Discovery and preclinical characterization of the antagonist anti-PD-L1 monoclonal antibody LY3300054

Discovery and preclinical characterization of the antagonist anti-PD-L1 monoclonal antibody... Background: Modulation of the PD-1/PD-L1 axis through antagonist antibodies that block either receptor or ligand has been shown to reinvigorate the function of tumor-specific T cells and unleash potent anti-tumor immunity, leading to durable objective responses in a subset of patients across multiple tumor types. Results: Here we describe the discovery and preclinical characterization of LY3300054, a fully human IgG1λ monoclonal antibody that binds to human PD-L1 with high affinity and inhibits interactions of PD-L1 with its two cognate receptors PD-1 and CD80. The functional activity of LY3300054 on primary human T cells is evaluated using a series of in vitro T cell functional assays and in vivo models using human-immune reconstituted mice. LY3300054 is shown to induce primary T cell activation in vitro, increase T cell activation in combination with anti-CTLA4 antibody, and to potently enhance anti-tumor alloreactivity in several xenograft mouse tumor models with reconstituted human immune cells. High-content molecular analysis of tumor and peripheral tissues from animals treated with LY3300054 reveals distinct adaptive immune activation signatures, and also previously not described modulation of innate immune pathways. Conclusions: LY3300054 is currently being evaluated in phase I clinical trials for oncology indications. Background inhibitory (i.e. “checkpoint”) receptors are CTLA-4 T cell activation occurs when T-cells receive two positive (CD152) and PD-1 (CD279), and the regulatory approval signals from antigen-presenting cells (APC): an antigen- of agents that target CTLA-4 (ipilimumab, Yervoy™), specific signal presented in the context of major histo- and PD-1 (nivolumab (Opdtivo™), pembrolizumab (Key- compatibility complex (MHC) which engages the T-cell truda™), has been key to bringing forth the modern era receptor (TCR), and a co-stimulatory signal from B7–1/ of immunotherapy. B7–2 (CD80/CD86) to the CD28 receptor on T-cells [1]. Two ligands have been described for PD-1: PD-L1 Initial T cell activation is followed by the surface expres- ((B7-H1, CD274), and PD-L2 (B7DC, CD273). While sion of a set of co-activating receptors such as CD137, baseline expression of PD-L2 is relatively limited to sub- OX40, GITR, and CD27 which enhance T-cell function, sets of dendritic cells, macrophages, B cells, mast cells and a set of T-cell inhibitory receptors which initiate in- and Th2 cells and tumor cells [3], expression of PD-L1 hibitory pathways that function to prevent uncontrolled is substantially broader with expression by APC, myeloid T-cell proliferation and function, and ultimately restore cells, subsets of activated T cells, endothelium, as well as T-cell functional homeostasis [2]. The prototypic T-cell a broad range of tumors (reviewed in [4–6]). While one physiological role of PD-L1 is believed to involve the * Correspondence: yiwen.li@lilly.com; MKALOS@ITS.JNJ.COM suppression of T-cell activation to minimize damage to Lilly Research Laboratories, Department of Cancer Immunobiology, New normal tissues by activated T cells [7, 8], more recent York, NY, USA Full list of author information is available at the end of the article © The Author(s). 2018, corrected publication May 2018. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 2 of 14 evidence suggests that PD-L1 might also play important avelumab and the anti-PD-L1 antibody BMS-93559 have roles to modulate innate immunity by sensing hypoxic been also shown to structurally overlap the binding site [9] and metabolic [10] stress. PD-L1 also binds to a sec- for PD-1 on PD-L1 [19, 20]. ond receptor B7–1 (CD80), which is the inhibitory lig- To-date preclinical development of agents that target and for CTLA-4 and is expressed on dendritic cells, the PD-1/PD-L1 axis has been based on the use of in macrophages, activated T and B cells and some non- vitro human assay systems which have revealed the po- hematopoietic cells (liver stromal cells and keratino- tential for blockade of PD-1 and PD-L1 to increase cytes) [6], raising the to-date untested possibility that T-cell activation and function, and in vivo preclinical ex- the PD-L1 ligand may play a role to modulate both the periments with surrogate, murine-specific antagonist PD-1 and CTLA-4 T cell inhibitory pathways. antibodies and syngeneic immune competent mouse The PD-L1/PD-1 axis is often subjugated by tumors to models. These models have demonstrated that blocking evade anti-tumor immune response; indeed, PD-L1 ex- the murine PD-1/PD-L1 axis can result in immune- pression in tumor tissues has been an important predictive mediated anti-tumor activity and in a number of cases biomarker of response for PD-1 pathway inhibitors across cooperativity and/or synergy with other T cell modulat- multiple cancers and molecules in clinical development. ing agents [21–23]. However, these studies have largely PD-L1 is genetically dysregulated in a variety of tumor failed to explore the functional and mechanistic proper- types, and increased expression of PD-L1 by tumors cor- ties of the clinical agents on human immune cells, relates with a poor prognosis in patients with lung, ovar- including understanding how blockade of the PD-1/PD- ian, renal and other solid tumors [11–13]. PD-L1 L1 axis might impact immune functions beyond T cells. expression can also be up-regulated in the tumor More recently, an anti-PD-1 antibody (REGN2810) has microenvironment as a result of immune activation and been characterized in a mouse model with human PD-1 production of pro-inflammatory cytokines such as knock-in, and shown to enhance anti-tumor murine T interferon-gamma (IFNγ), contributing to the establishment cell immunity in that setting [24]. of an “adapted” T-cell immunosuppressive milieu [14]. Here we describe the discovery, characterization, and The clinical validation of targeting the PD-1/PD-L1 axis preclinical development of LY3300054, an antagonist demonstrated by inhibition of the PD-1 receptor, has also anti-human PD-L1 antibody isolated from a human ScFv led to the clinical development and regulatory approval of phage display library. LY3300054 is a fully human IgG-1 multiple molecules that block the PD-L1. To-date this list of antibody engineered with ablated Fc-mediated immune approved PD-L1-targeting agents includes atezolizumab effector functions. LY3300054 potently blocks inter- (Tecentriq™), avelumab (Bavencio™), and durvalumab action of PD-L1 with its cognate receptors PD-1 and (Imfinzi™) across multiple tumor types and lines of therapy CD80, and cross-reacts with cynomolgous PD-L1. (reviewed in [15]). Approved antibodies that target the PD1/ Sequence alignment and mutagenesis demonstrates that PD-L1 axis include both effector competent and Fc effector- residue N63 on PD-L1, which is part of the PD-L1/PD-1 ablated molecules, without a to-date clear picture about how interphase, is a key residue for the target and species this variable might impact activity in the clinical setting. specificity of LY3300054. LY3300054 is capable of modu- Despite the to-date approval of a number of medicines lating T cell effector functions in a variety of in vitro im- that target the PD-1/PD-L1 axis, considerable efforts to de- mune cell functional assays, and to enhance T cell velop additional agents that target this pathway are ongoing functional activation and T cell-mediated anti-tumor ac- both clinically and pre-clinically, with multiple agents cur- tivity in three different mouse xenograft tumor models rently at various stages of development; these efforts reflect with reconstituted human immune cell compartments. the broad recognition and acceptance that targeting the High-content molecular analysis of tumor and peripheral PD-1/PD-L1 axis is likely to be a foundational component tissues collected from these animals reveals a broad for future immunotherapy-based strategies to treat cancer. spectrum of immune-related intra- and extra-tumoral ef- Analysis of the PD-1/PD-L1 co-complex structure has fects for LY3300054, including previously not described revealed a protein-protein interface that is largely devoid effects on innate immune pathways. LY3300054 is cur- of deep pockets [16], an interface typically amenable to rently under clinical evaluation in monotherapy and targeting by antibodies. High resolution crystallographic combination with other therapeutic modalities in mul- data sets have demonstrated that the epitopes of the tiple tumor types (NCT02791334; NCT03099109; PD-1 blocking antibodies nivolumab and pembrolizu- NCT02791334; NCT02791334). mab structurally cluster on the flat PD-L1-binding surface of PD-1 and overlap with the binding site for Methods PD-L1 [17–19]. While no structural data is publically Phage screening available for the crystal complexes of atezolizumab, A human scFv phage display library (AbCheck, Czech durvalumab, or avelumab with PD-L1, the epitope of Republic) was used to identify phage antibodies that Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 3 of 14 bound recombinant human PD-L1-Fc protein. Phage 100 μl of serially diluted anti-PD-L1 antibody or control that bound to human Fc, CD80 and CD86 were depleted IgG was then added and incubated at room temperature from the libraries by pre-incubation steps throughout for 2 h. After washing, the plate was incubated with goat the panning process. In some cases, libraries were heated anti-human IgG F(ab’)2-HRP conjugate (Jackson Immu- to 65 °C for 15 min prior to the panning step to select noResearch, West Grove, PA) at room temperature for for heat-stable scFv. Enrichment of PD-L1 specific scFv 1 h. The plates were washed and then incubated with 3, was tested with bacterial extracts containing soluble scFv 3′,5,5′-tetramethylbenzidine. The absorbance at in ELISA. Panned phages were screened for the presence 450 nm was read on a microplate reader. The half max- of scFv that blocked the interaction of PD-L1 with both imal effective concentration (EC50) was calculated using PD-1 and CD80. Clone ABC110 (LY3300054) was GraphPad prism software. selected from a large number of functional hits based on binding, blocking, and in-vitro functional properties, and Binding to canine PD-L1 variants its DNA sequence was cloned into a human expression Ninety-six well Immulon 4HBX ELISA plate was coated vector with an IgG1 effector-null backbone (IgG1-EN), overnight with 50 ng each of the wild type and mutant ca- containing the following residue changes; L234A, L235E, nine PD-L1-ECD-Fc in 100 μl of PBS, pH 7.2 with mild G237A, A330S, and P331S (11520463), and CHO cells agitation at 4 °C. After blocking and wash, a five-fold dilu- that stably expressed LY3300054 were established. tion series (0.0017–133 nM) of LY3300054 was added in LY3300054 IgG was purified from the culture super- duplicate and incubated with mild agitation for 1 h at natant by protein A affinity chromatography (Poros A, room temperature. The wells were washed and a 1:10,000 Applied Biosystems, Foster City, CA). By flow cytometry dilution of HRP-conjugated goat anti-Fab antibody LY3300054 was shown to specifically bind to the surface of (Jackson ImmunoResearch) was added and incubated at the PD-L1-positive (H292, HCC827) but not the PD-L1– room temperature following standard protocol. TMB per- negative A204 cell lines (Additional file 1:FigureS1). oxidase chromogenic substrate and stop solution were used according to manufacturer’s instruction for visualization Protein expression and purification and detection of signals. Absorbance readings were plotted The extracellular domain (ECD) of human PD-L1 was in GraphPad Prism software. EC values were calculated cloned into an Fc (human IgG1) construct (GS vector) by nonlinear regression curve fit analysis of the software’s that contained a Factor Xa cleavage site at the One Site-Specific Binding function. N-terminus of the hinge region. Human PD-L1-Fc was expressed in human 293-Freestyle cells (Invitrogen ELISA blocking assays on PD-L1 interaction with PD-1 or CD80 Corp., Carlsbad, CA) that were cultivated and trans- Serially diluted LY3300054 or control IgG were mixed fected according to manufacturer’s specifications. with the equal volume of a fixed concentration of bio- Human PD-L1-Fc was purified via standard ProA affinity tinylated PD-L1-Fc (100 ng/mL for PD-1 blocking and columns; human PD-L1 monomer was cleaved from the 500 ng/mL for CD80 blocking), and then incubated at purified Fc construct with Factor Xa enzyme. Cleaved Fc room temperature for 1 h. 100 μl of the mixture was and undigested PD-L1-Fc were purified out of the sam- transferred to 96-well plates pre-coated with human ple via standard ProA affinity column. Purified proteins PD-1-Fc or with human CD80-Fc at 100 ng/well (R&D were buffer exchanged into PBS, quantified and evalu- Systems) and then incubated at room temperature for ated by SDS-PAGE and analytical SEC analysis to con- an additional 1 h. After washing, Streptavidin-HRP con- firm structural integrity. Canine PD-L1-Fc and its jugate was added, and absorbance at 450 nm was read. mutants were expressed transiently in Expi293F cells IC50 represents the antibody concentration required for following transfection using ExpiFetamine 293. The ca- 50% inhibition of PD-L1 binding to PD-1 or to CD80. nine PD-L1-Fc and its mutants in addition to the cyno- molgus, murine and rat PD-L1-Fc were generated in a SPR binding to recombinant human, murine or cynomolgus manner similar to that of the human PD-L1-Fc. PD-L1 Surface plasmon resonance (SPR) (Biacore T200, GE ELISA binding assays Healthcare) was used to determine the binding kinetics Binding to recombinant PD-L1 of LY3300054 to human, cynomolgus, murine and rat Ninety-six-well plate (Immulon 2HB) was coated with PD-L1-Fc at 37 °C. Approximately 40 response units 100 ng of human PD-L1-Fc, murine PD-L1-Fc, or cyno- (RU) of LY3300054 were immobilized onto a CM5 chip molgus PD-L1-Fc (R&D Systems, Minneapolis, MN) using the standard amine coupling procedure. HBS-EP overnight at 4 °C. Wells were blocked for 2 h with block- buffer (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, ing buffer (PBS containing 5% nonfat dry milk) and then and 0.005% surfactant p20) was utilized as a running washed three times with PBS containing 0.1% Tween-20. buffer during binding kinetic measurements. The Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 4 of 14 PD-L1-Fc gradients were comprised of seven 3× dilu- hGM-CSF and 500 IU/ml hIL-4 for 4 days. CD4 T cells tions. Starting concentrations were 9 nM for the human were purified from fresh human PBMC of a different and cynomolgus PD-L1-Fc gradients and were 90 nM healthy donor (AllCells) using Human CD4 T Cell for the mouse and rat PD-L1-Fc. PD-L1-Fc proteins Isolation Kit (Miltenyi). The two types of cells were then 4 + were injected for 180 s (contact time) over the immobi- mixed in 96-well V-bottom plates with 5 × 10 CD4 T lized LY3300054 at a flow rate of 30 μl/min. The dissoci- cells and 5 × 10 immature DC in 100 μl of complete ation times for those measurements were 1500 s for the AIM-V medium per well. 100 μl of 2× serially diluted four top concentrations of the gradient and 240 s for the LY3300054 or human IgG1 was added into a well of the rest of the gradient. After dissociation, regeneration of plates. LY3300054 was also tested in combination with the LY3300054 surface was achieved with a single 18 s anti-CTLA4 antibody (Ipilimumab) at equimolar concen- injection of 0.75 M NaCl/25 mM NaOH at 30 μl/min trations ranging from 0.003 to 67 nM. After incubation for followed by a 30 s wash with HBS-EP to stabilize the 72 h at 37 °C at 5% CO2, supernatants and cell pellets were surface. Biacore T200 Evaluation Software (version 1.0) harvested and subjected to immunoassay (human IFN-γ was used to analyze the results from the kinetic experi- ELISA (R&D Systems) or 41-plex Milliplex MAP Human ments. After double referencing to remove artifacts from Cytokine/Chemokine Immunoassay Panel (Millipore, Bur- nonspecific binding, simultaneous global fitting of the data lington, MA) (analytes are listed in Additional file 2)and a for each concentration gradient to a 1:1 L model was per- custom-made Quantigene Plex gene expression analysis formed to determine the association rate (k ), dissoci- (see below). MLR studies of LY3300054 were repeated with on ation rate (k ), and dissociation constant (KD= koff/ at least four different CD4 T cell donors. off kon). At least four different concentration gradients were used to compute the kinetic parameters and their corre- Antigen recall assay sponding sample standard deviation. Frozen PBMCs were thawed, cultured in 10% FBS RPMI overnight at 37 °C at 5% CO , and seeded in a 96-well In vitro functional assays flat bottom tissue culture plate at 1 × 10 cells per well PD-1 reporter assay in 100 μl of 10% FBS/RPMI-1640. Antibodies were pre- + − PD-L1 aAPC/CHO-K1 (Promega) or PD-L1 aAPC/ pared at 4× concentrations and added to the cells at CHO-K1 (Promega part# CS187110) human T-activator 50 μl per well. After 1-h incubation, Tetanus Toxoid cells were plated in a 96-well white opaque tissue culture (50uL; 0.8μg/ml) (TT; #191A LIST Biological Laborator- plate at 40,000 cells per well in 100 μl of medium (10% ies Inc.) was added to wells with LY3300054 or medium FBS F-12, 0.2 mg/ml Hygromycin-B and 0.2 mg/ control. After 5 days in culture, supernatant was col- ml G418) and incubated overnight at 37 °C at 5% CO lected and an IFNγ ELISA (R&D Systems SIF50) was Medium was removed from the assay plate the following performed according to manufacturer’s instructions. day and serially diluted test and control antibodies were added at 40 μl per well in the assay buffer. GloResponse Effector function assays NFAT-luc2/PD1 Jurkat cells (Promega) were re- Antigen-dependent cell-mediated cytotoxicity (ADCC) assay suspended in assay buffer at a concentration of 1.25 × The ability of LY3300054 to mediate ADCC was tested 6 + 10 /ml and added to the plate at 40 μl per well. After in a Jurkat-FcγRIIIa reporter gene assay using a PD-L1 6 h of co-culture, assay plates were removed from the HEL cell line (ATCC TIB-180) as previously described incubator and equilibrated at room temperature for [26]. Anti-CD20 antibody rituximab (wild type IgG1) 5 min. Bio-Glo™ Reagent (Promega) was prepared ac- was tested as a positive control in the same assay against cording to manufacturer’s instructions and added to the CD20-positive WIL2-S cell line. Briefly, 1 × 10 each well at 80 μl per well. Plates were then incubated target cells at 50 μl and serially diluted antibodies at for 5 min at room temperature. Luminescence was mea- 4× concentrations at 25 μl were added per well. sured in a plate reader and data was analyzed using Jurkat-FcγRIIIa (V158) cells were added as effector GraphPad Prism software [25]. cells at the effector/target ratio of 15:1 at 25ul/well, and followed by 6 h incubation in a humidified 37 °C incu- Mixed leukocyte reaction (MLR) bator. Plates were removed and equilibrated to room CD14 monocytes were isolated from frozen human per- temperature for 5 min. Luciferase reagent was added at ipheral blood mononuclear cells (PBMC) obtained from 100 μl/well and luminescence was detected. a healthy donor (AllCells, Alameda, CA) with Human Monocyte Isolation Kit II (Miltenyi, Auburn, CA). Complement dependent cytotoxicity (CDC) assay Immature dendritic cells (DCs) were generated by cul- LY3300054 was tested using the PD-L1 HEL cells as turing these monocytes in complete RPMI-1640 medium targets. Rituximab was used as a positive control against containing 10% FBS in the presence of 1000 IU/ml WIL2-S cell line in the same experiment. Target cells Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 5 of 14 were treated with 1:3 titrations of the various antibodies (15–17 weeks of age, female) were obtained from and incubated for 30 min at 37 °C. Human complement Jackson Laboratories. Animal well-being and behavior, was added into the assay plates and incubated for 1 h at including grooming and ambulation were monitored at 37 °C. Alamar Blue reagent was then added to the wells least twice per week. Body weight and tumor volumes and incubated for an additional 24 h at 37 °C before fluor- were measured twice a week starting 1–2 weeks post im- escence was determined, as an indication of cell viability. plantation. Tumor volumes were calculated according to formula (vol = π/6 * l * w ) and plotted as geometric PBMC cytokine release assay means ± standard error of the mean (SEM). Statistical Fresh unstimulated human PBMC isolated from six analysis of tumor volume data was performed by two- healthy donors were incubated with plate bound way ANOVA on repeated measurements. LY3300054 antibody or control antibodies for 24 h, pre- coated over a broad titration range from 0.003 to 100 μg/ Co-implantation of human NCI-H292 tumor cells and human ml. Anti-CD3 antibody OKT3 (eBioscience, San Diego, PBMC (Winn model) CA) was used as a positive control. Using a commercially Freshly isolated human PBMCs were combined with available multiplex assay based on the Luminex platform freshly cultured NCI-H292 tumor cells (ATCC, Manassas, (Luminex Corporation, Austin, TX), 21 cytokines includ- VA) at a 1:4 E:T ratio and co-implanted subcutaneously ing Fractalkine, GM-CSF, IFNγ,IL-1β, IL-2, IL-4, IL-5, into the flanks of female NSG mice (groups of 8 mice per IL-6, IL-7, IL-8, IL-10, IL-12 (p70), IL-13, IL-17A, IL-21, treatment arm). One day later, weekly intraperitoneal (IP) IL-23, ITAC, MIP-1α,MIP-1β,MIP-3α,and TNF-α were treatments of either human IgG1 or LY3300054 at 10 mg/ measured in cell culture supernatants [27]. kg began and continued for a total of four doses. Tumor growth was monitored by caliper measurements. PD-L1 and HLA class I staining of human tumor lines NCI-H292, HCC827, OV79, and A204 (ATCC) tumor cells Established HCC827 xenograft tumor model with infused were cultured for approximatelly36 hours prior to non- human T cells enzymatic harvest. NCI-H292, HCC827, and A204 cells Mice were implanted subcutaneously into the flanks of fe- were stained for PD-L1 using FITC-conjugated anti-human male NSG mice with 10 × 10 freshly cultured HCC827 PD-L1 commercial antibody (clone MIH1, BD Biosciences), tumor cells (ATCC). When tumors reached volumes of ~ 3 6 Alexa Fluor® 488-conjugated LY3300054, or appropriate 300 mm (~ 4–5 weeks), 2.5 × 10 expanded human T isotype controls. NCI-H292, HCC827, and OV79 cells were cells were administered intravenously (IV) and mice were stained separately for HLA Class I expression using an treated with weekly IP injections of human IgG1 or APC-conjugated antibody (clone W6/32, RnDSystems, LY3300054 at 10 mg/kg for a total of four doses. Minneapolis, MN) Samples were collected on a 5-laser Established xenograft tumor models in CD34 hHSC- Fortessa X-20 cytometer (BD Biosciences) and analyzed engrafted mice: Cord blood derived CD34 hHSC with FlowJo V10 software (TreeStar). transplanted NSG mice were implanted subcutaneously with serially passaged HCC827 tumor fragments (4–5mm In vivo models in diameter) at 15–17 weeks of age. When the tumors All animal studies were approved by the Institutional reached volumes of approximately 200 mm (~ 30 days), Animal Care and Use Committee and performed in ac- weekly IP treatments of human IgG1 or LY3300054 at cordance with current regulations and standards of the 10 mg/kg began for a total of three doses. Fetal liver-der- United States Department of Agriculture and the ived CD34 hHSC transplanted NOG mice were National Institute of Health. All experiments with adop- implanted subcutaneously with serially-passaged OV79.F- tively transferred human PBMC or expanded human T Fluc2A–gfp tumor fragments (4–5mmindiameter) at scid tm1Wjl cells utilized NOD.Cg-Prkdc Il2rg /SzJ (NSG) ani- 15–17 weeks of age. OV79.FFLuc-2A-gfp tumor cells are mals (6–7 weeks of age, female, from Jackson Laborator- an ovarian carcinoma line transduced with lentivirus en- ies, Bar Harbor, MN), and were maintained in a 12 h coding firefly luciferase and green fluorescent protein light/dark cycle facility under pathogen-free conditions from a bicistronic transcript [28] and will be hereafter be in microisolator cages with standard laboratory chow referred to as OV79. When tumor volumes reached ~ + 3 and water ad libitum. Cord blood-derived CD34 150 mm (18 days), weekly IP treatments of human IgG1 hematopoietic stem cell (HSC) engrafted mice used for or LY3300054 at 10 mg/kg began for a total of four doses. scid tm1Sug the OV79 model utilized NOD.Cg-Prkdc Il2rg / JicTac animals (NOG, 15–17 weeks of age, female) and Immune phenotyping of peripheral blood from were obtained from Taconic BioSciences (Rensselaer, tumor-bearing mice in humanized models NY). Fetal liver derived CD34 HSC transplanted mice Peripheral human immune cell engraftment and pheno- used for the HCC827 model in NSG background type was assessed using Trucount™ tubes according to Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 6 of 14 manufacturer’s instruction (BD Biosciences, San Jose, superfamily, such as PD-L2, B7–1, B7–2, PD-1, CD28, CA). Briefly, 50 μl of blood from hHSC-transplanted TIGIT, TIM3, or VISTA (data not shown). mice (day 18, pre-treatment; day 34, after three treat- Biacore-based surface plasmon resonance analysis was ment doses; day 46, after four treatment doses), was performed to evaluate the affinity of LY3300054 for bind- added to the tubes and stained with antibodies against ing to Fc-tagged PD-L1. These analyses revealed an − 11 6 − 1 − 1 human CD45-FITC (BD Biosciences), human CD3- affinity of 8.19 × 10 M(k , = 1.40 × 10 M s ;k on off − 4 − 1 BV786 (Biolegend), human CD4-BV650 (BD Biosci- =1.14 ×10 s ). LY3300054 showed cross-reactivity to ences), human CD8-BV605 (Biolegend, San Diego, CA), cynomolgus PD-L1 with a similar affinity (K 1.22 × 10 10 6 − 1 − 1 − 4 − 1 and human PD-1-PEeFluor610 (eBiosciences, San Diego, ;k =1.51 × 10 M s ;k =1.84 ×10 s ), but on off CA) cell surface markers. Samples were subsequently not with the murine or rat PD-L1. fixed and collected on a 5-laser Fortessa X-20 cytometer To assess ligand blocking properties of LY3300054, solid (BD Biosciences) and analyzed with FlowJo V10 software phase blocking ELISA assays were performed. LY3300054 (TreeStar). Briefly, approximately 5000 fluorescent beads blocked PD-L1 binding to both PD-1 and CD80 ligands in were collected and enumerated. Human CD45 cells a concentration-dependent manner, with IC50 of 0.95 nM were also gated and enumerated, followed by subsequent and 2.4 nM, respectively (Fig. 1d, e). + + + gating on CD3 cells, followed by CD4 cell and CD8 To evaluate the ability of LY3300054 to bind to PD-L1 cell gating and enumeration, and finally PD-1 express- physiologically expressed on the surface of cells we per- ing cells were identified using appropriate IgG control. formed flow cytometry analyses on the tumor cell lines The absolute number of T cells and CD4+ and CD8+ with known surface PD-L1status. For these studies we subsets were calculated based on relative beads collected employed the NCI-H292, and HCC827 tumor cell lines compared to total number provide by manufacturer. evaluated in the in-vivo studies described below, as well Statistical analysis for human T cell engraftment and as the PD-L1-negative muscle rhabdomyosa cell line phenotype was performed using a two-way ANOVA on A204 (ATCCCRL-7900), and stained with either the repeated measurements. commercially available anti-PD-L1 antibody M1H1or with alexa647 fluor-conjugated LY3300054; as shown in Gene expression analysis of tumor and peripheral tissues Additional file 1: Figure S1, the PD-L1 –positive NCI- in humanized tumor models H292 and HCC827 stained robustly with either M1H1 Total RNA was isolated from snap-frozen tumor tissue or LY3300054, with the PD-L1 negative A204 cell line (day 15 from H292 model and day 15 post T cell infusion failed to stain with either reagent.. Finally, the OV79 cell from HCC827 tumor model) or from snap frozen white line employed for the in-vivo studies also stained posi- blood cell pellets, spleens, or bone marrow (hHSC- tive for PD-L1 (data not shown). engrafted models), using the MagMAX 96 Total RNA isolation (Life Technologies, Carlsbad, CA) and RNeasy Position N63 on human PDL-1 is a specificity anchor for mini (Qiagen, Hilden, Germany) kits, respectively. LY3300054 For QuantiGene Plex analysis, 500 ng of total RNA Since LY3300054 binds human PD-L1 but not human from tumor tissues were subjected to a custom-designed PD-L2, or murine and canine PD-L1, we performed multiplex assay (targets are listed in Additional file 2) ac- sequence alignments across each of these proteins to iden- cording to manufacturer (Affymetrix, Santa Clara, CA) tify key linear residues which might contribute to the protocol. For nCounter analysis, 100 ng of total RNA specificity of LY3300054 for human PD-L1. The multiple from white blood cells were analyzed with the Human sequence alignment analysis suggested that residues Immunology v2 (targets are listed in Additional file 2) 59–72 of the PD-L1 sequence (59-MEDKNIIQFVHGEE- nCounter codeset following manufacturer recommenda- 72) contribute to the human specificity of LY3300054, tions (NanoString Technologies, Seattle, WA). One- or since this sequence is missing in its entirety from the two-way ANOVA was used for statistical analysis. otherwise homologous sequence of human PD-L2 se- quence and is an area of relative divergence across the Results three tested species (Fig. 2a). In particular, we considered Binding and blocking properties of LY3300054 that positions 63 and 69 - side chains, which are exposed ELISA binding assays were performed to assess the se- to solvent according to PDB: 5C3T [16], might be pivotal lective binding and blocking properties of LY3300054. for the species specificity of LY3300054 because their cor- While LY3300054 bound to human and cynomolgous responding amino acid substitutions diverge among the PD-L1 with similar affinities (EC50 of 0.075 nM and three PD-L1 sequences (marked with * in the alignment of 0.085 nM, respectively) (Fig. 1a, b), LY3300054 did not Fig. 2a). We pursued two mutational strategies to test this bind to murine PD-L1 (Fig. 1c); furthermore, LY3300054 hypothesis. The first strategy focused on rescuing the did not bind to other proteins of the immunoglobulin binding of LY3300054 to canine PD-L1 by introducing Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 7 of 14 Fig. 1 Binding and blocking properties of LY3300054. Panels a-c: 96-well plates were coated with recombinant human (a), cynomolgus (b), or murine (c) PD-L1-Fc fusion protein (100 ng/well each). Bound LY3300054 was detected using HRP-conjugated anti-human Fab antibody and addition of chromogenic substrate (OD at 450nm). 96-well plates were coated with 100 ng/well of recombinant PD-1 (d) or B7-1 protein (e), then incubated with a mixture of biotin-conjugated PD-L1 and either LY3300054 or human IgG1 antibodies. Plate bound PD-L1 was detected using HRP-conjugated streptavidin and addition of chromogenic substrate (OD at 450 nm). In all experiments, each data point is the average of two replicates. Data (a-e) are representative of multiple independent experiments ab Fig. 2 Identification of LY3300054 epitope residues in human PD-L1. Panel a: CLUSTALW multiple sequence alignment of domain 1 of human (hu), canine (ca), and murine (mu) PD-L1 and hu-PD-L2 to identify the LY3300054 species specificity anchors on hu-PD-L1. Underlined is the human PD-1 6Å binding site on hu-PD-L1 (according to PDB: 4ZQK (26602187)). An alignment position is marked with (*) if both mu-PD-L1 and ca-PD-L1 substitutions differ from the hu-PD-L1 sequence. An alignment position is marked with (:) if either the mu-PD-L1 or ca-PD-L1 substitution differs from the hu-PD-L1 sequence. Panel b: Position N63 on human PD-L1 is a specificity anchor for LY3300054. Canine-to-human mutation K63N (▲) rescues the ELISA binding of LY3300054 to canine PD-L1. Like wild type ca-PD-L1-Fc (●), canine-to-human mutant N69H (△) does not bind LY3300054 Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 8 of 14 canine-to-human mutations at positions 63 and 69 of the IFNγ and IL2 secretion was substantially enhanced in canine PD-L1-Fc. As shown in Fig. 2b, only the variant the combination treatment compared to each of the sin- K63 N and not N69H rescued the binding of LY3300054 gle agents (Fig. 4a). High-content gene expression to ca-PD-L1-Fc. Notably, neither mutations (K63 N or analysis revealed overlapping gene expression changes N69H) compromise the structural integrity of the ca-PD- across all treatment groups, and also, in agreement with L1-Fc protein since the Size Exclusion Chromatography previous reports on the combination of PD-1 and ipilu- (SEC) profiles of both variants were identical to the SEC mimab therapy ([29]), distinct gene profiles in the com- profile of the wild type ca-PD-L1-Fc in Additional file 1: bination group, with the maximum treatment effect Figure S2. The second mutational strategy focused on ab- observed at 67.5 nM (Fig. 4b). LY3300054 single agent rogating the binding of LY3300054 to human PD-L1 by treatment induced gene expression changes indicative of introducing human-to-murine mutations at positions 63 immune activation, exemplified by increased expression and 69 in human PD-L1. Only N63Q and not H69A abro- of IFNG, IL2, IDO1, GZMB, IL1B, IL6, while ipilimumab gated the binding of LY3300054 to hu-PD-L1-Fc single agent treatment resulted in enhanced T cell Additional file 1: Figure S3. Thus, both mutational strat- activation, exemplified by enhanced ICOS, and IFNG ac- egies demonstrated that the N63 residue is pivotal for the companied by downregulation of myeloid genes (CD68, species specificity of LY3300054. Furthermore, analysis of CD14, HLA-DRA). The combination of LY3300054 and the co-crystal structure of the human PD-1/PD-L1 co- ipilimumab further upregulated T-cell specific genes crystal indicates that the N63 residue is part of the 6Ǻ reflecting a Th1 response (IFNG, IL2, TBX21), T-cell ac- binding site of PD-1 on PD-L1 ([16]). Hence, LY3300054 tivation (IL2, IFNG, ICOS) and downregulation of blocks the PD-1/PD-L1 interaction because its binding epi- myeloid genes (CD68, CD14, HLA-DRA). tope overlaps the binding site of PD-1 on PD-L1. LY3300054 ADCC and CDC functions Functional activity of LY3300054 in vitro LY3300054 was engineered to ablate Fc-gamma receptor The ability of LY3300054 to enhance T cell functional engagement and associated immune effector functions, activity was assessed using a variety of in vitro assays. In specifically ADCC and CDC. LY3300054 was evaluated a PD-1 reporter assay, using Jurkat cells engineered to by SPR and solid phase ELISA to lack binding to FcgRI, stably express human PD-1 and an NFAT-luciferase re- FcgRIIa, and FcgRIIIa (F158) within the limit of porter construct and CHO-K1 cells engineered to stably detection titrated to 10 μM antibody concentration (data express human PD-L1, addition of LY3300054 resulted not shown). Ablation of ADCC and CDC functions of in a concentration-dependent increase in NFAT-driven LY3300054 was evaluated in cell-based assays, using the expression of luciferase, overcoming the inhibitory ef- HEL PD-L1 positive tumor cell lines. In both ADCC and fects of PD-L1 expressed by CHO cells (Fig. 3a). In CDC assays LY3300054 did not direct detectable effector mixed leukocyte reactions (MLR) using allogeneic hu- function activity against HEL target cells, while man DC and T cells, addition of LY3300054 enhanced rituximab was shown to mediate significant ADCC and the allogeneic T cell response in a dose-dependent man- CDC response against CD20-positive Wil2-S cells ner, with activity observed at concentrations as low as (Additional file 1: Figure S4A, B). 0.05 nM as measured by IFN-γ secretion and mRNA ex- pression (Fig. 3b, e). Additional analysis of MLR cultures by 41-plex microbead-based cytokine and gene expres- LY3300054 does not trigger non-specific cytokine sion panels revealed enhanced secretion and transcrip- production by PBMC tion of multiple immune factors in response to We evaluated the ability of LY3300054 to result in LY3300054 treatment exemplified by increased levels of non-specific cytokine release from unstimulated IL-6, TNFα, MCP-1, MIP-1α,MIP-1β, RANTES, IP-10, human PBMC using plate-bound cytokine release as- IL-4, IL-13, IL-12 in culture medium (Fig. 3d) and up- says. While incubation of PBMC with anti-CD3ε or regulation of IL2, IL1B, IL21 genes (Fig. 3e). Finally, the anti-CD28 agonist antibody TGN1412 antibodies LY3300054 was also shown to enhance T cell activity in resulted in robust cytokine production for a number the tetanus-toxoid recall assay (TTRA) which measures of cytokines, including those associated with cytokine the ability to stimulate antigen-specific memory T cells releasesyndrome(CRS) (TNFα,IL-6,IL-2,IFNγ,and in PBMC. (Fig. 3c). IL-1β), incubation of donor PBMC with LY3300054 In MLR assays, LY3300054 also displayed biological did not result in significant levels of cytokine release activity in combination with the anti-CTLA4 antibody for any of the evaluated cytokines over a broad range ipilimumab. In these experiments, which utilized equi- of concentrations from 0.003 to100 μg/ml molar concentrations of LY3300054 and ipilimumab, (Additional file 1:FigureS5). Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 9 of 14 Fig. 3 LY3300054 enhances T cell activation in vitro. Panel a: Jurkat-NFAT reporter assay: Each data point represents the average of two technical replicates, with error bars representing the SD. Data are representative of three independent experiments. Panel b: Mixed leukocyte reactions. Supernatants were measured for IFN-γ production by ELISA. Each data point represents the average of 8 replicates, with error bars representing the SEM. Data are representative of multiple experiments and donor T cells/DC pairs. Panel c: Tetanus toxoid recall assay: Supernatants were measured for IFN-γ production by ELISA. Each data point represents the average of 4 replicates, with error bars representing the SD. Data are representative of two experiments with PBMC obtained from different donors. Panels d and e: Gene expression analysis of cell lysate (E) and cytokine level analysis of cell culture supernatant (F) from the mixed leukocyte reactions using QuantiGene Plex and microbead-based immunoassay panels, respectively. Volcano plots show Log2 fold change of gene expression (E) or cytokine levels (F) in the LY3300054 treated group compared to control group. The highlighted circles correspond to differentially expressed genes (DEG) or cytokines that display fold change > 1.7 (black solid vertical line) and p value < 0.05 (horizontal dotted line). Circle sizes are proportional to the level of expression in LY3300054 group. One-way ANOVA was used for statistical analysis. Human IgG1 LY3300054 Biological activity of LY3300054 in humanized murine model, we evaluated the effects of LY3300054 on anti- models in vivo tumor alloreactivity, and also performed detailed intra- We evaluated the functional activity of LY3300053 in tumoral and peripheral immune pharmacodynamic vivo using human tumor xenograft models and immune assessments to evaluate how LY3300054 therapy modu- deficient NSG animals reconstituted with human im- lated human immune cell activities. Notably, each of the mune cells. We performed these studies in different tumor cell lines evaluated in these studies robustly ex- models, including a preventative (co-implantation) presses HLA class I (Additional file 1: Figure S6). model, a therapeutic model with established tumors and For the preventative studies, mice were co-implanted animals reconstituted with allogeneic human T cells, and with a mixture of human PBMC and NCI-H292 tumor two therapeutic models with established tumors and ani- cells, followed by treatment with LY3300054 or control mals engrafted with human sHSC), each designed to human IgG1. Compared to untreated and human IgG1- evaluate different functional attributes of anti-PD-L1 im- treated animals, treatment with LY3300054 resulted in munotherapy. Each of these models evaluates the ability significant tumor inhibition (p <0.001) (Fig. 5a). The to modulate the inherent alloreactivity of the engrafted therapeutic potential of LY3300054 to modulate T cell- human immune system against the tumor. In each mediated anti-tumor activity in an established tumor Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 10 of 14 Fig. 4 Combination of LY3300054 and ipilimumab enhances T cell activation in vitro. Panel a: Mixed leukocyte reactions. Allogeneic DC were co-cultured with purified CD4 T cells for 72 h in the presence of increasing (two-fold increments) concentrations of LY3300054, ipilimumab or a combination of both antibodies ranging from 0.0003 to 67 nM. Supernatants were measured for IFN-γ and IL-2 production by ELISA. Each data point represents the average of 8 replicates, with error bars representing the SEM. Data were generated with four different PBMC donors. Panel b: Gene expression analysis of cell lysate from the mixed leukocyte reactions was performed using QuantiGene Plex assay. Venn diagram showing the number of shared (overlap circle) and treatment-specific (no overlap) DEGs across the different treatments. Tables list the Log2 fold-change of LY3300054 vs control group for genes with fold-change > 1.5, p value < 0.05. One-way ANOVA was used for statistical analysis setting was evaluated using the HCC827 xenograft mouse To evaluate the intra-tumoral and peripheral immune- tumor model, and animals reconstituted with ex-vivo ex- pharmacodynamic effects of LY3300054 therapy on panded CD3+ T cells. While infusion of expanded human engrafted human immune cells, tissues (tumor, spleen, T cells alone modestly delayed tumor growth, reflecting bone marrow, peripheral blood) were collected from the the baseline anti-tumor reactivity of the engrafted allor- above models and subjected to mechanistic analyses. In eactive T cells, treatment with LY3300054 significantly en- CD34+ hHSC-reconstituted, tumor-bearing models, hanced this effect resulting in potent anti-tumor activity treatment with LY3300054 resulted in an increase of the (Fig. 5b). The therapeutic potential and activity of absolute number of human T cells, an enhanced CD8/ LY3300054 in the context of more completely human CD4 T cell ratio, and increased frequency of PD1+ CD8 immune-replete animals was assessed in immunodeficient + and CD4+ T cells, indicative of T cell activation NSG or NOG mice engrafted with HSCs of human origin (Fig. 6a-d). Furthermore, LY3300054 therapy resulted in (CD34+ huHSCs), and two xenograft mouse tumor prominent immune-related gene expression changes models, using the HCC827 and OV79 tumor cell lines. consistent with IFNγ pathway and T cell activation in These experimental models display a more replete human the tumor tissue (Fig. 6e), as well as in spleen immune compartment exemplified by differentiation of (Additional file 1: Figure S7A) and peripheral blood cells both lymphoid and myeloid cells ([30]). In both models (Additional file 1: Figure S7B), and to a much lesser ex- LY3300054 therapy strongly enhanced the alloreactive tent in the bone marrow (Additional file 1: Figure S7C). anti-tumor response (Fig. 5c-d). In tumor tissue of CD34+ hHSC-reconstituted mice, Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 11 of 14 Fig. 5 LY3300054 demonstrates anti-tumor efficacy in xenograft tumor models reconstituted with human immune cells. Antibody treatment (indi- cated by red arrows), either human IgG or LY3300054, was delivered by intra-peritoneal injection in each case at 10 mg/kg. Tumor growth was monitored by caliper, and results are represented as a geometric mean of tumor volumes ± SEM. Panel a: Co-implantation model: NCI-H292 tumor cells and freshly isolated human PBMC were co-implanted subcutaneously into the flanks of NSG mice. n = 8 for all groups. Panel b: Estab- lished tumor model: HCC827 tumor cells were implanted subcutaneously into the flanks of NSG mice. When tumors reached volumes of ~ 300 mm (approximately 5 weeks), mice were infused with previously expanded human T cells (black arrow). n = 8 for all groups. Panels c and d: Established tumor models in CD34+ hHSC-reconstituted animals: HCC827 tumors (NSG, panel |C) or OV79 tumors (NOG, panel D). Mice were im- planted subcutaneously with either HCC827 or OV79 tumor fragments at ~ 15–17 weeks of age (~ 13–15 weeks post HSC engraftment). HCC827 3 3 tumors were allowed to grow to ~ 200 mm (4 weeks) and OV79 were allowed to grow to ~ 150 mm (18 days) before starting weekly treat- ments of either human IgG1 or LY3300054 at 10 mg/kg. n =5–9 per group. Statistically significant difference is indicated* (two-way repeated measurement ANOVA, RM-ANOVA) LY3300054-induced increased gene expression changes of expression were also apparent, suggesting tissue- and reflected T cell infiltration and activation (CD3E, CD8B, tumor-specific effects of LY3300054 (Fig. 6f). Further- CD4, PRF1, GZMB, TBX21, EOMES), myeloid cell infil- more, LY3300054 also upregulated genes indicative of T tration and differentiation (ITGAM, ITGAX, CD14, cell activation (upregulated CD274, PDCD1LG2, IDO1, CD68, ARG1) upregulation of co-inhibitory/co-stimula- CXCL9, CXCL10, CD3E, CD8B, CD4, ICOS, CD27 etc) tory receptors and ligands (TNFRSF9, TNFRS18, in tumor tissues collected from the established HCC827 TNFRSF4, CD28, CD27, ICOS, CD226, CD200R1, tumor model implanted with ex vivo expanded human T PDCD1, CD274, PDCD1LG2, TIGIT, HAVCR2, LAG3), cells and the co-implantation NCI-H292 model, al- cytokines and their receptors (IFNG, IL2, IL2RA, IL21, though the overall effect of the antibody was less pro- CCL3, CCL4, CCL5), interferon type I response (IFNA2, nounced in these models compared to HCC827-bearing IFNB1), antigen presentation and MHC class I and II CD34+ hHSC-reconstituted NSG mice (Additional file 1: (HLA-A, HLA-B, HLA-C, B2M, HLA-DRA), and down- Figure S8). regulation of only two genes from the tested panel, TGFB2 and IL1B (Fig. 6e). Three genes indicative of T Discussion cell activation CD274 (PD-L1), CCL5 (RANTES), and We describe the discovery and pre-clinical characterization ITGAL (LFA-1) were upregulated in all 4 tissues evalu- of LY3300054, an Fc effector function-silenced fully human ated, demonstrating a systemic effect of PD-L1 blockade antibody that binds to the PD-L1 checkpoint ligand and in HCC827-bearing CD34+ HSC-reconstituted NOG blocks interactions with both PD-1 and CD80 receptors. mice. While the gene profiles between tumor, spleen, LY3300054 is shown to lack ADCC and CDC effector func- and peripheral blood (the three tissues with robust gene tions, and to not trigger TCR-independent T cell activation modulation) showed patterns of overlap, unique patterns as assessed by non-specific cytokine production by human Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 12 of 14 Fig. 6 LY3300054 enhances peripheral T cell engraftment and activation and induces T cell inflamed phenotype in tumor tissues of CD34+ hHSC- engrafted NOG and NSG mice. Panels a, b, c, d: Blood from OV79-bearing CD34+ hHSC-engrafted NOG mice was analyzed for human T cell engraftment and phenotype using TruCount tubes on day 18 (pre-dose), day 34 (after three antibody doses), and day 46 (after four antibody doses). Peripheral T cell engraftment (A); CD8:CD4 ratio (B); PD-1 expression in CD4+ (C) and CD8+ T cells (D) cells. Results are represented as a geometric mean of engraftment + SEM with n = 9 mice on day 18 and day 34, and n = 5 mice on day 46. Two-way repeated measurements ANOVA was used for statistical analysis. Panel e: Gene expression analysis of tumor sample was performed using QuantiGene Plex assay. Volcano plots show Log2 fold-change of gene expression in the LY3300054 treated group compared to control group. The highlighted circles correspond to differentially expressed genes that display fold change > 1.7 (black solid vertical line) and p value < 0.05 (horizontal dotted line). Circle sizes are proportional to the level of expression in LY3300054 group. One-way ANOVA was used for statistical analysis. Panel f: Venn diagram showing the number of shared and tissue-specific DEGs (LY3300054 vs human IgG treatment) across various tissue types. Table on the right lists shared DEGsacross various tissues with fold-change > 1.7, p value < 0.05 for LY3300054 vs control PBMC in vitro. From a structural perspective, using se- setting. The paucity of information is related at least in quence analysis and mutagenesis we identify a key residue part to the relatively recent availability of humanized on PD-L1, N63, which is part of the binding site for PD-1, murine tumor models with reconstituted human im- may play important roles in the function of the PD-1/PD- mune cell compartment(s). We chose to thoroughly L1 interaction, and which contributes to the species specifi- evaluate the functional activity of LY3300054 in a variety city of LY3300054. In vitro LY3300054 is shown to block of immune-humanized mouse models, to begin to PD-L1-mediated T cell suppression of primary human T understand how anti-PD-L1 therapy might modulate cells in both primary MLR and tetanus-toxoid recall assays, anti-tumor T cell immunity in biologically complex and and to reverse TCR-engagement mediated activation of the relevant matrices, more representative of the clinical set- NFAT pathway using signal reporter Jurkat cells with ting. These models represented both prophylactic and ectopic expression of human PD-1. In both preventative therapeutic intervention, reconstituted with human and therapeutic xenograft tumor models reconstituted with PBMC, expanded human T cells, or CD34 HSC. In each human T cells or HSC, LY3300054 therapy resulted in of the tested models, LY3300054 treatment resulted in robust anti-tumor activity, accompanied by the develop- enhanced anti-tumor alloreactivity and robust anti- ment of distinct T cell inflamed signatures in the tumor tumor effects, demonstrating the potency of the agent as and peripheral tissues. well as the relevance of the PD-1/PD-L1 axis in the Although multiple agents that block the PD-1/PD-L1 context of humanized mouse models. axis have been described and evaluated in the clinic, We took advantage of the ability to collect tumor and relatively little information exists about the functional tissues from the animal models and performed high con- properties of these molecules in the pre-clinical in vivo tent flow cytometry and molecular immuno- Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 13 of 14 pharmacodynamic profiling to obtain insights about the Acknowledgements We thank Dr. Vera Molkenthin and her team at AbCheck for phage library mechanism of action of LY3300054-mediated anti-PD- antibody screening, members of Biologics Technology, Quantitative Biology L1 blockade in each of the humanized models. These and Non-Clinical Safety groups (Marshall Snavely, Krishnadatt Persaud, analyses clearly demonstrated the ability of LY3300054 Yongqing Jia, Liz Navarro, Sagit Hindi, Prabakaran Naryanasamy, Michele Iacolina, Ana Cardoso, Nicole Covino, Heather Griffith, Julia Lee, Nada Alakhras) to modulate human T cell functions as reflected by in- for technical assistance; Thompson Doman, Cindy Wang and Jason Manro for creased peripheral T cell numbers and in particular CD8 computational and statistical analysis, Yung-Mae Yao for helpful input and T cells, and enhanced activation status of peripheral discussions, Tim Bailey for antibody conjugations, Yvonne Vandenburg for project management; Mythili Koneru for insightful discussions, and Carl June allo- and/or xeno-reactive T cells triggered through TCR (University of Pennsylvania) for providing the OV79 cell line. engagement. High-content gene expression profiling re- vealed that LY3300054-induced a T cell inflamed pheno- Authors’ contributions Conception and design: YL, DLL, RN, MK. Development of methodology: YL, type in tumor tissues across all models tested, and GW, CC, JNH, GH, RN. Acquisition of data: YL, GW, LS, MM, CC, DS, AF, MM, almost exclusively resulted in upregulation of gene ex- AS, DC, II, DB, AP, and LM. Analysis and interpretation of data (e.g., produce pression; these data demonstrate the ability of figures, statistical analysis, biostatistics, computational analysis): YL, GW, LS, CC, DS, AF, JNH, RN and MK. Writing, review, and/or revision of the LY3300054 to effectively block the PD-L1/PD-1 axis and manuscript: YL, GW, CC, RN, JNH, MK. Study supervision: YL, DL, RN, MK. to activate T cells to drive more effective anti-tumor T All authors read and approved the final manuscript. cell immunity. Within a model, while a core set of differ- Competing interests entially expressed genes were shared across tumor and LY3300054 is an Eli Lilly and Company investigational drug currently in clinical normal tissues including spleen, peripheral blood and testing. All authors are or have been employees and shareholders of Eli Lilly bone marrow, a considerable number of genes were up- and Company. regulated in a tissue-specific manner suggesting that LY3300054 activity might be context-dependent. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in In more immune-replete animals, blockade of the published maps and institutional affiliations. PD-L1/PD-1 axis activated additional pathways beyond T lymphocytes, including genes associated with adaptive im- Author details Lilly Research Laboratories, Department of Cancer Immunobiology, New munity such as co-stimulatory and co-inhibitory receptors, York, NY, USA. Lilly Research Laboratories, Department of Preclinical cytokines and transcription factors, and innate immune 3 Pharmacology, New York, NY, USA. Lilly Research Laboratories, Department pathways such as interferon, MHC, and myeloid pathways. of Biologics Technology, New York, NY, USA. Lilly Research Laboratories, Department of Non-Clinical Safety, Indianapolis, IN, USA. Lilly Research Whether these observations reflect a direct effect of PD-L1 Laboratories, Department of Quantitative Biology, New York, NY, USA. Eli blockade on innate immune cells or an indirect effect 7 Lilly and Company, 450 East 29th Street, New York, NY 10016, USA. Janssen resulting from T cell activation,theyunderscorethe diverse Pharmaceutical Companies of Johnson and Johnson, Springhouse, PA, USA. and to-date not fully appreciated role for PD-L1 blockade Received: 14 December 2017 Accepted: 20 February 2018 to modulate the activity and function of cell types beyond T lymphocytes, and highlight the need to both integrate References and evaluate innate immune modulation and cell subsets in 1. Baxter AG, Hodgkin PD. Activation rules: the two-signal theories of immune the context of anti-PD-L1-based immunotherapies. activation. Nat Rev Immunol. 2002;2(6):439–46. 2. Chen L, Flies DB. Molecular mechanisms of T cell co-stimulation and co-inhibition. Nat Rev Immunol. 2013;13(4):227–42. Conclusion 3. Yearley JH, et al. PD-L2 expression in human tumors: relevance to anti-PD-1 LY3300054 is currently in phase I clinical studies to therapy in cancer. Clin Cancer Res. 2017;23(12):3158–67. evaluate activity as monotherapy and in combination 4. Francisco LM, Sage PT, Sharpe AH. The PD-1 pathway in tolerance and autoimmunity. Immunol Rev. 2010;236:219–42. with other agents. 5. Okazaki T, Honjo T. The PD-1-PD-L pathway in immunological tolerance. Trends Immunol. 2006;27(4):195–201. 6. Sharpe AH, Pauken KE. The diverse functions of the PD1 inhibitory pathway. Nat Rev Immunol. 2017;18:153–67. Additional files 7. Lin DY, et al. The PD-1/PD-L1 complex resembles the antigen-binding Fv domains of antibodies and T cell receptors. Proc Natl Acad Sci U S A. 2008; 105(8):3011–6. Additional file 1: Figure S1. PD-L1 expression on tumor cell lines. 8. Hunig T. The storm has cleared: lessons from the CD28 superagonist Figure S2. SEC profiles of dog PD-L1-Fc and its dog-to-human variants K63 TGN1412 trial. Nat Rev Immunol. 2012;12(5):317–8. N and N69H. Figure S3. Position N63 on human PD-L1 is a specificity 9. Noman MZ, et al. PD-L1 is a novel direct target of HIF-1alpha, and its anchor for LY3300054. Figure S4. LY3300054 does not induce ADCC or blockade under hypoxia enhanced MDSC-mediated T cell activation. J Exp CDC in vitro. Figure S5. LY3300054 does not induce non-specific cytokine Med. 2014;211(5):781–90. production by human PBMCs in vitro. Figure S6. HLA Class I expression 10. Watanabe R, et al. Pyruvate controls the checkpoint inhibitor PD-L1 and profile on tumor cell lines. Figure S7. LY3300054 enhances expression of suppresses T cell immunity. J Clin Invest. 2017;127(7):2725–38. immune-related genes in peripheral tissue of CD34+ HSC transplanted mice 11. Callahan MK, Postow MA, Wolchok JD. Targeting T cell co-receptors for bearing OV79 tumors (NOG). Figure S8 LY3300054 enhances expression of cancer therapy. Immunity. 2016;44(5):1069–78. immune-related genes in NCI-H292 and HCC827 tumors from humanized 12. Hamanishi J, et al. Programmed cell death 1 ligand 1 and tumor-infiltrating mouse tumor models. (PPTX 1085 kb) CD8+ T lymphocytes are prognostic factors of human ovarian cancer. Proc Additional file 2: Supplementary Methods. (DOCX 34 kb) Natl Acad Sci U S A. 2007;104(9):3360–5. Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 14 of 14 13. Okazaki T, et al. A rheostat for immune responses: the unique properties of PD-1 and their advantages for clinical application. Nat Immunol. 2013;14(12):1212–8. 14. Taube JM, et al. Colocalization of inflammatory response with B7-h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape. Sci Transl Med. 2012;4(127):127ra37. 15. (FDA), F.a.D.A. Hematology/Oncology (Cancer) Approvals & Safety Notifications. 2017. Available from: https://www.fda.gov/Drugs/ InformationOnDrugs/ApprovedDrugs/ucm279174.htm. 16. Zak KM, et al. Structure of the complex of human programmed death 1, PD-1, and its ligand PD-L1. Structure. 2015;23(12):2341–8. 17. Horita S, et al. High-resolution crystal structure of the therapeutic antibody pembrolizumab bound to the human PD-1. Sci Rep. 2016;6:35297. 18. Na Z, et al. Structural basis for blocking PD-1-mediated immune suppression by therapeutic antibody pembrolizumab. Cell Res. 2017;27(1):147–50. 19. Lee JY, et al. Structural basis of checkpoint blockade by monoclonal antibodies in cancer immunotherapy. Nat Commun. 2016;7:13354. 20. Liu K, et al. Structural basis of anti-PD-L1 monoclonal antibody avelumab for tumor therapy. Cell Res. 2017;27(1):151–3. 21. He YF, et al. Blocking programmed death-1 ligand-PD-1 interactions by local gene therapy results in enhancement of antitumor effect of secondary lymphoid tissue chemokine. J Immunol. 2004;173(8):4919–28. 22. Hirano F, et al. Blockade of B7-H1 and PD-1 by monoclonal antibodies potentiates cancer therapeutic immunity. Cancer Res. 2005;65(3):1089–96. 23. Iwai Y, et al. Involvement of PD-L1 on tumor cells in the escape from host immune system and tumor immunotherapy by PD-L1 blockade. Proc Natl Acad Sci U S A. 2002;99(19):12293–7. 24. Burova E, et al. Characterization of the anti-PD-1 antibody REGN2810 and its antitumor activity in human PD-1 knock-in mice. Mol Cancer Ther. 2017; 16(5):861–70. 25. Cong M, et al. Reporter bioassays to assess therapeutic antibodies in development for immunotherapy programs. Madison: Promega Corporation; 2015. 26. Parekh BS, et al. Development and validation of an antibody-dependent cell-mediated cytotoxicity-reporter gene assay. MAbs. 2012;4(3):310–8. 27. Stebbings R, et al. “Cytokine storm” in the phase I trial of monoclonal antibody TGN1412: better understanding the causes to improve preclinical testing of immunotherapeutics. J Immunol. 2007;179(5):3325–31. 28. Bertozzi CC, et al. Multiple initial culture conditions enhance the establishment of cell lines from primary ovarian cancer specimens. In Vitro Cell Dev Biol Anim. 2006;42(3–4):58–62. 29. Das R, et al. Combination therapy with anti-CTLA-4 and anti-PD-1 leads to distinct immunologic changes in vivo. J Immunol. 2015;194(3):950–9. 30. Shultz LD, et al. Humanized mice for immune system investigation: progress, promise and challenges. Nat Rev Immunol. 2012;12(11):786–98. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal for ImmunoTherapy of Cancer Springer Journals

Loading next page...
 
/lp/springer_journal/discovery-and-preclinical-characterization-of-the-antagonist-anti-pd-QSLZ0Bqz7Z
Publisher
BioMed Central
Copyright
Copyright © 2018 by The Author(s).
Subject
Medicine & Public Health; Oncology; Immunology
eISSN
2051-1426
D.O.I.
10.1186/s40425-018-0329-7
Publisher site
See Article on Publisher Site

Abstract

Background: Modulation of the PD-1/PD-L1 axis through antagonist antibodies that block either receptor or ligand has been shown to reinvigorate the function of tumor-specific T cells and unleash potent anti-tumor immunity, leading to durable objective responses in a subset of patients across multiple tumor types. Results: Here we describe the discovery and preclinical characterization of LY3300054, a fully human IgG1λ monoclonal antibody that binds to human PD-L1 with high affinity and inhibits interactions of PD-L1 with its two cognate receptors PD-1 and CD80. The functional activity of LY3300054 on primary human T cells is evaluated using a series of in vitro T cell functional assays and in vivo models using human-immune reconstituted mice. LY3300054 is shown to induce primary T cell activation in vitro, increase T cell activation in combination with anti-CTLA4 antibody, and to potently enhance anti-tumor alloreactivity in several xenograft mouse tumor models with reconstituted human immune cells. High-content molecular analysis of tumor and peripheral tissues from animals treated with LY3300054 reveals distinct adaptive immune activation signatures, and also previously not described modulation of innate immune pathways. Conclusions: LY3300054 is currently being evaluated in phase I clinical trials for oncology indications. Background inhibitory (i.e. “checkpoint”) receptors are CTLA-4 T cell activation occurs when T-cells receive two positive (CD152) and PD-1 (CD279), and the regulatory approval signals from antigen-presenting cells (APC): an antigen- of agents that target CTLA-4 (ipilimumab, Yervoy™), specific signal presented in the context of major histo- and PD-1 (nivolumab (Opdtivo™), pembrolizumab (Key- compatibility complex (MHC) which engages the T-cell truda™), has been key to bringing forth the modern era receptor (TCR), and a co-stimulatory signal from B7–1/ of immunotherapy. B7–2 (CD80/CD86) to the CD28 receptor on T-cells [1]. Two ligands have been described for PD-1: PD-L1 Initial T cell activation is followed by the surface expres- ((B7-H1, CD274), and PD-L2 (B7DC, CD273). While sion of a set of co-activating receptors such as CD137, baseline expression of PD-L2 is relatively limited to sub- OX40, GITR, and CD27 which enhance T-cell function, sets of dendritic cells, macrophages, B cells, mast cells and a set of T-cell inhibitory receptors which initiate in- and Th2 cells and tumor cells [3], expression of PD-L1 hibitory pathways that function to prevent uncontrolled is substantially broader with expression by APC, myeloid T-cell proliferation and function, and ultimately restore cells, subsets of activated T cells, endothelium, as well as T-cell functional homeostasis [2]. The prototypic T-cell a broad range of tumors (reviewed in [4–6]). While one physiological role of PD-L1 is believed to involve the * Correspondence: yiwen.li@lilly.com; MKALOS@ITS.JNJ.COM suppression of T-cell activation to minimize damage to Lilly Research Laboratories, Department of Cancer Immunobiology, New normal tissues by activated T cells [7, 8], more recent York, NY, USA Full list of author information is available at the end of the article © The Author(s). 2018, corrected publication May 2018. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 2 of 14 evidence suggests that PD-L1 might also play important avelumab and the anti-PD-L1 antibody BMS-93559 have roles to modulate innate immunity by sensing hypoxic been also shown to structurally overlap the binding site [9] and metabolic [10] stress. PD-L1 also binds to a sec- for PD-1 on PD-L1 [19, 20]. ond receptor B7–1 (CD80), which is the inhibitory lig- To-date preclinical development of agents that target and for CTLA-4 and is expressed on dendritic cells, the PD-1/PD-L1 axis has been based on the use of in macrophages, activated T and B cells and some non- vitro human assay systems which have revealed the po- hematopoietic cells (liver stromal cells and keratino- tential for blockade of PD-1 and PD-L1 to increase cytes) [6], raising the to-date untested possibility that T-cell activation and function, and in vivo preclinical ex- the PD-L1 ligand may play a role to modulate both the periments with surrogate, murine-specific antagonist PD-1 and CTLA-4 T cell inhibitory pathways. antibodies and syngeneic immune competent mouse The PD-L1/PD-1 axis is often subjugated by tumors to models. These models have demonstrated that blocking evade anti-tumor immune response; indeed, PD-L1 ex- the murine PD-1/PD-L1 axis can result in immune- pression in tumor tissues has been an important predictive mediated anti-tumor activity and in a number of cases biomarker of response for PD-1 pathway inhibitors across cooperativity and/or synergy with other T cell modulat- multiple cancers and molecules in clinical development. ing agents [21–23]. However, these studies have largely PD-L1 is genetically dysregulated in a variety of tumor failed to explore the functional and mechanistic proper- types, and increased expression of PD-L1 by tumors cor- ties of the clinical agents on human immune cells, relates with a poor prognosis in patients with lung, ovar- including understanding how blockade of the PD-1/PD- ian, renal and other solid tumors [11–13]. PD-L1 L1 axis might impact immune functions beyond T cells. expression can also be up-regulated in the tumor More recently, an anti-PD-1 antibody (REGN2810) has microenvironment as a result of immune activation and been characterized in a mouse model with human PD-1 production of pro-inflammatory cytokines such as knock-in, and shown to enhance anti-tumor murine T interferon-gamma (IFNγ), contributing to the establishment cell immunity in that setting [24]. of an “adapted” T-cell immunosuppressive milieu [14]. Here we describe the discovery, characterization, and The clinical validation of targeting the PD-1/PD-L1 axis preclinical development of LY3300054, an antagonist demonstrated by inhibition of the PD-1 receptor, has also anti-human PD-L1 antibody isolated from a human ScFv led to the clinical development and regulatory approval of phage display library. LY3300054 is a fully human IgG-1 multiple molecules that block the PD-L1. To-date this list of antibody engineered with ablated Fc-mediated immune approved PD-L1-targeting agents includes atezolizumab effector functions. LY3300054 potently blocks inter- (Tecentriq™), avelumab (Bavencio™), and durvalumab action of PD-L1 with its cognate receptors PD-1 and (Imfinzi™) across multiple tumor types and lines of therapy CD80, and cross-reacts with cynomolgous PD-L1. (reviewed in [15]). Approved antibodies that target the PD1/ Sequence alignment and mutagenesis demonstrates that PD-L1 axis include both effector competent and Fc effector- residue N63 on PD-L1, which is part of the PD-L1/PD-1 ablated molecules, without a to-date clear picture about how interphase, is a key residue for the target and species this variable might impact activity in the clinical setting. specificity of LY3300054. LY3300054 is capable of modu- Despite the to-date approval of a number of medicines lating T cell effector functions in a variety of in vitro im- that target the PD-1/PD-L1 axis, considerable efforts to de- mune cell functional assays, and to enhance T cell velop additional agents that target this pathway are ongoing functional activation and T cell-mediated anti-tumor ac- both clinically and pre-clinically, with multiple agents cur- tivity in three different mouse xenograft tumor models rently at various stages of development; these efforts reflect with reconstituted human immune cell compartments. the broad recognition and acceptance that targeting the High-content molecular analysis of tumor and peripheral PD-1/PD-L1 axis is likely to be a foundational component tissues collected from these animals reveals a broad for future immunotherapy-based strategies to treat cancer. spectrum of immune-related intra- and extra-tumoral ef- Analysis of the PD-1/PD-L1 co-complex structure has fects for LY3300054, including previously not described revealed a protein-protein interface that is largely devoid effects on innate immune pathways. LY3300054 is cur- of deep pockets [16], an interface typically amenable to rently under clinical evaluation in monotherapy and targeting by antibodies. High resolution crystallographic combination with other therapeutic modalities in mul- data sets have demonstrated that the epitopes of the tiple tumor types (NCT02791334; NCT03099109; PD-1 blocking antibodies nivolumab and pembrolizu- NCT02791334; NCT02791334). mab structurally cluster on the flat PD-L1-binding surface of PD-1 and overlap with the binding site for Methods PD-L1 [17–19]. While no structural data is publically Phage screening available for the crystal complexes of atezolizumab, A human scFv phage display library (AbCheck, Czech durvalumab, or avelumab with PD-L1, the epitope of Republic) was used to identify phage antibodies that Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 3 of 14 bound recombinant human PD-L1-Fc protein. Phage 100 μl of serially diluted anti-PD-L1 antibody or control that bound to human Fc, CD80 and CD86 were depleted IgG was then added and incubated at room temperature from the libraries by pre-incubation steps throughout for 2 h. After washing, the plate was incubated with goat the panning process. In some cases, libraries were heated anti-human IgG F(ab’)2-HRP conjugate (Jackson Immu- to 65 °C for 15 min prior to the panning step to select noResearch, West Grove, PA) at room temperature for for heat-stable scFv. Enrichment of PD-L1 specific scFv 1 h. The plates were washed and then incubated with 3, was tested with bacterial extracts containing soluble scFv 3′,5,5′-tetramethylbenzidine. The absorbance at in ELISA. Panned phages were screened for the presence 450 nm was read on a microplate reader. The half max- of scFv that blocked the interaction of PD-L1 with both imal effective concentration (EC50) was calculated using PD-1 and CD80. Clone ABC110 (LY3300054) was GraphPad prism software. selected from a large number of functional hits based on binding, blocking, and in-vitro functional properties, and Binding to canine PD-L1 variants its DNA sequence was cloned into a human expression Ninety-six well Immulon 4HBX ELISA plate was coated vector with an IgG1 effector-null backbone (IgG1-EN), overnight with 50 ng each of the wild type and mutant ca- containing the following residue changes; L234A, L235E, nine PD-L1-ECD-Fc in 100 μl of PBS, pH 7.2 with mild G237A, A330S, and P331S (11520463), and CHO cells agitation at 4 °C. After blocking and wash, a five-fold dilu- that stably expressed LY3300054 were established. tion series (0.0017–133 nM) of LY3300054 was added in LY3300054 IgG was purified from the culture super- duplicate and incubated with mild agitation for 1 h at natant by protein A affinity chromatography (Poros A, room temperature. The wells were washed and a 1:10,000 Applied Biosystems, Foster City, CA). By flow cytometry dilution of HRP-conjugated goat anti-Fab antibody LY3300054 was shown to specifically bind to the surface of (Jackson ImmunoResearch) was added and incubated at the PD-L1-positive (H292, HCC827) but not the PD-L1– room temperature following standard protocol. TMB per- negative A204 cell lines (Additional file 1:FigureS1). oxidase chromogenic substrate and stop solution were used according to manufacturer’s instruction for visualization Protein expression and purification and detection of signals. Absorbance readings were plotted The extracellular domain (ECD) of human PD-L1 was in GraphPad Prism software. EC values were calculated cloned into an Fc (human IgG1) construct (GS vector) by nonlinear regression curve fit analysis of the software’s that contained a Factor Xa cleavage site at the One Site-Specific Binding function. N-terminus of the hinge region. Human PD-L1-Fc was expressed in human 293-Freestyle cells (Invitrogen ELISA blocking assays on PD-L1 interaction with PD-1 or CD80 Corp., Carlsbad, CA) that were cultivated and trans- Serially diluted LY3300054 or control IgG were mixed fected according to manufacturer’s specifications. with the equal volume of a fixed concentration of bio- Human PD-L1-Fc was purified via standard ProA affinity tinylated PD-L1-Fc (100 ng/mL for PD-1 blocking and columns; human PD-L1 monomer was cleaved from the 500 ng/mL for CD80 blocking), and then incubated at purified Fc construct with Factor Xa enzyme. Cleaved Fc room temperature for 1 h. 100 μl of the mixture was and undigested PD-L1-Fc were purified out of the sam- transferred to 96-well plates pre-coated with human ple via standard ProA affinity column. Purified proteins PD-1-Fc or with human CD80-Fc at 100 ng/well (R&D were buffer exchanged into PBS, quantified and evalu- Systems) and then incubated at room temperature for ated by SDS-PAGE and analytical SEC analysis to con- an additional 1 h. After washing, Streptavidin-HRP con- firm structural integrity. Canine PD-L1-Fc and its jugate was added, and absorbance at 450 nm was read. mutants were expressed transiently in Expi293F cells IC50 represents the antibody concentration required for following transfection using ExpiFetamine 293. The ca- 50% inhibition of PD-L1 binding to PD-1 or to CD80. nine PD-L1-Fc and its mutants in addition to the cyno- molgus, murine and rat PD-L1-Fc were generated in a SPR binding to recombinant human, murine or cynomolgus manner similar to that of the human PD-L1-Fc. PD-L1 Surface plasmon resonance (SPR) (Biacore T200, GE ELISA binding assays Healthcare) was used to determine the binding kinetics Binding to recombinant PD-L1 of LY3300054 to human, cynomolgus, murine and rat Ninety-six-well plate (Immulon 2HB) was coated with PD-L1-Fc at 37 °C. Approximately 40 response units 100 ng of human PD-L1-Fc, murine PD-L1-Fc, or cyno- (RU) of LY3300054 were immobilized onto a CM5 chip molgus PD-L1-Fc (R&D Systems, Minneapolis, MN) using the standard amine coupling procedure. HBS-EP overnight at 4 °C. Wells were blocked for 2 h with block- buffer (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, ing buffer (PBS containing 5% nonfat dry milk) and then and 0.005% surfactant p20) was utilized as a running washed three times with PBS containing 0.1% Tween-20. buffer during binding kinetic measurements. The Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 4 of 14 PD-L1-Fc gradients were comprised of seven 3× dilu- hGM-CSF and 500 IU/ml hIL-4 for 4 days. CD4 T cells tions. Starting concentrations were 9 nM for the human were purified from fresh human PBMC of a different and cynomolgus PD-L1-Fc gradients and were 90 nM healthy donor (AllCells) using Human CD4 T Cell for the mouse and rat PD-L1-Fc. PD-L1-Fc proteins Isolation Kit (Miltenyi). The two types of cells were then 4 + were injected for 180 s (contact time) over the immobi- mixed in 96-well V-bottom plates with 5 × 10 CD4 T lized LY3300054 at a flow rate of 30 μl/min. The dissoci- cells and 5 × 10 immature DC in 100 μl of complete ation times for those measurements were 1500 s for the AIM-V medium per well. 100 μl of 2× serially diluted four top concentrations of the gradient and 240 s for the LY3300054 or human IgG1 was added into a well of the rest of the gradient. After dissociation, regeneration of plates. LY3300054 was also tested in combination with the LY3300054 surface was achieved with a single 18 s anti-CTLA4 antibody (Ipilimumab) at equimolar concen- injection of 0.75 M NaCl/25 mM NaOH at 30 μl/min trations ranging from 0.003 to 67 nM. After incubation for followed by a 30 s wash with HBS-EP to stabilize the 72 h at 37 °C at 5% CO2, supernatants and cell pellets were surface. Biacore T200 Evaluation Software (version 1.0) harvested and subjected to immunoassay (human IFN-γ was used to analyze the results from the kinetic experi- ELISA (R&D Systems) or 41-plex Milliplex MAP Human ments. After double referencing to remove artifacts from Cytokine/Chemokine Immunoassay Panel (Millipore, Bur- nonspecific binding, simultaneous global fitting of the data lington, MA) (analytes are listed in Additional file 2)and a for each concentration gradient to a 1:1 L model was per- custom-made Quantigene Plex gene expression analysis formed to determine the association rate (k ), dissoci- (see below). MLR studies of LY3300054 were repeated with on ation rate (k ), and dissociation constant (KD= koff/ at least four different CD4 T cell donors. off kon). At least four different concentration gradients were used to compute the kinetic parameters and their corre- Antigen recall assay sponding sample standard deviation. Frozen PBMCs were thawed, cultured in 10% FBS RPMI overnight at 37 °C at 5% CO , and seeded in a 96-well In vitro functional assays flat bottom tissue culture plate at 1 × 10 cells per well PD-1 reporter assay in 100 μl of 10% FBS/RPMI-1640. Antibodies were pre- + − PD-L1 aAPC/CHO-K1 (Promega) or PD-L1 aAPC/ pared at 4× concentrations and added to the cells at CHO-K1 (Promega part# CS187110) human T-activator 50 μl per well. After 1-h incubation, Tetanus Toxoid cells were plated in a 96-well white opaque tissue culture (50uL; 0.8μg/ml) (TT; #191A LIST Biological Laborator- plate at 40,000 cells per well in 100 μl of medium (10% ies Inc.) was added to wells with LY3300054 or medium FBS F-12, 0.2 mg/ml Hygromycin-B and 0.2 mg/ control. After 5 days in culture, supernatant was col- ml G418) and incubated overnight at 37 °C at 5% CO lected and an IFNγ ELISA (R&D Systems SIF50) was Medium was removed from the assay plate the following performed according to manufacturer’s instructions. day and serially diluted test and control antibodies were added at 40 μl per well in the assay buffer. GloResponse Effector function assays NFAT-luc2/PD1 Jurkat cells (Promega) were re- Antigen-dependent cell-mediated cytotoxicity (ADCC) assay suspended in assay buffer at a concentration of 1.25 × The ability of LY3300054 to mediate ADCC was tested 6 + 10 /ml and added to the plate at 40 μl per well. After in a Jurkat-FcγRIIIa reporter gene assay using a PD-L1 6 h of co-culture, assay plates were removed from the HEL cell line (ATCC TIB-180) as previously described incubator and equilibrated at room temperature for [26]. Anti-CD20 antibody rituximab (wild type IgG1) 5 min. Bio-Glo™ Reagent (Promega) was prepared ac- was tested as a positive control in the same assay against cording to manufacturer’s instructions and added to the CD20-positive WIL2-S cell line. Briefly, 1 × 10 each well at 80 μl per well. Plates were then incubated target cells at 50 μl and serially diluted antibodies at for 5 min at room temperature. Luminescence was mea- 4× concentrations at 25 μl were added per well. sured in a plate reader and data was analyzed using Jurkat-FcγRIIIa (V158) cells were added as effector GraphPad Prism software [25]. cells at the effector/target ratio of 15:1 at 25ul/well, and followed by 6 h incubation in a humidified 37 °C incu- Mixed leukocyte reaction (MLR) bator. Plates were removed and equilibrated to room CD14 monocytes were isolated from frozen human per- temperature for 5 min. Luciferase reagent was added at ipheral blood mononuclear cells (PBMC) obtained from 100 μl/well and luminescence was detected. a healthy donor (AllCells, Alameda, CA) with Human Monocyte Isolation Kit II (Miltenyi, Auburn, CA). Complement dependent cytotoxicity (CDC) assay Immature dendritic cells (DCs) were generated by cul- LY3300054 was tested using the PD-L1 HEL cells as turing these monocytes in complete RPMI-1640 medium targets. Rituximab was used as a positive control against containing 10% FBS in the presence of 1000 IU/ml WIL2-S cell line in the same experiment. Target cells Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 5 of 14 were treated with 1:3 titrations of the various antibodies (15–17 weeks of age, female) were obtained from and incubated for 30 min at 37 °C. Human complement Jackson Laboratories. Animal well-being and behavior, was added into the assay plates and incubated for 1 h at including grooming and ambulation were monitored at 37 °C. Alamar Blue reagent was then added to the wells least twice per week. Body weight and tumor volumes and incubated for an additional 24 h at 37 °C before fluor- were measured twice a week starting 1–2 weeks post im- escence was determined, as an indication of cell viability. plantation. Tumor volumes were calculated according to formula (vol = π/6 * l * w ) and plotted as geometric PBMC cytokine release assay means ± standard error of the mean (SEM). Statistical Fresh unstimulated human PBMC isolated from six analysis of tumor volume data was performed by two- healthy donors were incubated with plate bound way ANOVA on repeated measurements. LY3300054 antibody or control antibodies for 24 h, pre- coated over a broad titration range from 0.003 to 100 μg/ Co-implantation of human NCI-H292 tumor cells and human ml. Anti-CD3 antibody OKT3 (eBioscience, San Diego, PBMC (Winn model) CA) was used as a positive control. Using a commercially Freshly isolated human PBMCs were combined with available multiplex assay based on the Luminex platform freshly cultured NCI-H292 tumor cells (ATCC, Manassas, (Luminex Corporation, Austin, TX), 21 cytokines includ- VA) at a 1:4 E:T ratio and co-implanted subcutaneously ing Fractalkine, GM-CSF, IFNγ,IL-1β, IL-2, IL-4, IL-5, into the flanks of female NSG mice (groups of 8 mice per IL-6, IL-7, IL-8, IL-10, IL-12 (p70), IL-13, IL-17A, IL-21, treatment arm). One day later, weekly intraperitoneal (IP) IL-23, ITAC, MIP-1α,MIP-1β,MIP-3α,and TNF-α were treatments of either human IgG1 or LY3300054 at 10 mg/ measured in cell culture supernatants [27]. kg began and continued for a total of four doses. Tumor growth was monitored by caliper measurements. PD-L1 and HLA class I staining of human tumor lines NCI-H292, HCC827, OV79, and A204 (ATCC) tumor cells Established HCC827 xenograft tumor model with infused were cultured for approximatelly36 hours prior to non- human T cells enzymatic harvest. NCI-H292, HCC827, and A204 cells Mice were implanted subcutaneously into the flanks of fe- were stained for PD-L1 using FITC-conjugated anti-human male NSG mice with 10 × 10 freshly cultured HCC827 PD-L1 commercial antibody (clone MIH1, BD Biosciences), tumor cells (ATCC). When tumors reached volumes of ~ 3 6 Alexa Fluor® 488-conjugated LY3300054, or appropriate 300 mm (~ 4–5 weeks), 2.5 × 10 expanded human T isotype controls. NCI-H292, HCC827, and OV79 cells were cells were administered intravenously (IV) and mice were stained separately for HLA Class I expression using an treated with weekly IP injections of human IgG1 or APC-conjugated antibody (clone W6/32, RnDSystems, LY3300054 at 10 mg/kg for a total of four doses. Minneapolis, MN) Samples were collected on a 5-laser Established xenograft tumor models in CD34 hHSC- Fortessa X-20 cytometer (BD Biosciences) and analyzed engrafted mice: Cord blood derived CD34 hHSC with FlowJo V10 software (TreeStar). transplanted NSG mice were implanted subcutaneously with serially passaged HCC827 tumor fragments (4–5mm In vivo models in diameter) at 15–17 weeks of age. When the tumors All animal studies were approved by the Institutional reached volumes of approximately 200 mm (~ 30 days), Animal Care and Use Committee and performed in ac- weekly IP treatments of human IgG1 or LY3300054 at cordance with current regulations and standards of the 10 mg/kg began for a total of three doses. Fetal liver-der- United States Department of Agriculture and the ived CD34 hHSC transplanted NOG mice were National Institute of Health. All experiments with adop- implanted subcutaneously with serially-passaged OV79.F- tively transferred human PBMC or expanded human T Fluc2A–gfp tumor fragments (4–5mmindiameter) at scid tm1Wjl cells utilized NOD.Cg-Prkdc Il2rg /SzJ (NSG) ani- 15–17 weeks of age. OV79.FFLuc-2A-gfp tumor cells are mals (6–7 weeks of age, female, from Jackson Laborator- an ovarian carcinoma line transduced with lentivirus en- ies, Bar Harbor, MN), and were maintained in a 12 h coding firefly luciferase and green fluorescent protein light/dark cycle facility under pathogen-free conditions from a bicistronic transcript [28] and will be hereafter be in microisolator cages with standard laboratory chow referred to as OV79. When tumor volumes reached ~ + 3 and water ad libitum. Cord blood-derived CD34 150 mm (18 days), weekly IP treatments of human IgG1 hematopoietic stem cell (HSC) engrafted mice used for or LY3300054 at 10 mg/kg began for a total of four doses. scid tm1Sug the OV79 model utilized NOD.Cg-Prkdc Il2rg / JicTac animals (NOG, 15–17 weeks of age, female) and Immune phenotyping of peripheral blood from were obtained from Taconic BioSciences (Rensselaer, tumor-bearing mice in humanized models NY). Fetal liver derived CD34 HSC transplanted mice Peripheral human immune cell engraftment and pheno- used for the HCC827 model in NSG background type was assessed using Trucount™ tubes according to Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 6 of 14 manufacturer’s instruction (BD Biosciences, San Jose, superfamily, such as PD-L2, B7–1, B7–2, PD-1, CD28, CA). Briefly, 50 μl of blood from hHSC-transplanted TIGIT, TIM3, or VISTA (data not shown). mice (day 18, pre-treatment; day 34, after three treat- Biacore-based surface plasmon resonance analysis was ment doses; day 46, after four treatment doses), was performed to evaluate the affinity of LY3300054 for bind- added to the tubes and stained with antibodies against ing to Fc-tagged PD-L1. These analyses revealed an − 11 6 − 1 − 1 human CD45-FITC (BD Biosciences), human CD3- affinity of 8.19 × 10 M(k , = 1.40 × 10 M s ;k on off − 4 − 1 BV786 (Biolegend), human CD4-BV650 (BD Biosci- =1.14 ×10 s ). LY3300054 showed cross-reactivity to ences), human CD8-BV605 (Biolegend, San Diego, CA), cynomolgus PD-L1 with a similar affinity (K 1.22 × 10 10 6 − 1 − 1 − 4 − 1 and human PD-1-PEeFluor610 (eBiosciences, San Diego, ;k =1.51 × 10 M s ;k =1.84 ×10 s ), but on off CA) cell surface markers. Samples were subsequently not with the murine or rat PD-L1. fixed and collected on a 5-laser Fortessa X-20 cytometer To assess ligand blocking properties of LY3300054, solid (BD Biosciences) and analyzed with FlowJo V10 software phase blocking ELISA assays were performed. LY3300054 (TreeStar). Briefly, approximately 5000 fluorescent beads blocked PD-L1 binding to both PD-1 and CD80 ligands in were collected and enumerated. Human CD45 cells a concentration-dependent manner, with IC50 of 0.95 nM were also gated and enumerated, followed by subsequent and 2.4 nM, respectively (Fig. 1d, e). + + + gating on CD3 cells, followed by CD4 cell and CD8 To evaluate the ability of LY3300054 to bind to PD-L1 cell gating and enumeration, and finally PD-1 express- physiologically expressed on the surface of cells we per- ing cells were identified using appropriate IgG control. formed flow cytometry analyses on the tumor cell lines The absolute number of T cells and CD4+ and CD8+ with known surface PD-L1status. For these studies we subsets were calculated based on relative beads collected employed the NCI-H292, and HCC827 tumor cell lines compared to total number provide by manufacturer. evaluated in the in-vivo studies described below, as well Statistical analysis for human T cell engraftment and as the PD-L1-negative muscle rhabdomyosa cell line phenotype was performed using a two-way ANOVA on A204 (ATCCCRL-7900), and stained with either the repeated measurements. commercially available anti-PD-L1 antibody M1H1or with alexa647 fluor-conjugated LY3300054; as shown in Gene expression analysis of tumor and peripheral tissues Additional file 1: Figure S1, the PD-L1 –positive NCI- in humanized tumor models H292 and HCC827 stained robustly with either M1H1 Total RNA was isolated from snap-frozen tumor tissue or LY3300054, with the PD-L1 negative A204 cell line (day 15 from H292 model and day 15 post T cell infusion failed to stain with either reagent.. Finally, the OV79 cell from HCC827 tumor model) or from snap frozen white line employed for the in-vivo studies also stained posi- blood cell pellets, spleens, or bone marrow (hHSC- tive for PD-L1 (data not shown). engrafted models), using the MagMAX 96 Total RNA isolation (Life Technologies, Carlsbad, CA) and RNeasy Position N63 on human PDL-1 is a specificity anchor for mini (Qiagen, Hilden, Germany) kits, respectively. LY3300054 For QuantiGene Plex analysis, 500 ng of total RNA Since LY3300054 binds human PD-L1 but not human from tumor tissues were subjected to a custom-designed PD-L2, or murine and canine PD-L1, we performed multiplex assay (targets are listed in Additional file 2) ac- sequence alignments across each of these proteins to iden- cording to manufacturer (Affymetrix, Santa Clara, CA) tify key linear residues which might contribute to the protocol. For nCounter analysis, 100 ng of total RNA specificity of LY3300054 for human PD-L1. The multiple from white blood cells were analyzed with the Human sequence alignment analysis suggested that residues Immunology v2 (targets are listed in Additional file 2) 59–72 of the PD-L1 sequence (59-MEDKNIIQFVHGEE- nCounter codeset following manufacturer recommenda- 72) contribute to the human specificity of LY3300054, tions (NanoString Technologies, Seattle, WA). One- or since this sequence is missing in its entirety from the two-way ANOVA was used for statistical analysis. otherwise homologous sequence of human PD-L2 se- quence and is an area of relative divergence across the Results three tested species (Fig. 2a). In particular, we considered Binding and blocking properties of LY3300054 that positions 63 and 69 - side chains, which are exposed ELISA binding assays were performed to assess the se- to solvent according to PDB: 5C3T [16], might be pivotal lective binding and blocking properties of LY3300054. for the species specificity of LY3300054 because their cor- While LY3300054 bound to human and cynomolgous responding amino acid substitutions diverge among the PD-L1 with similar affinities (EC50 of 0.075 nM and three PD-L1 sequences (marked with * in the alignment of 0.085 nM, respectively) (Fig. 1a, b), LY3300054 did not Fig. 2a). We pursued two mutational strategies to test this bind to murine PD-L1 (Fig. 1c); furthermore, LY3300054 hypothesis. The first strategy focused on rescuing the did not bind to other proteins of the immunoglobulin binding of LY3300054 to canine PD-L1 by introducing Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 7 of 14 Fig. 1 Binding and blocking properties of LY3300054. Panels a-c: 96-well plates were coated with recombinant human (a), cynomolgus (b), or murine (c) PD-L1-Fc fusion protein (100 ng/well each). Bound LY3300054 was detected using HRP-conjugated anti-human Fab antibody and addition of chromogenic substrate (OD at 450nm). 96-well plates were coated with 100 ng/well of recombinant PD-1 (d) or B7-1 protein (e), then incubated with a mixture of biotin-conjugated PD-L1 and either LY3300054 or human IgG1 antibodies. Plate bound PD-L1 was detected using HRP-conjugated streptavidin and addition of chromogenic substrate (OD at 450 nm). In all experiments, each data point is the average of two replicates. Data (a-e) are representative of multiple independent experiments ab Fig. 2 Identification of LY3300054 epitope residues in human PD-L1. Panel a: CLUSTALW multiple sequence alignment of domain 1 of human (hu), canine (ca), and murine (mu) PD-L1 and hu-PD-L2 to identify the LY3300054 species specificity anchors on hu-PD-L1. Underlined is the human PD-1 6Å binding site on hu-PD-L1 (according to PDB: 4ZQK (26602187)). An alignment position is marked with (*) if both mu-PD-L1 and ca-PD-L1 substitutions differ from the hu-PD-L1 sequence. An alignment position is marked with (:) if either the mu-PD-L1 or ca-PD-L1 substitution differs from the hu-PD-L1 sequence. Panel b: Position N63 on human PD-L1 is a specificity anchor for LY3300054. Canine-to-human mutation K63N (▲) rescues the ELISA binding of LY3300054 to canine PD-L1. Like wild type ca-PD-L1-Fc (●), canine-to-human mutant N69H (△) does not bind LY3300054 Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 8 of 14 canine-to-human mutations at positions 63 and 69 of the IFNγ and IL2 secretion was substantially enhanced in canine PD-L1-Fc. As shown in Fig. 2b, only the variant the combination treatment compared to each of the sin- K63 N and not N69H rescued the binding of LY3300054 gle agents (Fig. 4a). High-content gene expression to ca-PD-L1-Fc. Notably, neither mutations (K63 N or analysis revealed overlapping gene expression changes N69H) compromise the structural integrity of the ca-PD- across all treatment groups, and also, in agreement with L1-Fc protein since the Size Exclusion Chromatography previous reports on the combination of PD-1 and ipilu- (SEC) profiles of both variants were identical to the SEC mimab therapy ([29]), distinct gene profiles in the com- profile of the wild type ca-PD-L1-Fc in Additional file 1: bination group, with the maximum treatment effect Figure S2. The second mutational strategy focused on ab- observed at 67.5 nM (Fig. 4b). LY3300054 single agent rogating the binding of LY3300054 to human PD-L1 by treatment induced gene expression changes indicative of introducing human-to-murine mutations at positions 63 immune activation, exemplified by increased expression and 69 in human PD-L1. Only N63Q and not H69A abro- of IFNG, IL2, IDO1, GZMB, IL1B, IL6, while ipilimumab gated the binding of LY3300054 to hu-PD-L1-Fc single agent treatment resulted in enhanced T cell Additional file 1: Figure S3. Thus, both mutational strat- activation, exemplified by enhanced ICOS, and IFNG ac- egies demonstrated that the N63 residue is pivotal for the companied by downregulation of myeloid genes (CD68, species specificity of LY3300054. Furthermore, analysis of CD14, HLA-DRA). The combination of LY3300054 and the co-crystal structure of the human PD-1/PD-L1 co- ipilimumab further upregulated T-cell specific genes crystal indicates that the N63 residue is part of the 6Ǻ reflecting a Th1 response (IFNG, IL2, TBX21), T-cell ac- binding site of PD-1 on PD-L1 ([16]). Hence, LY3300054 tivation (IL2, IFNG, ICOS) and downregulation of blocks the PD-1/PD-L1 interaction because its binding epi- myeloid genes (CD68, CD14, HLA-DRA). tope overlaps the binding site of PD-1 on PD-L1. LY3300054 ADCC and CDC functions Functional activity of LY3300054 in vitro LY3300054 was engineered to ablate Fc-gamma receptor The ability of LY3300054 to enhance T cell functional engagement and associated immune effector functions, activity was assessed using a variety of in vitro assays. In specifically ADCC and CDC. LY3300054 was evaluated a PD-1 reporter assay, using Jurkat cells engineered to by SPR and solid phase ELISA to lack binding to FcgRI, stably express human PD-1 and an NFAT-luciferase re- FcgRIIa, and FcgRIIIa (F158) within the limit of porter construct and CHO-K1 cells engineered to stably detection titrated to 10 μM antibody concentration (data express human PD-L1, addition of LY3300054 resulted not shown). Ablation of ADCC and CDC functions of in a concentration-dependent increase in NFAT-driven LY3300054 was evaluated in cell-based assays, using the expression of luciferase, overcoming the inhibitory ef- HEL PD-L1 positive tumor cell lines. In both ADCC and fects of PD-L1 expressed by CHO cells (Fig. 3a). In CDC assays LY3300054 did not direct detectable effector mixed leukocyte reactions (MLR) using allogeneic hu- function activity against HEL target cells, while man DC and T cells, addition of LY3300054 enhanced rituximab was shown to mediate significant ADCC and the allogeneic T cell response in a dose-dependent man- CDC response against CD20-positive Wil2-S cells ner, with activity observed at concentrations as low as (Additional file 1: Figure S4A, B). 0.05 nM as measured by IFN-γ secretion and mRNA ex- pression (Fig. 3b, e). Additional analysis of MLR cultures by 41-plex microbead-based cytokine and gene expres- LY3300054 does not trigger non-specific cytokine sion panels revealed enhanced secretion and transcrip- production by PBMC tion of multiple immune factors in response to We evaluated the ability of LY3300054 to result in LY3300054 treatment exemplified by increased levels of non-specific cytokine release from unstimulated IL-6, TNFα, MCP-1, MIP-1α,MIP-1β, RANTES, IP-10, human PBMC using plate-bound cytokine release as- IL-4, IL-13, IL-12 in culture medium (Fig. 3d) and up- says. While incubation of PBMC with anti-CD3ε or regulation of IL2, IL1B, IL21 genes (Fig. 3e). Finally, the anti-CD28 agonist antibody TGN1412 antibodies LY3300054 was also shown to enhance T cell activity in resulted in robust cytokine production for a number the tetanus-toxoid recall assay (TTRA) which measures of cytokines, including those associated with cytokine the ability to stimulate antigen-specific memory T cells releasesyndrome(CRS) (TNFα,IL-6,IL-2,IFNγ,and in PBMC. (Fig. 3c). IL-1β), incubation of donor PBMC with LY3300054 In MLR assays, LY3300054 also displayed biological did not result in significant levels of cytokine release activity in combination with the anti-CTLA4 antibody for any of the evaluated cytokines over a broad range ipilimumab. In these experiments, which utilized equi- of concentrations from 0.003 to100 μg/ml molar concentrations of LY3300054 and ipilimumab, (Additional file 1:FigureS5). Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 9 of 14 Fig. 3 LY3300054 enhances T cell activation in vitro. Panel a: Jurkat-NFAT reporter assay: Each data point represents the average of two technical replicates, with error bars representing the SD. Data are representative of three independent experiments. Panel b: Mixed leukocyte reactions. Supernatants were measured for IFN-γ production by ELISA. Each data point represents the average of 8 replicates, with error bars representing the SEM. Data are representative of multiple experiments and donor T cells/DC pairs. Panel c: Tetanus toxoid recall assay: Supernatants were measured for IFN-γ production by ELISA. Each data point represents the average of 4 replicates, with error bars representing the SD. Data are representative of two experiments with PBMC obtained from different donors. Panels d and e: Gene expression analysis of cell lysate (E) and cytokine level analysis of cell culture supernatant (F) from the mixed leukocyte reactions using QuantiGene Plex and microbead-based immunoassay panels, respectively. Volcano plots show Log2 fold change of gene expression (E) or cytokine levels (F) in the LY3300054 treated group compared to control group. The highlighted circles correspond to differentially expressed genes (DEG) or cytokines that display fold change > 1.7 (black solid vertical line) and p value < 0.05 (horizontal dotted line). Circle sizes are proportional to the level of expression in LY3300054 group. One-way ANOVA was used for statistical analysis. Human IgG1 LY3300054 Biological activity of LY3300054 in humanized murine model, we evaluated the effects of LY3300054 on anti- models in vivo tumor alloreactivity, and also performed detailed intra- We evaluated the functional activity of LY3300053 in tumoral and peripheral immune pharmacodynamic vivo using human tumor xenograft models and immune assessments to evaluate how LY3300054 therapy modu- deficient NSG animals reconstituted with human im- lated human immune cell activities. Notably, each of the mune cells. We performed these studies in different tumor cell lines evaluated in these studies robustly ex- models, including a preventative (co-implantation) presses HLA class I (Additional file 1: Figure S6). model, a therapeutic model with established tumors and For the preventative studies, mice were co-implanted animals reconstituted with allogeneic human T cells, and with a mixture of human PBMC and NCI-H292 tumor two therapeutic models with established tumors and ani- cells, followed by treatment with LY3300054 or control mals engrafted with human sHSC), each designed to human IgG1. Compared to untreated and human IgG1- evaluate different functional attributes of anti-PD-L1 im- treated animals, treatment with LY3300054 resulted in munotherapy. Each of these models evaluates the ability significant tumor inhibition (p <0.001) (Fig. 5a). The to modulate the inherent alloreactivity of the engrafted therapeutic potential of LY3300054 to modulate T cell- human immune system against the tumor. In each mediated anti-tumor activity in an established tumor Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 10 of 14 Fig. 4 Combination of LY3300054 and ipilimumab enhances T cell activation in vitro. Panel a: Mixed leukocyte reactions. Allogeneic DC were co-cultured with purified CD4 T cells for 72 h in the presence of increasing (two-fold increments) concentrations of LY3300054, ipilimumab or a combination of both antibodies ranging from 0.0003 to 67 nM. Supernatants were measured for IFN-γ and IL-2 production by ELISA. Each data point represents the average of 8 replicates, with error bars representing the SEM. Data were generated with four different PBMC donors. Panel b: Gene expression analysis of cell lysate from the mixed leukocyte reactions was performed using QuantiGene Plex assay. Venn diagram showing the number of shared (overlap circle) and treatment-specific (no overlap) DEGs across the different treatments. Tables list the Log2 fold-change of LY3300054 vs control group for genes with fold-change > 1.5, p value < 0.05. One-way ANOVA was used for statistical analysis setting was evaluated using the HCC827 xenograft mouse To evaluate the intra-tumoral and peripheral immune- tumor model, and animals reconstituted with ex-vivo ex- pharmacodynamic effects of LY3300054 therapy on panded CD3+ T cells. While infusion of expanded human engrafted human immune cells, tissues (tumor, spleen, T cells alone modestly delayed tumor growth, reflecting bone marrow, peripheral blood) were collected from the the baseline anti-tumor reactivity of the engrafted allor- above models and subjected to mechanistic analyses. In eactive T cells, treatment with LY3300054 significantly en- CD34+ hHSC-reconstituted, tumor-bearing models, hanced this effect resulting in potent anti-tumor activity treatment with LY3300054 resulted in an increase of the (Fig. 5b). The therapeutic potential and activity of absolute number of human T cells, an enhanced CD8/ LY3300054 in the context of more completely human CD4 T cell ratio, and increased frequency of PD1+ CD8 immune-replete animals was assessed in immunodeficient + and CD4+ T cells, indicative of T cell activation NSG or NOG mice engrafted with HSCs of human origin (Fig. 6a-d). Furthermore, LY3300054 therapy resulted in (CD34+ huHSCs), and two xenograft mouse tumor prominent immune-related gene expression changes models, using the HCC827 and OV79 tumor cell lines. consistent with IFNγ pathway and T cell activation in These experimental models display a more replete human the tumor tissue (Fig. 6e), as well as in spleen immune compartment exemplified by differentiation of (Additional file 1: Figure S7A) and peripheral blood cells both lymphoid and myeloid cells ([30]). In both models (Additional file 1: Figure S7B), and to a much lesser ex- LY3300054 therapy strongly enhanced the alloreactive tent in the bone marrow (Additional file 1: Figure S7C). anti-tumor response (Fig. 5c-d). In tumor tissue of CD34+ hHSC-reconstituted mice, Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 11 of 14 Fig. 5 LY3300054 demonstrates anti-tumor efficacy in xenograft tumor models reconstituted with human immune cells. Antibody treatment (indi- cated by red arrows), either human IgG or LY3300054, was delivered by intra-peritoneal injection in each case at 10 mg/kg. Tumor growth was monitored by caliper, and results are represented as a geometric mean of tumor volumes ± SEM. Panel a: Co-implantation model: NCI-H292 tumor cells and freshly isolated human PBMC were co-implanted subcutaneously into the flanks of NSG mice. n = 8 for all groups. Panel b: Estab- lished tumor model: HCC827 tumor cells were implanted subcutaneously into the flanks of NSG mice. When tumors reached volumes of ~ 300 mm (approximately 5 weeks), mice were infused with previously expanded human T cells (black arrow). n = 8 for all groups. Panels c and d: Established tumor models in CD34+ hHSC-reconstituted animals: HCC827 tumors (NSG, panel |C) or OV79 tumors (NOG, panel D). Mice were im- planted subcutaneously with either HCC827 or OV79 tumor fragments at ~ 15–17 weeks of age (~ 13–15 weeks post HSC engraftment). HCC827 3 3 tumors were allowed to grow to ~ 200 mm (4 weeks) and OV79 were allowed to grow to ~ 150 mm (18 days) before starting weekly treat- ments of either human IgG1 or LY3300054 at 10 mg/kg. n =5–9 per group. Statistically significant difference is indicated* (two-way repeated measurement ANOVA, RM-ANOVA) LY3300054-induced increased gene expression changes of expression were also apparent, suggesting tissue- and reflected T cell infiltration and activation (CD3E, CD8B, tumor-specific effects of LY3300054 (Fig. 6f). Further- CD4, PRF1, GZMB, TBX21, EOMES), myeloid cell infil- more, LY3300054 also upregulated genes indicative of T tration and differentiation (ITGAM, ITGAX, CD14, cell activation (upregulated CD274, PDCD1LG2, IDO1, CD68, ARG1) upregulation of co-inhibitory/co-stimula- CXCL9, CXCL10, CD3E, CD8B, CD4, ICOS, CD27 etc) tory receptors and ligands (TNFRSF9, TNFRS18, in tumor tissues collected from the established HCC827 TNFRSF4, CD28, CD27, ICOS, CD226, CD200R1, tumor model implanted with ex vivo expanded human T PDCD1, CD274, PDCD1LG2, TIGIT, HAVCR2, LAG3), cells and the co-implantation NCI-H292 model, al- cytokines and their receptors (IFNG, IL2, IL2RA, IL21, though the overall effect of the antibody was less pro- CCL3, CCL4, CCL5), interferon type I response (IFNA2, nounced in these models compared to HCC827-bearing IFNB1), antigen presentation and MHC class I and II CD34+ hHSC-reconstituted NSG mice (Additional file 1: (HLA-A, HLA-B, HLA-C, B2M, HLA-DRA), and down- Figure S8). regulation of only two genes from the tested panel, TGFB2 and IL1B (Fig. 6e). Three genes indicative of T Discussion cell activation CD274 (PD-L1), CCL5 (RANTES), and We describe the discovery and pre-clinical characterization ITGAL (LFA-1) were upregulated in all 4 tissues evalu- of LY3300054, an Fc effector function-silenced fully human ated, demonstrating a systemic effect of PD-L1 blockade antibody that binds to the PD-L1 checkpoint ligand and in HCC827-bearing CD34+ HSC-reconstituted NOG blocks interactions with both PD-1 and CD80 receptors. mice. While the gene profiles between tumor, spleen, LY3300054 is shown to lack ADCC and CDC effector func- and peripheral blood (the three tissues with robust gene tions, and to not trigger TCR-independent T cell activation modulation) showed patterns of overlap, unique patterns as assessed by non-specific cytokine production by human Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 12 of 14 Fig. 6 LY3300054 enhances peripheral T cell engraftment and activation and induces T cell inflamed phenotype in tumor tissues of CD34+ hHSC- engrafted NOG and NSG mice. Panels a, b, c, d: Blood from OV79-bearing CD34+ hHSC-engrafted NOG mice was analyzed for human T cell engraftment and phenotype using TruCount tubes on day 18 (pre-dose), day 34 (after three antibody doses), and day 46 (after four antibody doses). Peripheral T cell engraftment (A); CD8:CD4 ratio (B); PD-1 expression in CD4+ (C) and CD8+ T cells (D) cells. Results are represented as a geometric mean of engraftment + SEM with n = 9 mice on day 18 and day 34, and n = 5 mice on day 46. Two-way repeated measurements ANOVA was used for statistical analysis. Panel e: Gene expression analysis of tumor sample was performed using QuantiGene Plex assay. Volcano plots show Log2 fold-change of gene expression in the LY3300054 treated group compared to control group. The highlighted circles correspond to differentially expressed genes that display fold change > 1.7 (black solid vertical line) and p value < 0.05 (horizontal dotted line). Circle sizes are proportional to the level of expression in LY3300054 group. One-way ANOVA was used for statistical analysis. Panel f: Venn diagram showing the number of shared and tissue-specific DEGs (LY3300054 vs human IgG treatment) across various tissue types. Table on the right lists shared DEGsacross various tissues with fold-change > 1.7, p value < 0.05 for LY3300054 vs control PBMC in vitro. From a structural perspective, using se- setting. The paucity of information is related at least in quence analysis and mutagenesis we identify a key residue part to the relatively recent availability of humanized on PD-L1, N63, which is part of the binding site for PD-1, murine tumor models with reconstituted human im- may play important roles in the function of the PD-1/PD- mune cell compartment(s). We chose to thoroughly L1 interaction, and which contributes to the species specifi- evaluate the functional activity of LY3300054 in a variety city of LY3300054. In vitro LY3300054 is shown to block of immune-humanized mouse models, to begin to PD-L1-mediated T cell suppression of primary human T understand how anti-PD-L1 therapy might modulate cells in both primary MLR and tetanus-toxoid recall assays, anti-tumor T cell immunity in biologically complex and and to reverse TCR-engagement mediated activation of the relevant matrices, more representative of the clinical set- NFAT pathway using signal reporter Jurkat cells with ting. These models represented both prophylactic and ectopic expression of human PD-1. In both preventative therapeutic intervention, reconstituted with human and therapeutic xenograft tumor models reconstituted with PBMC, expanded human T cells, or CD34 HSC. In each human T cells or HSC, LY3300054 therapy resulted in of the tested models, LY3300054 treatment resulted in robust anti-tumor activity, accompanied by the develop- enhanced anti-tumor alloreactivity and robust anti- ment of distinct T cell inflamed signatures in the tumor tumor effects, demonstrating the potency of the agent as and peripheral tissues. well as the relevance of the PD-1/PD-L1 axis in the Although multiple agents that block the PD-1/PD-L1 context of humanized mouse models. axis have been described and evaluated in the clinic, We took advantage of the ability to collect tumor and relatively little information exists about the functional tissues from the animal models and performed high con- properties of these molecules in the pre-clinical in vivo tent flow cytometry and molecular immuno- Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 13 of 14 pharmacodynamic profiling to obtain insights about the Acknowledgements We thank Dr. Vera Molkenthin and her team at AbCheck for phage library mechanism of action of LY3300054-mediated anti-PD- antibody screening, members of Biologics Technology, Quantitative Biology L1 blockade in each of the humanized models. These and Non-Clinical Safety groups (Marshall Snavely, Krishnadatt Persaud, analyses clearly demonstrated the ability of LY3300054 Yongqing Jia, Liz Navarro, Sagit Hindi, Prabakaran Naryanasamy, Michele Iacolina, Ana Cardoso, Nicole Covino, Heather Griffith, Julia Lee, Nada Alakhras) to modulate human T cell functions as reflected by in- for technical assistance; Thompson Doman, Cindy Wang and Jason Manro for creased peripheral T cell numbers and in particular CD8 computational and statistical analysis, Yung-Mae Yao for helpful input and T cells, and enhanced activation status of peripheral discussions, Tim Bailey for antibody conjugations, Yvonne Vandenburg for project management; Mythili Koneru for insightful discussions, and Carl June allo- and/or xeno-reactive T cells triggered through TCR (University of Pennsylvania) for providing the OV79 cell line. engagement. High-content gene expression profiling re- vealed that LY3300054-induced a T cell inflamed pheno- Authors’ contributions Conception and design: YL, DLL, RN, MK. Development of methodology: YL, type in tumor tissues across all models tested, and GW, CC, JNH, GH, RN. Acquisition of data: YL, GW, LS, MM, CC, DS, AF, MM, almost exclusively resulted in upregulation of gene ex- AS, DC, II, DB, AP, and LM. Analysis and interpretation of data (e.g., produce pression; these data demonstrate the ability of figures, statistical analysis, biostatistics, computational analysis): YL, GW, LS, CC, DS, AF, JNH, RN and MK. Writing, review, and/or revision of the LY3300054 to effectively block the PD-L1/PD-1 axis and manuscript: YL, GW, CC, RN, JNH, MK. Study supervision: YL, DL, RN, MK. to activate T cells to drive more effective anti-tumor T All authors read and approved the final manuscript. cell immunity. Within a model, while a core set of differ- Competing interests entially expressed genes were shared across tumor and LY3300054 is an Eli Lilly and Company investigational drug currently in clinical normal tissues including spleen, peripheral blood and testing. All authors are or have been employees and shareholders of Eli Lilly bone marrow, a considerable number of genes were up- and Company. regulated in a tissue-specific manner suggesting that LY3300054 activity might be context-dependent. Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in In more immune-replete animals, blockade of the published maps and institutional affiliations. PD-L1/PD-1 axis activated additional pathways beyond T lymphocytes, including genes associated with adaptive im- Author details Lilly Research Laboratories, Department of Cancer Immunobiology, New munity such as co-stimulatory and co-inhibitory receptors, York, NY, USA. Lilly Research Laboratories, Department of Preclinical cytokines and transcription factors, and innate immune 3 Pharmacology, New York, NY, USA. Lilly Research Laboratories, Department pathways such as interferon, MHC, and myeloid pathways. of Biologics Technology, New York, NY, USA. Lilly Research Laboratories, Department of Non-Clinical Safety, Indianapolis, IN, USA. Lilly Research Whether these observations reflect a direct effect of PD-L1 Laboratories, Department of Quantitative Biology, New York, NY, USA. Eli blockade on innate immune cells or an indirect effect 7 Lilly and Company, 450 East 29th Street, New York, NY 10016, USA. Janssen resulting from T cell activation,theyunderscorethe diverse Pharmaceutical Companies of Johnson and Johnson, Springhouse, PA, USA. and to-date not fully appreciated role for PD-L1 blockade Received: 14 December 2017 Accepted: 20 February 2018 to modulate the activity and function of cell types beyond T lymphocytes, and highlight the need to both integrate References and evaluate innate immune modulation and cell subsets in 1. Baxter AG, Hodgkin PD. Activation rules: the two-signal theories of immune the context of anti-PD-L1-based immunotherapies. activation. Nat Rev Immunol. 2002;2(6):439–46. 2. Chen L, Flies DB. Molecular mechanisms of T cell co-stimulation and co-inhibition. Nat Rev Immunol. 2013;13(4):227–42. Conclusion 3. Yearley JH, et al. PD-L2 expression in human tumors: relevance to anti-PD-1 LY3300054 is currently in phase I clinical studies to therapy in cancer. Clin Cancer Res. 2017;23(12):3158–67. evaluate activity as monotherapy and in combination 4. Francisco LM, Sage PT, Sharpe AH. The PD-1 pathway in tolerance and autoimmunity. Immunol Rev. 2010;236:219–42. with other agents. 5. Okazaki T, Honjo T. The PD-1-PD-L pathway in immunological tolerance. Trends Immunol. 2006;27(4):195–201. 6. Sharpe AH, Pauken KE. The diverse functions of the PD1 inhibitory pathway. Nat Rev Immunol. 2017;18:153–67. Additional files 7. Lin DY, et al. The PD-1/PD-L1 complex resembles the antigen-binding Fv domains of antibodies and T cell receptors. Proc Natl Acad Sci U S A. 2008; 105(8):3011–6. Additional file 1: Figure S1. PD-L1 expression on tumor cell lines. 8. Hunig T. The storm has cleared: lessons from the CD28 superagonist Figure S2. SEC profiles of dog PD-L1-Fc and its dog-to-human variants K63 TGN1412 trial. Nat Rev Immunol. 2012;12(5):317–8. N and N69H. Figure S3. Position N63 on human PD-L1 is a specificity 9. Noman MZ, et al. PD-L1 is a novel direct target of HIF-1alpha, and its anchor for LY3300054. Figure S4. LY3300054 does not induce ADCC or blockade under hypoxia enhanced MDSC-mediated T cell activation. J Exp CDC in vitro. Figure S5. LY3300054 does not induce non-specific cytokine Med. 2014;211(5):781–90. production by human PBMCs in vitro. Figure S6. HLA Class I expression 10. Watanabe R, et al. Pyruvate controls the checkpoint inhibitor PD-L1 and profile on tumor cell lines. Figure S7. LY3300054 enhances expression of suppresses T cell immunity. J Clin Invest. 2017;127(7):2725–38. immune-related genes in peripheral tissue of CD34+ HSC transplanted mice 11. Callahan MK, Postow MA, Wolchok JD. Targeting T cell co-receptors for bearing OV79 tumors (NOG). Figure S8 LY3300054 enhances expression of cancer therapy. Immunity. 2016;44(5):1069–78. immune-related genes in NCI-H292 and HCC827 tumors from humanized 12. Hamanishi J, et al. Programmed cell death 1 ligand 1 and tumor-infiltrating mouse tumor models. (PPTX 1085 kb) CD8+ T lymphocytes are prognostic factors of human ovarian cancer. Proc Additional file 2: Supplementary Methods. (DOCX 34 kb) Natl Acad Sci U S A. 2007;104(9):3360–5. Li et al. Journal for ImmunoTherapy of Cancer (2018) 6:31 Page 14 of 14 13. Okazaki T, et al. A rheostat for immune responses: the unique properties of PD-1 and their advantages for clinical application. Nat Immunol. 2013;14(12):1212–8. 14. Taube JM, et al. Colocalization of inflammatory response with B7-h1 expression in human melanocytic lesions supports an adaptive resistance mechanism of immune escape. Sci Transl Med. 2012;4(127):127ra37. 15. (FDA), F.a.D.A. Hematology/Oncology (Cancer) Approvals & Safety Notifications. 2017. Available from: https://www.fda.gov/Drugs/ InformationOnDrugs/ApprovedDrugs/ucm279174.htm. 16. Zak KM, et al. Structure of the complex of human programmed death 1, PD-1, and its ligand PD-L1. Structure. 2015;23(12):2341–8. 17. Horita S, et al. High-resolution crystal structure of the therapeutic antibody pembrolizumab bound to the human PD-1. Sci Rep. 2016;6:35297. 18. Na Z, et al. Structural basis for blocking PD-1-mediated immune suppression by therapeutic antibody pembrolizumab. Cell Res. 2017;27(1):147–50. 19. Lee JY, et al. Structural basis of checkpoint blockade by monoclonal antibodies in cancer immunotherapy. Nat Commun. 2016;7:13354. 20. Liu K, et al. Structural basis of anti-PD-L1 monoclonal antibody avelumab for tumor therapy. Cell Res. 2017;27(1):151–3. 21. He YF, et al. Blocking programmed death-1 ligand-PD-1 interactions by local gene therapy results in enhancement of antitumor effect of secondary lymphoid tissue chemokine. J Immunol. 2004;173(8):4919–28. 22. Hirano F, et al. Blockade of B7-H1 and PD-1 by monoclonal antibodies potentiates cancer therapeutic immunity. Cancer Res. 2005;65(3):1089–96. 23. Iwai Y, et al. Involvement of PD-L1 on tumor cells in the escape from host immune system and tumor immunotherapy by PD-L1 blockade. Proc Natl Acad Sci U S A. 2002;99(19):12293–7. 24. Burova E, et al. Characterization of the anti-PD-1 antibody REGN2810 and its antitumor activity in human PD-1 knock-in mice. Mol Cancer Ther. 2017; 16(5):861–70. 25. Cong M, et al. Reporter bioassays to assess therapeutic antibodies in development for immunotherapy programs. Madison: Promega Corporation; 2015. 26. Parekh BS, et al. Development and validation of an antibody-dependent cell-mediated cytotoxicity-reporter gene assay. MAbs. 2012;4(3):310–8. 27. Stebbings R, et al. “Cytokine storm” in the phase I trial of monoclonal antibody TGN1412: better understanding the causes to improve preclinical testing of immunotherapeutics. J Immunol. 2007;179(5):3325–31. 28. Bertozzi CC, et al. Multiple initial culture conditions enhance the establishment of cell lines from primary ovarian cancer specimens. In Vitro Cell Dev Biol Anim. 2006;42(3–4):58–62. 29. Das R, et al. Combination therapy with anti-CTLA-4 and anti-PD-1 leads to distinct immunologic changes in vivo. J Immunol. 2015;194(3):950–9. 30. Shultz LD, et al. Humanized mice for immune system investigation: progress, promise and challenges. Nat Rev Immunol. 2012;12(11):786–98. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit

Journal

Journal for ImmunoTherapy of CancerSpringer Journals

Published: Apr 30, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off