0 200 400 600 800 IL-2 IL -2 [ p g /m l] 0 5000 10000 15000 Granzyme A G ra n z y m e A [ p g /m l] 0 2000 4000 6000 8000 10000 Perforin P e rf o ri n [ p g /m l] 0 1000 2000 3000 4000 Granzyme B G ra n z y m e B [ p g /m l] Simultaneous costimulatory T-cell engagement and checkpoint inhibition by PRS- 344/S095012, a 4-1BB/PD-L1 bispecific compound for tumor-localized activation of the immune system Aizea Morales-Kastresana1*, Lucia Pattarini2*, Marina Pavlidou1*, Janet K. Peper-Gabriel1*, Christian Barthels1, Eva-Maria Hansbauer1, Rachida Bel Aiba1, Birgit Bossenmaier1, Alix Scholer-Dahirel2, Thomas Jaquin1, Catherine Gallou2, Véronique Blanc2, Christine Rothe1, Shane A Olwill1 1Pieris Pharmaceuticals GmbH, Zeppelinstrasse 3, 85399 Hallbergmoos - Germany 2Institut de Recherches Servier Oncology R&D Unit, Croissy Sur Seine, France *Co-authors / equally contributing authors LB135 INTRODUCTION • 4-1BB (CD137) is a key co-stimulatory immunoreceptor and a promising oncology target. • Peripheral immune activation by 4-1BB agonistic antibodies has been associated with on-target toxicity and a limited therapeutic window. • To overcome 1st generation 4-1BB agonist safety and efficacy drawbacks, we have generated PRS-344/S095012, a 4-1BB/PD-L1 bispecific Anticalin® protein/mAb fusion protein (Figure 1) designed to have a 4-1BB localized activity, while also offering the benefit of checkpoint inhibition (Figure 2). • Here we describe the preclinical in vitro and in vivo activity of PRS- 344/S095012. This program is part of the strategic alliance between Pieris and Servier. PRS-344/S095012 is capable of dual target engagement ▪ PRS-344/S095012 binds to 4-1BB and PD-L1 in a comparable way to the respective single building blocks and can bind both targets simultaneously. ▪ PRS-344/S095012 effectively blocks the PD-1/PD-L1 binding and shares an overlapping 4-1BB-binding epitope with a clinically active anti-4-1BB benchmark mAb. A B C Figure 3. Binding A,B) Direct binding to human recombinant 4-1BB and PD-L1 C) Simultaneous binding of PRS- 344/S095012 to 4-1BB and PD-L1. D) Blockade of PD-1/PD-L1 interaction. E) Competition with an anti-4-1BB benchmark mAb. All experiments were conducted with an ELISA-based approach D E PRS-344/S095012 stimulates activated T cells in a PD-L1-dependent fashion and enhances their proinflammatory and cytotoxic potential ▪ PRS-344/S095012-mediated co-stimulation is strictly PD-L1 dependent and only occurs upon TCR engagement, reducing the risk of peripheral toxicity. ▪ PRS-344/S095012 stimulates the release of cytotoxic molecules and cytokines from activated antigen-specific CD8 T cells or polyclonal T cells. ▪ The in vitro functional activity of PRS-344/S095012 is superior to single agent anti-PD-L1 or benchmark anti-4-1BB mAb. ▪ Engagement of PDL1 and 4-1BB through PRS-344/S095012 bispecific is superior to combination of PD-L1 and 4-1BB mAbs. B Figure 4. In vitro activity A-C) Co-culture assay: human T cells with coated anti-CD3 mAb and tested constructs co-cultured with A) hPD-L1 positive CHO cells or B) control CHO cells and C) W/O anti- CD3 as a negative control. n.d., not detected. D-E) Recall assay of human PBMCs stimulated with a peptide pool with the indicated constructs or (D), pre-expanded with a peptide pool and re-stimulated with the peptide pool plus the indicated constructs (E). Data is shown as mean ± SEM, n.s= non-significant, ***, P < 0.01. F) Mixed lymphocyte reaction: CD8 T cells from healthy blood donors co-cultured with monocyte-derived dendritic cells from another healthy blood donor. A 0 50 100 150 200 250 300 IF N g [ p g /m L ] *** *** *** *** 0 1000 2000 3000 4000 IF N g [ p g /m L ] n.s. *** *** *** *** *** *** PRS-344/S095012 displays Ab-like PK in mice and drives a strong anti-tumoral activity superior to anti-PD-L1 mAb ▪ The mAb-like half-life of the anti-PD-L1 mAb building block is preserved within PRS-344/S095012. ▪ PRS-344/S095012 triggers a dose-dependent antitumoral response that leads to a significant extension of survival in a humanized KI model. ▪ Complete regression of implanted tumors is observed in 5 out of 10 mice treated with the highest dose of PRS-344/S095012. ▪ PRS-344/S095012 is superior to equimolar doses of anti-PD-L1 mAb treatment alone. Construct t1/2 [h] Anti-PD-L1 building block 390 PRS-344/S095012 295 • PRS-344/S095012 is a 4-1BB / PD-L1 bispecific, generated by the genetic fusion of a 4-1BB-binding Anticalin® protein and an anti-PD-L1 mAb. • PRS-344/S095012-mediated 4-1BB activation is PD-L1-dependent, potentially reducing the risk of peripheral toxicity. Furthermore, 4-1BB co-stimulation only occurs in combination with simultaneous TCR signaling, focusing co- stimulation to antigen-specific T cells. • PRS-344/S095012 induces an effective CD8 T cell response, leading to secretion of inflammatory cytokines and cytotoxic molecules. • PRS-344/S095012 displays mAb-like pharmacokinetics in mice. • PRS-344/S095012 induces a dose-dependent anti-tumor response in a mouse model setup refractory to anti-PD-L1 and significantly extends the survival of mice. • Preclinical data support clinical evaluation of PRS-344/S095012. ACKNOWLEDGEMENTS - We would like to especially thank to the scientists and research assistants that performed these experiments: Maximilien Grandclaudon, Celine Sancerne, Matthieu Riviere, Christina Grasmüller, Nicole Andersen, Linda Schnapp, Markus Rehle and Nicolas Quilitz as well as Marlon Hinner and Louis Matis for their support with original concept. Conclusions Figure 2. PRS-344/S095012 dual MoA: selective activation of 4-1BB+ T cells in PD-L1+ tumor and/or antigen-presenting cells in the tumor microenvironment or tumor-draining lymph node (dLN) and blocking of the PD-1 / PD-L1 interaction. No clustering of 4-1BB is expected in the periphery. Tumor cell / APC PD-1/PD-L1 axis inhibition High PD-L1 expression PD-1 Blocking PD-1/PD-L1 interaction Healthy cell Low/No PD- L1 expression PRS-344/ S095012 4-1BB T cell No 4-1BB clustering + co-stimulation No activity in the periphery Tumor-specific T cell Tumor cell / APC High PD-L1 expression 4-1BB clustering + T cell co-stimulation Activity in the tumor and tumor draining lymph nodes T cells with PD-L1- cell line Co-culture in absence of aCD3 Antigen-recall T cell assay CD8 T cells in mixed lymphocyte reaction T cells with PD-L1+ cell line C A Figure 5. In vivo activity A) CD1 mice were injected with 10 mg/kg of the constructs. Plasma concentration of constructs is plotted, and half-life estimated by non-compartmental analysis. B-D) h-4-1BB KI mice subcutaneously implanted with MC-38-huPD- L1 cells were treated with equimolar doses of constructs biweekly for 3 weeks. Tumor growth (B,C) and survival (D) are shown.. * = p < 0.05; ** = p < 0.01; *** = p < 0.001 vs vehicle group 0 10 20 30 40 50 0 1000 2000 3000 4000 5000 Vehicle control 0 10 20 30 40 50 PRS-344 / S095012 10 mg/kg 0 10 20 30 40 50 PRS-344 / S095012 1 mg/kg 0 10 20 30 40 50 PRS-344 / S095012 0.1 mg/kg 0 10 20 30 40 50 PRS-344 / S095012 0.01 mg/kg 0 10 20 30 40 50 anti-PD-L1 mAb 7.7 mg/kg 0 10 20 30 40 50 anti-PD-L1 mAb 0.8 mg/kg 0 10 20 30 40 50 anti-4-1BB mAb 3 mg/kg Days after randomization 0 5 10 15 20 25 30 35 0 1000 2000 3000 4000 *** *** *** *** * ** * B 0 10 20 30 40 50 0 20 40 60 80 100 D C anti-PD-L1 building block PRS-344/S095012 C o n ce n tr a ti o n [ u g /m l] M e a n t u m o r vo lu m e ( m m 3 ) P e rc e n t su rv iv a l Days after implantationDays after implantation Tu m o r v o lu m e ( m m 3 ) D E F nd nd nd nd nd Figure 1. Structure of PRS-344/S095012 4-1BB-targeting Anticalin® proteins PD-L1-targeting Ab FALA mutant IgG4 Tumor-specific T cell PRS-344/S095012 anti-PD-L1 building block anti-4-1BB mAb atezolizumab + anti-4-1BB mAb hIgG4 PRS-344/S095012 anti-PD-L1 building block anti-4-1BB mAb atezolizumab + anti-4-1BB mAb hIgG4 Vehicle hIgG4 atezolizumab anti-4-1BB mAb atezolizumab + anti-4-1BB mAb PRS-344/S095012 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 2000 4000 6000 8000 10000 12000 Construct [nM] IL -2 [ p g /m L ] Competition with anti-4-1BB benchmark to 4-1BB binding 0.001 0.01 0.1 1 10 100 1000 0.0 0.2 0.4 0.6 0.8 1.0 1.2 PRS-344/S095012 anti-4-1BB mAb anti-PD-L1 building block Concentration [nM] f r e e t r a c e r [ n M ] Blockade of PD-1/PD-L1 binding 0.001 0.01 0.1 1 10 100 1000 0 5 10 15 20 PRS-344/S095012 atezolizumab Concentration [nM] fr e e t r a c e r [ n M ] anti-PD-L1 building block 4-1BB binding 0.001 0.01 0.1 1 10 100 1000 0 10000 20000 30000 Concentration [nM] R F U PRS-344/S095012 anti-4-1BB Anticalin PD-L1 binding 0.001 0.01 0.1 1 10 100 1000 0 10000 20000 30000 Concentration [nM] R F U anti-PD-L1 building block PRS-344/S095012 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 .0 0 3 0 .0 0 8 0 .0 2 3 0 .0 7 0 .2 0 .6 1 .9 5 .6 1 6 .7 5 0 0 2000 4000 6000 8000 10000 12000 Construct [nM] IL -2 [ p g /m L ] 5 0 5 0 5 0 5 0 5 0 0 2000 4000 6000 8000 10000 12000 IL -2 [ p g /m L ] Construct [nM]