This invention relates to novel monoclonal anti-RhD antibodies prepared by recombinant DNA methods.

The Rhesus blood group system is a major antigenic constituent of the human red blood cell membrane; of this group, the RhD antigen is of particular clinical importance in relation to isoimmune reactions. An Rh D- individual with c.nti-RhD who receives RhD+ blood is liable to suffer substantial red blood cell (RBC) destruction due to the RhD phenotype incompatibility, and thus blood of donors must routinely be classified as RhD+ or RhD-. Anti RhD monoclonal antibodies (antiD

Mabsj are capable of providing blood-typing reagents of high specificity and reliability.

The RhD antigen is also responsible for haemolytic disease of the newborn (HDN) . This condition arises in newborn RhD+ infants of RhD- mothers previously sensitised to RhD antigen as a result of IgG anti-RhD antibodies crossing the placenta ⁴ -during pregnancy and causing foetal red blood cell (RBC) destruction. Sensitization of the RhD- mother to RhD antigen often occurs during the birth of an earlier RhD+ child due to some ioetal RBCs entering the maternal circulation and being recognised as foreign by the maternal immune system. To reduce the incidence of HDN, it is routine practice in the United Kingdom and many other countries to give anti-RhD antibodies to RhD- mothers immediately after the birth of an RhD+ infant so that any RhD+ RBCs which may have entered the maternal circulation are rapidly removed.

The search for the most effective anti D Mabs has proved to be extremely time consuming, involving the isolation of B-lymphocytes from humans immunised against RhD, usually Rh-ve mothers who have given birth to Rh+ve

children. Such lymphocytes are subjected to EBV treatment to provide an immortalised cell-line directly or the EBV-treated cells are hybridised with suitable mouse myeloma cells to provide a hydridoma: The cell- line or hybridoma may then be used to produce the anti-D Mab in the conventional way.

However, there are significant differences between anti-D Mabs in terms of their binding affinities for red cells, their ability to recognise D-variants such as D ^u and D ^I , and their ability to destroy target cells by phagocyi osis or cell-mediated lysis. It is desirable, therefore, to have available a method of combining the favourable parameters of different anti-D Mabs or, indeed of combining the most favourable features of selected anti-D Mabs wjjth Mabs of quite different specificities which present particular advantages, in order to produce so-called chimaeric Mabs.

The concept of building chimaeric Mabs, has been described by Jones e _^ al (Nature 321. 522-525 (1986)) and Riechiαann et al (Nature 332. 323-327 (1988)). Three dimensional studies have shown that immunoglobulins comprise essentially constant regions common to most Mabs and terminally situated variable domains associated with antigen binding. It has been shown that the variable domains consist of two .-.sheets joined by a disulphide bridge with their hydrophobi.c faces in contact. Sequence comparisons among heavy- and light-chain variable domains (V _H and V _L respectively) have revealed that each of these domains comprises three hypervariable domains or complementarity determining regions (CDRs) set in a framework of four relatively conserved regions, the framework regions (FRs) . The CDRs are primarily responsible for the recognition of specific antigens. The structure of the /3-sheet framework is similar in different antibodies, as the packing together of V _L and V _H FRs is conserved and therefore the orientation of V _L with respect to V _H is

fixed.

Genes coding Zor a number of Mabs are now available and the sequences coding for the variable regions V _L -and V _H have been determined. It is thus possible to replace the latter sequences by DNA coding for V _L and V _H from different Mabs and indeed to construct the latter by incorporating DNA coding for chosen CDRs into DNA coding for a standard set of FRs. It is thus possible to construct genes coding for chimeric anti-D Mabs having the CDRs from anti-D Mabs possessing particularly desirable specificities or other properties and framework and constant regions derived from Mabs having other desirable properties.

It is a prerequisite of such construction that the amino acid sequences of the CDR regions of the chosen anti-D Mabs and/or the genes coding for them, should be known. The specific CDR gene sequences can then be synthesised, conveniently by chemical synthesis of the appropriate oligonucleotides, and incorporated into DNA sequences coding for a standard set of FRs and the human or other) constant region. Of course, the FRs may be 'Identical with those of the Mab providing the constant region or, more conveniently, they may be a standard set of FRs which can be used generally in the synthesis of chimeric Mabs.

We have produced a number of anti-D Mabs of particular interest and have determined their amino acid sequences, thus making it possible for DNA sequences corresponding to their CDRs to be synthesised and incorporated into V _H and V _L sequences as described above. These may then be combined with DNA coding for the constant region to enable novel anti-D Mabs to be synthesised which may have lower, the same or higher binding ability. Thus, according to one aspect we provide DNA sequences comprising oligonucleotides encoding CDR1, CDR2, and CDR3 regions of _H and _L domains of antibodies

against the human RhD antigen, and functional equivalents thereof. In particular, we have investigated and sequenced eleven Mabs, namely a) FOG-B, b) PAG-1, c) MAD-2, d) FOG-1, e) FOM-1, f) FOM-A, g) BRAD-3, h) JAC-10, i) GAD-2, J) REG-A, K) HAM-B, whose heavy and light chain sequences are represented in figures 2-14, of the accompanying drawings, and which have both varied and particularly useful binding specificities. The figures 2 and 3 show the nucleotide and amino acid sequences of the light chain variable domains of the Mabs FOG-B and PAG-1. Corresponding sequences for the heavy chain variable domains of these two Mabs are shown in figures 4 and 5, and sequences of the heavy chain variable domains of the Mabs MAD-2, FOG-1, FOM-A, BRAD-3, JAC-10, GAD-:., REG-A and HAM-B are shown in figures 6-14.

Synthetic genes, for both heavy and light chains may be created by combining selected CDR 1, 2, and 3 regions, which may be selected from different antibody molecules having varied binding specificities.

Thus according to a.further aspect, we provide DNA molecules coding for the heavy or light chain fragments of a monoclonal antibody or fragment thereof comprising CDR1, CDR2 and CDR3 encoding oligonucleotides from antibodies FOG-B, PAG-1, MAD-2, ^' FOG-1, FOM-1, FOM-A,

BRAD-3, JAC-10, GAD-2, REG-A and HAM-B as illustrated in ^• figures 2-14.

In order to create functional genes, such oligonucleotides must be incorporated into a backbone sequence such that when expressed, functional proteins result.

Thus according to a further aspect, we provide DNA molecules comprising a gene coding for the framework regions of a human antibody light or heavy chain having inserted therein in the correct CDR region, oligonucleotides encoding CDR1, CDR2 and CDR3 regions according to the present invention.

In the synthesis of a chimeric Mab in accordance with the invention, single stranded DNA coding for the V _H region of a chosen Mab (not necessarily an anti-D Mab) is incorporated in single stranded form into a vector capable of producing single stranded DNA, such as the M13 bacteriophage. Fig. 1 shows diagrammatically the structure of a single stranded V _H DNA including framework regions FR1 to FR4 with complementarity determining regions CDR1 to CDR3 of a Mab. Theae steps can be accomplished by conventional techniques such as those described in Riech ann et al (Nature, 332. 323-327, (1988)).

Three oligonucleotides may then be prepared corresponding to the CDR regions of the chosen anti-D Mab variable domain, eg the V _H regior of FOG-B as shown in Fig. 4, and will include several nucleotides on either side of each CDR region to permit hybridisation with the framework regions FR1 to FR4 (see figure 1) . The sequences of the latter will normally be substantially homologous with those of the anti-D Mab (e.g. FOG-B) but since the oligonucleotides will normally be synthesised -chemically, hybridisation may be ensured by matching the overlapping nucleotides exactly to the FRs 1 to 4. It may also be beneficial to modify the oligonucleotides to express the CDRs more efficiently in the eventual host cells.

The three oligonucleotides, shown in Fig 1 as oligo 1 to oligo 3, may then be annealed to a single stranded V _H DNA in the M13 vector and used as primers to synthesise second strand DNA containing the anti-D V _H CDR sequences. This may be achieved conventionally using a suitable polymerase. Since the antibody specificity is determined solely by the three CDR regions, the actual V _H gene chosen for the framework template is immaterial. All that is required is that there is sufficient homology of the three chosen oligonucleotides with the template. This is ensured by appropriate design of the

terminal nucleotides of the synthetic oligonucleotide primers. Thus the second strand may contain sequences from substantially any human antibody heavy chain gene, so long as the resulting expressed protein posesses the desired binding parameters.

The double stranded M13 vector may then be used to transform a suitable host microorganism e.g. a conventional E. coli and one or more clones selected which contain the required anti-D V _H specificity. The correct clone may be identified by DNA seqencing.

The corresponding V _L DNA (e.g. for FOG-B) may be prepared in the same way.

The DNA coding for the V _H and V _L regions may then be excised from the above vectors and introduced into other vectors. _t

According to a further aspect, we provide DNA molecules being synthetic genes for chimaeric antibody, heavy or light chains when incorporated into vectors capable of expressing such antibody chains. Preferred vectors include mammalian expression vectors, such as - pSV2gpt (heavy chains) and pSV2neo (light chains) containing DNA coding for the human constant region. Such vectors are readily available from a number of laboratories, or can readily be prepared by incorporating DNA coding for human constant region into known mammalian vectors.

The expression vectors so constructed may then be co-transfected into an appropriate cell-line e.g. a non- secreting IgG myeloma, for large scale production. Thus according to a yet further aspect, the present invention provides each of the CDR polypeptides of the Mabs FOG-B, PAG-1, MAD-2, FOG-1, FOM-1, FOM-A, BRAD-3, JAC-IO, GAD-2, REG-A and HAM-B shown in Figs. 2-14 of .the accompanying drawings in single stranded or double stranded form in the absence of the constant and or framework regions of said Mabs.

According to a yet further aspect, the invention

provides chimaeric antibody heavy and light chains of the variable domains comprising CDR polypeptide sequences of the present invention.

Knowledge of the antibody sequences according to the invention enables new chimaeric anti-D antibody molecules to be prepared, having appropriately designed binding specificities. These antibodies may be used for both therapy and diagnosis using presently known techniques. According to a yet further aspect, we provide anti- RhD reagents comprising at least one antibody molecule according to the invention.

According to a still yet further aspect, we provide pharmaceutical compositions for use in passive immunisation to prevent haemolytic disease of the newborn comprising an antibody of the present invention together with at least one phamacologically acceptable carrier or diluent.

A sterile solution of such an antibody for human injection may be formulated in any physiologically acceptable aqueous medium, for example isotonic phosphate buffered saline or se um. Alternatively, the antibody may be supplied in a freeze-dried" formulation ready for reconstitution prior to use.

EXAMPLE

(1) Construction of Chimaeric Antibody Genes

Three oligonucleotide primers are synthesised using an Applied Biosystems machine according to the manufacturer's instructions and purified on an 8 M U ea/polyacrylamide gel (Sanger & Coulson, Febs Lett. , • JLZ. 107-110, 1978). The primers are designed to comprise in their central regions sequences complementary to the CDR1, CDR2 and CDR3 regions of the . anti-RhD antibody PAG-1 heavy chain gene, as identified according to the criteria described by Kabat et al. (Sequences of Proteins of Immunological Interest, US Department of Health and Social Services, 1987) .

The central sequences are flanked at both their 5' and 3 ' -termini by sequences of 10 nucleotides which hybridise to the termini of the corresponding framework region sequences adjacent to the CDR sequence of the heavy chain antibody gene NEWM (Poljack et al. ,

Biochemistry 1J5, 3412-3420, 1977) . The primers are then hybridised to the derived NEWM single stranded DNA heavy chain sequence in the M13 bacteriophage and the complementary strand of the heavy chain variable region extended using DNA polymerase (Neuberger et al., Nature 314. 268-270 (1985), Jones et al., Nature 321, 522-5 (1986)). The M13 vector also contains an appropriate arrangement for ultimate expression, i.e. a leader sequence, and unique Hindlll and BamHI restriction sites.

A similar construct is prepared from oligonucleotide primers homologous to the CDR regions of the PAG-1 anti-RhD antibody light chain genes, and utilising the M13 vector in which V _L and J _L regions of the antibody gene PAV1 (Sun et al., Nucleic Acids Research 22, 4921-4934, 1985) are cloned.

(2) Expression of Antibody Polypeptides

The cloned genes for the V _H domains are excised using Hindlll and BamHI and cloned into pSV2gpt

(Mulligan and Berg, 'PNAS 78 _. , 2072-6, 1981). The cloned light chain genes are similarly excised and cloned into pSV2neo (Southern and Berg, J. Molec. Appl. Genetics 327-381, 1981) . Sequences encloding IgGl constant regions are then inserted into the vectors (Riechmann -et al., Nature 312, 323-7, (1988). Both vectors are then transfected by electroporation (Potter et al., PNAS 81. 7161-3, 1984) into the rat myeloma cell line YO (YB2/3.0 AG, 20) (Galfre and Milstein, Methods in Enzymology 3., 1-46, 1981) for antibody production.