Administration of biologics to enhance T-cell function is part of a rapidly growing field of cancer immunotherapy demonstrated by the unprecedented clinical success of several immunoregulatory receptor targeting antibodies. While these biologic agents confer significant anti-tumor activity through targeted immune response modulation, they can also elicit broad immune responses potentially including the production of anti-drug antibodies (ADAs). DTA-1, an agonist monoclonal antibody against GITR, is a highly effective anti-tumor treatment in preclinical models. We demonstrate that repeated dosing with murinized DTA-1 (mDTA-1) generates ADAs with corresponding reductions in drug exposure and engagement of GITR on circulating CD3+ CD4+ T cells, due to rapid hepatic drug uptake and catabolism. Mice implanted with tumors after induction of preexisting mDTA-1 ADA show no anti-tumor efficacy when given 3 mg/kg mDTA-1, an efficacious dose in naive mice. Nonetheless, increasing mDTA-1 treatment to 30 mg/kg in ADA-positive mice restores mDTA-1 exposure and GITR engagement on circulating CD3+ CD4+ T cells, thereby partially restoring anti-tumor efficacy. Formation of anti-mDTA-1 antibodies and changes in drug exposure and disposition does not occur in GITR−/− mice, consistent with a role for GITR agonism in humoral immunity. Finally, the administration of muDX400, a murinized monoclonal antibody against the checkpoint inhibitor PD-1, dosed alone or combined with mDTA-1 did not result in reduced muDX400 exposure, nor did it change the nature of the anti-mDTA-1 response. This indicates that anti-GITR immunogenicity may not necessarily impact the pharmacology of coadministered monoclonal antibodies, supporting combination immunomodulatory strategies.
Due to their long half-life, high specificity, and excellent therapeutic window, antibodies and antibody-based biologics constitute one of the more successful classes of recently developed products in the biotechnology industry (Beck et al., 2010; Reichert 2011). Significant focus has recently been directed toward immune system modulating biologics, with monoclonal antibodies directed against immune checkpoint receptors including PD-1 and CTLA-4 showing favorable pharmacokinetic (PK) and significant survival benefit in mouse models and clinical studies (Melero et al., 2007; Peggs et al., 2009; Hodi et al., 2010; Robert et al., 2011; Topalian et al., 2012; Hamid et al. 2013; Wolchok et al., 2013; Page et al., 2014).
In addition to targeting immune checkpoint inhibitors, biologics targeting members of the tumor necrosis factor receptor superfamily are emerging as an additional avenue of cancer immunotherapy (Sugamura et al., 2004; Murata et al. 2006; Croft et al., 2013). One of these targets is the GITR, a T-cell costimulatory receptor (Ronchetti et al., 2004; Schaer et al., 2012). It has been shown that GITR can regulate immunity by potentiating CD4+ and CD8+ effector T-cell function while modulating regulatory T cells (Tone et al., 2003; Kanamaru et al., 2004; Kohm et al., 2004; Nocentini and Riccardi, 2005; Nishikawa et al., 2008). In tumor-bearing mice, treatment with the GITR agonist monoclonal antibody murinized DTA-1 (mDTA-1) has been shown to increase intratumoral effector T-cell function and decrease regulatory T-cell stability, culminating in effective tumor regression (Cohen et al., 2006, 2010; Zhou et al., 2007; Coe et al., 2010).
While modulation of costimulatory pathways have shown great preclinical potential as immunotherapies, immune-related adverse events ranging from hyper-costimulation to cytokine dysregulation and immunogenicity have been observed (Chirino et al., 2004; Reuben et al., 2006; Suntharalingam et al., 2006; Brennan et al., 2010; Clarke 2010; Murphy et al., 2014). Ultimately, the administration of any biologic has the potential to generate anti-drug antibodies (ADAs) (Pepinsky et al., 2001; Shankar et al., 2006, 2007; Badylak and Gilbert 2008). Although the degree of immunogenicity against a particular therapeutic is typically based on inherent properties of the molecule being administered (Chirino et al., 2004; De Groot and Moise 2007; De Groot and Martin 2009; Badylak and Gilbert 2008), other factors such a dose, dose regimen, route of administration, and disease status of the host could also play a significant role. The presence of ADAs can dramatically alter the PK and pharmacodynamic properties of administered biologics due to immune-complex formation changes in disposition and in the case of neutralizing antibodies by modulating biologic activity. Furthermore, due to their unique mechanism of action, biologics targeting immunoregulatory receptors may enhance not only specific anti-tumor immune responses, but also lead to nonspecific immunologic activation and the possible emergence of immune disorders (Kohm et al., 2004; Brennan et al., 2010). Therefore, a complete characterization of the biologic responses to potent immunomodulatory drugs following extended exposure is critical to better predicting their therapeutic window. Here, we describe for the first time the dramatic GITR-mediated PK and pharmacological changes that follow repeated treatment with a murinized IgG2a version of the anti-GITR agonist monoclonal antibody DTA-1 (mDTA-1) when dosed alone or in combination with the PD-1 antagonist antibody muDX400.
Materials and Methods
All mouse procedures were performed in accordance with the Institutional Animal Care and Use Committee at Merck, and housed in a facility approved by the American Association for Assessment and Accreditation of Laboratory Animal Care, and maintained under a protocol in accordance with the guidelines of the American Association for Assessment and Accreditation of Laboratory Animal Care (http://www.aaalac.org/).
Mice, Tumor Cell Lines, and Anti-Tumor Efficacy Experiments.
Unless otherwise stated, all dosing was conducted as a s.c. bolus dose given in the upper scruff on the back of the neck. In tumor growth experiments, care was given to dose as far away from the tumor as possible. C57BL/6J mice were obtained from the Jackson Laboratory (Bar Harbor, ME), and BALB/c mice were obtained from Taconic (Hudson, NY). Mice were approximately 8–14 weeks of age with body weights ranging from 20 to 25 g. In the anti-tumor efficacy experiments, BALB/c mice were inoculated s.c. with 3.0 × 105 CT26 murine colon carcinoma cells (ATCC, Manassas, VA) and C57BL/6J mice were inoculated s.c. with 1.0 × 106 MC38 murine colon carcinoma cells (National Cancer Institute, Bethesda, MD). Cells were injected in 100 μl volume. For anti-tumor efficacy experiments, upon achieving a tumor burden of approximately 100 mm3, as gauged by electronic calipers, mice were given a single s.c. injection of 3 or 30 mg/kg mDTA-1 or IgG2a isotype control. Additional cohorts of naive and ADA-positive mice were given three separate s.c. doses of 3 mg/kg mDTA-1 every seven days (Q7D). Tumor growth was monitored twice per week for approximately 3 weeks, at which point the animals were euthanized. Tumor length and width were measured using the electronic caliper and tumor volume determined using the following formula:
mDTA-1 IgG2a (anti-GITR), muDX400 IgG1 D265A (anti-PD-1), and mouse anti-VP2 IgG2a (isotype control) were produced, purified, and characterized by the Protein Sciences Group at Merck & Co. Inc. Fluorescein isothiocyanate anti-mouse CD3, eFluor 450 anti-mouse CD4, and Alexa Fluor-488/Anti-F4/80 were purchased from BD Biosciences (San Jose, CA).
DyLight 650 N-hydroxysuccinimide ester labeling kits (Life Technologies, Carlsbad, CA) were used to conjugate an N-hydroxysuccinimide ester fluorophore (excitation/emission of 646/674) to label the experimental proteins of interest. Prior to labeling, antibodies were dialyzed into 50 mM sodium borate (Life Technologies) utilizing a 10-kDa mol. wt. cutoff Slide-A-Lizer dialysis cassette (Life Technologies). Each reaction contained protein at a final concentration of approximately 2 mg/ml. The conjugation was initiated by combining the protein with the dimethylsulfoxide-dissolved dye and incubating for 1 hour at room temperature. Unreacted DyLight 650 was removed using a Thermo Fisher Scientific Dye Removal Column (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s protocols. If needed, protein samples were concentrated with a 30-kDa mol. wt. cutoff Amicon ultracentrifugation filter unit (EMD Millipore Corp., Billerica, MA) and sterile filtered using a 0.22-μm Millex syringe filter (EMD Millipore). Labeled antibodies were characterized using a NanoDrop UV-visible spectrophotometer (Thermo Fisher Scientific) to determine the protein concentration and degree of labeling as dye/protein (mol/mol) ratio. The integrity of the labeled antibodies was additionally assessed by size exclusion chromatography (SEC) and high-performance liquid chromatography (HPLC), and by extended incubation in phosphate-buffered saline (PBS) and wild-type BALB/c mouse plasma (data not shown). All reagents were analyzed for purity and activity by several methods prior to administration (data not shown). All DyLight 650–labeled antibodies had an average degree of labeling of 2.1 dye molecules per antibody.
Pretreatment of BALB/c Mice.
For generating ADA-positive mice, either two or three doses of indicated antibody were administered with 7 or 14 day intervals. Previous experiments have shown no measurable differences between ADA titers of animals pretreated with 2, 3, or 4 doses of mDTA-1. All antibodies were unlabeled and administered by s.c. injection in either 100 or 200 μl of sterile PBS or appropriate buffer. In PK experiments, mice were dosed twice, 7 days apart, with either 3 mg/kg single agent or 1.5 mg/kg of each agent in combination in 100 μl PBS. For tumor model experiments, mice were dosed three times, 14 days apart, with 3 mg/kg mDTA-1. In all cohorts, there was a 10–20 day washout period prior to experimentation to allow for the removal of unlabeled mDTA-1 or muDX400 (confirmed by enzyme-linked immunosorbent assay).
Blood and Tissue Collection.
Animals were euthanized with carbon dioxide followed by terminal cardiac puncture. Blood was collected into EDTA-K2 CapiJect microcollection tubes (Terumo Medical Corporation, Somerset, NJ) and placed on ice. Plasma was separated from whole blood via centrifugation for 6 minutes at 6000g at 4°C, and stored at −80°C. Likewise, whole blood samples were aliquoted into polypropylene vials (Corning Glassworks, Corning, NY) and stored at −80°C. For tissue lysates, organ samples were collected and immediately placed into 2-ml Precellys Lysing Tubes (Bertin Technologies, Rockville, MD) and placed into dry ice. Upon partial thawing, a 1:10 dilution of Dulbecco’s PBS containing 1% Triton X-100 (MP Biomedicals, Solon, OH) and 1× Halt Protease Inhibitor single-use cocktail (Thermo Fisher Scientific) was added to the samples. Then, tissue slurries were produced using a Precellys Evolution Homogenizer. The slurry was centrifuged (10,000g at 4°C for 10 minutes) and the tissue lysate supernatants were collected and either processed immediately or stored at −80°C until analysis.
Samples were applied onto a BioSep-SEC-S 3000 column with an inline SecurityGuard filter (Phenomenex, Torrance, CA) and analyzed using an Agilent 1200 HPLC system (Agilent Technologies, Santa Clara, CA), utilizing an integrated UV and fluorescence detector (Hamamatsu Corperation, Bridgewater, NJ). The size exclusion protocol used a 15-minute isocratic run with Dulbecco’s PBS buffer (HyClone, GE Life Sciences, Little Chalfont, United Kingdom) as the mobile phase, and was run at a constant flow rate of 1 ml/min−1 and kept at room temperature. The effluent was monitored continuously by optic observation at 280 nm and by the net fluorescence intensity at an excitation wavelength of 646 nm and emission of 674 nm. For mDTA-1 incubation experiments, plasma from naive and mDTA-1–pretreated mice (after a 15-day washout period) was collected as previously described. Defined quantities of mDTA-1 or an IgG2a isotype control (50, 100, or 250 μg/ml) was spiked into 20 μl of neat plasma, incubated for 15 minutes at 37°C, and applied to the BioSep-SEC-S 3000 column. The analysis and data collection were performed using Agilent Technologies’ ChemStation software as well as EXCEL (Microsoft Corporation, Seattle, WA) and GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA). The column performance was assessed prior to and following each batch of samples by running 20 μl of a standard set of molecular weight markers (Bio-Rad Laboratories, Hercules, CA).
Fluorescence Emission–Linked Assay.
Net fluorescence intensity was quantified utilizing a Promega Glowmax plate reader (Promega, Madison, WI) equipped with a fluorescence optical filter with excitation/emission wavelengths of 625/660–720. Concentrations of fluorophore-labeled protein were calculated by generating individual calibration curves in naive BALB/c liver tissue, whole blood, and plasma for each DyLight 650–labeled test article. These calibration curves were analyzed using low fluorescence–background black 96-well polystyrene assay plates (Corning Glassworks), and blank tissue lysates, blood, and plasma were used for background corrections. A linear correlation function was fitted to the data using best-fit parameters (EXCEL). Concentrations of DyLight-labeled test articles were calculated as microgram equivalents per gram of wet tissue. Subsequent tissue/blood ratios were calculated by comparing the quantified concentrations of the test articles in the respective tissue types versus the concentration in blood at the indicated time point. Ratios were calculated using EXCEL.
Capillary Fluorescent Electrophoresis Analysis.
Tissue lysates were analyzed using a Caliper Lifesciences LabChip GX II microfluidic capillary electrophoresis apparatus (Perkin Elmer, Hopkinton, MA) according to the manufacturer’s protocol. Briefly, the tissue lysates (previously diluted 1:10 in tissue lysate buffer) were diluted an additional 2- to 10-fold in the caliper sample buffer, denatured at 75°C for 5 minutes, centrifuged at 2000g for 2 minutes, and then run. Electropherograms were generated by LabChip GXII Touch software (Perkin Elmer).
Tissue Distribution and PK Analysis.
For PK analysis and biodistribution of mDTA-1 and muDX400 antibodies, mice were separately given a single s.c. bolus dose of either 5 or 30 mg/kg DyLight 650–labeled mDTA-1 or muDX400. The animals were sacrificed at the previously indicated time points of 2, 4, 6, 8, 24, and 96 hours (5 mg/kg) or 4, 8, 16, 48, and 96 hours (30 mg/kg). The concentrations of test articles in serum, whole blood, and selected tissues were quantified by a fluorescence emission–linked assay, with lower detection limits of approximately 5 ng/ml for DyLight 650–labeled experimental proteins. The PK parameters were calculated in WinNonlin (Pharsight, Mountain View, CA) using noncompartmental methods on the concentration-time data quantified in the fluorescence emission–linked assay. The concentration-time data for all antibodies were plotted using GraphPad Prism 6 (GraphPad Software, Inc.), and are presented with their corresponding S.D. values.
Fluorescent Analysis of Immune Complex (IC) Uptake in Mouse Kupffer Cells (KCs).
Naive and mDTA-1–pretreated wild-type BALB/c mice (previously inoculated with 3 mg/kg mDTA-1 Q7D × 2) were injected s.c. with 30 mg/kg DyLight-labeled mDTA-1 and KCs were enriched from liver-associated cells as previously described (Zeng et al., 2013; Li et al., 2014). Briefly, 24 hours postdose mice were euthanized and underwent cardiac perfusion with 3 ml × min−1 PBS for 3 minutes. The livers were removed, minced, and collected in 30 ml RPMI media + 10% fetal bovine serum + 0.1% collagenase IV, and subsequently incubated at 37°C for 30 minutes. The KCs were enriched from hepatocytes, nonparenchymal, and liver sinusoidal endothelial cells by gradient centrifugation. Heavy sediments were removed by centrifugation at 20g for 2 minutes and the pellet was discarded. Larger cells were removed by centrifugation at 50g for 3 minutes and the pellet was discarded. Cells were clarified by two alternating rounds of centrifugation of 50g for 3 minutes (discarding the pellet) and 300g for 3 minutes (discarding the supernatant), and always resuspended in 5 ml of RPMI media. On the final spin, cells were resuspended with 500 μl RPMI media and incubated with 5 μg/ml Alexa Fluor 488/anti-F4/F80 antibody for 30 minutes at 4°C. Cells were then plated directly atop microscopy slides and allowed to adhere for 2 hours. Nonadherent cells were removed by three washes with PBS. Cells were then visualized with an EVOS FL Cell Imaging System (Life Technologies) containing both the green fluorescent protein and Cy5 light cubes. Images were acquired and overlaid utilizing the onboard EVOS FL software.
ADA Detection and Free-Drug Detection Assays.
Mice were pretreated with the indicated antibody (3 mg/kg Q7D × 3) and plasma samples were collected as described previously. Six mice were used per time point. Plasma samples were then run on a mesoscale discovery (MSD) bridging immunogenicity assay. MSD plates were coated with 150 μl of Blocker A solution [5 g bovine serum albumin (5%) in 100 ml Dulbecco’s PBS] per well and stored at 4°C overnight (or up to 4 days). A mixture of 1 μg/ml biotin-DTA-1 and 1 μg/ml sulfo-DTA-1 was prepared in assay diluent buffer (0.5% bovine serum albumin, 0.05% Tween 20, 0.25% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate, 5 mM EDTA, 0.35 M NaCl, in PBS, pH7.4); 100 μl of the DTA-1 mix from the previous step was added to each well of the reaction plate (Corning Glassworks, nonbinding surface); 35 μl of the assay diluent buffer was added to each well; and 15 μl of neat sample plasma (or pooled BALB/c mouse serum for background wells) was added to each well and mixed and incubated on a plate shaker at 300 rpm for 2 hours at room temperature. Samples were washed, while MSD plates were blocked 3× with washing buffer (PBS with 0.5% Tween 20). Then, 50 μl of the sample mixture from the previous step was added to duplicate wells on the MSD read plate. The plate was incubated on the shaker at 300 rpm for 1 hour at room temperature. Plates were washed 3× with washing buffer. Then, 150 μl of MSD 1× Read Buffer T was added to each well and read on the MSD instrument (SI-6000 Meso Scale Discovery LLC, Gaithersburg, MD). To determine the concentration of free mDTA-1 or muDX400 in circulation, a drug-capture enzyme-linked immunosorbent assay utilizing a recombinant mouse GITR or PD-1 protein as the capture antigen was used, followed by a horseradish peroxidase–conjugated goat anti-mouse IgG as the detection antibody, adhering to the manufacturer’s protocols.
Flow Cytometry and Fluorescence-Assisted Cell Sorting.
Mouse peripheral blood mononuclear cells (PBMCs) were isolated from naive or mDTA-1–pretreated (ADA-positive) mice 4 hours after a bolus dose of either 3 or 30 mg/kg labeled mDTA-1 or DyLight 650/mIgG2a isotype control. Briefly, whole blood was collected by terminal cardiac puncture, placed into EDTA-treated collection tubes, and placed on ice. Ammonium-chloride-potassium red blood cell lysis buffer was added to whole blood at a 20-fold excess to lyse the RBCs according to the ammonium-chloride-potassium lysing buffer kit (Life Technologies) protocol. Samples were then centrifuged at 300g for 5 minutes at 4°C. The supernatant was aspirated and the isolated PBMCs were washed three times with ice-cold fluorescence-activated cell sorting (FACS) buffer (Sigma-Aldrich, St. Louis, MO).
Cells from mice treated with DyLight 650–labeled mDTA-1 or isotype control were divided into two groups; one group was incubated with anti-CD3 and anti-CD4 (experimental) and the other group was incubated with anti-CD3, anti-CD4, and 5 μg/ml DyLight 650/DTA-1 [maximum median fluorescence intensity (MFI) control]. Cells were then washed three times in ice-cold FACS buffer, followed by FACS and analysis.
To establish the relative maximum and minimum MFIs of mDTA-1 GITR engagement, negative controls received no ex vivo DTA-1 staining (negative control) and (maximum) GITR engagement was accomplished through ex vivo incubation with 5 μg/ml DyLight 650/DTA-1 for 30 minutes. For labeled mDTA-1 or IgG2a isotype control, the percent GITR binding was calculated using the following formula:
DyLight 650–Labeled mDTA-1 Antibody Plasma Concentration Decreases Rapidly over Time in BALB/c Mice Previously Treated with mDTA-1.
Figure 1 shows the dramatic changes in the concentration time profiles of circulating mDTA-1 antibodies when administered to mice previously exposed to the drug relative to naive control mice. By 96 hours following s.c. administration the plasma concentrations in naive mice were 28 and 161 μg/ml (5 and 30 mg/kg doses) versus 0.1 and 1.5 μg/ml, respectively in mDTA-1–pretreated mice. The PK parameters calculated from the concentration time profiles of both cohorts (Table 1) indicated that previous treatment with mDTA-1 resulted in a nonlinear reduction in exposure [area under curve (AUC)]; 24-fold (5mg/kg) and 7-fold (30 mg/kg). Corresponding changes in clearance and half-life were consistent with these reductions in the AUC in pretreated mice, and the PK parameters in naive mice assume the elimination phase kinetics remain constant after the 96-hour collection point.
DTA-1 Forms High Molecular Weight (HMW) Complexes in Plasma from mDTA-1–Pretreated Mice.
To gain better understanding of the increased mDTA-1 elimination in pretreated versus naive mice, mDTA-1 stability and potential protein interactions were characterized by fluorescent SEC-HPLC analysis. In the plasma of naive mice, DyLight 650/mDTA-1 elutes as a single monomeric peak at all time points (Fig. 2A), consistent with intact IgG. No HMW complexes, aggregates, or catabolic products were observed, whereas in pretreated BALB/c mice we observed the presence of HMW complexes with retention times of approximately 6 and 7 minutes corresponding to approximate molecular weights of 1 MDa and 350 kDa, respectively (Fig. 2B). Consistent with the fast elimination of mDTA-1 in pretreated mice (Fig. 1B), the DyLight 650–labeled mDTA-1 chromatographic signal was essentially depleted by 48 hours. The formation of HMW complexes in vivo is indicative of potential drug/anti-drug ICs.
Furthermore, a major shift was observed in the size exclusion profile of labeled mDTA-1 when incubated in vitro in plasma from mice previously exposed to mDTA-1 changing from monomeric IgG to mostly HMW forms (Fig. 3). This is in agreement with the presence of high ADA levels and the rapid formation of drug/anti-drug ICs. As expected, no changes in the chromatographic profiles were observed in naive mouse plasma spiked with labeled mDTA-1 or plasma from pretreated mice spiked with an isotype control DyLight 650–labeled antibody. The relatively low level of mDTA-1 HMW complexes in circulation versus in spiked pretreated plasma (Fig. 2B versus Fig. 3) is consistent with a mechanism for rapid hepatic elimination of ICs (Løvdal et al., 2000; Ganesan et al., 2012).
Hepatic Uptake of DyLight 650–Labeled DTA-1 in Pretreated BALB/c Mice.
Total fluorescence profiles (microgram equivalents per milliliter of tissue lysate) of labeled mDTA-1 after s.c. injection of 30 mg/kg in naive and pretreated BALB/c mice were measured over time in several organs and whole blood. A tissue/blood ratio greater than 1 is indicative of tissue uptake and accumulation. In naive animals tissue/blood ratios remained below 0.5 (Fig. 4A), with no evidence of drug accumulation in any organ analyzed (Fig. 4B). In contrast, mice previously exposed to mDTA-1 experienced intense mDTA-1 organ uptake after dosing. Tissue/blood ratios in the liver, and to a lesser degree in the kidneys, reached their highest values (∼150 and 20, respectively) at 96 hours for the liver and 48 hours for the kidney (Fig. 4C). The total levels of DyLight 650/DTA-1 (μg/g tissue equivalents) were 10-fold higher by 24 hours postdose in the liver of pretreated mice relative to naive animals (150 versus 17.5 μg/g) (Fig. 4, B and D). Overall, these results demonstrate that pretreatment with mDTA-1 dramatically changes the PK and disposition of fluorescent mDTA-1 resulting in dramatic liver uptake, atypical of normal IgGs.
Hepatic Catabolism of DyLight 650/mDTA-1.
To better characterize the dramatic changes in antibody exposure and liver uptake in pretreated mice, tissue lysates of naive and pretreated BALB/c mice dosed with 30 mg/kg fluorescent mDTA-1 were analyzed by capillary fluorescent electrophoresis analysis to determine drug integrity (Fig. 5). The major IgG band at approximately 150 kDa corresponds to intact mDTA-1 antibody and it is the main component in plasma from naive mice (Fig. 5A). Only trace amounts of degradation products were found in any organ analyzed, including the liver (Fig. 5A).
Conversely, consistent with the increased clearance, only low levels of intact mDTA-1 were present in plasma from BALB/c mice previously treated with mDTA-1 (Fig. 5B). In the liver of these mice the majority of the experimental antibody was catabolized as indicated by the abundance of small molecular weight degradation products denoted by the arrows. These results support the presence of circulating ADAs in pretreated mice that upon retreatment form drug/anti-drug ICs. Then, the ICs likely undergo rapid fragment-crystallizable gamma receptor–mediated hepatic clearance and degradation as previously shown by others (Bogers et al., 1991; Løvdal et al., 2000; Ganesan et al., 2012). The efficient removal of potential ICs is in agreement with the low levels of circulating DTA-1 containing HMW complexes in pretreated mice as shown in Fig. 2B.
KC Uptake of DyLight 650/mDTA-1.
KCs are one of the major cell types involved in removal of ICs (Løvdal et al., 2000). To investigate their role in the increased hepatic catabolism of mDTA-1 in pretreated mice, KCs were isolated from livers of naive and mDTA-1–pretreated animals and analyzed 24 hours after DyLight 650/mDTA-1 s.c. administration (Fig. 6). Intense DyLight 650/mDTA-1 staining was prevalent in KCs from pretreated mice (Fig. 6, red area). No mDTA-1 fluorescence signal was observed in KCs from naive mice, suggesting that the increased elimination of mDTA-1 in pretreated mice was due in part to KC uptake of anti-mDTA-1/mDTA-1 ICs. Furthermore, ex vivo staining with an Alexa Fluor-488–conjugated anti-F4/80 antibody, used to identify KCs (Fig. 6, left panels), confirmed that all KCs were also positive for DyLight 650/mDTA-1 signal.
DTA-1 Treatment Elicits a Target-Mediated ADA Response.
To assess the relative levels of ADAs and the potential role of the target (GITR) in anti-mDTA-1 immunogenicity and altered clearance, plasma samples from naive or mDTA-1–pretreated wild-type and GITR−/− C57BL/6 mice were collected at 3, 30, 60, and 90 days following a s.c. mDTA-1 dose of 3mg/kg mDTA-1.
Treatment with mDTA-1 in wild-type mice led to a robust ADA response by day 30 following the most recent mDTA-1 administration and then gradually declined but remained positive through at least day 120. The ADA signal was over 105-fold higher than background levels observed in GITR−/− knockout mice (Fig. 7A). Likewise, there was no detectable mDTA-1 in plasma from wild-type mice (lower limit of quantification of 0.09 μg/ml) after 3 days postdosing, whereas at days 3, 30, 60, and 90 the GITR−/− mice had average mDTA-1 concentrations of 31, 13, 3, and 0.6 μg/ml, respectively (Fig. 7B). These results clearly demonstrate that the agonistic antibody mDTA-1 critically requires the presence of the target (GITR) to induce a specific anti-drug response.
Restoration of DTA-1 Anti-Tumor Efficacy in ADA-Positive Mice.
The impact of ADAs on mDTA-1 anti-tumor efficacy was evaluated in naive and ADA-positive BALB/c (Fig. 8, A and B) and C57BL/6J mice (Fig. 8, C and D) implanted with CT26 and MC38 syngeneic tumors, respectively. Mice with established tumors (average 100 mm3) were dosed s.c. with single or multiple doses of 3 mg/kg mDTA-1 at day 0 or at days 0, 7, and 14, respectively, or with a single 30 mg/kg dose. The average tumor volume was recorded every 3 to 4 days (Fig. 8, A and C) and the final tumor volumes were individually plotted and statistically analyzed (Fig. 8, B and D). In naive BALB/c and C57BL/6J mice, a 3 mg/kg single dose was sufficient to induce tumor regression in both the CT26 and MC38 tumor models. Efficacy was comparable between cohorts treated with a single dose of 3 or 30 mg/kg mDTA-1 or three weekly doses of 3 mg/kg. Conversely, in ADA-positive mice, neither a single nor multiple doses of 3 mg/kg mDTA-1 showed any anti-tumor activity. However, a single 30 mg/kg dose of mDTA-1 did slow tumor progression in mice with pre-existing ADA, particularly in the MC38 model. The same response, although to a lesser degree, was observed in the CT26 model. The presence of ADAs in these mice was confirmed by the MSD assay and SEC-HPLC (data not shown).
Efficient mDTA-1 Target Engagement in ADA-Positive Mice Can Be Achieved with a 10-Fold Increase in Dose.
The presence of ADAs against mDTA-1 dramatically decreased tumoricidal activity. This was likely due in part to the enhanced elimination of the drug upon IC formation. To determine if the ADAs would also neutralize mDTA-1 binding, we evaluated the effects of ADAs on mDTA-1 GITR engagement 4 hours after dosing using flow cytometry and FACS.
A s.c. dose of 3 mg/kg of DyLight 650–labeled mDTA-1 engaged approximately 95% of circulating CD3+ CD4+ GITR cells in naive mice in comparison with only ∼5% in ADA-positive mice (Fig. 9A). Increasing the mDTA dose to 30 mg/kg restored the mDTA-1 target engagement to 91%. The MFI from the negative controls (no antibody) were comparable to the MFI observed from s.c. dosing of an IgG2a isotype control. Purified PBMCs were also visualized for mDTA-1 binding in all groups using a fluorescent microscope (Fig. 9B). Abundant DyLight 650 cellular staining was observed in both PBMCs from naive mice dosed with 3 mg/kg and ADA-positive mice dosed with 30 mg/kg. No fluorescence signal was detected in PBMC from ADA-positive mice dosed with 3 mg/kg of mDTA-1 or with 3 or 30 mg/kg of isotype control antibody (data not shown). Flow cytometry analysis of circulating CD3+ CD4+ T-cell populations in naive and ADA-positive mice (Fig. 9C) demonstrated strong mDTA-1-GITR engagement in naive mice dosed with 3 mg/kg. However, when ADA-positive mice were given the same 3mg/kg DyLight 650/mDTA dose no MFI changes were observed, consistent with decreased concentrations of available mDTA-1 in the circulation due either to fast IC elimination and/or direct ADA neutralization of mDTA-1 binding. Importantly, mDTA-1 target engagement on peripheral blood CD4+ CD3+ T cells was almost restored to naive levels in ADA-positive mice treated with 30 mg/kg.
Effect of Pretreatment in the Disposition of PD-1 Antagonist muDX400 in BALB/c Mice.
Drug combinatorial approaches are rapidly becoming a hallmark of cancer treatment. Therefore, better understanding of potential pharmacological drug-drug interactions is critical to improving rational combination therapy design. To that end, in vivo PK and disposition were used to assess whether pretreatment with mDTA-1 dosed in combination with the PD-1 antagonist antibody muDX400 would further accelerate mDTA-1 disposition and/or immunogenicity (Hirsch et al., 2015; Pinheiro et al., 2015). In addition, we compared the effects of mDTA-1 versus muDX400 pretreatment, dosed alone or in combination, in the PK and disposition of muDX400.
Figure 10 shows muDX400 concentration time profiles in naive, muDX400, and muDX400 + mDTA-1–pretreated BALB/c mice. Fluorescently labeled muDX400 concentration time profiles and PK parameters were indistinguishable irrespective of pretreatment combinations (Table 2). In contrast, there is a dramatic reduction in drug exposure for mDTA-1 in pretreated mDTA-1/muDX400 combo mice. Interestingly, the changes in mDTA-1 PK parameters in mice pretreated with mDTA-1 alone (Table 1) were similar to those pretreated in combination with muDX400. This suggests limited cross reactivity between these two immunoregulatory targets in the generation of an anti-drug response. Likewise, analysis of muDX400 disposition in muDX400-pretreated mice (Fig. 11, A and B) and muDX400 + mDTA-1–pretreated mice (Fig. 11, C and D) revealed that tissue/blood ratios of muDX400 remained below 0.5 in both cohorts at all collected time points (Fig. 11, A and C). This is indicative of negative tissue uptake or accumulation. Furthermore, distribution of muDX400 revealed no evidence of drug accumulation in any organ analyzed (Fig. 11, B and D), demonstrating that muDX400 is highly stable in circulation with a nonspecific organ disposition pattern typical of normal IgGs. In addition, plasma samples from a separate group of BALB/c mice pretreated Q7D × 2 with 3.0 mg/kg mDTA-1 (alone), 1.5 mg mDTA-1 + muDX400, or 3.0 mg/kg muDX400 (alone) were collected 21 days after the second dose, spiked with labeled muDX400, and examined by SEC-HPLC. The resulting chromatographs showed a single monomeric peak corresponding to intact IgG and no detectible HMW complexes (data not shown).
This study characterized the impact of ADAs on the pharmacology of the GITR agonist monoclonal antibody mDTA-1. We compared the PK, disposition, target engagement, and anti-tumor efficacy of mDTA-1 in naive and ADA-positive mice. This study demonstrated that treatment with mDTA-1 in naive wild-type mice generated anti-mDTA-1 antibodies. Subsequent doses of mDTA-1 formed HMW complexes in circulation, which led to hepatic deposition and drug elimination, thereby limiting drug exposure and reducing target engagement and anti-tumor efficacy of mDTA-1. Moreover, we showed that this ADA response was GITR mediated, mDTA-1 specific, and did not alter the PK of a coadministered PD-1 antagonist, muDX400.
Interestingly, the lack of an ADA response against mDTA-1 in GITR−/− mice following repeated treatment strongly supports the hypothesis that direct agonistic engagement of GITR by mDTA-1 is responsible for the immunogenicity of the mDTA-1 molecule in wild-type mice, as opposed to an unusually strong immunogenicity of the molecule itself. An investigation into this target-mediated ADA mechanism might clarify whether mDTA-1 engagement of GITR on T or B cells (Zhou et al., 2010) is potentiating mDTA-1 immunogenicity.
The PK, disposition, and target engagement of biotherapeutics are typically evaluated through the use of either radioactively labeled molecules (Rojas et al., 2005; Kelley et al., 2013), fluorescently labeled molecules (Alvarez et al., 2012), or anti-drug assays (Bourdage et al., 2007; Kelley et al., 2013). In our study of mDTA-1 in naive and ADA-positive mice, we used fluorescent labeling of the mDTA-1 and muDX400 monoclonal antibodies to follow their fate in vivo at relevant pharmacological doses. Direct quantitation of labeled antibodies circumvents potential assay interference issues when assessing drug levels in the presence of ADAs. SEC-HPLC analysis of mDTA-1 from plasma of ADA-positive mice dosed or incubated with fluorescently labeled mDTA-1 revealed the presence of HMW complexes consistent with the presence of anti-mDTA-1/mDTA-1 ICs (Løvdal et al., 2000; Rojas et al., 2005; Ganesan et al., 2012). The relatively low concentration of ICs detected in blood by this method in ADA-positive mice treated with DyLight 650–labeled DTA-1 versus the total conversion to HMW complexes formed by spiking DyLight 650/mDTA-1 ex vivo in plasma from ADA-positive mice suggests that ICs were efficiently removed from circulation. The nonlinear reduction in DTA-1 exposure observed in pretreated animals dosed with 5 versus 30 mg/kg suggests saturation of IC formation or a plateau in the elimination capacity of DTA-1/anti-DTA-1 ICs by the reticuloendothelial system. To ascertain the clearance mechanisms of mDTA-1 ICs in BALB/c mice, we conducted various tissue distribution analyses of DyLight 650/mDTA-1 administered to naive and ADA-positive cohorts. Our results suggested that, following administration to mice with a preexisting ADA response against mDTA-1, the ADAs bound to and formed ICs with mDTA-1. These ICs were then rapidly cleared through hepatic uptake, at least partially by KCs, followed by catabolism and subsequent elimination. While this is consistent with previous studies (Løvdal et al., 2000; Ganesan et al., 2012) it does not preclude the possibility of IC uptake by other liver-associated cells, such as liver sinusoidal endothelial cells, which in our studies showed sporadic DyLight 650/mDTA-1 staining (data not shown). Additionally, the renal uptake observed in ADA-positive mice may be partially due to the excretion of free fluoroprobe in circulation resulting from ADA-related drug processing (Nlend et al., 2014).
We demonstrated that while single- and multiple-dose regimens of 3 mg/kg mDTA-1 provided equivalent anti-tumor efficacy in naive mice, neither was sufficient to achieve anti-tumor efficacy in mice with preexisting ADA. This revealed that mDTA-1 anti-tumor efficacy was ablated by the anti-mDTA-1 antibody response. Furthermore, we showed that in ADA-positive mice, a 3 mg/kg mDTA-1 dose was insufficient to engage GITR on circulating CD3+ CD4+ T cells. This lack of target engagement was likely responsible for the diminished anti-tumor efficacy of mDTA-1 in ADA-positive mice. When the mDTA-1 dose in ADA-positive mice was increased to 30 mg/kg, GITR engagement on circulating CD3+ CD4+ T cells was comparable to that of naive mice dosed with 3 mg/kg. Anti-tumor efficacy was observed, but not to the same degree as in ADA-negative mice dosed with 3 mg/kg mDTA-1. The lack of complete anti-tumor efficacy may be due to limited intratumoral receptor engagement, compared with receptor engagement in circulation. Overall, this supports a correlation between GITR engagement and anti-tumor efficacy, and of critical importance, that restoration of anti-tumor efficacy was possible.
Recently, the development of novel routes of administration, such as intratumoral dosing (Marabelle et al., 2013, 2014), have been evaluated as an option to limit systemic T-cell activation while maintaining intratumoral immune activation and anti-tumor efficacy. Given that the mDTA-1 ADA response can be overcome in vivo, localized GITR activation in the tumor may be sufficient to confer anti-tumor efficacy. Therefore, our results warrant the preclinical exploration of localized treatments with agonistic therapeutics to explore the dynamics of anti-tumor efficacy versus ADA responses.
Given recent trends of cancer immunotherapy toward employing combinations of various immunoregulatory receptor agonists and antagonists (Blank 2014; Lu et al., 2014; Westin et al., 2014), our study sought to characterize whether repeated administration of mDTA-1, an agonist anti-GITR antibody, would enhance ADA to a coadministered anti-PD-1 antibody (muDX400). Our results suggest that while mDTA-1 elicits a rapid anti-drug response, it is specific to mDTA-1, and muDX400 PK and disposition are not impacted, at least in the duration of the study. This does not preclude the possibility of an eventual anti-muDX400 response, but we can infer that treatment with mDTA-1 will not necessarily induce a similarly enhanced ADA response to molecules administered simultaneously.
Of importance in our studies is whether the emergence of ADA responses against GITR agonists is translatable to humans. Anaphylaxis induced by ADAs against mDTA-1 and OX-86 have already been characterized in preclinical and clinical studies (Werier et al., 1991; Abramowicz et al., 1992; Baudouin et al., 2003; Murphy et al., 2014), Although, the manifestations of immunogenicity in preclinical models are generally not considered predictive to humans, our results suggest that close attention to the appearance of ADA responses should be given when administering T-cell agonists. Anti-GITR agonist antibodies are already in the clinic as monotherapy and in combination with anti-PD-1. Ultimately, understanding how the presence of ADAs changes the PK, drug exposure, and clearance of mDTA-1 and muDX400 in mouse models could be helpful for anticipating potential dose adjustments and safety concerns in the clinic, both as monotherapies and in combination.
In conclusion, we showed that exposure to mDTA-1 in mice results in GITR-mediated mDTA-1 immunogenicity leading to the generation of neutralizing ADAs, formation of drug/anti-drug ICs, rapid mDTA-1 clearance, hepatic disposition, and catabolism. Likewise, we observed diminished target engagement and loss of mDTA-1 anti-tumor efficacy in mice with pre-existing ADA. Of critical importance, mDTA-1 anti-tumor efficacy in CT26 and MC38 murine syngeneic tumor models could be partially restored at pharmacological doses that reestablish target engagement on circulating CD3+ CD4+ T cells in ADA-positive mice. We also demonstrated that mDTA-1 did not impact the pharmacology of muDX400 when dosed in combination. Taken together, our results facilitate better understanding of ADA responses against immunomodulatory monoclonal antibodies and the role of target engagement in regulating drug pharmacology.
The authors acknowledge the contributions of Dr. Laurence Fayadat-Dilman and Dr. Brian Reardon at the Department of Protein Sciences for providing the mDTA-1, muDX400, and VP2 antibodies; Priscilla Lapresca for Animal Facility support; Dr. Wolfgang Seghezzi at the Department of Bioanalytics for MSD and enzyme-linked immunosorbent assay support; Dr. Terri McClanahan, Dr. Fernando Ugarte, and Dr. Doug Wilson at the Department of Profiling and Expression for support of flow cytometry facilities and equipment; and Dr. Jennifer Yearley, Dr. Meric Ovacik, and Dr. Mohammad Tabrizifard for valuable discussions.
Participated in research design: Brunn, Beebe, Escandón.
Conducted experiments: Brunn, Mauze, Ueda, Gu, Wiswell.
Contributed new reagents or analytic tools: Brunn, Mauze, Gu, Zhang, Hodges, Escandón.
Performed data analysis: Brunn, Mauze, Gu, Zhang.
Wrote or contributed to the writing of the manuscript: Brunn, Escandón.
- Received October 7, 2015.
- Accepted December 14, 2015.
This work was supported by Merck Research Laboratories (MRL), Merck & Co. Inc. The MRL Postdoctoral Research Program provided financial support to N.D.B.
- anti-drug antibody
- fluorescence-activated cell sorting
- high molecular weight
- high-performance liquid chromatography
- immune complex
- Kupffer cell
- murinized DTA-1
- median fluorescence intensity
- mesoscale discovery
- peripheral blood mononuclear cell
- phosphate-buffered saline
- every 7 days
- size exclusion chromatography
- Copyright © 2016 by The American Society for Pharmacology and Experimental Therapeutics