Integration vectors for antibody chimerization by homologous recombination in hybridoma cells

Production of recombinant granulocyte colony-stimulating factor by knocking into the active immunoglobulin heavy chain gene locus in the hybridoma cell line

The hybridoma cell line KM50 originally produces a monoclonal antibody at a concentration of approximately 40 mg ml(-1) in ascites. To investigate the possibility to apply this expression system to the production of useful proteins, the cDNA encoding human granulocyte colony-stimulating factor was inserted by homologous recombination into just downstream of the promoter of the active immunoglobulin heavy chain gene of KM50. Site directed integration of targeting DNAs resulted in the disruption of expression of the immunoglobulin heavy chain proteins with a frequency of 1 in 10 approximately 100 G418-resistance transfectants. One of the monoclonal antibody-deficient transfectants produced25 ng ml(-1) of granulocyte colony-stimulating factor in the supernatant of its cell culture the number of molecules of which corresponds to that of the monoclonal antibody originally produced by KM50. However, when this transfectant was injected intraperitoneally, it produced only a 9 mug ml(-1) concentration of granulocyte colony-stimulating factor in ascites, which is approximately 3 orders of magnitude less than the monoclonal antibody. This method may be applicable to production of other recombinant proteins, although further optimization in the conditions of production would be needed in order to reach much higher yields.

Instability of chimaeric antibody secretion by anti-carcinoembryonic antigen producing hybridoma cells after gene targeting

OBJECTIVE

To produce a chimaeric version of the 11-285-14 anti-carcinoembryonic antigen (CEA) monoclonal antibody using a gene targeting approach.

MATERIALS AND METHODS

A replacement vector was constructed to insert the human constant gamma1 gene within the mouse heavy chain locus of 11-285-14 hybridoma cells. The mouse constant gamma1 gene (1.5 kb) and the mouse mu intron fragment (2.2 kb) were amplified by PCR and cloned into a pKO Scrambler vector. The human constant gamma1 gene fragment (2.2 kb) was cloned next to the intron fragment. Resistant colonies were screened by ELISA for the presence of the human isotype in their supernatants.

RESULTS

Of the 4,370 resistant colonies obtained, 87 colonies showed secretion of the human isotype at levels between 4 and 32 ng/ml. PCR and Southern blot results confirmed the correct integration of the human gene by homologous recombination within the heavy chain locus. Most of the producers ceased to express the human isotype within a few weeks after the initial positive ELISA results. Instability of secretion could not be explained by genetic instability in all the clones, which suggests the presence of other undefined epigenetic or physiologic mechanisms.

CONCLUSION

Gene targeting resulted in transformants with unstable and low production rates of chimaeric anti-CEA antibody.

Use of a simple, general targeting vector for replacing the DNA of the heavy chain constant region in mouse hybridoma cells

It is often necessary to modify the constant region of the immunoglobulin (Ig) heavy chain in order to produce Ig with optimal properties. In the case of Ig production by mouse hybridoma cells, it is possible to modify the Ig heavy chain (IgH) locus by gene targeting to achieve the desired changes. DNA segments from the JH-S micro region and from the region 3' of Calpha are normally present in the functional IgH gene of all hybridomas, regardless of the heavy chain class which is expressed. Consequently, these DNA segments could in principle serve as 5' and 3' homology regions to create a "universal" targeting vector for replacing the constant region exons in the IgH locus of any hybridoma cell. The practicality of this vector design has been uncertain. That is, the extent of the chromosomal DNA which would be replaced by a universal targeting vector would be as little as 5 kb (in a cell producing the alpha heavy chain) and as much as 180 kb (in a micro -producing cell), and it has been uncertain whether it would be practical to generate such long chromosomal deletions by gene targeting. Using a vector of this design, we found (a) that correctly targeted recombinant cells lacking the 180 kb DNA segment occurred at a low but usable frequency, (b) that these recombinants expressed the modified IgH locus at the same rate as the original hybridoma and (c) that IgH expression in these cell lines was stable. Our results thus indicate that this vector design is suitable for modifying IgH loci expressing any heavy chain, provided that an efficient selection or screening for targeted recombinants is available.

Improving the expression of chimeric antibodies following homologous recombination in hybridoma cells

Mouse/human chimeric antibodies can easily be generated by exchanging the immunoglobulin constant gene segments with human sequences in mouse hybridoma cells by gene targeting. This obviates the need to isolate the variable gene regions from the hybridoma and permits high-level expression of chimeric antibodies, because the production rate of the hybridoma is often maintained. Here we show that the efficiency of this strategy can be further improved by increasing the number of high-producer clones arising from a recombination experiment. In principle, antibody production can be enhanced by removing the selection marker genes from the targeted cells.

Induction of tumor immunity by autologous B lymphoma cells expressing a genetically engineered idiotype

A fusion protein containing a B cell lymphoma idiotype (Id) and granulocyte-macrophage colony-stimulating factor (GM-CSF) is a potent stimulator of tumor immunity. In three different tumor models we show that immunization with autologous lymphoma cells that have been engineered to express the Id in the context of GM-CSF is much more effective than immunization with an equivalent dose of the purified protein. The lymphoma Id could be modified by introducing the GM-CSF gene into the immunoglobulin (Ig) heavy chain locus via gene targeting. This approach circumvents the isolation of the rearranged immunoglobulin variable genes from the tumor and the preparation of tumor-specific vector constructs. The low production of Id/GM-CSF fusion proteins by transfected cells, which is a major obstacle in the use of purified fusion proteins for immunotherapy, is due to the presence of the cytokine gene in the immunoglobulin locus. Low production, however, is not limiting in the cell-based setting, because upon in vivo administration of the modified autologous cells, even minute expression levels are sufficient to induce tumor immunity.

Genetic stability of gene targeted immunoglobulin loci. I. Heavy chain isotype exchange induced by a universal gene replacement vector

Gene targeting at the immunoglobulin loci of B cells is an efficient tool for studying immunoglobulin expression or generating chimeric antibodies. We have shown that vector integration induced by human immunoglobulin G1 (IgG1) insertion vectors results in subsequent vector excision mediated by the duplicated target sequence, whereas replacement events which could be induced by the same constructs remain stable. We could demonstrate that the distribution of the vector homology strongly influences the genetic stability obtained. To this end we developed a novel type of a heavy chain replacement vector making use of the heavy chain class switch recombination sequence. Despite the presence of a two-sided homology this construct is universally applicable irrespective of the constant gene region utilized by the B cell. In comparison to an integration vector the frequency of stable incorporation was strongly increased, but we still observed vector excision, although at a markedly reduced rate. The latter events even occurred with circular constructs. Linearization of the construct at various sites and the comparison with an integration vector that carries the identical homology sequence, but differs in the distribution of homology, revealed the following features of homologous recombination of immunoglobulin genes: (i) the integration frequency is only determined by the length of the homology flank where the cross-over takes place; (ii) a 5' flank that does not meet the minimum requirement of homology length cannot be complemented by a sufficient 3' flank; (iii) free vector ends play a role for integration as well as for replacement targeting; (iv) truncating recombination events are suppressed in the presence of two flanks. Furthermore, we show that the switch region that was used as 3' flank is non-functional in an inverted orientation.

Immunosuppression by Fc region-mismatched anti-T cell antibody treatment

Formation of anti-antibodies (anti-Ab) is known to counteract immunotherapy with anti-T cell antibodies. Our previously described immunological approach prevented anti-Ab with the consequence of prolonged survival of fully mismatched skin grafts in C57BL/6 mice. These mice were treated with a single priming injection of a monoclonal anti-T cell Ab followed by repeated injections of anti-T cell mAb differing in species origin from the priming mAb. We now show prolonged tolerance to discordant xenogeneic, to bispecific, and even to polyclonal Ab, and demonstrate that the underlying immunosuppressive principle is due to a difference in heavy chain constant region between first and second antibodies, independent of whether or not they share the same idiotype. To examine this phenomenon, a panel of mAb was generated which share the same mouse anti-Thy-1.2 idiotype, but carry a human IgG1(T23), IgG3(T212C8), or mouse IgG2a(MmT1) constant heavy chain region. We found that sequential injection of MmT1 and T23 according to the above treatment schedule induced huIgG1 isotype-specific tolerance to T23, which was similar to that seen when using a primary mAb (MmT5) that was, instead, fully mismatched with T23 in both idiotype and constant region. Thus, differences of idiotype between primary and booster Ab were inconsequential for their ability to inhibit anti-Ab formation. This novel form of induced specific tolerance to anti-T cell Ig survived graft rejection and was still evident 230 days after termination of the T cell depletion protocol. Taken together, these results demonstrate that rechallenge with Fc region-mismatched Ab opens an immunological window that allows for induction of tolerance to immunogenic anti-T cell Ab and prolonged immunosuppression.