1 h at 37°C. Following a final series of washes, the immune complexes were detected using a streptavidin-horseradish peroxidase conjugate (Thermo Fisher Scientific) and TMB chromogen (eBiosciences). A specific preparation of 3E10scFv, quantified by biochemical methods, was used as a reference standard. AR binding affinities A competitive sandwich-type ELISA was used to estimate the relative binding affinities of the various antibodies to AR. The wells of a 96-well polystyrene plate were coated with capture antibody (anti-AR antibody, clone PG21; 2.5 μg/ml) for 2 h at room temperature. Cell lysates containing various AR forms were then added. Both LNCaP and transiently transfected HEK293 cells were used as a source of wild-type human AR. Lysates from a clone of HEK293 cells stably expressing AR Q640X, as well as from HEK293 cells transiently transfected with the AR splice variant ARv7, were used as a source of mutant or variant AR proteins. Cell lysates were prepared in 1× lysis buffer (cat # 9803; Cell Signaling Technology) and their total protein content was measured by the BCA protein concentration assay. Cell lysates were diluted in 1× ELISA sample buffer (eBiosciences) and added to the wells at a protein concentration of 0.1 mg/ml (LNCap lysate) or 0.05 mg/ml (transfected HEK lysate). The cell lysates were incubated with capture antibody for 2 h at 37°C and washed 5× times with 200 μl 0.1% PBST (wash buffer). The AR441 mAb labeled with peroxidase (AR441-HRP, Santa Cruz Biotechnologies) was mixed with increasing concentrations of competitor antibodies (AR441scFv, 3E10-AR441, 3E10-Myc-his-AR441 biAbs, or AR441 parental mAb), then added to the wells. After 90 min incubation at 37°C and thoroughly washing, TMB substrate was added to the wells. The reaction was stopped after 10 min at 37°C and OD450 values were read (Spectramax GEMINI plate reader, Molecular Devices). Optical density values were plotted against competitor concentration. We assumed a single binding site. Relative binding affinity (KD) (Njus et al., 2009) values of the competitor/inhibitor antibody were calculated from the sigmoidal response curve fitting generated using GraphPad v6. Measurement of intracellular calcium LNCaP cells (1.3 × 104/well) were grown in clear-bottom, polylysine-coated black 96-well plates for 24 h. LNCaP cells were pretreated with serum-free, phenol red-free media (DMEM) during 24 h previous to the treatment with 3E10-AR441 biAb, in order to decrease fluorescence background signal. Cells were labeled with calcium five dye diluted in 1× DPBS (Molecular Devices) for 2.5 h at 37°C. Inhibitors or biAbs were added in parallel. The cells were treated with increasing amounts of 3E10-AR441, 3E10scFv, or media for 2.5 h at 37°C. Intracellular calcium release was triggered by the addition of DHT (final 100 nM), ionomycin (positive control; final 10 μM) or DMEM only (negative control). Fluorescence from the wells was measured using a FLIPR Tetra High Throughput Cellular Screening System (Molecular Devices) during a total time of 20 s. The intracellular Ca2+ levels after treatment were expressed as relative total fluorescence when background autofluorescence was subtracted for each well. The fluorescence intensity is directly proportional to the amount of Ca2+ released (Takahashi et al., 1999; Sun et al., 2006; Vicencio et al., 2006). Luciferase reporter assay LNCaP cells were stably transduced with lentivirus encoding a firefly luciferase gene under the control of four tandem copies of a consensus ARE (Open Biosystems) to create LNCap/ARELuc cells. To measure AR-dependent transcriptional activity, LNCap/ARELuc cells were plated in 96-well plates and allowed to grow for 24 h to 80% confluency. The medium was then changed to a medium lacking androgen, RPMI1640 medium with 5% KO serum (Invitrogen) for 18 h. Inhibitors or biAbs were then added for 1 h, followed by DHT (final 10 nM) or vehicle for an additional 7 h. Cells were then lysed, and firefly luciferase activity was measured with a commercial kit (Promega). For measurement of ligand-independent transcriptional activity, LNCaP/ARELuc cells were transiently transfected with a mammalian expression plasmid encoding the constitutively active ARv7 splice variant, or pCDNA3 plasmid as a negative control. Transfections were performed with Mirus Prostate Transfection reagent. Twenty-four hours after transfection, the cells were plated in white 96-well plates and allowed to adhere for 24 h in RPMI1640 + 10% FCS. The next evening the medium was changed to RPMI1640 + 5% KO serum × 18 h. Bispecific antibodies or inhibitors, diluted in RPMI1640 + 5% KO serum were then added and culture was continued for another 24 h. Cells were then lysed and luciferase activity was measured as above. Determination of prostate-specific antigen (KLK3) mRNA levels by real-time PCR LNCaP cells were incubated in RPMI1640 medium containing 5% KO serum overnight to lower AR signaling. Inhibitors or biAbs were added for 1 h, followed by the addition of DHT (final 10 nM) or vehicle for 5 h. The cells were then lysed and the RNA was isolated and transcribed into cDNA (Superscript II, Promega). Relative PSA (KLK3) mRNA expression was quantified using quantitative real-time PCR, relative to that of GAPDH. The primer sequences for GAPDH were sense 5′- ACC ACA GTC CAT GCC ATC AC-3′ and antisenses 5′-TCC ACC ACC CTG TTG CTG TA-3′. KLK3 primer sequences were sense 5′-AGG CCT TCC CTG TAC ACC AA-3′ and antisense 5′-CCT TTA CTG GTC CGG TTC TG-3′. The relative level of expression of PSA (KLK3) mRNA was quantified by using the comparative ΔΔCt method with GAPDH as internal standard (Heinlein and Chang, 2002). Supplementary data Supplementary data are available at Protein Engineering, Design and Selection online. Conflict of interest The authors have no financial conflicts of interest to report. Funding This research work was supported by grant [PC111456] from the Department of Defense/Congressionally-directed Medical Research Program’s Prostate Cancer Research Program, award to M.B.L. and R.W. REFERENCES Banuelos , C.A. , Lal , A., Tien , A.H. et al. ( 2014 ) PLoS ONE , 9 , e107991 . Crossref Search ADS PubMed Blakkisrud , J. , Løndalen , A., Dahle , J., Turner , S., Holte , H., Kolstad , A. and Stokke , C. ( 2017 ) J. Nucl. Med. , 58 , 55 – 61 . Crossref Search ADS PubMed Brand , L.J. , Olson , M.E., Ravindranathan , P. et al. ( 2015 ) Oncotarget , 6 , 3811 – 3824 . Crossref Search ADS PubMed Cao , Y. and Lam , L. ( 2003 ) Adv. Drug. Deliv. Rev. , 55 , 171 – 197 . Crossref Search ADS PubMed Cetin , M. , Evenson , W.E., Gross , G.G., Jalali-Yazdi , F., Krieger , D., Arnold , D., Takahashi , T.T. and Roberts , R.W. ( 2017 ) J. Mol. Biol. , 429 , 562 – 573 . Crossref Search ADS PubMed Chan , S.C. , Li , Y. and Dehm , S.M. ( 2012 ) J. Biol. Chem. , 287 , 19736 – 19749 . Crossref Search ADS PubMed Claessens , F. , Denayer , S., Van Tilborgh , N., Kerkhofs , S., Helsen , C. and Haelens , A. ( 2008 ) Nucl. Recept. Signal , 6 , 1 – 13 . Dotan , E. , Cohen , S.J., Starodub , A.N. et al. ( 2017 ) J. Clin. Oncol. , 35 , 3338 – 3346 . Crossref Search ADS PubMed Drake , C.G. , Sharma , P. and Gerritsen , W. ( 2014 ) Oncogene , 33 , 5053 – 5064 . Crossref Search ADS PubMed Eisermann , K. , Wang , D., Jing , Y., Pascal , L.E. and Wang , Z. ( 2013 ) Transl. Androl. Urol. , 2 , 137 – 147 . PubMed Estanqueiro , M. , Amaral , M.H., Conceição , J. and Sousa Lobo , J.M. ( 2015 ) Colloids Surf. B Biointerfaces , 126 , 631 – 648 . Crossref Search ADS PubMed Gerlier , D. and Thomasset , N. ( 1986 ) J. Immunol. Methods , 94 , 57 – 63 . Crossref Search ADS PubMed Godbole , A.M. and Njar , V.C. ( 2011 ) Prostate Cancer , 2011 , 1 – 13 . Crossref Search ADS Guo , Z. , Yang , X., Sun , F. et al. ( 2009 ) Cancer Res. , 69 , 2305 – 2313 . Crossref Search ADS PubMed Hansen , J.E. , Fischer , L.K., Chan , G. et al. ( 2007 a) Cancer Res. , 67 , 1769 – 1774 . Crossref Search ADS PubMed Hansen , J.E. , Tse , C.M., Chan , G., Heinze , E.R., Nishimura , R.N. and Weisbart , R.H. ( 2007 b) J. Biol. Chem. , 282 , 20790 – 20793 . Crossref Search ADS PubMed Hansen , J.E. , Chan , G., Liu , Y. et al. ( 2012 ) Sci. Transl. Med. , 4 , 157ra142 . Crossref Search ADS PubMed Heinlein , C.A. and Chang , C. ( 2002 ) Mol. Endocrinol. , 16 , 2181 – 2187 . Crossref Search ADS PubMed Hu , D.G. , Hickey , T.E., Irvine , C., Wijayakumara , D.D., Lu , L., Tilley , W.D., Selth , L.A. and Mackenzie , P.I. ( 2014 ) Horm Cancer , 5 , 61 – 71 . Crossref Search ADS PubMed Jiang , H. , Xie , Y., Burnette , A., Roach , J., Giardina , S.L., Hecht , T.T., Creekmore , S.P., Mitra , G. and Zhu , J. ( 2013 ) Appl. Microbiol. Biotechnol. , 97 , 621 – 632 . Crossref Search ADS PubMed Karantanos , T. , Corn , P.G. and Thompson , T.C. ( 2013 ) Oncogene , 32 , 5501 – 5511 . Crossref Search ADS PubMed Lapouge , G. , Erdmann , E., Marcias , G. et al. ( 2007 ) Int. J. Cancer , 121 , 1238 – 1244 . Crossref Search ADS PubMed Lee , W.R. , Jang , J.Y., Kim , J.S., Kwon , M.H. and Kim , Y.S. ( 2010 ) Nucleic Acids Res. , 38 , 1596 – 1609 . Crossref Search ADS PubMed Li , T. , Bourgeois , J.P., Celli , S. et al. ( 2012 ) FASEB J. , 26 , 3969 – 3979 . Crossref Search ADS PubMed Libertini , S.J. , Tepper , C.G., Rodriguez , V., Asmuth , D.M., Kung , H.J. and Mudryj , M. ( 2007 ) Cancer Res. , 67 , 9001 – 9005 . Crossref Search ADS PubMed Libyh , M.T. , Goossens , D., Oudin , S. et al. ( 1997 ) Blood , 90 , 3978 – 3983 . PubMed Lim , K.J. , Sung , B.H., Shin , J.R., Lee , Y.W., Kim , D.J., Yang , K.S. and Kim , S.C. ( 2013 ) PLoS ONE , 8 , e66084 . Crossref Search ADS PubMed Liu , H.H. , Tsai , Y.S., Lai , C.L., Tang , C.H., Lai , C.H., Wu , H.C., Hsieh , J.T. and Yang , C.R. ( 2014 ) Biomedicine (Taipei) , 4 , 2 . Crossref Search ADS PubMed Maleki , L.A. , Baradaran , B., Majidi , J., Mohammadian , M. and Shahneh , F.Z. ( 2013 ) Hum. Antibodies , 22 , 9 – 13 . PubMed Myung , J.K. , Banuelos , C.A., Fernandez , J.G. et al. ( 2013 ) J. Clin. Invest. , 123 , 2948 – 2960 . Crossref Search ADS PubMed Njus , B.H. , Chigaev , A., Waller , A. et al. ( 2009 ) Assay Drug Dev. Technol. , 7 , 507 – 515 . Crossref Search ADS PubMed Noble , P.W. , Young , M.R., Bernatsky , S., Weisbart , R.H. and Hansen , J.E. ( 2014 ) Sci. Rep. , 4 , 5958 . Crossref Search ADS PubMed Parhi , P. and Sahoo , S.K. ( 2015 ) J. Colloid Interface. Sci. , 451 , 198 – 211 . Crossref Search ADS PubMed Peterziel , H. , Mink , S., Schonert , A., Becker , M., Klocker , H. and Cato , A.C. ( 1999 ) Oncogene , 18 , 6322 – 6329 . Crossref Search ADS PubMed Quayle , S.N. , Mawji , N.R., Wang , J. and Sadar , M.D. ( 2007 ) Proc. Natl. Acad. Sci. USA , 104 , 1331 – 1336 . Crossref Search ADS Rinaldi , A.S. , Freund , G., Desplancq , D., Sibler , A.P., Baltzinger , M., Rochel , N., Mély , Y., Didier , P. and Weiss , E. ( 2013 ) Exp. Cell. Res. , 319 , 838 – 849 . Crossref Search ADS PubMed Sadar , M.D. ( 2011 ) Cancer Res. , 71 , 1208 – 1213 . Crossref Search ADS PubMed Sun , Y.H. , Gao , X., Tang , Y.J., Xu , C.L. and Wang , L.H. ( 2006 ) J. Androl. , 27 , 671 – 678 . Crossref Search ADS PubMed Sun , A. , Tang , J., Terranova , P.F., Zhang , X., Thrasher , J.B. and Li , B. ( 2010 ) Int. J. Cancer , 126 , 764 – 774 . Crossref Search ADS PubMed Takahashi , A. , Camacho , P., Lechleiter , J.D. and Herman , B. ( 1999 ) Physiol. Rev. , 79 , 1089 – 1125 . Crossref Search ADS PubMed Tan , M.H. , Li , J., Xu , H.E., Melcher , K. and Yong , E.L. ( 2015 ) Acta Pharmacol. Sin. , 36 , 3 – 23 . Crossref Search ADS PubMed Vicencio , J.M. , Ibarra , C., Estrada , M., Chiong , M., Soto , D., Parra , V., Diaz-Araya , G., Jaimovich , E. and Lavandero , S. ( 2006 ) Endocrinology , 147 , 1386 – 1395 . Crossref Search ADS PubMed Wang , R. , Sun , Y., Li , L. et al. ( 2017 ) Eur. Urol. , 72 , 835 – 844 . Crossref Search ADS PubMed Weisbart , R.H. , Noritake , D.T., Wong , A.L., Chan , G., Kacena , A. and Colburn , K.K. ( 1990 ) J. Immunol. , 144 , 2653 – 2658 . PubMed Weisbart , R.H. , Miller , C.W., Chan , G., Wakelin , R., Ferreri , K. and Koeffler , H.P. ( 2003 ) Cancer Lett. , 195 , 211 – 219 . Crossref Search ADS PubMed Weisbart , R.H. , Wakelin , R., Chan , G., Miller , C.W. and Koeffler , P.H. ( 2004 ) Int. J. Oncol. , 25 , 1113 – 1118 . PubMed Weisbart , R.H. , Gera , J.F., Chan , G., Hansen , J.E., Li , E., Cloninger , C., Levine , A.J. and Nishimura , R.N. ( 2012 ) Mol. Cancer Ther. , 11 , 2169 – 2173 . Crossref Search ADS PubMed Weisbart , R.H. , Chan , G., Jordan , G., Noble , P.W., Liu , Y., Glazer , P.M., Nishimura , R.N. and Hansen , J.E. ( 2015 ) Sci. Rep. , 5 , 12022 . Crossref Search ADS PubMed Wolska-Washer , A. , Robak , P., Smolewski , P. and Robak , T. ( 2017 ) Expert. Opin. Emerg. Drugs. , 22 , 259 – 273 . Crossref Search ADS PubMed Yang , C.S. , Xin , H.W., Kelley , J.B., Spencer , A., Brautigan , D.L. and Paschal , B.M. ( 2007 ) Mol. Cell. Biol. , 27 , 3390 – 3404 . Crossref Search ADS PubMed Zack , D.J. , Stempniak , M., Wong , A.L., Taylor , C. and Weisbart , R.H. ( 1996 ) J. Immunol. , 157 , 2082 – 2088 . PubMed Zalba , S. , Contreras , A.M., Haeri , A., Ten Hagen , T.L., Navarro , I., Koning , G. and Garrido , M.J. ( 2015 ) J. Control. Release , 210 , 26 – 38 . Crossref Search ADS PubMed Zegarra-Moro , O.L. , Schmidt , L.J., Huang , H. and Tindall , D.J. ( 2002 ) Cancer Res. , 62 , 1008 – 1013 . PubMed Zhan , X. , Ander , B.P., Liao , I.H., Hansen , J.E., Kim , C., Clements , D., Weisbart , R.H., Nishimura , R.N. and Sharp , F.R. ( 2010 ) Stroke , 41 , 538 – 543 . Crossref Search ADS PubMed Zhu , Y.S. , Cai , L.Q., You , X., Cordero , J.J., Huang , Y. and Imperato-McGinley , J. ( 2003 ) J. Androl. , 24 , 681 – 687 . Crossref Search ADS PubMed Author notes † " All work was performed in Charleston, SC and Sepulveda, CA Edited by Valerie Daggett © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com
TI - Development of cell-penetrating bispecific antibodies targeting the N-terminal domain of androgen receptor for prostate cancer therapy
JF - Protein Engineering, Design and Selection
DO - 10.1093/protein/gzx058
DA - 2017-12-01
UR - https://www.deepdyve.com/lp/oxford-university-press/development-of-cell-penetrating-bispecific-antibodies-targeting-the-n-gWYgTamRJs
SP - 785
VL - 30
IS - 12
DP - DeepDyve
ER -