Costimulatory T-cell engagement by PRS-342, a GPC3/4-1BB bispecific molecule, leads to activation of T cells AACR Annual and tumor growth inhibition in a HCC humanized mouse model Meeting 2019 Birgit Bossenmaier, Corinna Schlosser, Rachida-Siham Bel Aiba, Eva-Maria Hansbauer, Thomas Jaquin, Christian Barthels, Janet Peper, Markus Zettl, Benjamin Weiche, Thibaut Angevin, Michelle Yegres, Reno Winter, Stefan Grüner, Christine Rothe, Shane A. Olwill Abstract # 4302 Pieris Pharmaceuticals, Inc., 255 State Street, Boston, Massachusetts Pieris Pharmaceuticals, GmbH, Lise-Meitner Strasse 30, 85354 Freising, Germany Reporter Assay PRS-342 leads to tumor-localized increase of TILs Background PRS-342 reporter cell assay Pharmacokinetic profile of PRS-342 in mice in a humanized HCC xenograft model 4-1BB (CD137) is a key costimulatory immunoreceptor and a highly promising therapeutic • PRS-342 costimulated T cells in a Jurkat Nf-kB reporter cell assay only in the • Preliminary mouse PK was performed in male CD-1 mice to compare PRS-342 with an • FFPE embedded xenograft tumor were analyzed histologically (HE) and immuno- target in cancer. To overcome toxicity and efficacy limitations of current 4-1BB-targeting presence of GPC3-positive tumor cell lines. a-GPC3 antibody. histologically (IHC) for T-cell infiltration. antibodies, we have developed 4-1BB Anticalin®/tumor-targeting mAb bispecifics that • PRS-342 has a typical antibody like PK profile in mice comparable to the a-GPC3 • Tumor IHC staining for human CD3, CD4 and CD8 shows a dose-dependent increase activate T cells in a tumor localized fashion. We have previously reported on the generation antibody used as building block in the bispecific PRS-342 construct. in the frequency of human tumor associated T cells (TILs) for PRS-342 vs controls, and characterization of PRS-343, a clinical-stage 4-1BB/HER2 bispecific molecule, with suggesting tumor-localized T-cell activation. regard to preclinical proof-of-concept and basic drug-like properties (1). Here, we describe PK in CD1 mice the preclinical dataset for PRS-342, a 4-1BB/GPC3 bispecific based on the Anticalin® technology. GPC3 is an oncofetal protein with high tumor selectivity and high expression in B % TIL frequency hCD3, hCD4 and hCD8 by IHC (necrotic areas are excluded) not only hepatocellular carcinomas, but also in a variety of other tumors with high medical need. Intratumoral T-cell infiltration % TILs of total tumor area - necrotic area Anticalin® therapeutics are 18 kD proteins derived from human lipocalins. We utilized phage display to generate an Anticalin® protein binding to 4-1BB with high affinity and TILs %CD3 T cells %CD4 T cells %CD8 T cells specificity. The PRS-342 bispecific construct was generated by genetic fusion of the 4-1BB- Vehicle control 0.8 0.7 0.4 specific Anticalin® protein to a humanized high affinity GPC3-targeting monoclonal antibody • PRS-342 PRS-342 0.8 µM 6.2 4.7 4.5 • a-4-1BB antibody PRS-342 13 M 4.1 3.3 3.0 with an engineered IgG4 backbone. EC50 [nM] EC50 [nM] EC50 [nM] µ • a-GPC3 antibody PRS-342 has excellent drug-like properties and can be produced with high yields. PRS-342 Hep-G2 Hep-GB NCI-N87 PRS-342 67 µM 10.7 4.9 7.3 • Isotype control -GPC3 antibody 0.8 M 1.6 1.3 0.7 was designed to be strictly dependent on tumor binding, which is necessary for clustering a-4-1BB antibody 0.94 1.39 1.77 a µ a-GPC3 antibody 13 µM 0.8 0.6 0.5 of 4-1BB, to elicit 4-1BB costimulation and T-cell activation. This was confirmed using a-GPC3 antibody - - different in vitro T-cell costimulation assays based on mixed culture of human T cells and a-4-1BB antibody 13 µM 0.8 0.7 0.2 GPC3-expressing tumor cell lines. These data further demonstrate the ability of PRS-342 PRS-342 0.38 0.78 - to bind both targets simultaneously. PRS-342 was also evaluated for activity in a PRS-342 induces 4-1BB clustering and downstream signaling in a Jurkat Nf-kB reporter cell line in humanized HepG2 mouse xenograft model, with results supporting its differentiated MoA + the presence of GPC3-positive HepG2 and Hep3B cells with low nM EC50 values. GPC3-negative NCI C TIL frequency (hCD8 ) by IHC compared to relevant benchmark controls. –N87 cell are used as control. Only the 4-1BB antibody can activate the Jurkat Nf-kB reporter cell in Concept: tumor-specific and tumor-localized the absence of GPC3 positive tumor cells. costimulatory activation of T cells PRS-342 induces 4-1BB engagement and T-cell T cell costimulation activation in a GPC3 dependent manner Tumor cell Concept of costimulatory T-cell engagement by An analysis of the pharmacokinetic properties of PRS-342 as well as of an a-GPC3 antibody was in tumor PRS-342: Within a patient´s tumor, tumor-specific performed in mice. Male CD-1 mice approximately 5 weeks of age (2 mice per timepoint) were T cells are bridged with tumor cells by the • Pan T cells were coincubated with GPC3high Hep-G2, Hep-GB and MKN-45 cells and injected into a tail vein with a dose 2 mg/kg. Plasma samples from the mice were obtained at the Vehicle a-4-1BB mAb (3.9 mg/kg; 13 µM) costimulatory bispecific PRS-342 which PRS-342. timepoints of 5 min, 24 h, 168 h, and 336 h. Plots of the plasma concentration over time for the simultaneously binds the tumor target GPC3 and anti-GPC3 antibody and PRS-342 are shown. Both the antibody and the bispecific-construct show Costimulatory the immune receptor 4-1BB. The resulting • Supernatant concentrations were determined for IL-2. bispecific Tumor target typical antibody pharmacokinetics profiles. GPC3 MHC- clustering of 4-1BB -provides a local coactivatory peptide signal to the T cell, further enhancing its T-cell Hep-G2 (GPC3 positive) Clustering Hep-GB (GPC3 positive) MKN-45 (GPC3 negative) 2 5 0 0 2 5 0 0 Costimulatory T cell receptor receptor (TCR)-mediated activity and leading to 2 0 0 0 receptor tumor destruction. Toxic side effects are expected 2 0 0 0 ] ] L Tumor-specific 4-1BB L to be manageable, as PRS-342 does not induce m 1 5 0 0 m Signal 2 Signal 1 / PRS-342 leads to tumor growth inhibition in a 1 5 0 0 / g g T Cell p p [ [ clustering and activation of 4-1BB in the absence 1 0 0 0 2 2 [pg/ml] 1 0 0 0 2 - - - L L Activation I IL of target-positive cells, and healthy tissue is I humanized HCC xenograft model 5 0 0 spared by tumor costimulated T cells due to the 5 0 0 a-GPC3 mAb (3.9 mg/kg; 13 µM) PRS-342 (3.9 mg/kg; 13mM) 0 absence of a primary, TCR-mediated signal. 0 1 1 0 1 1 0 c o n c e n t r a t i o n [ n M ] c o n c e n t r a t i o n [ n M ] PRS-342 design, target binding and activity in • PRS-342 (B) FFPE Xenograft tumors taken from the in vivo study described in (A) were stained for HE (not • a-4-1BB antibody • Immunocompromised mice (NOG) engrafted with GPC3-positive tumor cells (HepG2) were injected with human PBMC and treated weekly with PRS-342 at three dose levels. shown) and for the T-cell marker CD3, CD4 and CD8. Percentage of TILs per tumor area excluding reporter and T-cell costimulation assay • a-GPC3 antibody the necrotic area were calculated for all groups (BioSiteHisto). (C) Representative pictures for CD8 • Isotype control • Control molecules were an a-GPC3 antibody (IgG4 variant) in equimolar doses, an a-4- staining of Hep-G2 tumors demonstrating significant increased TIL infiltration for PRS-342 tumors 1BB benchmark antibody in equimolar doses, and vehicle control. compared to all controls (vehicle, a-GPC3 antibody and a-4-1BB antibody). A Anti-4-1BB B PRS-342 Design C PRS-342 Design (DNA) IL-2 induced by human Pan T cells costimulated by PRS-342 in the presence of GPC3-positive Anticalin Protein (Ac) HepG2 and Hep-GB cells in a coculture assay. No PRS-342 dependent activation was observed in • PRS-342 showed dose-dependent tumor growth inhibition (TGI) comparable to a-GPC3 α-GPC3 antibody presence of GPC3-negative MKN-45 cells. IL-2 levels in the culture supernatants were measured antibody, indicating that TGI is dominated by GPC3 inhibition in this model. by an electrochemiluminescence (ECL) immunoassay. Summary PRS-342 leads to dose dependent T-cell mediated TM=74°C (DSC) PRS-342 Median tumor growth inhibition in a humanized HepG2 xenograft model KD=02.3nM (SPR) A • PRS-342 was designed to elicit 4-1BB costimulatory effects in a tumor- α-4-1BB Ac cytolysis of GPC3 expressing tumor cells localized manner. D SPR Affinity • PRS-342 induced 4-1BB costimulation results in a dose-dependent T-cell killing of • PRS-342 is a 4-1BB/GPC3 bispecific genetic fusion of a high-affinity 4-1BB- ® kon koff KD GPC3 expressing tumor cells measured with an impedance based method. binding Anticalin and a high affinity a-GPC3 antibody. target name [M-1*s-1] [s-1] [nM] • No increase of T-cell mediated killing was observed with equimolar doses of anti-GPC3 • PRS-342 has excellent drug like properties and can be produced with high Fc-a-4-1BB antibody, Fc-4-1BB Anticalin fusion, anti-4-1BB-antibody and isotype control. yields. hu-4-1BB-His 3.5E+04 1.5E-04 4.1 Anticalin fusion Fc-a-4-1BB Anticalin fusion • PRS-342 has a pharmacokinetic profile comparable to classical antibodies. T c e ll k illin g H e p G 2 a-GPC3 antibody huGPC-3-His 3.6E+05 2.5E-04 0.7 • T-cell costimulation by PRS-342 leads to: 1 0 0 G C 3 3 -h Ig G 4 v 1 -L 1 -S 0 5 7 5 .0 4 J 1 0 ( H C ) 0 .0 8 n M o Nf-kB activation in a reporter cell assay. PRS-342 huGPC-3-His 4.1E+05 2.7E-04 0.7 a-GPC3 antibody G C 3 3 -h Ig G 4 v 1 - L 1 - S 0 5 7 5 .0 4 J 1 0 ( H C ) 0 .4 n M o Increased production of IL-2, a pro-inflammatory cytokine associated s i G C 3 3 - h Ig G 4 v 1 - L 1 -S 0 5 7 5 .0 4 J 1 0 ( H C ) 2 n M with anti-tumor immune response in a co-culture assay. E Binding ELISA s 5 0 y l G C 3 3 - h Ig G 4 v 1 - L 1 - S 0 5 7 5 .0 4 J 1 0 (H C ) 1 0 n M o Dose dependent cytolysis in impedance based real time killing assay. 4-1BB/anti-Anticalin-HRP GPC3/anti-human IgG-HRP GPC3/4-1BB-Bio/Extravidin-HRP c i a-GPC3 antibody f G C 3 3 - h Ig G 4 v 1 1 0 n M o TIL infiltration in tumors of a HCC xenograft in humanized mice. i PRS-342 c e h Ig G 4 - S 0 5 7 5 .0 4 J 1 0 -S 2 2 8 - F A L A 1 0 n M • The preclinical studies reported here demonstrate potent T-cell activation p 0 s that is strictly dependent on the presence of GPC3-positive tumor cells. U re lu m a b 1 0 n M % h Ig G 4 v 1 1 0 n M • GPC3-dependent activation of tumor-specific T cells is expected to result in an improved safety profile. -5 0 2 0 4 0 6 0 8 0 • Collectively our in vitro and in vivo data support the continued 4-1BB ELISA GPC3 ELISA Dual binding ELISA development of PRS-342. PRS-342 Design (A, B, C) and target binding (D, E). (D) shows binding of the Fc-a-4-1BB Anticalin tim e [h ] (A) Immunocompromised female NOG mice carrying established HepG2 xenograft tumors were fusion with a KD of 4.1 nM. The GPC3 arm binds with 0.7 nM to GPC3. On and off-rate kinetic T-cell mediated cytolysis of GPC3 expressing HepG2 tumor cells was assessed using the engrafted with 5 × 106 fresh human PBMC, followed by weekly i.p. treatment with PRS-342, a-4-1BB binding constant for a-GPC3 antibody and PRS-342 are similar. (E) ELISA data demonstrate PRS- xCELLigence RTCA HT system. Non-adherent CD8+ T cells were cocultured with HepG2 cells in benchmark antibody, a-GPC3 antibody or isotype control at 0.5 mg up to 20 mg/kg doses (i.p.) 342 binds GPC3 with comparable behavior to the a-GPC3 parental antibody. In a dual binding presence of test constructs. 4-1BB costimulation of cytotoxic T cells results in an increased (Charles River). Mice (n=15 per gr. remained on the study until spontaneous death or if ethical ELISA setting PRS-342 (4-1BB/GPC3 bispecific), is capable of binding both targets simultaneously. cytolysis of HepG2 cells which is measured over time. sacrifice was required, read out median tumor growth). References: (1) Cancer Immunol Res 2016;4(11 Suppl):Abstract nr B016.