If you are referring to reducing the titer in the immunized animal, you could give the animal a drug that has immunosuppressive properties. Velcade (PS-341) is a drug that effectively reduces plasma cell counts. There are many immunosuppressive drugs available.
If you want to reduce the antibodies affinity, you are undertaking a much more delicate process. Affinity can be transiently affected by introducing a unfavorable environment for the antibody. You can accomplish this by changing the temp, salt conditions, pH or introducing a denaturing agent like guanidine HCL.
Changing any of these conditions in just a little excess could also result in a permanent reduction in affinity.
If you are attempting to tailor the affinity constant of the Ab by changing the architecture of the antigen, you are correct this can be done using site directed mutagenesis. However I think it would be more effective to perform a random mutagenesis and then screen the mutants for the Ka value you want.
Lastly the antigen could also be post-translationally modified. After performing in vitro transcription and translation you could modify your antigen using different strategies such as ubiquination, phosphorylation, prenylation etc... All these could potentially affect Ka as well.
A Hybridoma system to produce a monoclonal antibody should also be explored. Here is a reputable company that produces monoclonal antibodies:
http://www.sdix.com/Products/Custom-Antibody-Services/Ab-Production-and-Services/mAb-Antibody-Development/Hybridoma-Development.aspx
