WO1991005856A1

WO1991005856A1 - Chimeric monoclonal antibodies generated by trans-splicing

Info

Publication number: WO1991005856A1
Application number: PCT/US1990/005797
Authority: WO
Inventors: Robert Steven Becker
Original assignee: Loyola University Of Chicago
Priority date: 1989-10-13
Filing date: 1990-10-10
Publication date: 1991-05-02

Abstract

Chimeric monoclonal antibodies may be produced by a trans-splicing mechanism. Such antibodies may include selected variable or constant regions, or other biological acitivities as combined by trans-splicing. Chimeric monoclonal antibodies or nucleic acids encoding them may be purified and isolatd from cells in which they are generated by trans-splicing.

Description

CHIMERIC MONOCLONAL ANTIBODIES GENERATED BY TRANS-SPLICING

Robert Steven Becker

TECHNICAL FIELD

The present invention relates in general to the production of monoclonal antibodies and in particular to the production of chimeric monoclonal antibodies utilizing the iji vivo trans-splicing of immunoglobulin gene products.

BACKGROUND

A "chimeric monoclonal antibody" is a monoclonal antibody having the antigen binding sites of an immunoglobulin produced by a B cell hybridoma but the heavy chain is changed from the region normally associated therewith to a region of a different heavy chain class. The new constant region may be a heavy chain isotope from the same animal species or from a different animal species; or a fusion protein between a immunoglobulin heavy chain and another biologically active protein.

Monoclonal antibodies are extremely useful for a large number of applications because of their binding specificity for antigens. Monoclonal and polyclonal antibodies are utilized in basic research, medicine and commerce to detect, quantify and localize antigens. Often these assays require that another molecular moiety, such as an enzyme, be associated with the antibody molecules. Commonly, the antibodies and enzymes are conjugated by chemical reactions. This conjugation, however, may also be accomplished using molecular genetic techniques. Genetically engineered chimeric immunoglobulin molecules may be generated using recombinant DNA technology. These molecules may have variable regions from a mouse monoclonal antibody and an immunoglobulin constant region of another species, such as human, or another protein not naturally associated with immunoglobulins.

Immunoglobulin Gene Structure and Transcription Immunoglobulins are composed of two types of protein chains: heavy chains and light chains. Each of these chains has a variable region which is different for each antibody produced and a constant region which is conserved for each class of light and heavy chain. The variable regions of these chains are responsible for the specific binding of an antigen and the constant regions are responsible for the biological activities of immunoglobulins.

The variable and constant regions for heavy and light chains are encoded by different sets of genes (reviewed by Honjo, Annual Review of Immunology, 499, (1983). The two classes of light chain constant regions, kappa and lambda, are encoded by different constant region genes. Each light chain constant region gene has a unique set of variable region genes. Heavy chain constant region genes encode the mu (μ), delta (δ), gamma (γ), epsilon (ε) and alpha (α) classes of immunoglobulin heavy chain molecules. In some species, multiple γ and/or α heavy chain classes exist, unlike light chains, all of the heavy chain constant region genes share the same set of variable region genes. The variable region genes are actually composed of a set of gene segments which rearrange and form functional genes. The heavy chain variable region gene is composed of three types of gene segments: VH, DH and JH segments; there are multiple members of each of these gene segment families. A heavy chain variable region gene is generated by rearranging single members of the VH, DH and JH segment families to form a VDJ gene.

Genes encoding immunoglobulin heavy chains have a number of exons and introns. The variable and constant regions are encoded on separate exons. The last exon of the variable region and the first exon of the constant region are separated by several kilobases. This intron contains the immunoglobulin switch region and an enhancer which potentiates the transcript of the heavy chain genes in B cells.

Transcription of the immunoglobulin heavy chain genes is initiated at the promoter region which is immediately upstream of the variable region gene. After transcription, the exons of the variable and constant regions are spliced together, deleting the introns, to form a functional mRNA molecule encoding the heavy chain [reviewed by Rogers et al., Adv. Immunol. , 35, 39 (1984)]. It is believed in the art that immunoglobulins are produced only by cis-splicing.

Monoclonal Antibodies

Monoclonal antibodies were first generated by Kohler, Nature, 256, 495 (1975) using a HAT-sensitive plasmacytoma as a fusion partner. By fusing normal mouse B cells with the plasmacytoma fusion partner and selecting for a hybrid with HAT medium, hybridoma cell lines are generated which continuously grow and produced monoclonal antibodies of a desired specificity. The plasmacytoma used as a fusion partner, P3X63.Ag8, was derived from the cell line M0PC21. This fusion partner produces a γl heavy chain and kappa light chain of unknown specificity. Hybridomas generated from fusions with P3X63.Ag8 consequently produce two immunoglobulin molecules, one contributed by the normal B cell in the fusion and the other by the P3X63.Ag8 fusion partner. The technical disadvantage of having hybridoma cells producing two immunoglobulin molecules is overcome by selecting for immunoglobulin heavy and light chain loss variants of P3X63.Ag8, such as SP2/0 and Ag8.653 [Kearney et al., J. Immunol. , 123, 1548 (1979); Schulman et al. Nature, 276, 269 (1978)]. These new fusion partners generated by selection allowed for the generation of hybridomas producing only the immunoglobulin molecules contributed by the normal B cells.

Genetically Engineered Chimeric Antibodies

Mouse immunoglobulin variable regions may be associated with human [Morrison et al. Proc. Natl. Sci. (USA), 81, 6851 (1984); Boulianne et al., Nature, 312, 643 (1984); Tan et al., J. Immunol. , 135, 3564 (1985); Sahagan et al., J. Immunol., 137, 1066 (1986)]; rat [Alters et al., J. Immunol. , 142, 2018 (1989)]; and rabbit [Dangl et al., EMBO J., 7, 1989 (1988)] constant regions in chimeric antibody molecules using recombinant DNA technology. Enzymatic activities such as nucleases may be associated with monoclonal variable regions in similar chimeric molecules [Neuberger et al., Nature, 312, 604 (1984)]. The genetically engineered antibodies have characteristics which are useful for basic research and commercial interests. For example, mouse-rat chimeric antibodies specific for the cell surface antigen CD4 are more efficient at killing CD4⁺ cells than are normal rat antibodies with the same specificity (Alters et al., supra.) . Also, genetically engineered immunoglobulins may be used to study the structural and biological characteristics of the various domains within the heavy chain constant regions (Schneider et al., Proc. Natl. Acad. Sci. (USA), 85, 2509 (1988); Dangl et al., supra) . Ultimately, chimeric antibodies may be constructed to have the optimal combination of specificity and effector functions [reviewed by Morrison et al., Ann. N.Y. Acad. Sci., 57, 187 (1987) and Morrison et al., Cancer Invest., 6 , 185 (1988)]. However, the generation of genetically engineered immunoglobulin molecules has been both time consuming and technically difficult [reviewed by Morrison et al., Ann. N.Y. Acad. Sci., 57, 187 (1987)]. The variable region genes for the immunoglobulin heavy and light chains must first be isolated and cloned from the genome of a hybridoma which produces an antibody of desired antigen specificity. Subsequently, these variable region genes are inserted into an expression vector containing the chimeric constant region. This isolation and cloning of variable region genes greatly complicates and slows the generation of chimeric immunoglobulins. These DNA constructs are then transfected into B cell lines and these cell lines consequently produce the chimeric immunoglobulins of interest. Thus, with the current technology, the making of each chimeric immunoglobulin molecule requires the isolation and splicing of both the immunoglobulin variable and constant region genes at the DNA level.

SUMMARY OF THE INVENTION

The present invention provides a process for producing a chimeric antibody which includes the steps of providing a cell expressing DNA encoding a first portion of a chimeric antibody, introducing into the cell DNA encoding a second portion of a chimeric antibody, combining the first portion and the second portion within the cell by trans-splicing, culturing the cell in a medium, and selecting a chimeric antibody comprising the first portion and the second portion from the medium. More particularly, the process preferably includes the step of fusing the cell with a B cell including the DNA encoding the second portion. Any appropriate fusion procedure may be used according to the present invention, but the polyethylene glycol fusion procedure of Oi et al., in Selected Methods in Cellular Immunology, Mishell et al. eds., Freeman and Co., San Francisco, 351-372 (1980) is currently preferred. The process may include the step of transfecting the cell with a vector comprising the DNA encoding the second portion.

The present invention also provides a purified and isolated product of the process according to the present invention. More particularly, the product includes a portion of a chimeric antibody encoding an enzyme or a toxin. The product according to the present invention may include a second portion of a region of a type of antibody different from the type of antibody encoded by a first portion or may include a second portion of a region of an antibody from of a different species from the species of the antibody encoded by a first portion.

The present invention further provides a chimeric antibody wherein a second portion encodes a variable region of an antibody or a constant region of an antibody.

Also according to the present invention provides a process for producing a chimeric antibody- encoding nucleic acid which includes the steps of providing a cell expressing DNA encoding a first portion of a chimeric antibody, introducing into the cell DNA encoding a second portion of a chimeric antibody, combining the first portion and the second portion within the cell by trans-splicin , culturing the cell in a medium, and selecting a chimeric antibody comprising the first portion and the second portion from the medium. More particularly, the process preferably includes the step of fusing the cell with a B cell including the DNA encoding the second portion. The process may include the step of transfecting the cell with a vector comprising the DNA encoding the second portion. A purified and isolated product of the process according to the present invention includes nucleic acid encoding an enzyme or a toxin. The product according to the present invention may include nucleic acid encoding a second portion of a region of a type of antibody different from the type of antibody encoded by a first portion or may include nucleic acid encoding a second portion of a region of an antibody from of a different species from the species of the antibody encoded by a first portion. The present invention further provides a chimeric antibody-encoding nucleic acid wherein a second portion encodes a variable region of an antibody or a constant region of an antibody.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a graph of the results of a radioimmunoassay used to detect the types of constant region in immunoglobulins produced according to the present invention;

FIG. 2 is a graph of the results of a radioimmunoassay used to confirm the specificity of an anti-γl antiserum employed to generate FIG. 1; FIG. 3 is a graph of the results of a radioimmunoassay used to confirm the specificity of an anti-γl antiserum employed to generate FIG. 1;

FIG. 4 is a graph of the results of a radioimmunoassay used to detect the types of constant region present in immunoglobulins produced according to the present invention; FIG. 5 is a graph of the results of a radioimmunoassay used to detect the types of constant region present in immunoglobulins produced according to the present invention; and FIG. 6 is a reproduction of a photograph of an

SDS-PAGE gel illustrating immunoprecipitation studies of immunoglobulins produced according to the present invention.

DETAILED DESCRIPTION

The present invention indicates that chimeric immunoglobulin molecules may be generated by splicing variable and constant region genes subsequent to gene transcription. The utilization of trans-splicing for the generation of chimeric immunoglobulin molecules eliminates the necessity of isolating immunoglobulin variable region genes. Consequently, this technology greatly facilitates basic research and commercial interests which utilize chimeric antibodies.

Trans-splicing of Gene Products

Exons on RNA transcripts are generally spliced together with exons on the same strand of RNA. The splicing of exons on the same strand of RNA is termed cis-splicing. The iτι vivo splicing together of exons on different strands of RNA is reported to occur in several types of eucaryotes including: Caenorhabditis elegans [reviewed by Blumenthal et al., Trends Genet., , 305 (1988)]; Trypanosoma brucei [Sather et al., Proc. Natl. Acad. Sci. (USA)., 2, 5695 (1985); Sutton et al., Cell., 47, 527 (1986)]; Leishmania enriettii, [Miller et al., Nucleic Acids Res., 14, 7341 (1986); Miller et al., Mol. Cell. Biol., 8_, 2597 (1988)], and the chloroplasts of plants [Hildebrand et al., Proc. Natl. Acad. Sci.

(USA), 85, 372 (1988)]. This splicing together of exons from different RNA strands has been termed trans- splicing. Trans-splicing of RNA has also been demonstrated _in vitro. The efficiency of iri vitro trans-splicing may be increased by the presence of mutually complimentary sequences in the introns of the two RNA strands [Solnick, Cell, 42, 157 (1985); Konarska et al., Cell, 42, 165 (1985)]. These complimentary sequences allow the two RNA strands, which trans-splice, to anneal and form a duplex. Complimentary sequences were not, however, a strict requirement for trans- splicing.

Another form of trans-splicing has been demonstrated to take place in human red blood cells. The enzyme glucose-6-phosphate dehydrogenase has been shown to be encoded in two parts on different chromosomes, the chromosome numbered 6 and the X chromosome. Transcripts of these two genes are never trans-spliced at the RNA level. Rather, it is believed that, during the translation of the gene transcripts, the translation process jumps from one mRNA transcript to the second transcript producing a protein which has portions encoded by two different genes. Though the mechanism of this type of trans-splicing is different from that of the trans-splicing of RNA, the resulting protein has the same characteristic of being encoded by more than one, genetically-unlinked gene.

As used herein, the term "trans-splicing" is intended to encompass trans-splicing of RNA, trans- splicing at the level of translation and post- translational trans-splicing.

The trans-splicing of immunoglobulin gene products allows immunoglobulin variable and constant region genes which are not associated at the DNA level to be joined at the RNA and/or protein level. The immunoglobulin chains which result from trans-splicing are chimeric antibodies having the variable region from one heavy chain gene and the constant region from a second heavy chain gene. The trans-splicing of immunoglobulin heavy chains requires that a B cell simulateously express two immunoglobulin heavy chains. Such a B cell, which produce two immunoglobulin heavy chains, may be generated by fusing normal B cells with the P3X63.Ag8 plasmacytoma cell line or by transfecting B cell lines with expression vectors containing functional immunoglobulin heavy chain genes. There is no limitation inherent in the procedure for trans-splicing as to the species of heavy chain that may be associated with a variable region using trans-splicin . For example, using the right DNA constructs, a human or rabbit constant region gene may be associated with a mouse variable region. Moreover, constructs containing constant region genes fused with genes for enzymes or toxins may be used for the production of fusion proteins having a mouse variable region and an enzymatic or toxin activity. The term "constant region" as used herein is intended to include a constant region of an immunoglobulin and any other protein in which may be combined with a variable region by trans-splicin . Chimeric antibodies may also have designed effector functions which will be useful in diagnostic tests and as new therapeutic agents. These types of chimeric immunoglobulin molecules potentially have many purposes in basic research, medicine and industry.

The utilization of trans-splicing greatly improves the efficiency of generating chimeric immunoglobulin molecules. Utilizing trans-splicing, a single DNA construct containing the desired immunoglobulin constant region gene can be used to produce any number of chimeric immunoglobulins with different antigen specificities. The utilization of trans-splicing eliminates the necessity of isolating and cloning the variable region genes to be expressed in association with a desired constant region.

These cells expressing two immunoglobulin heavy chain genes may be generated by cell fusion between two expressing cell lines and by transfection. There are a number of ways of transfecting cells: electroporation, calcium phosphate precipitation, and protoplast fusion. Where the vectors are properly generated, two immunoglobulin genes may be transfected into a non-B cell or a B cell no longer producing immunoglobulin and obtain trans-splicing. This technique may be used to generate chimeric molecules not related to immunoglobulins.

Example 1

Binding Specificity of μ and γl Molecules Produced By 9AE10

Experiments were performed to determine whether the variable and constant region exons of immunoglobulin heavy chains trans-splice in vivo during RNA processing. To address this issue, B cells which expressed two different immunoglobulin heavy chain genes were utilized. Mouse hybridomas producing a μ heavy chain of a known antigen specificity generated by employing the fusion partner P3X63.Ag8 were used for this purpose. These hybridoma cells have two functional immunoglobulin heavy chain rearrangements; one rearrangement, encoding γl, is contributed by the fusion partner P3X63.Ag8 and the other rearrangement, encoding μ, is contributed by the normal B cell in the fusion. These immunoglobulin heavy chain genes have different variable regions which are transcribed and expressed simultaneously by the hybridoma cells. Because the variable region of known antigen specificity is associated with the μ constant region gene and not with the γl gene at the DNA level, it was of interest to determine whether the antigen binding specificity known to be associated with the μ heavy chain was also associated with γl heavy chains. The presence of antigen binding specificity on both μ and γl constant regions may be used as an indicator that the variable region contributed by the normal B cell had trans- spliced with the γl heavy chain contributed by P3X63.Ag8.

The hybridoma 9AE10 was generated by fusing P3X63.Ag8 cells with spleen cells from a Balb/c mouse which had been immunized with rabbit lymphocytes. The hybridoma cells were selected from the non-fused P3X63.Ag8 cells and normal B cells using HAT medium

[reviewed by Oi et al., in Selected Methods in Cellular Immunology, Mishell et al. eds., Freeman and Co., San Francisco, 351-372 (1980)]. The hybridoma clones which grew up were screened for the production of antibodies which specifically bound rabbit T cells.

Of the 543 clones which grew out, 26 were found to have stably-produce antibodies against anti- rabbit lymphocytes. This was determined by mixing the supernatants from the cultures of the hybridoma cells with the rabbit lymphocytes. After washing, ¹²⁵ι- labeled goat anti-mouse Ig was added and the number of bound counts were determined. The monoclonal with anti- lymphocyte activity was tested for the ability to a immunoprecipitate rabbit cell surface protein. The hybridoma 9AE10 was found to efficiently immunoprecipitate a 25,000 dalton protein. Subsequently, the specificity of 9AE10 was tested using immunofluorescent analysis of rabbit T cell and B cell preparations. These analyses demonstrated that the molecule which 9AE10 reacted with was antigen on the surface of rabbit T cells. The hybridoma 9AE10 was found to produce IgM with anti-rabbit T cell specificity, as well as the IgGl which came from the P3X63.Ag8 cells [McNicholas et al.. Immunology, 43, 635 (1981)]. To determine whether the anti-rabbit T cell specificity was associated with both the μ and γl heavy chains, rabbit thymocytes, which are predominantly T cells, were mixed with the 9AE10 culture supernatant. After washing away unbound monoclonal antibody, varying amounts of rabbit anti-mouse μ antiserum, rabbit anti- mouse γl antiserum, or normal rabbit serum were added to the cells. After washing, ¹²⁵I-goat-anti-rabbit immunoglobulin was added to the cells. The protocol utilized in these experiments generally follows that of Tsu et al., in Selected Methods in Cellular Immunology, Mishell et al. eds.. Freeman and Co., San Francisco, 390-394 (1980). The cells were again washed and the number of bound H-i counts were determined using a gamma counter. The association of ¹²⁵ι counts with cells treated with the anti-μ or anti-γl above the counts observed with normal rabbit serum treated cells was used to indicate that μ or γl molecules, respectively, had anti-T cell activity.

In FIG. 1, binding of γl and μ molecules produced by the hybridoma 9AE10 to rabbit T cells is illustrated. In FIG. 1, counts per minute as determined on the gamma counter are read on the ordinate and microliters of antiserum may be determined on the abscissa. Dots indicate the average results of experiments using rabbit anti-μ-antisera, pluses indicate the results of experiments using rabbit anti-γ antisera and asterisks indicate the results of experiments using normal rabbit serum. The bar indicates the standard deviation for each result calculated on the basis of 3 data points. The results of this assay indicated that both the μ and γl molecules produced by the hybridoma had anti-T cell binding activity as shown in FIG. 1. Thus, the immunoglobulin heavy chain transcripts within the hybridoma 9AE10 appear to be trans-splicing.

Example 2

Binding Specificity Of Rabbit Anti-γl Antiserum

To establish that the rabbit anti-γl was specifically detecting γl bound to the antigen, and not cross-reacting with IgM, the anti-γl antiserum of Example 1 was assayed for its ability to bind mouse

IgM. Purified mouse IgM produced by the mouse myeloma TEPC 183 (commercially available from Bionetics, Charleston, South Carolina) was coated on wells of a 96 well plate by physical adherence. After washing the unbound IgM from the wells, varying amounts of rabbit anti-μ antiserum, rabbit anti-γl antiserum and normal rabbit serum were added to the wells. The wells were then washed and ¹²^I-labeled goat anti-rabbit immunoglobulin was added to each well. The wells were again washed and the number of ¹²⁵I-bound counts were determined for each well.

In FIG. 2, binding of rabbit anti-γl antiserum, rabbit anti-μ antiserum and normal rabbit serum to mouse μ molecules produced by the cell line TEPC 183 is illustrated. Along the ordinate may be read counts per minute times 1000 and along the abscissa may be read microliters of antiserum. The error bars and the identity of each antiserum are indicated by the same symbols which correspond to the same number of data points and type of materials as they do in FIG. 1. As shown in FIG. 2, the results indicated that the anti-γl antiserum does not react with the mouse μ. The specificity of the anti-γl antiserum of Example 1 was also tested against an IgM monoclonal antibody which was specific for mouse Thy 1 antigen, an antigen on the surface of mouse T cells. This monoclonal antibody was produced by a hybridoma, 20-10-5, generated from a fusion with the SP2/0 cells [Auchincloss et al., J. Immunol., 128, 1584 (1982)], a heavy and light chain loss variant of P3X63.Ag8; thus, 20-10-5 only produces μ heavy chain molecules. This hybridoma was obtained from the American Type Culture Collection (A.T.C.C.) under the accession number HB 23. The assay was similar to the assay analyzing the culture supernatant of 9AE10 except that mouse thymocytes were used as the antigen.

Thymocytes from Balb/C mice were mixed with supernatant from an HB 23 culture. After subsequent washing, varying amounts of rabbit anti-γl antiserum, rabbit anti-μ antiserum, or normal rabbit serum were mixed with the cells. ¹²⁵I-labeled goat anti-rabbit immunoglobulin was added after washing the cells. After again washing the cells, the number of ¹²^ι counts bound to the cells was determined using a gamma counter. In FIG. 3, binding of rabbit anti-γl antiserum, rabbit anti-μ antiserum and normal rabbit serum to μ molecules produced by the hybridoma HB 23. Along the ordinate, counts per minute are indicated, while along the abscissa, microliters of antiserum may be read. The error bars and the identity of each antiserum are indicated by the same symbols which correspond to the same number of data points and type of materials as they do in FIG. 1.

The results as shown in FIG= 3, indicate that the anti-γl antiserum did not cross-react with μ and that the HB 23 hybridoma did not produce γl molecules with specificity for the antigen. The inability of the anti-γl reagent to react with μ molecules, confirmed that 9AE10 culture supernatants did contain γl molecules with binding specificity for rabbit T cells.

Example 3

Binding Specificity of μ And γl Molecules Produced By Other Hybridomas

To determine whether the binding of γl molecules to the antigen was peculiar to the hybridoma 9AE10, we similarly tested other hybridomas which were obtained from the A.T.C.C. These hybridomas were generated with the fusion partner P3X63.Ag8 and, like 9AE10, produced μ heavy chains of a defined specificity. A clone of the cell line P3X63.Ag8 was deposited on October 12, 1989 as accession number ATCC CRL 10260 with the American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland. The hybridoma TIB 99, obtained from the A.T.C.C, has specificity for the cell surface antigen Thy 1.2, an antigen found on T cells within the C57B1 strain of mice [this hybridoma was originally called HO-13-4; Marshak- Rothstein et al., J. Immunol., 122, 2491 (1979)]. Another hybridoma, identified as TIB 150 by the A.T.C.C, has specificity for the cell surface antigen Lyt 2.2, an antigen found on a subset of T cells within the Balb/C strain of mouse [this hybridoma was originally called HO-2.2; Gottlieb et al., Immunogenetics, 10, 545 (1980)].

The assays for analyzing these TIB 99 and TIB 150 hybridoma culture supernatants were similar to the assay for 9AE10 supernatant except that C57B1 and Balb/C thymocytes were utilized as antigen, respectively.

C57B1 mouse thymocytes were mixed with supernatant from a TIB 99 culture. After washing, varying amounts of rabbit anti-γl antiserum, rabbit anti-μ antiserum, and normal rabbit serum were mixed with the cells. 2~ι- labeled goat anti-rabbit immunoglobulin was added after washing the cells. After again washing the cells, the number of ¹²⁵ι counts bound to the cells was determined using a gamma counter.

In FIG. 4, binding of γl and μ molecules produced by the hybridoma TIB 99 to mouse T cells is illustrated. Along the ordinate, counts per minute are indicated, while microliters of antiserum may be read along the abscissa. The antisera used in conjunction with the generation of the results identified by the points in FIG. 4 are identified by the same symbols which correspond to the same number of data points and type of materials as they do in FIG. 1. The error bars in FIG. 4 are derived in the same way as are the error bars in FIG. 1.

In another experiment, Balb/C mouse thymocytes were mixed with supernatant from a TIB 150 culture. After subsequent washing, varying amounts of rabbit anti-γl antiserum, rabbit anti-μ antiserum, and normal rabbit serum were mixed with the cells. ¹²⁵I-labeled goat anti-rabbit immunoglobulin was added after washing the cells. After again washing the cells, the number of ¹²⁵I counts bound to the cells was determined using a gamma counter.

In FIG. 5, binding of γl and μ molecules produced by the hybridoma TIB 150 to mouse T cells is illustrated. Along the ordinate, counts per minute are indicated, while along the abscissa, microliters of antiserum may be read. The antisera used in conjunction with the generation of the results identified by the points in FIG. 5 are identified by the same symbols which correspond to the same number of data points and type of materials as they do in FIG. 1. The error bars in FIG. 5 are derived in the same way as are the error bars in FIG. 1.

Like the hybridoma 9AE10, the hybridomas TIB 99 and TIB 150 produced γl and μ molecules which bound to the antigen for which the hybridoma was specific, as shown in Figs. 4 and 5. Unexpectedly, TIB 150 produced more γl molecules with binding specificity than μ molecules (the possible significance of this observation is discussed below) . The detection of γl molecules with the same binding specificity as the μ molecules within the culture supernatants of TIB 99 and TIB 150 indicated that the results obtained using supernatants from 9AE10 cultures were not peculiar to one hybridoma or to the clone of P3X63.Ag8 used in the fusion.

Example 4

Binding Specificity Of The γl Molecules Produced By P3X63.Ag8

To determine whether the antigen binding specificities of the γl molecules within the culture supernatants 9AE10, TIB 99 and TIB 150 were contributed by the fusion partner P3X63.Ag8, supernatants from

P3X63.Ag8 cultures were tested for γl molecules which bound to rabbit, Balb/C and C57B1 thymocytes in a procedure generally the same as that used for 9AE10, TIB 99 and TIB 150 and HB 23 in the Examples above. The results of these experiments indicated that P3X63.Ag8 did not produce γl molecules with binding activity for T cells (data not shown). Thus, the anti-T cell binding specificities of the γl molecules produced by the hybridomas must have been contributed by the normal B cells in the fusions. Example 5

Molecular Analysis Of μ And γl Molecules Produced By 9AE10

The immunoglobulin molecules from the culture supernatant of 9AE10 were labeled with ^{1 5}ι and were immunoprecipitated with the anti-μ and anti-γl reagents. The immune precipitates were reduced and alkylated, and subsequently separated by SDS- polyacrylamide gel electrophoresis (SDS-PAGE).

The immunoglobulin molecules were precipitated with 18% sodium sulfate. The precipitate was subsequently dialyzed against saline and was labeled with ¹²^I. One hundred thousand counts of the labeled protein were mixed with the 20 μl of the anti-μ and anti-γl antisera. The antigen-antibody complexes were subsequently brought down with S. aureus, Cowans strain, which has protein A on its surface. The S. aureus- antigen-antibody spellet was reduced, alkylated and the soluble portion run on a 10% SDS-PAGE gel. The markers were purified mouse IgGl and IgM. The gels run at 150 volts and 25 mA for approximately 2 hrs. The gels were stained with Coomassie blue, dried and set up on auoradiographies for 1 to 2 days.

An autoradiograph of the gel was used to identify the heavy and light chains that were precipitated. As controls, ¹²⁵I-labeled IgM molecules produced by the cell line MOPC104E were immunoprecipitated with the anti-μ and anti-γl reagents.

FIG. 6 is a reproduction of a photograph of an SDS-PAGE gel illustrating μ and γl molecules immunoprecipitated with anti-γl and anti-μ immunological reagents. In FIG. 6, ¹²⁵I-labeled γl and μ molecules produced by the hybridoma 9AE10 (in lanes 1 and 2) and 1251-labeled μ molecules produced by the cell line

MOPC104E (in lanes 3 and 4) were immunoprecipitated with rabbit anti-γl (lanes 1 and 4) and rabbit anti-μ (in lanes 2 and 3) antisera. The immunoprecipitates were ~ separated by SDS-PAGE and the labeled peptide chains were visualized by autoradiography. The μ, γl and light chains are indicated by M, G and LC, respectively. The sizes of the heavy chain bands are indicated in kilodaltons (kd) along the ordinate.

The anti-μ reagent immunoprecipitated a μ heavy chain of approximately 70 kd and light chains, as shown in FIG. 6. This reagent did not precipitate any γl heavy chains which are 53 kd in size. The inability of anti-μ to immunoprecipitate γl chains indicated that ¹⁵ the anti-μ did not cross-react with the γl molecules and that the γl and μ chains produced by 9AE10 were not associated with each other.

The anti-γl reagents, however, precipitated heavy chains of sizes indicative of μ and γl chains, as

20 shown in FIG. 6. The precipitation of μ-sized peptide chains by the anti-γl was not the result of cross- reactivity with the μ chains, since the anti-γl reagent did not precipitate μ molecules produced by MOPC104E, a mouse B cell line. These results suggested that some μ ²⁵ molecules produced by 9AE10 have γl antigenic determinants, but the γl chains lacked any μ antigenic determinants.

The presence of μ and γl determinants of Ig chains of 70 kd was confirmed by sequential in immunoprecipitation experiments (data not shown) . Similar analysis of the immunoglobulins produced by TIB 99 and TIB 150 produced results which were similar to the immunoprecipitations of 9AE10 immunoglobulins. These results suggest that some trans-splicing between ~- constant region exons may be taking place in these hybridomas. Example 6

cDNA Copies of Trans-splice Products

The results from the analyses of γl and μ molecules for binding specificity suggest that immunoglobulin heavy chain transcripts trans-splicing during RNA processing. To conclusively demonstrate that trans-splicing does take place, γl mRNA may be analyzed within P3X63.Ag8 cells which are expressing a transfected μ gene. The μ gene, which may be isolated from the hybridoma SP6 [Kohler et al., Eur. J. Immunol. , (5, 511 (1976)], has a variable region of known nucleotide sequence (nucleotide sequence obtained from Dr. Hozumi, unpublished data) and cloned within an expression vector, pR-SP6 [Ochi et al., Proc. Natl. Acad. Sci (USA), 80, 6351 (1983)]. P3X63.Ag8 cells transfected with this construct may express both the endogenous γl gene and the transfected μ gene. The variable region gene associated with the transfected μ gene is present in both μ and γl mRNA.

Other polypeptides of interest may be combined with a variable or constant region by transfecting a cell containing DNA encoding a polypeptide or protein to which the polypeptide or protein of interest is to be spliced with DNA encoding the polypeptide or protein of interest. Fusion proteins may be generated by cloning immunoglobulin exons and exon of enzymes together so that both genes are in the same reading frame. This eliminates any problems with not having the proper RNA splice sites between immunoglobulin and enzyme exons.

Polymerase chain reactions (PCR) [reviewed by White et al., Trends Genet. , 5_, 185 (1989)] may be used to detect and clone the products of trans-splicin . By knowing the nucleotide sequence of the heavy chain variable region genes contributed by P3X63.Ag8 [Bothwell et al., Cell, 24, 625 (1981)] and the transfected gene, the nucleotide sequence of the products of the trans- splicing between the γl and μ transcripts may be obtained.

Oligonucleotide primers which specifically anneal to the first exons of the constant regions of μ and γl and the heavy chain variable region genes from transfected gene and P3X63.Ag8 may be generated. Using the four combinations of the two VH primers with the two CH primers, cDNA copies of the four possible cis/trans- splice products of the μ and γl transcript may be amplified and cloned [Roth et al., Science, 241, 1354 (1988); Lacy et al., Proc. Natl. Acad. Sci. (USA), 86, 1023 (1989)]. The detection, isolation and nucleotide sequencing of these cDNA's demonstrates that trans- splicing does take place in P3X63.Ag8 cells.

Example 7

Transfection with Vector Encoding Transcript for Trans-splicing

The production of antigen specific IgM and IgGl by hybridomas suggests that immunoglobulin heavy chain transcripts trans-splice during RNA processing. To conclusively demonstrate that trans-splicing does take place, mRNA may be analyzed within the hybridoma

SP6. The SP6 hybridoma was generated by fusing an IgM- producing B cell specific for the antigen, trinitrophenol (TNP), with the P3 cell line (Kohler et al., 1976, supra) . Thus, this hybridoma is analogous to the hybridomas, 9AE10, TIB 99 and TIB 150 previously analyzed. The IgGl and IgM molecules produced by this hybridoma may be analyzed for their TNP binding specificity. It is anticipated that both IgM and IgGl bind. This hybridoma may be useful for further analysis because the μ heavy chain and kappa light chain genes which were contributed by the B cell and encode the immunoglobulin chains which are specific for TNP have been cloned and their nucleotide sequence determined (nucleotide sequence obtained from Dr. Hozumi, unpublished data) . The nucleotide sequence of the γl heavy chain gene contributed by P3 is also known (Bothwell et al., 1981, supra) . By knowing the nucleotide sequence of the both the heavy chain genes present within the SP6 hybridoma, the nucleotide sequences of the mRNA molecules which result from the trans-splicing of the two heavy chain transcripts may be predicted.

Polymerase chain reactions (PCR) (reviewed by White et al., 1989, supra) may be used to detect and clone the products of this trans-splicing. PCR utilizes oligonucleotide segments which anneal specifically to the regions upstream and downstream of the gene to be amplified. These annealing segments act as primers for the Taq polymerase; by repeatedly running the polymerase reaction, the region between the two primers is exponentially amplified. Because the nucleotide segments which act as primers specifically anneal to the regions upstream and downstream of the gene of interest, no other regions upstream and downstream of the gene of interest, no other regions of DNA are amplified.

The nucleotide sequences of the heavy chain genes within SP6 are known so that primers may be generated which specifically anneal to the γl and μ constant regions and the two heavy chain variable region genes within SP6. Using these primers, cDNA copies of the cis/trans-splice immunoglobulin mRNA molecules generated in this hybridoma may be amplified and cloned.

The oligonucleotide primers which specifically anneal to the TNP binding variable region gene and the γl constant region gene may be used into amplify the cDNA from the SP6 hybridoma. The combination of these two primers only produces an amplified product in the PCR reactions if the variable region contributed by the μ gene of the B cell is associated with the γl constant gene contributed by P3. Thus, if these primers amplify a cDNA segment, it may be concluded that trans-splicing has taken place. The nucleotide sequence of this amplified DNA may further confirm the existence of trans-splicing in these hybridomas. Transcripts of heavy chain genes introduced into a B cell line by transfection may be demonstrated to trans-splice with transcripts of endogenous heavy chain genes. The trans-splicing of transcripts form endogenous and transfected genes supplies an efficient means of generating chimeric antibodies. The immunoglobulin genes which encode the TNP binding IgM molecule of the hybridoma SP6 may be used for these experiments. Both of these heavy and light chain genes are cloned into the expression vector pR-HL_TNp (Ochi et al., 1983, supra) . This construct may be expressed when transfected into P3X63.Ag8 cells (Ochi et al., 1983, supra) . P3X63.Ag8 cells transfected with this construct express both the endogenous IgGl genes and the transfected IgM genes. Trans-splicing between the two heavy chain transcripts is indicated by the binding to

TNP of both the IgM and IgGl molecules produced by these transfected cells.

Should the SP6 hybridoma and the transfected cells produce IgGl molecules with TNP binding specificity but lack mRNA molecules having the TNP binding variable region gene associated with the γl constant region gene. This may be an indication of trans-splicing at the transcriptional or post- transcriptional level.

Glucose-6-phosphate dehydrogenase (G6PD) from human red blood cells is encoded by two individual genes on separate chromosomes (Kanno et al., 1989 , supra; Henikoff et al., 1989, supra) . Even though these two genes encode portions of a single peptide chain, their RNA transcripts are never spliced together (Kanno et al., 1989, supra) . As an alternative to the trans- splicing of the RNA messages, the protein may be generated by the translation of the first mRNA strand jumping to the second mRNA strand. Another alternative, is that two different proteins are initially produced but these proteins are cleaved and subsequently spliced together. It may be the case that the IgGl molecules with TNP binding specificity are the result of a similar process.

The existence of γl molecules with the TNP binding variable region may be confirmed using a monoclonal anti-idiotypic antibody which is available from Dr. Hozumi. This monoclonal antibody binds specifically to the TNP binding heavy chain variable region of SP6 hybridoma antibodies. Whether this variable region is associated with γl molecules may be addressed by purifying IgGl molecules from the supernatant of the SP6 hybridoma and/or the pR-HL_TNp transfected P3X63.Ag8 cells.

Association of the TNP binding variable region with the γl constant region at the protein but not the mRNA level is an indication that the phenomenon is most likely analogous to the production of G6PD. Trans- splicing at the RNA, translational or protein (post- translational) levels of expression are all intended to come within the scope of the present invention. IgM-producing hybridomas generated by fusion with P3X63.Ag8 cells secrete both IgM and IgG with binding specificity. Thus, one hybridoma mass be used to produce two class of immunoglobulin with binding specificity. This approach is applicable to hybridomas producing immunoglobulin classes other than IgM.

The transfection of heavy chain genes into B cell hybridomas or other B cell lines may be used to generate chimeric antibodies. These chimeric antibodies may have any compatible heavy chain.

The constant regions may be members of a class of mouse heavy chain not originally produced by the hybridoma. This allows efficient generation of monoclonal antibodies of different heavy chain classes all having the same antigen specificity.

The constant regions of heavy chains from other animal species may be utilized to produce chimeric immunoglobulins. These antibodies may have the mouse heavy chain variable regions but the constant regions of, for example, human immunoglobulins. This generation of chimeric antibodies with mouse variable regions and the heavy chain of another species, such as man, involves preparing an expression vector which has a functional human heavy chain constant region. This vector is transfected into immunoglobulin producing B cells or B cell hybridomas, forcing the transfected cells to express both mouse an human immunoglobulins. Trans-splicing within these transfected cells would generate chimeric antibodies. The chimeric antibodies may have numerous diagnostic and therapeutic applications human and veterinary medicine.

Expression vectors which are transfected into the B cells may have chimeric constant regions which are a combination of an immunoglobulin heavy chain and an enzymatically or biologically active protein. These constructs may be used to generate chimeric antibodies which have the antigen binding specificity of an antibody and the biological activity of, for example, an enzyme or toxin or a biologically active substance lymphokine or colony stimulating factor, or of cellular receptor. These types of chimeric antibodies may have uses in diagnostics, clinical medicine, or in industry. These types of chimeric molecules may be useful for any circumstance where a biological activity, such as an enzyme, needs to be associated specifically with a certain chemical entity.

Useful agents which may be conjugated according to the fluorescent agents; toxins (such as ricin) ; enzymes, i.e., besides enzymes which can be used in colorimetric tests, the chimeric antibody with associated enzymes could be used industrially in situations where enzyme used in the processing of a product need to be associated with a specific substance or surface. By making monoclonal antibodies to that substance and subsequently generating chimeric antibodies with the enzyme these enzymatic agents could be generated growth factors, growth factors which may be targeted to specific cells molecules affecting enzymatic activities, pH and ion concentrations. These types of molecules could affect the processes that take place in cytoplasmic vesicles. Possibly affecting such thing as antigen processing and presentation which could be useful in the study and therapy of autoimmune diseases, etc. The antigen binding end would allow us to specifically affect only specific populations of cells. Peptide hormes, including target hormones to specific cells, and DNA and RNA binding factors to allow the binding factor to bind to the DNA or RNA and retain the ability to bind to another molecule via the antigen binding site. While the present invention is described in terms of preferred embodiment, it is expected the variations and modifications will occur to one skilled in the art upon consideration of the present disclosure.

For example, this technique provides the ability to associate any protein with an antigen binding site. Proteins unrelated to immunoglobulins may also be spliced together according to the present invention. Research on the _in vitro trans-splicing indicates that trans-splicing is facilitated by the two RNA strands which are to trans-splice having complementary sequence within the intron which allow them to anneal. Annealing is postulated to associate the two strand non-covalently and allowing for the interaction and splicing of the two exons. A preliminary analysis of the switch sites which are in the intron between the variable and constant region exons and upstream of every other constant region gene indicates that switch regions have this complementary structure. Generating similar structures in the introns of other genes within expression vectors may permit trans-splicing between other cellular proteins.

Besides the generation of unique new molecules this technique may be utilized for gene therapy. The problem with gene therapy at present is the lack of regulation of the gene expression. By putting in a gene via an expression vector which by itself cannot be expressed by a cell but can be trans-spliced with the endogenous defective cell product, two things may be done: 1) Generating of the expression of a functional gene by replacing defective exons by trans-splicing; and 2) obtaining a regulated system because only the cells which would normally express the endogenous protein trans-splice with the gene product of the transfected gene. Moreover, nucleic acid and protein constructions according to the present invention may be isolated by techniques well known to those skilled in the art, such as affinity purification which targets a known affinity of a contruction to be isolated.

Accordingly, it intended that the invention include all such modifications and variations which come within the scope of the claims.

Claims

1. A process for producing a chimeric antibody comprising the steps of: providing a cell expressing DNA encoding a first portion of a chimeric antibody; introducing into the cell DNA encoding a second portion of a chimeric antibody; combining the first portion and the second portion within the cell by trans-splicing; culturing the cell in a medium; and selecting a chimeric antibody comprising the first portion and the second portion from the medium.

2. The process as recited in claim 1 wherein said introducing step comprises the step of fusing the cell with a B cell comprising the DNA encoding the second portion.

3. The process as recited in claim 1 wherein said introducing step comprises the step of transfecting the cell with a vector comprising the DNA encoding the second portion.

4. A purified and isolated product of the process of claim 1.

5. The product as recited in claim 4 wherein a portion of a chimeric antibody encodes an enzyme.

6. The product as recited in claim 4 wherein a portion of a chimeric antibody encodes a toxin.

7. The product as recited in claim 4 wherein a second portion of a chimeric antibody is a region of a type of antibody different from the type of antibody encoded by a first portion.

8. The product as recited in claim 4 wherein a second portion of a chimeric antibody is a region of an antibody from of a different species from the species of the antibody encoded by a first portion.

9. The product as recited in claim 4 wherein a second portion of a chimeric antibody is a variable region of an antibody.

10. The product as recited in claim 4 wherein a second portion of a chimeric antibody is a constant region of an antibody.

11. A process for producing a chimeric antibody-encoding nucleic acid comprising the steps of: providing a cell expressing DNA encoding a first portion of a chimeric antibody; introducing into the cell DNA encoding a second portion of a chimeric antibody; combining the first portion and the second portion within the cell by trans-splicing; culturing the cell in a medium; and isolating a chimeric antibody-encoding nucleic acid from the cells.

12. The process as recited in claim 11 wherein said introducing step comprises the step of fusing the cell with a B cell comprising the DNA encoding the second portion.

13. The process as recited in claim 11 wherein said introducing step comprises the step of transfecting the cell with a vector comprising the DNA encoding the second portion.

14. A purified and isolated product of the process of claim 11.

15. The product as recited in claim 14 wherein a portion of a chimeric antibody-encoding nucleic acid encodes an enzyme.

16.^' The product as recited in claim 14 wherein a portion of a chimeric antibody-encoding nucleic acid encodes a toxin.

17. The product as recited in claim 14 wherein a second portion of a chimeric antibody-encoding nucleic acid encodes a region of a type of antibody different from the type of antibody encoded by a first portion.

18. The product as recited in claim 14 wherein a second portion of a chimeric antibody-encoding nucleic acid encodes a region of an antibody from of a different species from the species of the antibody encoded by a first portion.

19. The product as recited in claim 14 wherein a second portion of a chimeric antibody-encoding nucleic acid encodes a variable region of an antibody.

20. The product as recited in claim 14 wherein a second portion of a chimeric antibody-encoding nucleic acid encodes a constant region of an antibody.