Couto J R, Christian R B, Peterson J A, Ceriani R L
Cancer Research Fund of Contra Costa, Walnut Creek, California 94596, USA.
Cancer Res. 1995 Dec 1;55(23 Suppl):5973s-5977s.
A humanized antibody retains from the original murine antibody the variable region amino acid residues that are required for antigen binding. These generally include the grafted complementarity determining regions, as well as a few key framework residues. Although the remainder of the framework sequences are imported from a human antibody, they nevertheless differ at a few positions from the human consensus sequences. These atypical residues, which arose by somatic mutation during the affinity maturation of the chosen human antibody, could elicit an immune response in some of the patients receiving the humanized antibody. Thus, ideally one should, instead, choose human consensus frameworks for humanizing murine antibodies. Because there is a different consensus sequence for each of the subclasses of variable light and heavy chains, a method is needed to choose the most appropriate one. We are developing a minimal positional template for such a purpose. A minimal positional template indicates which positions in the variable region frameworks are absolutely required for maintaining the integrity of the binding domains. Therefore, to choose a human framework for humanization, one screens the available human consensus sequences for those that are most similar to the original murine sequence, specifically at the positions indicated by the template. In the subsequent humanization protocol, one then retains all of the murine residues found in the positions indicated in the template while humanizing the residues found at all other positions. A conservative positional template has been applied to the humanizations of the antibreast epithelial mucin antibodies BrE-3 and KC4-G3 without loss of binding affinity. Now we are using progressive cycles of computer modeling and laboratory testing to develop a minimal template. The first of such cycles produced template B, which has been used successfully in the humanization of the antibreast epithelial antigen BA46 antibody Mc3. This prompted us to design template C, which further approaches the desired minimal template. Future constructs will test the validity of this template as well as the validity of this novel humanization approach.
人源化抗体保留了原始鼠源抗体中抗原结合所需的可变区氨基酸残基。这些通常包括移植的互补决定区以及一些关键的骨架残基。尽管骨架序列的其余部分是从人抗体导入的，但它们在一些位置上与人类共有序列仍存在差异。这些非典型残基是在所选人抗体亲和力成熟过程中通过体细胞突变产生的，可能会在一些接受人源化抗体的患者中引发免疫反应。因此，理想情况下，应该选择人类共有骨架来对鼠源抗体进行人源化。由于可变轻链和重链的每个亚类都有不同的共有序列，因此需要一种方法来选择最合适的序列。我们正在为此目的开发一种最小位置模板。最小位置模板表明可变区骨架中的哪些位置对于维持结合域的完整性是绝对必需的。因此，为了选择用于人源化的人骨架，人们会在可用的人类共有序列中筛选那些与原始鼠源序列最相似的序列，特别是在模板指示的位置。在随后的人源化方案中，在对所有其他位置发现的残基进行人源化时，会保留在模板指示位置发现的所有鼠源残基。一种保守的位置模板已应用于抗乳腺上皮粘蛋白抗体BrE - 3和KC4 - G3的人源化，且未丧失结合亲和力。现在我们正在使用计算机建模和实验室测试的递进循环来开发一种最小模板。第一个这样的循环产生了模板B，它已成功用于抗乳腺上皮抗原BA46抗体Mc3的人源化。这促使我们设计模板C，它更接近所需的最小模板。未来的构建体将测试该模板的有效性以及这种新型人源化方法的有效性。
