SubscribeMonoclonal antibodies (MABs) have been an indispensable tool for research laboratories over decades. This is all the more true in the era of proteomics, where specific and sensitive analytical reagents at protein level are essential.
Until recently, production of innovative MABs was time consuming and laborious. Initially, sufficient amounts of the antigen have to be produced for injection into an animal. When the animal eventually develops an immune response, B cells are isolated and immortalized, finally isolated and then screened for the production of specific MABs.
Fig. 1: Bacteriophages that carry a suitable antibody fragment on their surface are isolated in successive selection rounds. The expression of the antibody genes in E.coli enables rapid production of the antibodies.
This process typically means a wait of 6-9 months before the researchers can work with the MAB. During this period, important projects are delayed and it is by no means sure that the MAB will in fact work in the necessary assays. The production of such MABs can thus become a bottleneck for the research project.
Shorten Unnecessary Waiting Periods
The use of in vitro selection technologies such as the phage display technology (Fig. 1) in combination with large synthetic antibody libraries, has become an established method for the generation of human antibodies that are essential for antibody drug development [1]. In recent years, several biotechnology companies have developed these recombinant technologies and provide them for the development of therapeutic antibodies. This technology also offers benefits beyond the fact that these recombinant antibodies consist of human sequences.
It is considerably faster: After 4-6 weeks, the antibodies are already isolated and after a further 2-4 weeks, purified mg-amounts are available for testing. Following the selection, the antibody genes are available on E.coli expression plasmids which can easily and rapidly be modified, e.g. to alter the valence or to fuse the antibodies with peptide tags or enzymes. Finally, this in vitro method can be automated and allows for the production of antibodies against many antigens simultaneously.
Fig. 2: Jamshidi trepan with representative bone marrow of all three lines of maturity: immunohistochemical detection of the CD33 antigen using #97, 1:500. Strong positive reaction on myeloic precursor cells and monohistiocytic cells; promelocytes are stained most strongly, the more mature cells of the granulopoiesis show an increasingly lower expression of CD33. In addition, individual young megakaryocytes were stained, erythropoiesis was not stained.
Urgently Required: Anti-CD33 Antibodies
A typical project for the generation of antibodies through phage display is outlined here using the CD33 surface antigen as an example. CD33 is a leukocyte differentiation antigen, which is highly expressed in hematopoietic cells and demonstrates a prominent expression pattern on early granulocytopoietic precursors. The expression of CD33 also enables differentiation between myeloic and lymphatic forms of leukemia. This antigen is therefore extremely interesting in regard to tissue-based tumor diagnostics. Although the CD33 gene is cloned, its function is virtually unknown so far.
CD33 belongs to the family of sialoadhesins and has therefore carbohydrate-binding properties. The protein has a total length of 363 amino acids with an extracellular immunoglobulinlike domain of 241 amino acids in length and forms a homodimer. As CD33 is expressed on a lot of myeloic leukemias, it is obvious that antibodies against CD33 are currently being investigated in clinical trials regarding their use in therapy. A well-known example is the humanized antibody Mylotarg which is conjugated with the cytotoxic agent calicheamicin.
So far, however, it has not been possible to detect CD33 specifically on human formalin-fixed tissues using conventional methods, since there are no MABs available that effectively recognize CD33 in this important diagnostic technique. The antibodies developed for this purpose so far proved to be unsuitable possibly due to post-translational modification with several sugar side chains which may mask potential epitopes. A phage display approach was thus attempted to isolate antibody which would allow a modern diagnosis of different types of leukemia by immunohistochemistry (IHC).
Fig. 3: Tissue section of a human tonsil: immunohistochemical proof of CD33 with #97, 1:500. Individual
histiocytes, so-called starry sky macrophages demonstrate a distinct cytoplasmatic grainy reaction.
It is only rarely that granulocytes in small vessels or in tissue are stained weakly positive. All other tissue components are negative for CD33.
Week -2 to 0: Production of the Antigen
The cDNA coding for the amino acids 154 to 259 of the human CD33 precursor (Swissprot P20138) was amplified by using PCR, transferred into an E.coli expression plasmid and purified after expression [see 2 for the method]. A total of 6 mg purified antigen protein were obtained from 125 ml E.coli culture.
Week 1 to 4: Isolation of the Antibodies
The antigen was immobilized on a micro titer plate and panned against the Hu-CAL Gold library (HuCAL, Human Combinatorial Antibody Library, see [3]). This is a collection of more than 15 billion human antibodies in which the antibody genetic information is linked to the respective antibody protein. Non-binding antibodies were removed through washing and the binding antibodies then eluted and the corresponding genetic information was transferred from the phage to E.coli by injection.
Transformed bacterial cells were isolated and several hundred colonies were transferred to a micro titer plate for growth and antibody expression. Following lysis of the cells, the extracts were analysed for binding to CD33 in ELISA. Approximately 20 bacterial expression clones produced an antibody with ELISA signals that were significantly (more than 10 times) above the control (no antigen).
Week 5 to 8: Production and QC
After sequencing of the ELISA-positive antibodies, three candidates, #97, #98 and #99 - each formatted as a bivalent Fab molecule with a Myc and His tag on the C-terminus of the heavy chain - were chosen and expressed in shaking flasks, then purified using NiNTA chromatography. We obtained 1.7 mg (#97), 0.6 mg (#98) and 2.2 mg (#99) purified antibodies from 250 ml of culture. Their specificity was verified and confirmed in ELISA.
Result
The purified antibodies were then tested by IHC on various formalin-fixed tissue segments. One of the three antibodies tested (#97) provided very good results (Figs: 2 and 3). Therefore, by using recombinant antibody technology, a monoclonal IHC-positive antibody could be developed within a few weeks. This CD33 phage antibody can now be used to provide reliable proof of the presence of myeloic cells in formalin-fixed tissue.
Achim Knappik | Hartmut Merz
References
[1] Kretzschmar T., and von Rüden T. Current Opinion in Biotechnology 13, 598-602, 2002
[2] Frisch C. et al.: Journal of Immunological Methods 275, 205-214, 2003
[3] Knappik A. et al.: The Journal of Molecular Biology, 296, 57-86, 2000
Prof. Dr. Hartmut Merz
Medizinische Universität Schleswig Holstein
Campus Lübeck
Ratzeburger Allee 160
23538 Lübeck, Germany
Tel.: +49 451 5002714
merz@patho.mu-luebeck.de
Dr. Achim Knappik
MorphoSys AG
Lena-Christ-Str. 48
82152 Martinsried/Planegg, Germany
Tel.: +49 89 89927 304
knappik@morphosys.com