Discovery of a novel anti-LAG3 antagonist antibody

Wenqing Jiang, Lei Fang, Zhengyi Wang, Taylor B. Guo Suite 500, 2275 Research BLVD, Rockville Maryland 20850, I-Mab Biopharma US I-MAB Biopharma, Inc., Shanghai, China

ABSTRACT In vitro functional assays of TJ-A3

LAG3 is a negative regulator of T cells. Persistent T-cell activation in a chronic A LAG3/MHCII blocking B inflammatory environment, as in a tumor or during chronic viral infection, results in L A G 3 /M H C II b lo c k in g a s s a y Jurkat-hLAG3 IL2 release 5 2 5 F 7

sustained LAG3 expression, contributing to a state of exhaustion manifest in impaired 4 T J A 3 2

h 3

C

/ 1

proliferation and cytokine production. Clinical studies showed striking synergy between h 2 C TJ-A3: 0.4nM TJ-A3: 4.4nM LAG3 and PD1 in multiple settings, highlighting the therapeutic potential of LAG3. Here we 1 25F7: 0.9nM 25F7: 12.7nM

0 0 .0 0 1 0 .0 1 0 .1 1 1 0 1 0 0 1 0 0 0 reported the discovery of a novel LAG3 antibodies (TJ-A3) with sub-nanomolar binding C o n c e n tr a tio n (n M ) affinity. Upon binding, it blocked the interaction of LAG3 to its receptor MHCII, leading to increased production of IL-2 in Jurkat cells overexpressing LAG3 and in activated human C Primary T assay Donor 1 Donor 2 1 0 3 0 1 0 0 n M 1 0 0 n M 2 5

primary T cells. TJ-A3 can significantly inhibit tumor growth in combination with a PD- CD3/28 8 1 0 n M 1 0 n M

) l ) 1 nM CD4+ T cells l 2 0 1 nM

6 m

m

/ /

blocking antibody in a syngeneic mouse tumor model by using LAG3-humanized mice. g IL-2 g 0 .1 n M

1 5

n

n

(

(

SAg 4

IL-4 2

2

L L I 1 0 Together with favorable cynomolgus monkey PK and cell line development profile, these GM-CSF MLR I 2 5 studies support further clinical development of TJ-A3. DCs 0 0 monocyte h Ig G 2 5 F 7 T J -A3 Ig G T J -A3 P D -L 1 T J -A3 + m Ab P D -L 1 m Ab MATERIALS AND METHODS (A) Receptor blocking assay. (B) Left, assay principle of Jurkat IL2 assay. Right, Stimulation of IL2 release in Jurkat-hLAG3 by LAG3 mAbs. (C) Left, assay principle of primary assay. Middle, stimulation of IL2 release in activated CD4+ T cells by LAG3 mAbs or PD- Antibody generation. A mouse lead monoclonal antibody Mu147H was originally L1 mAb or combination treatment. obtained through standard immunization and hybridoma process using the extracellular In vivo efficacy of TJ-A3 domain (ECD) of human LAG3 as the antigen. Following sequencing, the molecule was 800

humanized (using CDR-grafting and back mutation strategy) and affinity matured, which ) 800

3 700

) VehicleVehicle

3 700 resulted in an antibody named TJ-A3, with a human IgG4 Fc domain with a serine to 600 TecentriqTecentriq(1mg/kg)Vehicle 600 500 Tecentriq+TJTecentriq(1mg/kg)+TJ007477(10mg/kg)Tecentriq(1mg/kg)-A3 proline mutation at No.228 position to avoid arm exchange. 500 400 Tecentriq+25F7Tecentriq(1mg/kg)+25F7(10mg/kg)Tecentriq(1mg/kg)+TJ007477(10mg/kg) Ab 400 In vitro receptor blocking assay. LAG3 antibodies were incubated with Tag1-LAG3 and 300 Tecentriq(1mg/kg)+25F7(10mg/kg) MC38 hLAG3 mice 200 300 200

Tag2-MHCII for 15 min at room temperature. Pre-mixed detection antibodies containing (mm Volume Tumor 100

0 (mm Volume Tumor 100 anti-Tag1-Tb3+ and anti-Tag2-XL665 antibodies were added and incubated for another 1 00 3 6 9 12 15 18 Days0 after3 administration6 9 12 15 18 hr at room temperature. The plate was analyzed by PerkinElmer Envision plate reader. Days after administration

Vehicle Jurkat-LAG3 cell based IL2 release assay. LAG3 antibodies were incubated with Raji Vehicle TecentriqTecentriq Tecentriq+TJTecentriq + 25F7-A3 Tecentriq+25F7Tecentriq + TJ007477 1000

1000 1000 1000

3

3 3 cells and Jurkat-hLAG3 cells for 48 hours in the presence of TCR stimulation. Culture 800 428 3 800 378 800 416 800 351 382 442 362 390 600 medium was collected and incubated with anti-IL2 antibody pairs. IL-2 concentration was 481 600 350 600 393 600 366 386 349 452 360 400 400 400 400 388 391 347 488 383 387 407 445

further analyzed by PerkinElmer Envision plate reader. 200 200 200 200

Tumor Volume mm Volume Tumor

Tumor Volume mm Volume Tumor Tumor Volume mm Volume Tumor 384 mm Volume Tumor 411 374 486 0 0 0 0 0 3 6 9 12 15 0 3 6 9 12 15 0 3 6 9 12 15 0 3 6 9 12 15 Primary T functional assay. LAG3 antibodies were incubated with MHC-mismatched DC Days post 1st dose Days post 1st dose Days post 1st dose Days post 1st dose and CD4+ T cells for 5 days in the presence of TCR stimulation. Culture medium was In vivo efficacy of combo treatment of TJ-A3 and anti-PDL1 in an MC38 syngeneic mouse model. MC38 cells were subcutaneously implanted into huLAG3 mice. When the tumor grew to around 50 mm3, mice were intraperitoneally treated with control, anti-PDL1 only collected and incubated with anti-IL2 antibody pairs. IL-2 concentration was further or anti-PDL1 plus the indicated LAG3 mAb. Tumor growth was monitored by volumetric measurement. Shown are means±SEM (top) analyzed by PerkinElmer Envision plate reader. and individual mouse data (bottom). In vivo antibody treatment in MC38 huLAG3 syngeneic model. Humanized-Lag3 mice Pilot pharmacokinetics of TJ-A3 were subcutaneously implanted with 1×106 MC38 cells on day 0. On day 6, mice with an 10mg/kg TJ-A3 average tumor volume of 52mm3 were randomized into four treatment groups (N=7/group). PK parameters Mouse were intraperitoneally administered saline, anti-PDL1 (Tecentriq) at 1 mg/kg, and t 1/2 (hr) 156 anti-PDL1 (Tecentriq) at 1 mg/kg plus anti-LAG3 antibodies (10 mg/kg) every 3 days for 4 Tmax (hr) 0.08 doses. Tumor volumes were monitored by caliper measurement twice per week for the Vss (mL/kg) 55.53 CL (mL/hr/kg) 0.24 duration of the experiment. Cmax (ug/mL) 255.50 RESULTS AUC (h*mg/mL) 40.59 Naïve cynomolgus monkeys were treated with a single i.v. injection of TJ-A3 (10 mg/kg, n=2). Serum samples from each individual monkey were collected at various time points and measured for the levels of anti-LAG3 mAbs by a generic ELISA method using Binding properties of TJ-A3 recombinant LAG3 ECD as the capture antigen.

A Biacore Ka = 6.1E+06 (1/Ms) B ELISA CONCLUSIONS Kd = 2.8E-03 (1/s) KD = 4.6E-10 (M) • TJ-A3 is an antagonist antibody against LAG3 with sub-nanomolar affinity.

25F7: 0.25nM • TJ-A3 binds to human and cynomolgus LAG3, blocking the binding of LAG3 to its TJ-A3: 0.06nM receptor MHCII molecule. • Upon binding, TJ-A3 leads to the increased production of IL-2 in Jurkat cells overexpressing LAG3 and in activated human primary T cells. Cell based binding Cell based binding on Jurkat-LAG3 cells • TJ-A3 showed strong synergistic effects on the activation in combo with anti-PDL1 C Jurkat-hLAG3 binding D Activatedon activated CD4+ CD4 T bindingT cells E Cyno LAG3 binding 400 600 25F7 25F7 antibody. Consistently, when combined with PD-L1 antibody, TJ-A3 can significantly TJ007477 300 TJ007477 inhibit tumor growth in a syngeneic MC38 tumor model using LAG3-humanized mice. IgG 400 IgG 200

MFI • Time-concentration profiles indicated that the PK of TJ-A3 was consistent with normal MFI 25F7: 0.57nM 200 25F7: 0.51nM 25F7: 4.0nM 100 TJ-A3: 0.06nM TJ-A3: 0.32nM TJ-A3: 7.7nM human IgG.

0 0 • Developability assessment of the TJ-A3 sequence indicates that it is a reasonably 0.001 0.01 0.1 1 10 100 0.001 0.01 0.1 1 10 100 Concentration (nM) Concentration (nM) developable molecule to be taken forward. CMC-wise, a cell line with decent titer have TJ- 25F7 25F7 IgGTJ007477 25F7 TJ007477 EC50 0.5175 0.3213 been developed. A3 EC50 0.5715 0.06450 (A) Affinity of TJ-A3 to LAG3 ECD by Biacore. (B) Binding of LAG3 mAbs to soluble LAG3 by ELISA. (C) Binding of LAG3 mAbs to Jurkat-hLAG3 cells by FACS. (D) Binding of LAG3 mAbs to activated human CD4+ T cells by FACS. (E) Binding of LAG3 mAbs to cyno LAG3 by ELISA.