Characterization of scFv-Ig constructs generated from the anti-CD20 mAb 1F5 using linker peptides of varying lengths.

The heavy (VH) and light (VL) chain variable regions of the murine anti-human CD20 mAb 1F5 were cloned, and four single-chain Ab (scFv) molecules were constructed using linker peptides of variable lengths to join the VH and VL domains. Three constructs were engineered using linker peptides of 15, 10, and 5 aa residues consisting of (GGGGS)3, (GGGGS)2, and (GGGGS)1 sequences, respectively, whereas the fourth was prepared by joining the VH and VL domains directly. Each construct was fused to a derivative of human IgG1 (hinge plus CH2 plus CH3) to facilitate purification using staphylococcal protein A. The aggregation and CD20 binding properties of these four 1F5 scFv-Ig derivatives produced were investigated. Both size-exclusion HPLC column analysis and Western blots of proteins subjected to nonreducing SDS-PAGE suggested that all four 1F5 scFv-Ig were monomeric with m.w. of approximately 55 kDa. The CD20 binding properties of the four 1F5 scFv-Ig were studied by ELISA and flow cytometry. The 1F5 scFv-Ig with the 5-aa linker (GS1) demonstrated significantly superior binding to CD20-expressing target cells, compared with the other scFv-Ig constructs. Scatchard analysis of the radiolabeled monovalent GS1 scFv-Ig revealed a binding avidity of 1.35 x 108 M-1 compared with an avidity of 7.56 x 108 M-1 for the native bivalent 1F5 Ab. These findings suggest that the GS1 scFv-Ig with a short linker peptide of approximately 5 aa is the best of the engineered constructs for future studies.

Mitogenic properties of a bispecific single-chain Fv-Ig fusion generated from CD2-specific mAb to distinct epitopes.

The combination of anti-CD2 mAb 9.6 and 9-1, specific for distinct epitopes, induces proliferation of resting human T cells. The mitogenic activity of this mAb mixture depends upon accessory cells and the 9-1 mAb Fc domain. To further study the functional properties of these mAb, their variable regions were cloned and expressed as monospecific single-chain Fv (scFv) proteins fused to the human IgG1 Fc domain (scFvIg). A novel bispecific scFvIg was constructed by cloning the two monospecific scFv binding sites in tandem, with the 9.6 scFv placed N-terminal to the 9-1 scFvIg. Monospecific scFvIg binding to CD2 was comparable to that of the corresponding parental mAb, while the bispecific scFvIg exhibited binding activity similar to that of the 9-1 scFvIg. The combination of 9.6 scFvIg and 9-1 mAb was mitogenic, whereas mixtures including the 9-1 scFvIg were non-stimulatory, confirming the unique properties of the 9-1 IgG3 Fc. Without the IgG3 tail, the bispecific 9.6/9-1 scFvIg was directly mitogenic and was a more potent mitogen than the mAb mixture, but was accessory cell dependent. Unlike the combination of mAb, the bispecific reagent did not directly mobilize calcium in T cells. In comparison to the mAb mixture, bispecific 9.6/9-1 scFvIg-mediated stimulation of a mixed lymphocyte reaction was significantly more resistant to inhibition of the CD28 co-stimulatory pathway by the inhibitor CTLA-4-Ig. These results show that expression of the 9.6 and 9-1 binding sites together on a bispecific scFvIg increased the mitogenic properties of the mAb and altered the degree of accessory cell signals required for T cell activation.

Costimulation by CD28 sFv expressed on the tumor cell surface or as a soluble bispecific molecule targeted to the L6 carcinoma antigen.

Interaction of the CD80 (B7-1) and CD86 (B7-2) molecules on antigen presenting cells with the receptors CD28 and CTLA-4 on T cells generates signals important in the regulation of immune responses. Because this receptor system involves multiple receptor-ligand interactions, determining the function for individual receptors has been difficult. One approach is the use of antibodies and their derivatives with singular specificity as substitute ligands to explore the activities of these molecules. We have constructed recombinant mono- and bi-specific sFv molecules specific for the CD28 receptor that are capable of binding and generating costimulatory signals to activate T cells. We demonstrate that these soluble molecules are capable of higher levels of costimulation than soluble CD80Ig at equivalent concentrations. We also constructed artificial adhesion receptors on the cell surface using two different CD28-specific sFvIgs fused to the CD80 cytoplasmic and transmembrane domains. In this report, we compared costimulation by a soluble bispecific (alpha CD28-alpha L6) single chain sFvIg fusion protein to that generated by L6 antigen positive (L6+) H3347 tumor cells transduced with cell surface expressed forms of alpha CD28 sFv's. We show that the bispecific protein can target potent CD28 costimulatory activity to L6+ tumor cells in vitro. We also show that transfection of the cell surface forms of the two different CD28 sFvIgs into H3347 tumor cells allows them to generate significant costimulatory signals to activated T cells. Finally, we demonstrate that tumor cell presentation of either the soluble bispecific or transduced cell surface sFv generate similar costimulatory effects resulting in T cell activation.

Rapid and reliable cloning of antibody variable regions and generation of recombinant single chain antibody fragments.

Single chain antibody variable region fragments (sFv), by virtue of their size and method of construction are potentially useful as therapeutic reagents and as tools for exploring cell surface receptor function. sFv offer several advantages over the intact immunoglobulin molecule. For instance, they are expressed from a single transcript and can be molecularly linked to other proteins to generate bispecific sFv molecules or single-chain immunotoxins. The relatively small size of sFv is an advantage in allowing for easier penetrance into tissue spaces, and their clearance rate is exceedingly rapid. sFv are useful for gene therapy since they can be directed to a specific cellular localization and can be fused to retroviral env genes to control viral host range. To prepare sFv to murine and human leukocyte CD antigens, we devised a method for rapid cloning and expression that can yield functional protein within 2-3 weeks of RNA isolation from hybridoma cells. The variable regions were cloned by poly-G tailing the first strand cDNA followed by anchor PCR with a forward poly-C anchor primer and a reverse primer specific for constant region sequence. Both primers contain flanking restriction sites for insertion into PUC19. Sets of PCR primers for isolation of murine, hamster and rat VL and VH genes were generated. Following determination of consensus sequences for a specific VL and VH pair, the VL and VH genes were linked by DNA encoding an intervening peptide linker [usually (Gly4Ser)3] and the VL-link-VH gene cassettes were transferred into the pCDM8 mammalian expression vector. The constructs were transfected into COS cells and sFvs were recovered from spent culture supernatant. We have used this method to generate functional sFv to human CD2, CD3, CD4, CD8, CD28, CD40, CD45 and to murine CD3 and gp39, from hybridomas producing murine, rat, or hamster antibodies. Initially, the sFvs were expressed as fusion proteins with the hinge-CH2-CH3 domains of human IgG1 to facilitate rapid characterization and purification using goat anti-human IgG reagents or protein A. We also found that active sFv could be expressed with a small peptide > or = tag > or = or in a tail-less form. Expression of CD3 (G19-4) sFv tail-less or Ig tailed forms demonstrated increased cellular signalling activity and suggested that sFv have potential for activating receptors.

Tyrosine phosphorylation of CD6 by stimulation of CD3: augmentation by the CD4 and CD2 coreceptors.

When T cells are activated via the T cell receptor (TCR) complex a number of cellular substrates, including some cell surface proteins, become phosphorylated on tyrosine (Tyr) residues. Phosphorylation of cytoplasmic Tyr renders these cell surface receptors competent to interact with proteins that link cell surface receptors to protein in the intracellular signaling pathways. Here we show that Tyr residues in the cytoplasmic domain of CD6 become phosphorylated upon T cell activation via the TCR complex. Tyr phosphorylation was observed when the T cells were activated by crosslinking CD3 or by cocrosslinking CD3 with CD2 or CD4, but not when the cells were stimulated by crosslinking CD2, CD4, or CD28 alone. Unlike other Tyr kinase substrates, such as the phospholipase C gamma 1-associated pp35/36 protein, whose level of Tyr phosphorylation is highest when T cells are activated by cocrosslinking CD3 with CD2, the levels of CD6 Tyr phosphorylation are highest when T cells were activated by cocrosslinking CD3 with CD4.

Functional roles of the V3 hypervariable region of HIV-1 gp160 in the processing of gp160 and in the formation of syncytia in CD4+ cells.

To study the roles of the V3 hypervariable region (amino acid residues 301-336) of the HIV-1 envelope glycoprotein (gp160) during infection, we constructed recombinant vaccinia viruses that expressed either wild-type gp160 (v-env10) or mutant gp160 in which the V3 region was deleted (v-dl29.1 and v-dl29.2). In v-dl29.1 the V3 loop, formed by disulfide bonding between cysteine residues 301 and 336, was deleted from cys301 to cys336 (inclusive) and replaced by one serine residue. In v-dl29.2 the V3 loop was deleted from arg303 to ala334 and replaced by three residues: gly-ala-gly. Cells infected with all three recombinant vaccinia viruses expressed gp160 on the cell surface, but v-dl29.1-derived gp160 was not cleaved into gp120 and gp41 and did not bind the CD4 glycoprotein. In contrast, gp160 produced by recombinant v-dl29.2 was cleaved normally, and the mutant gp120 produced was secreted and retained binding activity to CD4+ cells. However, both mutants failed to induce syncytia in HeLa CD4+ cells. Thus a disulfide loop at the V3 portion of gp160 is required for cleavage into gp120 and gp41, presumably because the loop is required for proper tertiary structure. The sequence within the V3 loop, however, is not required for cleavage and secretion of gp160, or for binding to CD4+, but this region is essential for gp120-mediated syncytia formation.

Signal transduction through CD4 receptors: stimulatory vs. inhibitory activity is regulated by CD4 proximity to the CD3/T cell receptor.

The binding of antibody to the CD4 molecule inhibits mobilization of cytoplasmic free calcium ([Ca2+]i) in response to CD3 cross-linking on resting T cells. Similarly, when CD3 and CD4 are independently and simultaneously cross-linked, calcium mobilization is inhibited when compared to that induced by cross-linking CD3 alone. In contrast, when anti-CD4 and anti-CD3 are cross-linked together, calcium mobilization is substantially higher than from CD3 cross-linking alone. A heteroconjugate consisting of covalently bound CD3 and CD4 monoclonal antibodies (mAb) retains the ability to mobilize [Ca2+]i in CD4 cells at protein concentrations approximately two orders of magnitude lower than the free CD3 mAb, and the activity of the heteroconjugate is inhibitable by free CD4 mAb. The CD3/CD4 heteroconjugate also shows significantly greater activity in stimulation of inositol phosphate IP1, IP2 and IP3 synthesis in T cells than the CD3 mAb alone, and again the activity is inhibited by free CD4 mAb. The activity of the CD3/CD4 heteroconjugate is not simply due to oligomerization, since CD3/CD3 or CD4/CD4 homoconjugates or homoconjugate mixtures did not show increased activity. Other heteroconjugates (CD3/CD5 and CD3/CD28) were not different than the CD3/CD3 homoconjugate in their ability to increase [Ca2+]i. Purified CD4 T cells that do not respond to CD3 mAb in solution do respond to the CD3/CD4 heteroconjugate in solution by proliferating in the presence of a CD28 mAb, with a significant fraction of CD4 cells entering the second cycle within the first three days of stimulation. The CD3/CD4 heteroconjugate co-modulates the CD3 and CD4 receptors, indicating that the heteroconjugate is not simply anchoring the T cell receptor to the T cell surface like anti-CD3 on a solid surface. These results suggest that CD4 plays an active role in signal transduction when brought into close physical proximity to the CD3/T cell receptor complex during major histocompatibility complex class II-restricted antigen presentation.

An immunoglobulin light chain dimer with CD4 antigen specificity.

A subclone that had lost its IgG1 heavy chain was derived from a hybridoma cell line G17-2 that produces an anti-CD4 monoclonal antibody. This subclone was found to secrete a kappa light chain dimer (LCD) that could bind to the CD4 antigen expressed on a subset of human T lymphocytes. The light chain dimer bound to the same or similar epitope as the parental antibody since it blocked the binding of the parenteral anti-CD4 MAb but not the binding of another anti-CD4 MAb G19-2 that recognizes a different epitope. A rabbit anti-idiotype prepared against G17-2 crossreacted with the LCD and could block the antigen binding of both G17-2 and the LCD. Scatchard analysis performed with 35S-methionine or 3H-leucine labelled LCD showed an association constant Ka of 2.2 x 10(7) M-1, whereas the G17-2 parental antibody showed an association constant Ka of 2.5 x 10(9) M-1. These data indicate that the antigen specificity of the G17-2 parental MAb is conferred to a large extent by its light chain. The LCD was expressed on the surface of the LCD-producing hybridoma cells. Southern blot analysis with C kappa and J kappa probes demonstrated a single kappa transcription units which does not have any unusual DNA rearrangements and is distinct from the NS-1 kappa genes. To our knowledge, this LCD is unique in its ability to bind to a large (55,000 mol. wt) glycoprotein antigen.

Antibodies to Tp67 and Tp44 augment and sustain proliferative responses of activated T cells.

We have shown previously that binding of a monoclonal antibody (MAb) to Tp44 molecules increased the proliferation of anti-CD3-activated T cells by causing enhanced IL 2 receptor expression and IL 2 release. We now show that anti-CD5 (Tp67) antibodies have a similar effect under conditions in which monocytes are suboptimally activated or where monocytes are not present. The activity did not depend on antibody isotype or on the precise CD5 epitope recognized. Functional experiments indicated that both IL 2 production and IL 2 receptor expression were enhanced by antibody binding. Anti-Tp67 and anti-Tp44 appear to augment proliferation through distinct mechanisms, because both antibodies together had greater activity than either antibody alone. In neither system is the Fc portion of the antibody required, because F(ab')2 fragments had activity equivalent to that of the intact antibody and were effective at concentrations as low as 10 ng/ml. Fab fragments of anti-Tp67 were active, but Fab fragments of anti-Tp44 had no effect. Anti-Tp67 and anti-Tp44 were able to sustain continuous proliferation of anti-CD3-Sepharose-stimulated T cells for up to 2.5 wk without exogenous IL 2 or feeder cells. These experiments suggest that Tp67 and Tp44 are receptors that play a critical regulatory role in the control of T cell proliferation.

Covalent association between human thymus leukemia-like antigens and CD8(Tp32) molecules.

Several TL-like molecules have been identified on human thymocytes. Some of these TL-like molecules are disulfide-bonded to CD8(Tp32), on which they form the thymus-specific 45 kilodalton (Kd) subunit of the CD8 complex. TL-like molecules are associated with beta 2 microglobulin. However, beta 2 microglobulin is not co-precipitated with the CD8-TL complexes, suggesting that CD8 and beta 2 microglobulin are alternative association structures for TL-like molecules. CD8 is present as high m.w. complexes of 72 and 135 Kd that are composed of 32 Kd subunits only, and as high m.w. complexes of 120 and greater than 150 Kd that are composed of both 32 and 45 Kd (TL-like) subunits. A third subunit of CD8 with an apparent m.w. of 52 Kd was seen as part of the 72 Kd complex, and may represent an incompletely reduced dimer of the 32 Kd subunit.

Lyt-2 and lyt-3 antigens are on two different polypeptide subunits linked by disulfide bonds. Relationship of subunits to T cell cytolytic activity.

Lyt-2 and Lyt-3 antigens are carried on separate disulfide-bonded subunits of the same cell surface macromolecules. These are present on thymocytes in a variety of multimeric forms consisting of disulfide-bonded dimers, tetramers, and hexamers of pairwise combinations of three subunits (30,000, 34,000, and 38,000 Mr). From reduced and alkylated Nonidet-P40 thymus extracts, a monoclonal anti-Lyt-3 precipitates only the 30,000 Mr subunit, but not the 30,000 Mr subunit. Almost all of the Lyt-2 and Lyt-3 subunits on the cell are covalently linked by disulfide bonds. However, small amounts of free Lyt-3 subunit was seen in some experiments. Similarly, small amounts of Lyt-2-3- material, consisting of dimers of the 38,000 and 34,000 Mr subunits were identified. Each of the three subunits migrated with a basic charge (pI greater than 8) on two-dimensional gels. Cytotoxic effector cells that are blocked by anti-Lyt-2 and anti-3 can be treated with trypsin and other arginine-specific proteases to remove these antigens. At low concentrations of these proteases, Lyt-3 antigens are selectively removed. After selective removal of Lyt-3 antigens, cytotoxic effector cells are still active and blocking of activity by anti-Lyt-3 is significantly reduced, whereas blocking of activity by anti-Lyt-2 is significantly increased. Neither Lyt-2 nor Lyt-3 is allelically excluded on thymocytes or T cells. These results suggested that the Lyt-2, Lyt-3 macromolecules are associated with but do not serve as the T cell antigen receptor.

Two stages of B-cell memory development with different T-cell requirements.

We present evidence here for two stages in B-memory cell development, the first of which is T independent and the second T dependent. For these studies, we use a new type of T-deficient mouse (allotype suppressed) which specifically lacks T-helper activity (Th) for a subset of memory B cells responsible for approximately 10% of the overall IgG antibody response. We have shown elsewhere that these mice (SJL X BALB/c hybrids suppressed for Ig-1b) lack Th capable of helping Ig-1b memory cells, although they have normal Th activity for all other IgG memory B cells. This selective Th deficiency allows study of the effects of T depletion on memory development and avidity maturation of one population of B cells under conditions where the bulk of the immune response in the animal is proceeding normally, thus obviating environmental problems due to secondary effects of T depletion. With this sytem, we show that after a single priming dose of 2,4-dinitrophenyl-keyhole limpet hemocyanin, the memory B-cell pool in suppressed and nonsuppressed donors is indistinguishable with respect to magnitude and avidity of the response for all IgG antibodies produced, including Ig-1b antibody, despite the fact that expression of Ig-1b memory cells is prevented in intact Ig-1b-suppressed mice by the absence of Th capbale of cooperating with these memory cells. We have shown elsewhere that virtually all of the Ig-1b memory is carried by Ig-1b bearing cells. In contrast with the lack of suppressor T-cell effect on initial Ig-1b memory cell development, our data show that continued Ig-1b memory development is selectively impaired in suppressed mice. When primed mice are boosted repeatedly with the priming antigen, the average avidity of most of the IgG memory cells increases over 100-fold while there is no avidity increase in the Ig-1b component. To explain these data, we suggest that the development of high avidity memory occurs in two stages. The first stage, which occurs as a result of primary antigenic exposure, is the creation of a pool of IgG-bearing memory cells with a relatively low average avidity for the antigen. The appearance of these first stage memory cells does not require help from (post-thymic) Th, although Th are required for the expression of these memory cells (antibody production). The second stage of B-memory development requires both further antigenic stimulation and B-memory cell interaction with competent Th. This is a continuing process in which the number of memory cells in the pool remains relatively constant but the average avidity of these cells increases with continued antigenic exposure.