5. FS118 mAb² induces an increase in total soluble PD-L1 in cynomolgus monkey

Conclusions

• mLAG-3/PD-L1 mAb² shows anti-tumour activity at low doses with an indication ofdose-response in a MC38 tumour model.

• mLAG-3/PD-L1 mAb² and single agent combination resulted in distinct modulations ofLAG-3 and PD-L1 cell surface expression on tumour-infiltrating T cells.

• mLAG-3/PD-L1 mAb² treatment reduces LAG-3 expression on T cells in the tumourmicroenvironment while single agent combination resulted in increase of LAG-3expression. LAG-3-positive T cells in the tumour microenvironment are reported to behighly immunosuppressive and described as one of the key component in a PD-(L)1resistance mechanism (Koyama et al. 2016). By co-engaging both targets, FS118 has thepotential to overcome LAG-3-mediated resistance pathway through a uniquemechanism different to the combination of monospecific agents.

• FS118 mAb² treatment of cynomolgus macaques shows similar biology to that of thesurrogate in mice with respect to soluble PD-L1 upregulation and is therefore atranslationally relevant mechanism of action to assess clinically.

• Both soluble LAG-3 and PD-L1 in the serum of treated animals are related to the uniquemechanism of target engagement of FS118 surrogate mAb² and warrant furtherinvestigation of these markers in the clinical setting.

Mustapha Faroudi, Matthew Kraman, Natalie Fosh, Claire Reader, Daniel Gliddon, Claire Seal, Christa Lucas, Alexander Koers, Mateusz Wydro, Michelle Morrow & Neil Brewis

F-star Biotechnology Ltd, Cambridge, UK

6. FS118 mAb² dual blockade of LAG-3/PD-L1 immunosuppressive pathways drives potential benefit over combination treatments to overcome acquired PD-1/PD-L1 resistance

Figure 4. Evidence for targetengagement. Binding of LAG-3 and PD-L1 on the cell surface accelerate theshedding of the proteins,demonstrating the potential to be usedas a target engagement biomarker.Fold of change of total soluble LAG-3and total soluble PD-L1 levels in theserum from MC38 tumour bearingmice treated with a single dose of themLAG-3/PD-L1 mAb² (1, 3 or 10mg/kg), the combination of the LAG-3/mock mAb² and PD-L1 mAb (1, 3 or10 mg/kg) or the isotype controlantibody (10 mg/kg). Data werenormalized to the IgG control groupand fold change was calculated. Dottedline represents the baseline as definedby isotype control.

AACR Annual Meeting 2019 | 29 Mar-03 Apr | Atlanta | Poster Number: 2399 Strictly for personal use - DO NOT POST ONLINE

Figure 2. Total LAG-3 and free PD-L1 expression in tumour-infiltrating lymphocytes (TILs) following a single dose of mLAG-3/PD-L1 mAb² or combination treatment. A - Schematic representation of the cell surface target detection: total LAG-3 wasmeasured using an anti-mouse LAG-3 antibody (LAG-3 mAb) that does not compete with the mAb² for binding LAG-3, freePD-L1 was measured using an anti-mouse PD-L1 antibody (PD-L1 mAb) that does compete with the mAb² for binding PD-L1.B - Total LAG-3 and free PD-L1 expression on CD4+ T cells TILs following single dosing of test antibodies in MC38 model asmeasured by flow cytometry. When subcutaneous tumours reached a palpable size mice were administrated with oneintraperitoneal injection of mLAG-3/PD-L1 mAb² or combination antibodies. At designated timepoints following dosing micewere sacrificed and tumours processed for flow cytometry. Data are presented as mean +/- SD

2. mLAG-3/PD-L1 mAb² and single agent combination result in distinct modulations of LAG-3 surface expression on tumour-infiltrating T cells

3. mLAG-3/PD-L1 mAb² target engagement directly leads to a decreased number of available receptor binding sites on tumour-infiltrating T cells

4. mLAG-3/PD-L1 mAb² increases levels of soluble LAG-3 and PD-L1

Tumour/APC

BackgroundDespite advances with therapies targeting the PD-1/PD-L1 pathway, many patients are refractoryto or relapse following treatment. Resistance to anti-PD-1 treatment is associated withupregulation of other checkpoint inhibitor receptors such as LAG-3. Co-treatment with antibodiestargeting LAG-3 and PD-L1 could potentially overcome this resistance. An alternative approach isthe development of a bispecific antibody encompassing binding sites for two antigens. FS118,currently tested in Phase I clinical trial in patients with advanced malignancies (NCT03440437), is atetravalent bispecific antibody (mAb2) targeting LAG-3 and PD-L1 that provides dual pathwayblockade with the potential to drive unique biology by simultaneously binding PD-L1 and LAG-3,and to unlock synergistic benefits to patients treated with current monotherapies.In a syngeneic model we used a surrogate mAb2 of FS118 (mLAG-3/PD-L1 mAb2 blocks both mousePD-L1 and mouse LAG-3 binding to their receptors) to determine the effect of dual checkpointblockade on tumour growth and to elucidate the modulation of the underlying immunemechanisms following treatment with the mAb2. FS118 is under option to Merck KGaA, Darmstadt,Germany.

LAG-3

PD-L1

Figure 5. Pharmacodynamic changes in total soluble PD-L1 in cynomolgus monkeys who received intravenous injectionof the FS118 mAb² dosed twice-weekly at 60 and 200 mg/kg in a 4-week GLP toxicity study. The capture of soluble PD-L1 is indicative of target engagement.

Figure 3. Treatment with the single agent combination increased the total number of available PD-L1 plus LAG-3receptors (red areas), whereas treatment with mLAG-3/PD-L1 mAb² appeared to not change or reduce the overallnumber of receptors (green areas) on the TILs.Percentage of CD4+ and CD8+ T cells with bound drug on TILs following single dosing of test antibodies in MC38 model asmeasured by flow cytometry. When subcutaneous tumours reached palpable size, mice were administrated with onintraperitoneal injection of mLAG-3/PD-L1 mAb² or combination of individual agents. At designated timepoints followingdosing mice were sacrificed and tumours processed into cell suspensions. Tumour cell suspensions were split in two,then saturated or not with mLAG-3/PD-L1 mAb² to measure the maximum available target binding sites. Sample werethen labelled with fluorescently conjugated antibodies to identified immune cells populations subsets. Test antibodiesbound on the cell surface were detected using an anti-human CH2 fluorescently labelled antibody. Data are presentedas mean +/- SEM.

1. mLAG-3/PD-L1 mAb² inhibits tumour growth in a MC38 syngeneic colon carcinoma model

0 5 10 15 20 25

0

600

1200

1800

Time post first dose (d)

Tum

ou

r vo

lum

e (m

ean

+/-

SEM

, mm

3)

mLAG-3/PD-L1 mAb2 (10 mg/kg)

mLAG-3/PD-L1 mAb2 (3 mg/kg)

mLAG-3/PD-L1 mAb2 (1 mg/kg) IgG ctrl

LAG-3/mock + anti-PD-L1 mAb (1 mg/kg)LAG-3/mock + anti-PD-L1 mAb (3 mg/kg)

LAG-3/mock + anti-PD-L1 mAb (10 mg/kg)

24 48 72 96 120 144 1680

10

20

30

40

50

Time post dosing (h)

% F

ree

PD

-L1

+ of

CD

4 T

cel

ls

Free PD-L1 expression on CD4+ T cells in TILs

24 48 72 96 120 144 1680

10

20

30

40

50

Time post dosing (h)

% L

AG

-3+ o

f C

D4

T c

ells

mLAG-3/PD-L1 mAb² (10mg/kg)mLAG-3/PD-L1 mAb² (3mg/kg)

mLAG-3/PD-L1 mAb² (1mg/kg)

LAG-3/mock mAb² + PD-L1 mab (10mg/kg)LAG-3/mock mAb² + PD-L1 mAb (3mg/kg)LAG-3/mock mAb² + PD-L1 mAb (1mg/kg)

IgG ctrl

Total LAG-3 expression on CD4+ T cells in TILs

CD

4+

T ce

lls

0

20

40

60

80

100

Time post dosing (h)

%C

D4

+ T c

ells

wit

h b

ou

nd

dru

g

24 48 96168

0

20

40

60

80

100

Time post dosing (h)

%C

D4

+ T c

ells

wit

h b

ou

nd

dru

g

24 48 96168

CD

8+

T ce

lls

mLAG-3/PD-L1 mAb² (10 mg/kg) LAG-3/mock mAb² + PD-L1 mAb (10 mg/kg)

0

20

40

60

80

100

Time post dosing (h)

% C

D8

+ T

ce

lls w

ith

bo

un

d d

rug

IgG ctrl

mLAG-3/PD-L1 mAb²

Maximum available mAb2 binding sites(mLAG-3/PD-L1 mAb² group)Maximum available mAb2 binding sites(IgG ctrl group)

24 48 96168

0

20

40

60

80

100

Time post dosing (h)

% C

D8

+ T

ce

lls w

ith

bo

un

d d

rug

IgG ctrl

LAG-3/mock mAb² + PD-L1 mAb

Maximum available mAb2 binding sites(combination group)

Maximum available mAb2 binding sites(IgG ctrl group)

24 48 96168

24h48h

72h96h

168h0

2

4

6

8

Solu

ble

LA

G-3

(Fo

ld c

han

ge)

24h48h

72h96h

168h

0

1

2

3

4

Solu

ble

PD

-L1

(Fo

ld c

han

ge)

*

* No data for 72 hrs due to low sample volumes

mLAG-3/PD-L1 mAb² (10mg/kg)

mLAG-3/PD-L1 mAb² (3mg/kg)

mLAG-3/PD-L1 mAb² (1mg/kg)

LAG-3/mock mAb² + PD-L1 mAb (10mg/kg)

LAG-3/mock mAb² + PD-L1 mAb (3mg/kg)

LAG-3/mock mAb² + PD-L1 mAb (1mg/kg)

IgG ctrl (10mg/kg)

T cell

Koyama et al - Nat Commun. 2016 Feb 17;7:10501Huang RY et al. Oncoimmunology. 2016 Oct 28;6(1):e1249561

Fcab™Fc TARGET B TARGET B

TARGET ATARGET A

mAb²™

Directed amino acid substitutions to

create new binding sites in Fc

Rapid conversion into a tetravalent IgG i.e.

bivalent for both targets

Plug-and-play into any mAb

T cell

Figure 1. Dual blockade of LAG-3 and PD-L1 withmLAG-3/PD-L1 mAb² in the MC38 tumour-bearingmice model resulted in an increased anti-tumouractivity comparable to a combination of the singleagents targeting LAG-3 and PD-L1.MC38 tumour cells were injected subcutaneouslyin C57/Bl6 mice until a palpable tumour formed(day 8 post-inoculation). The cohorts wereadministrated either with: isotype control (10mg/kg) or mLAG-3/PD-L1 mAb² or the combinationLAG-3/mock mAb² plus PD-L1 mAb at 1, 3, 10mg/kg. Red arrows represent three doses injectedat day 0, day 3, and day 6 post-dosing. X axis showstime (days) following the first dose. Y axis showstumour volume (mm3).

T cellPD-L1 LAG-3

mLAG-3/PD-L1 mAb² Fluorescently conjugated

LAG-3 mAb(non-competitive binding)

Fluorescently conjugated PD-L1 mAb(competitive binding)

Dose: 60 mg/kg Dose: 200 mg/kg

A LAG-3/PD-L1 mAb² can overcome PD-L1-mediated compensatory upregulation of LAG-3 induced by single-agent checkpoint blockade

A - Assessment of receptor occupancy by flow cytometry : Free PD-L1 and Total LAG-3

B – Single dose of mLAG-3/PD-L1 mAb2 transiently decreases LAG-3 expression on T cells while single agent combination increased LAG-3 expression

0 7 14 21 28 35 42 49 56 63 70

0

2000

4000

6000

8000

10000

Elapsed time (days)

Pla

sma

PD

-L1

(p

g/m

L)

0 7 14 21 28 35 42 49 56 63 70

0

2000

4000

6000

8000

10000

Elapsed time (days)

Pla

sma

PD

-L1

(p

g/m

L)

LAG-3 PD-1

MHC II PD-L1

Upregulation of LAG-3 increases resistance following PD-1/L1 therapy

Partial response

Tumour/APC

T cell

1. PD-(L)1 monotherapy 3. FS118 mAb²

PD-1 mAb FS118

+

Significant reduction in LAG-3 expression

FS118 suppresses resistance

Resistance

Differentiated MoA addressing LAG-3-mediated acquired resistance

Superior T cell activation

+ + +

2. PD-(L)1 + LAG-3 monotherapies

LAG-3 mAb + PD-1 mAb

LAG-3 expression increase

LAG-3 expression still increasing following treatment with combination

T cell activation

+ +

LAG-3 expression increase