An engineered humanized common light chain mouse model to develop bispecific antibodies

8 min read
February 8, 2024

Rong, Y et al. An Engineered Mouse Model That Generates a Diverse Repertoire of Endogenous, High-Affinity Common Light Chain Antibodies. Antibodies, 2024


Bispecific antibodies combining two Fab arms specific for different antigens have proved their efficiency in preclinical models and in patients. One tricky aspect of IgG-like bispecific antibody production is the correct pairing of two different heavy chains with two different light chains.

This study describes the use of a pre-arranged IgK model to partially overcome this issue and reduce the number of incorrect combinations by fixing the light chain. This common light chain (cLC) model, generated by genOway for 23andMe, was used to discover antibodies specific to two different antigens and sharing the exact same light chain. These antibodies were then successfully combined into new bispecific candidates for further preclinical studies.

cLC mice were generated through the replacement of the mouse IgKJ cluster by a mouse IGKV10-96/KJ1 pre-arranged sequence (Figure 1). IGKV10-96 was selected based on its ability to efficiently pair with various heavy chains, its lack of residues known to be associated with downstream risks, and its 75% identity with human IGKV1-33 and IGKV1-27 germlines. The IgKJ1 segment was selected since it is the most frequently used J-segment in mice and in combination with IGKV10-96.

Figure 1: Design of the cLC mouse model

IgM and IgG serum titers were similar in cLC and WT mice, while IgK concentration was slightly lower in cLC mice (Figure 2). Furthermore, the light chain repertoire of the cLC model is mainly composed of light chains derived from IGKV10-96 (99.7%).

 Figure 2: Antibody serum titers in wildtype and pre-arranged cLC C57BL/6 mice. (a) IgM, (b) IgG, and (c) kappa serum concentration of C57BL/6 mice (wildtype, blue, n = 6) and common light chain mice (cLC, red, n = 6). The difference in concentration of IgM and IgG was not significant using Wilcoxon rank tests. The kappa antibody concentration was determined to be significantly different (Wilcoxon rank test, * p-value < 0.05; n.s. not significant).

A robust immune response was observed in both WT and cLC mice when immunized with OVA, antigen A, or antigen B (Figure 3a), and it was also observed that ~90% of spleen B cells expressed a kappa light chain and that 50–70% of B cells were mature follicular (CD23+/CD21int) B cells (Figure 3b).

Figure 3: Characterization of B cell-specific immune response upon immunization; immune repertoire characterization of immunized mice (N = 2 for the antigen A immunized wildtype mice, N = 3 for the antigen A or antigen B immunized cLC mice, respectively). (a) Antigen A and antigen B serum antibody titration by ELISA. The error bar represents standard deviation. (b) Analysis of B cellpopulations in each experimental group. 

The Ig subclass distribution is similar in cLC and WT mice: the majority of anti-OVA antibodies are of the IgG2b and IgG2C isotypes, while approximately 10% were IgG1. 

Figure 4: Subclass and IgG isotype distribution in the sequenced single-cell immune repertoire of OVA-immunized wildtype C57BL/6 mice (blue, n = 6089) and cLC mice (red, n = 3716)

Finally, the authors showed that despite the expression of a single pre-arranged Vκ germline, a diverse heavy chain repertoire is maintained as all VH families were used. The repertoire is only slightly restricted in cLC mice compared to WT mice, while clonotype diversity is comparable between both lines.

These results establish that the cLC model is a promising tool to reduce the complexity of bispecific antibody development and manufacturing, by enabling the generation of high-affinity antibodies using a common light chain.