Antibodies that can be used in western blotting have to be able to recognise their antigen in denatured form - they predominantly recognise linear epitopes and may have been raised against synthetic peptides. To be able to work in vivo, the antibodies have to recognise the natively folded antigen or natively unfolded regions of the protein and have to recognise epitopes that are accessible - not buried in the interior of the protein, in the membrane  or in protein complexes. In immunohistochemistry, the structure of the antigens may have been altered by the fixing procedure.

Advantages and disadvantages of in vitro and mouse ascites methods for producing mAb are highlighted in this section. It should be noted that it is likely that in vitro methods will meet more than 90% of the needs for mAb. Some of the advantages and disadvantages are concerned with animal-welfare issues. Others deal with the economics of producing mAb. As noted below under the section titled “In Vivo and In Vitro Methods for Commercial Production of mAb”, in vitro methods can cost ½ to 6 times the mouse ascites method. Some of the factors that cause in vitro production to be expensive are labor and equipment costs that are usually due to poor hybridoma production of mAb in vitro. If the investigator must use several types of media or different equipment, as happens occasionally, labor costs rise and research is delayed (Moro and others 1994; Stoll and others 1996b; Butler and Huzel 1995).

https://www.ncbi.nlm.nih.gov/books/NBK100200/

Antibodies that can be used in western blotting have to be able to recognise their antigen in denatured form - they predominantly recognise linear epitopes and may have been raised against synthetic peptides. To be able to work in vivo, the antibodies have to recognise the natively folded antigen or natively unfolded regions of the protein and have to recognise epitopes that are accessible - not buried in the interior of the protein, in the membrane  or in protein complexes. In immunohistochemistry, the structure of the antigens may have been altered by the fixing procedure.

Thanks Mehdi and Annemarie! It's very helpful.  

btw, to make antibody work in vivo, what is the best way to prepare immunogen? According to your answer, the antigen for immunization is better to be the native form of the protein. How should I make the protein and purify it? And should I add foreign protein like KLH with the purified protein to immunize the animal?