Point mutations cause the somatic diversification of IgM and IgG2a antiphosphorylcholine antibodies.

The genetic mechanism responsible for the somatic diversification of two mAbs was determined. The two PC-binding hybridomas were representative of events early and late in the immune response. The P28 cell line that produces an IgM antibody and thus represents events early in the immune response, was found to have 3 bp changes in its heavy chain variable (VH) region, with some changes in antibody affinity or specificity. The RP93 cell line that produces an IgG2a antibody and thus represents later events in the immune response, was found to have 9 bp changes in its VH region resulting in decreased affinity for PC and altered specificity. Oligonucleotides specific for linked base changes in the second hypervariable regions of both of these antibodies were used to look for previously undescribed V regions or other donor sequences that could have been responsible for these base changes. Since no donor sequences were found, we have concluded that somatic point mutation rather than gene conversion, V region replacement or the expression of an unidentified germline VH region gene is truly responsible for at least some of the somatic diversification of these antibodies.

Monoclonal antibodies reveal the structural basis of antibody diversity.

Hybridoma technology has made it possible to introduce into continuous culture normal antibody-forming cells and to obtain large amounts of the immunoglobulin produced by each of these cells. Examination of the structure of a number of monoclonal antibodies that react with a single antigen has provided new information on the structural basis of the specificity and affinity of antibodies. Comparisons of families of monoclonal antibodies derived from a single germ line gene revealed the importance of somatic mutation in generating antibody diversity. Monoclonal antibodies that react with variable regions of other monoclonals allow the further dissection and modulation of the immune response. Finally, the continued somatic instability of immunoglobulin genes in cultured antibody-forming cells makes it possible to determine the rate of somatic mutation and to generate mutant monoclonal antibodies that may be more effective serological reagents.

The effect of the scid mutation on mechanism and control of immunoglobulin heavy and light chain gene rearrangement.

Most Abelson murine leukemia virus (A-MuLV)-transformed cell lines derived from scid (severe combined immune deficient) mice actively rearrange their endogenous immunoglobulin (Ig) heavy (H), but not light (L) chain variable region genes. Such cell lines express germline VH segments and other RNA transcripts that are characteristically produced by early precursor (pre)-B lymphocytes, but do not express high levels of transcripts from the germline kappa (k) constant region (C kappa) locus. However, we have derived scid A-MuLV transformants that express germline C kappa transcripts and attempt kappa gene assembly. In one case kappa gene expression and rearrangement occurred in the absence of mu H chain expression, and in another was not induced efficiently by introduction of a mu-expression vector. Although the vast majority of scid H and L chain coding sequence joins are grossly aberrant, scid A-MuLV transformants can form normal coding joints at a very low frequency. In contrast, these cells form generally normal signal sequence joins at an approximately normal efficiency. Thus, these findings mechanistically distinguish coding and signal join formation. Subcloning analyses suggest that scid A-MuLV transformants that do not attempt chromosomal coding sequence joining may have a relative survival advantage, and therefore that these events may often result in unrepaired chromosomal breakage and cell death.

Structural heterogeneity of murine IgD during ontogeny.

We have studied the expression of the two membrane delta heavy chains (delta 1 and delta 2) and the two native IgD structures (IgDI and IgDII) in neonatal mice. Both delta-chains appear simultaneously during development and neonatal mice, like adults, express equal amounts of delta 1 and delta 2. IgDI and IgDII also appear simultaneously during ontogeny and in the same ratio as expressed by adult mice of the same strain. Thus, when IgD first appears during ontogeny it shows the same structural heterogeneity as observed in adult mice.

Murine cell surface immunoglobulin: two forms of delta-heavy chain.

The heavy chain of isolated murine cell surface IgD is present in two forms, separable by electrophoresis on SDS-polyacrylamide gradient gels. Both forms of delta-heavy chain are present on the surface of intact spleen cells and have apparent m.w. of approximately 70,000 (delta1) and 68,000 (delta2). Treatment of surface IgD with neuraminidase before isolation results in a single IgD heavy chain band on SDS gels having an apparent m.w. of 65,000, indicating that delta 1 and delta 2 differ in sialic acid content. Delta 2 is removed from the cell surface by papain more readily than delta 1, suggesting a possible functional significance for the two forms.

Murine cell surface immunoglobulin: two native IgD structures.

Murine spleen cell surface IgD is found in two forms upon SDS-polyacrylamide gel electrophoresis under nonreducing conditions. Both IgDI and IgDII are present on the surface of intact cells and appear to be native structures. Neither form could be accounted for by proteolytic degradation or disulfide bond rearrangement. IgDII has an apparent m.w. of 96,000, suggesting an HL structure. IgDI has an apparent m.w. of 150,000, approximately 33,000 daltons lower than that expected for an H2L2 structure. Significant variation in the relative amounts of IgDI and IgDII was found when mice of different strains were examined.

Genetic control of murine IgD structural heterogeneity.

Murine B lymphocytes express two native forms of surface IgD: IgDI consists of two delta heavy chains and two light chains; IgDII consists of one heavy chain and one light chain. The relative amounts of IgDI and IgDII present on spleen cells were found to vary significantly among different strains of mice. Genetic evidence demonstrated that the IgDI/IgDII ratio is linked to the Igh-5 allotype. (The delta heavy chain is the product of the Igh-5 locus). Mice bearing the Igh-5e allotype have a low ratio, and mice bearing the Igh-5a or Igh-5b allotype have a high ratio. The Igh-5 locus and the gene controlling the IgDI/IgDII ratio appear to map to the region between the Igh-6 and Igh-V loci.

Fc receptors on cultured myeloma and hybridoma cells.

The specificity of the Fc gamma receptors on the X63.Ag8.653 nonproducing myeloma cell line has been examined for binding to IgG1-, IgG2a-, and IgG2b-containing antigen-antibody complexes. Complexes containing each of these subclasses bind, and the binding of each is inhibited by the others. Trypsin treatment did not inhibit the binding of any of these subclasses. Furthermore, the monoclonal anti-Fc receptor antibody 2.4G2 inhibits the binding of all three subclasses. These results, together with those of other investigators, suggest that there is a single FcR for IgG1, IgG2a, and IgG2b on mouse B cells which differs in its specificity from the macrophage Fc gamma R. This is confirmed by the fact that a mutant IgG2b myeloma protein which binds to the macrophage Fc gamma 1/gamma 2b receptor does not bind to the Fc gamma R on X63.Ag8.653.

Identification of mutant monoclonal antibodies with increased antigen binding.

Sib selection and an ELISA have been used to isolate hybridoma subclones producing mutant antibodies that bind antigen better than the parental monoclonal antibody. Such mutants arise spontaneously in culture at frequencies of 2.5-5 X 10(-5). The sequences of the heavy and light chain variable regions of the mutant antibodies are identical to that of the parent and the Ka values of the mutants and the parent are the same. The increase in binding is associated with abnormalities of the constant region polypeptide and probably reflect changes in avidity of these antibodies.

The molecular and biochemical characterization of mutant monoclonal antibodies with increased antigen binding.

Mutant mAb with increased Ag binding were generated from a hybridoma cell line, 36-65, that secretes an IgG1,kappa anti-p-azophenylarsonate-(Ars) specific antibody. The mutant antibodies were identified using an Ars-specific ELISA and sib selection so that approximately 10(6) cells could be analyzed. The ELISA used as Ag a low ratio of Ars coupled to BSA and was set up so that only those antibodies that had higher binding than the parent would be detected. Seven mutant producing cell lines were isolated from five independent clones of 36-65. The mutant antibodies bind Ag 20 to more than 200-fold better than the parent and have wild type V region sequences. All have C region mutations that result in an increased avidity. At least five different genetic events are responsible for the C region mutations.

Pharmacokinetic analysis of biosynthetically radiolabeled mouse monoclonal antibodies in normal rhesus monkeys.

Two mouse monoclonal IgG2b antibodies that differ in their CH2 domains have been compared for their blood mean residence times (MRTs) in normal female Rhesus monkeys. A biosynthetically incorporated radiolabel (S-35 methionine) was used to follow blood elimination of the antibodies and urinary excretion. Results indicate that the blood (plasma) MRTs for the Ar13.4 and the ArM16.1 antibodies were remarkably similar in normal Rhesus monkeys. This observation varies greatly from those obtained with these same antibodies in separate studies in normal Balb/c mice. In Balb/c mice, the ArM16.1 antibody was found to be removed from the blood six times faster than the Ar13.4. The CONSAM program was used to fit a multiexponential function to the plasma data obtained from mice and monkeys. The sum of exponentials was then used to calculate the MRT values. These results demonstrate that mouse monoclonal antibody pharmacokinetics significantly differ between normal primates and mice. Presumably the Rhesus monkey better reflects the clinical behavior and pharmacokinetics of mouse monoclonal antibodies than do rodents. Therefore, the use of normal primates for the preclinical evaluation of monoclonal antibodies for in vivo use is suggested.

The production of more useful monoclonal antibodies 2. The use of somatic-cell genetic and recombinant-DNA technology to tailor-make monoclonal antibodies.

In last month's issue(1) Matthew Scharff and his colleagues discussed recent improvements in the technique of making monoclonal antibodies by cell fusion. However, not all the monoclonal antibodies generated by hybridoma technology have all of the properties required for a particular task. This second part of a two-part review deals with ways in which these first-generation reagents can be improved by identifying somatic-cell mutants with the desired properties or by engineering new and even novel molecules using recombinant DNA technology.

Intravascular metabolism of normal and mutant mouse immunoglobulin molecules.

The metabolism of IgG immunoglobulins in the body is tightly regulated in order to maintain their intravascular concentration. Different subclasses may have different intravascular half-lives, and in the mouse, passively administered IgG2b disappears from the circulation more rapidly than IgG2a. We have attempted to localize the sequences in the constant region which are responsible for this difference by examining the intravascular metabolism of mutant immunoglobulins that were generated in tissue culture and have undergone deletions of individual constant region domains or contain different combinations of gamma 2b and gamma 2a CH2 and CH3 domains. Our results suggest that the regulation of intravascular metabolism is complex but indicate that sequences in the CH3 domain are important in determining the different intravascular half-lives of IgG2b and IgG2a antibodies in the mouse.

VH gene usage in humans: biased usage of the VH6 gene in immature B lymphoid cells.

Preferential usage of JH-proximal VH genes has been demonstrated in immature murine B cell repertoires. To determine whether this phenomenon is also evident in human repertoires, we studied utilization of VH6, the most JH-proximal human VH gene. Examination of VH gene usage in a panel of precursor B cell acute lymphoblastic leukemia samples indicated that 15% of the IgH rearrangements utilized VH6. VH6 is a single-member family in a total repertoire of 100-200 VH genes; thus, if usage were purely random, one would expect VH6 rearrangement frequency to be less than 1%. Analysis of VH gene usage in normal lymphoid tissues also revealed biased usage of VH6. VH6 was preferentially utilized in 16- to 24-week-old fetal liver as compared to adult peripheral blood mononuclear cells or spleen. Possible implications of the conservation of preferential usage of JH-proximal genes in both immature murine and human repertoires are discussed.

Studies on the somatic instability of immunoglobulin genes in vivo and in cultured cells.

We have examined the molecular mechanism and impact of somatic diversification on the T15 heavy chain variable region gene in vivo and in vitro. Somatic point mutation appears to be responsible for the changes we have observed in both hybridomas from early and late in the immune response and in the S107 myeloma cell line in culture. By identifying S107 mutants with decreases in antigen binding, we have shown that a single point mutation can cause the loss of binding to the eliciting antigen and the acquisition of binding to another antigen. Furthermore, in this case a point mutation of the T15 heavy chain variable region gene caused the conversion of an important protective antibody to an autoantibody. While the S107 cell line frequently generates both constant and variable region mutants, hybridomas appear to have relatively stable variable region genes and unstable constant region genes which in some cases result in mutants with increased binding.

Isolation of scid pre-B cells that rearrange kappa light chain genes: formation of normal signal and abnormal coding joins.

Consistent with an ordered immunoglobulin (Ig) gene assembly process during precursor (pre-) B cell differentiation, we find that most Abelson murine leukemia virus (A-MuLV)-transformed pre-B cells derived from scid (severe combined immune deficient) mice actively form aberrant rearrangements of their Ig heavy chain locus but do not rearrange endogenous kappa light chain variable region gene segments. However, we have identified several scid A-MuLV transformants that transcribe the germline Ig kappa light chain constant region and actively rearrange the kappa variable region gene locus. In one case progression to the stage of kappa light chain gene rearrangement did not require expression of Ig mu heavy chains; furthermore, this progression could not be efficiently induced following expression of mu heavy chains from an introduced vector. As observed in pre-B cell lines from normal mice, attempted V kappa-to-J kappa rearrangements in scid transformants occur by inversion at least as frequently as by deletion. The inverted rearrangements result in retention of both products of the recombination event in the chromosome, thus allowing their examination. scid kappa coding sequence joins are aberrant and analogous in structure to previously described scid heavy chain coding joins. In contrast, the recognition signals that flank involved coding segments frequently are joined precisely back-to-back in normal fashion. The scid VDJ recombinase defect therefore does not significantly impair recognition of, site-specific cutting at, or juxtaposition and appropriate ligation of signal sequences. Our finding that the scid defect prevents formation of correct coding but not signal joins distinguishes these events mechanistically.