Poster
The discovery of a CD3 Agonistic SdAb to Accelerate the Development of CD3 T cell Engager Multi-specific Antibodies
The discovery of a CD3 agonistic single domain antibody (sdAb) to
accelerate the development of CD3 T cell engager multi-specific antibodies
1
1
1
1
1
1
1
1
1
1
1
1
Yingyu Li , Yi Zhou , Cheng Luo , Xin Wang , Yang Xin , Shishuai Shao , Tinging Yang , Mingzhu Shao , Jin Chen , Ziyu Chen , Sujuan Hong , Hai Huang ,
1
1
1
1
2
2
2
1
1
Cheng Chen , Hui Wang , Miaomiao Song , Yuanxin Zhang , Shuai Yang , Xiaojie Tu , Sujuan WANG , Li Chen , Yu Liang
Author affiliation: 1 Probio Inc, Nanjing, China 2 Legend Biotech, Nanjing, China
INTRODUCTION
RESULTS
RESULTS
CD3 T cell engager (TCE) bispecific antibodies (BsAbs) is a clinically well validated and dominant
therapeutic modality of approved bispecific antibodies, contributing to 10 out of 14 marketed BsAb
drugs. To develop a TCE BsAb with desired potency and safety profile, a CD3 agonist Ab with appropriate
potency and ideal developability is critical. Currently, all approved TCE BsAbs are based on a CD3
agonist Ab in the format of conventional monoclonal antibodies containing both light chain and heavy
chain, which may limit the flexibility in multi-specific antibody construction and development. In contrast,
considering these limitations, single domain antibodies (sdAb, or VHH) are highly favored in the
construction and development of bi- or multi-specific antibodies, mainly due to their smaller size, good
stability and developability. Therefore, to ease the development of TCE multi-specific antibodies, we set
forth to discover an agonist sdAb of CD3.
Figure 3. Structural diagram of CD20xCD3 (AHF43673) BsAbs
Figure 6. Potent TGI in vivo as CD20xCD3 VHH (AHF43673)
NKG mice (6-8 weeks), were inoculated with 5
x 106
human PBMCs on Day -16, followed by
the inoculation of 5 x 106 Jeko-1 tumor cells on
Day -14. On Day 0, the mice were randomly
grouped (n=8) and treated with the said
amount of BsAbs, i.v., followed by treatment on
day 4, 7, 11 and 14. The tumor volume was
measured twice a week. Significant and potent
tumor growth inhibition was observed for F28,
F29, and F32, though with a weaker potency
for F30. (One-way ANOVA LSD (L) test, ****
P<0.0001)
We first identified a parental CD3 sdAb from an immunized camel phage display library. Through successful
humanization and affinity maturation, we obtained a series of variants with a spectrum of binding affinity
to CD3 and potency in activating T cells. These CD3 sdAb variants were then used to construct a
series of bi- or multi-specific antibodies with ease and flexibility, no matter the TAA (tumor associated
antigen) arm is a conventional monoclonal antibody (CD20, HER2, etc.), or sdAb (GPRC5D, etc.). The
optimization process focused on fine-tuning the affinity, valency, and geometry of both the CD3 arm
and TAA arm. This resulted in multi-specific antibodies with enhanced yield and developability, potent
tumor cell killing capabilities, and with reduced cytokine release in vitro comparing to clinical benchmark
antibodies. In preclinical in vivo studies, these TCE multi-specific antibodies showed favorable
pharmacokinetics (PK) and safety profiles along with robust tumor growth inhibition.
CD20xCD3 BsAbs were constructed by fusing CD3 VHH (variant AHF43673) to various positions of Rituximab as
depicted above, recombinantly expressed, purified and tested in vitro in RGA & TDCC assays and in vivo for
tumor growth inhibition.
Figure 4. T cell activation as CD20xCD3 VHH (AHF43673) by RGA
Figure 7. Mouse PK study for CD20xCD3 VHH (variant 9) BsAbs
Table 3: Pharmacokinetic analysis
RESULTS
Fig 1. Improved binding affinity & potency in T cell activation
T cell activation as GPRC5DxCD3 VHH
CD3 binding on GS-J2C cell by FACS
Reporter Gene Assay (RGA) was performed by co-culturing GC-J2C RGA cells with CD20+ tumor cell (Jeko-1)
or CD20- tumor cell (SKBR-3), in the presence of various concentrations of BsAbs or control. T cell activation was
measured by luciferase expression using a standard approach.
Figure 5. Comparable TDCC with lower cytokine release as CD20x CD3 VHH (AHF43673)
ELISA binding to Cyno CD3ED
Female C57BL/6 mice (6-8 weeks, N=3) were randomly grouped and administrated with said amount of BsAbs
on day 0, followed by serum sample collection at 0.5, 2, 4, 6, 24, 48, 72, 96, 192, 240, 336, 504, 672 and 840 h.
The serum concentration of BsAb were determined by capture ELISA, and plotted against time. The half-life and
other PK parameters were calculated by Phoenix WinNonlin with NCA (Non-compartmental analysis) and results
summarized in Table 3. Overall, three CD20xCD3 VHH BsAbs showed a half-life of 190~292 hours, a typical PK
profile as an Fc tagged Ab, and comparable to CD20 mAb or other Fc tagged VHH.
Affinity matured CD3 VHH variants were compared with the parental
clone (AS39678) in binding to human T cell line and cyno CD3E/D protein,
or potency in activating T cells as GPRC5DxCD3 BsAbs by Reporter Gene
Assay (RGA).
Table 1: Improved binding affinity to hu CD3E/D complex by SPR
CONCLUSION
• Humanization and affinity maturation (AM) was successfully performed for a CD3 VHH Ab originally identified
from camel (clone AS39678)
Figure 2. Binding to an overlapping epitope with
Table 2: Thermostability measured by DSF
• Totally 26 AM variants were characterized and 6 selected as CD3 VHH tool box, showing a spectrum of binding
affinity & potency in T cell activation
Sp34 by FACS binning assay
• CD20xCD3 VHH BsAbs of various formats were constructed using one of the selected CD3 VHH variants
(AHF43673). These BsAbs showed comparable tumor cell killing with lower cytokine release in TDCC assays,
and potent tumor growth inhibition in a CDX efficacy model.
TDCC assay was performed by treating co-cultured human PBMC (donor NF0068) and CD20+ tumor cell (Je-
ko-1), with various concentration of BsAbs or control at E:T of 25:1, for 24hr. The tumor cell killing was measured
by LDH release and cytokine release by HTRF assay for IFNg, TNFa and IL-2.
• As a cornerstone component, these CD3 VHH variants may be highly desirable for developing a portfolio of
CD3 TCE bsAbs, with great ease and flexibility
ProBioCDMO.com