Jianghu Upright Antibody Optimization

The process of in vitro affinity maturation requires a lot of time and effort.
The current strategies are usually based on structural principles or a small library approach.
The success of these methods depends on the following two factors:

1.
The existence of high-quality eutectic structure
.

2.
An algorithm that can calculate the energy change when a sudden change occurs
.

The traditional affinity optimization process may take at least 6 months, but when the calculation strategy is combined with the key test evaluation, it can save a lot of time.
Another is that if you rely on the classic crystal structure to simulate the affinity by computer Rather than making an actual antigen-antibody co-crystal structure, it can also save a lot of time
.
Of course, in some traditional methods such as libraries and mutagenesis, the co-crystal structure can even increase the affinity by more than 10-100 times.
In the absence of an antigen-antibody co-crystal structure, computer prediction is actually quite challenging
.

There is a thinking inertia here, that is, the higher the affinity is, the better, in fact, it is not, in most cases, we still need to adjust the affinity of the drug according to the actual mechanism of the drug, so as to maximize the efficacy of the drug
.
In fact, this is a bit like a quadratic function curve, we need to find that vertex
.

Improving affinity is an optimization step, but too high affinity can easily cause some troubles.
Let’s take a solid tumor as an example.
If the affinity of the antibody is too strong, it may cause the macromolecular antibody to only gather at a high concentration near the solid tumor, which hinders To prevent the spread of the antibody in its part, it is not conducive to the drug effect to achieve its maximum effect
.
Another point is to consider that antibodies have different affinities under different pH environments, which is conducive to some of its functions.
For example, some antibodies are easy to decompose under acidic conditions in cells such as lysosomes.
If diffusive, is it possible to release some of the receptors that bind to it, so that these released receptors can return to the cell surface to play a mediating role
?

Specific optimization

In order to reduce the cross-reactivity of antibodies to other antigens, expand the specificity of antibodies to related antigens or improve the binding ability of antibodies to antigens of different species, antibodies need to be engineered to optimize their specificity
.
Random mutagenesis and targeted mutagenesis are two commonly used methods to optimize antibody specificity
.

If the 3D structure of the antibody-antigen complex is available for reference, the human epitope can also be mutated to another germline epitope (for example, monkey epitope or mouse epitope), and prediction can be made by using the "affinity maturation function" of the software And introduce mutations that optimize the affinity of heterologous antigens
.

IgG is the most abundant class of antibodies, and they constitute approximately 75% of the serum immunoglobulin pool
.
Improving the antigen-binding ability of antibodies is mainly determined by the variable regions of IgG
.
IgG is composed of two light chains and two heavy chains.
The light chain contains variable (VL) and constant (CL) domains.
The heavy chain contains one variable (VH) and three constants (CH 1, CH 2 and CH).
3) Domain
.
The multi-domain nature of IgG subtly divides its biological activity into different subdomains
.
The antigen-binding fragment (Fab) contains two variable domains and mediates antigen recognition through six peptide loops called complementarity determining regions (CDR)
.
In contrast, the crystallizable fragment (Fc) contains constant domains (CH 2 and CH 3) that mediate effector functions by binding to immune receptor molecules such as complement proteins and Fc receptors
.

Of course, antibody optimization is not limited to these aspects.
In other aspects, such as conformational stability, solubility, biological effects, etc.
, are very important.
Due to the space limitation of this article, I will not discuss them one by one
.

If the specific proteins or molecules involved in the pathogenesis can be clarified, antibodies are undoubtedly one of the effective treatment options
.
After so many years of development, the therapeutic antibodies on the market can be roughly divided into these two categories-

The first type is the direct use of naked antibodies for disease treatment.
Such antibodies induce apoptosis by directly targeting cancer cells, target the tumor microenvironment (blocking the nutrient supply of cancer cells), and target immune checkpoints (prevent immune escape).
) And mediation pathways (such as ADCC/CDC, where other immune cells are recruited to kill cancer cells in the mediation pathway) to trigger cell death
.

The second category is to make additional modifications to the naked antibody to enhance its therapeutic effect
.
Examples of this are immune cytokines (selected cytokines are fused with antibodies to enhance delivery specificity), antibody-drug conjugates (ADC, enhance drug delivery, reduce non-specific toxic side effects to non-targeted tissues) , Antibody-radionuclide conjugates (ARC, targeted radiotherapy), bispecific antibodies (enhance specific targeting), immunoliposomes and CAR-T
.

When we return to the history of monoclonal antibody development, it is not difficult to find that in fact, new breakthroughs occur almost every year.
I believe that with the further development of science and technology, the improvement of computing power and speed, and the expansion of the three-dimensional protein structure database, precise treatment Efficiency and progress can be further improved and accelerated
.

references:

1.
https://academic.
oup.
com/abt/article/4/1/45/6142994?search result=1#229867976

2.
https://
3.
Cannon DA, Shan L, Du Q, Shirinian L, Rickert KW, Rosenthal KL, et al.
(2019) Experimentally guided computational antibody affinity maturation with de novo docking, modelling

and rational design.
PLoS Comput Biol 15(5):e1006980.

4.
https://