Clonal analysis of the primary and secondary B cell responses of neonatal, adult, and xid mice to (T,G)-A--L

The splenic focus assay was used to clone B cells from neonatal, adult and xid mice in order to examine their primary and secondary responses to (T,G)-A--L. Adult precursor cell frequencies to (T,G)-A--L were achieved late in neonatal ontogeny. Primary xid B cells responded to DNP-HY but not to (T,G)-A--L in the splenic focus assay. The frequency of secondary B cells from (T,G)-A--L-primed xid mice was less than or equal to 10% that of secondary B cells from wild-type (non-xid or X/Xxid heterozygous) mice. Although xid B cells were poorly responsive to (T,G)-A--L in the splenic focus assay, (T,G)-A--L-primed xid mice could provide help as recipients for stimulation of wild-type primary and secondary B cells. It seems likely that the B2 subset contributes most of the splenic focus response to (T,G)-A--L. The fine specificities of antibodies produced by neonatal, xid, and adult (wild-type) B cell clones were analyzed using analogues of (T,G)-A--L. A specificity shift was observed between the adult primary and secondary antibody responses to (T,G)-A--L. Less than 10% of adult primary clones produced antibodies cross-reactive on (Phe,G)-A--L (recognizing A--L determinants or Phe,Glu determinants), whereas more than 70% of primary clones produced Tyr,Glu side-chain specific antibodies cross-reactive on GT. The percentage of clones producing GT-binding antibodies diminished in the secondary response, while the percentage of clones producing antibodies cross-reacting on (Phe,G)-A--L increased. Neonatal clones also produced mostly GT-binding antibodies but gave a higher percentage of (Phe,G)-A--L-cross-reacting antibodies than adult primary clones. The specificities of secondary antibodies produced by xid and wild-type B cell clones were dissimilar. First, xid secondary clones were "primary-like" in that no anti-A--L antibodies were detected. Second, clones whose antibodies bound side-chain determinants but not GT were produced in higher frequency by xid than by wild-type secondary B cells. The differential responsiveness of B cell subsets to antigen and regulatory signals may influence memory B cell generation and the specificity of antibodies produced in the primary vs secondary response.     

Idiotypic analysis of hybridoma antibodies to branched synthetic polymer (Tyr, Glu)-Ala-Lys: idiotypic relationship with antibodies to linear random polymer (Glu60, Ala30, Tyr10)

The expression of three anti-GAT idiotypes, CGAT, Gte, and GA-1, on 17 C57BL/10 and four C3H.SW hybridoma anti-(T,G)-A--L antibodies was analyzed. These hybridoma anti-(T,G)-A--L antibodies exhibited two patterns of fine antigen binding specificity. The majority of the hybridoma antibodies bound the (T,G)-A--L, GT, and GAT polymers but not the GA polymer, and were designated as GT-reactive hybridoma antibodies. A minor population of hybridoma anti(T,G)-A--L antibodies bound to (T,G)-A--L but not to GT, GAT, or GA, i.e., (T,G)-A--L-specific. A complete correlation between fine antigen binding pattern and the expression of CGAT idiotype was demonstrated. None of the 21 hybridoma anti-(T,G)-A--L antibodies expressed the GA-1 idiotype. All of the GT-reactive and none of the GT-nonreactive hybridoma anti-(%,G)-A--L antibodies expressed the CGAT idiotype. Furthermore, the Gte idiotype was found on the majority of CGAT+-bearing C57BL/10 hybridoma anti-(T,G)-A--L antibodies. These results indicate that C57BL/10 anti-(T,G)-A--L antibody repertoire can be grouped into a minimum of three families; i.e., CGAT+ Gte+, CGAT+ Gte-, and CGAT- Gte- families, with the CGAT+ Gte+ family as the major compartment. This is confirmed by the high percentage idiotype binding of serum anti-(T,G)-A--L antibodies with anti-CGAT idiotypic antisera. Finally, anti-idiotypic antisera made against CGAT+ hybridoma anti-GAT or anti-(T,G)-A--L antibodies crossreact extensively with other CGAT+ hybridoma anti-GAT and anti-(T,G)-A--L antibodies. However, additional experiments demonstrated that CGAT+ hybridoma anti-(T,G)-A--L antibodies also possess private idiotypes.     

Xid and normal mice express a light chain-associated cross-reactive idiotype in response to (T,G)-A--L and (T,G)-A--L-mBSA

Rabbit anti-idiotypic (Id) antibodies were prepared against purified ascites anti-(T,G)-A--L antibodies (TGB5) that had been absorbed to remove A--L-specific antibodies and were specific for (T,G)-side chain determinants. Purified rabbit anti-TGB5 Id antibodies detected an allotype-independent, light chain-associated cross-reactive Id expressed by the majority of individual mice immunized with (T,G)-A--L, (T,G)-A--L coupled to methylated bovine serum albumin (mBSA), or the linear terpolymer GAT. Primary and secondary monoclonal hybridoma protein (HP) antibodies from X/Xxid heterozygous (wild-type) mice immunized with (T,G)-A--L and/or (T,G)-A--L-mBSA were analyzed for isotypy and were grouped into eight antibody fine specificity sets defined by the patterns of direct binding to the antigens (T,G)-A--L, (Phe,G)-A--L, (T,G)-Pro--L, GT, and A--L. Analysis of these primary and secondary HP for TGB5 idiotypy showed a preferential expression of the TGB5 Id among GT+-binding HP (antibody fine specificity sets 1 through 3). All of the primary GT+-binding HP and the majority of secondary GT+-binding HP (sets 1 through 3) were TGB5 Id+. Most but not all of the TGB5 Id+ HP bound GAT. Of the side-chain-specific HP (sets 1 through 7), 78% of primary HP vs 49% of secondary HP bound GT. By these criteria, the primary HP response appears more restricted than the secondary HP response, consistent with the idea that Id diversification and antibody heterogeneity are regulated and selected events occurring during memory B cell generation. Although xid mice produce less antibody than wild-type mice to (T,G)-A--L, the TGB5 Id was produced early in the primary response by both xid and wild-type mice immunized with (T,G)-A--L or (T,G)-A--L-mBSA, and was maintained as a detectable Id in equivalent amounts in their secondary serum antibody responses. These results support the idea that distinct B cell subsets, including the xid B cell subset, share the same immunoglobulin gene repertoire.     

Production of antisera specific for idiotype(s) of murine anti-(T,G)-A--L antibodies.

Antibodies to the synthetic polypeptide (T,G)-A--L were raised in C57BL/10 and C3H.SW mice. For each strain, the anti-(T,G)-A--L antibodies from 10 animals were pooled, affinity purified on a (T,G)-A--L-Sepharose column, and used to immunize Lewis rats. The resulting rat antisera were adsorbed with insolubilized normal mouse globulin in order to remove anti-isotypic and anti-allotypic antibodies. The residual antibodies specifically inhibited the binding of (T,G)-A--L by anti-(T,G)-A--L as measured by a radioimmunoassay. The specificity of this inhibition was demonstrated as follows: 1) failure of the anti-(T,G)-A--L anti-idiotype to inhibit the binding of nuclease to anti-nuclease antibody of the same allotype; 2) failure of Lewis anti-[B10 anti-(T,G)-A--L] to inhibit C3H.SW anti-(T,G)-A--L and vice versa; 3) ability to absorb anti-C3H.SW anti-idiotypic activity on insolubilized C3H.SW anti-(T,G)-A--L but not on B10 anti-(T,G)-A--L. The same or cross-reactive idiotype(s) was present in the majority of individuals of each of these strains.     

Cross-reactive idiotypic determinants on antibodies and antigen-specific helper T cell continuous lines

Anti-idiotypic antibodies produced in C57BL/6 mice (H-2b, Igh-1b) against (T,G)-A--L-specific antibodies of C3H.SW mice (H-2b, Igh-1j) were used to probe (T,G)-A--L-specific helper T cell lines and clones for the expression of idiotypic determinants on the cell surface of the monoclonal functional T cells. By using the fluorescence-activated cell sorter (FACS II), anti-idiotypic sera of individual mice that specifically bind C3H.SW anti-(T,G)-A--L antibodies were shown to stain significantly cells of the E-9M(+) continuous T helper line originated from C3H.SW (T,G)-A--L "educated" T cells. The same antisera did not react with a helper T cell line of C3H.SW origin specific to human gamma-globulin. They also did not stain a (T,G)-A--L-specific helper T cell line derived from CWB (H-2b, Igh-1b) mice, which differ from C3H.SW mice only in their heavy chain allotypes. Thus, the expression of the idiotypic determinants on the T cell lines appears to be antigen-specific and linked to the heavy chain allotypic marker as shown for the specific antibodies. Different clones derived from the E-9 M(+) line were tested their reactivity with the individual anti-idiotypic sera. All clones but one (1.11) were stained significantly. The clones were tested for their biologic activity and all of them except clone 1.11 were found to exert helper activity specific to (T,G)-A--L. Thus, individual anti-idiotypic sera against C3H.SW anti-(T,G)-A--L antibodies recognize cross-reactive idiotypic structures on the surface of antigen-specific monoclonal helper T cells.     

Inheritance in the rat of the antibody response to two different determinants of (T,G)-A--L

Two (T,G)-A-L preparations differing in molecular weight were tested for immunogenecity in two inbred rat strains, AS and BN. Both strains produced antibody with specificity towards different antigenic determinants on the (T,G)-A--L molecule. The inheritance of antibody response toward one of the determinants was autosomal, dominant and linked to the major histocompatibility complex. Although the antibody response toward the other determinant was also inherited as an autosomal dominant trait, it was not linked to the major histocompatibility complex.     

Genetic control of immune responsiveness to poly (LPro)--poly (LLys)-derived polypeptides by histocompatibility-linked immune response genes in the rat.

In rats responsiveness to branched synthetic polypeptides carrying a Pro--L backbone, such as (T,G)-Pro--L or (Phe,G)-Pro-L and to Pro--L itself is controlled by Ir genes which are linked to the major histocompatibility genes. The level of antibody production to these polypeptides does not fall into strict high or low responder categories but covers the range in between. (T,G)-pro--L and Pro--L elicit a very similar response pattern which, however, differs from that obtained with (Phe,G)-Pro--L. Anti-(T,G),Pro--L antibodies do not cross-react with (T,G)-A--L, but do so extensively with Pro--L. Anti-(Phe,G)-Pro--L antibodies show cross-reactivity to (Phe,G)-A--L only when the antibody-producing strain is a high responder to (Phe,G)-A--L. These results when considered in view of data obtained in mice on genetic control of the immune response to (T,G)-Pro--L suggest that at least two unlinked Ir genes are involved in controlling anti-Pro--L responsiveness.     

Anti-idiotype stimulation of antigen-specific antigen-independent antibody responses in vitro. II. Triggering of B lymphocytes by idiotype plus anti-idiotype in the absence of T lymphocytes

Antibodies specific for the idiotypes of B10 anti-(T,G)-A--L antibodies (anti-id) induced B lymphocytes to secrete anti-(T,G)-A--L antibodies in vitro in the absence of both antigen and T lymphocytes, provided either that the B lymphocytes were previously primed in vivo with (T,G)-A--L or that id specific for (T,G)-A--L was added to the cultures. These antigen- and T lymphocyte-independent responses were antigen specific and appeared not to require accessory cells. The results suggested that B lymphocyte activation occurred via the formation of id-anti-id complexes, and evidence was obtained that this activation involved two separate interactions between the B lymphocytes and the id-anti-id complexes. These studies document a previously undescribed regulatory function of anti-idiotype antibodies.     

T lymphocyte-enriched murine peritoneal exudate cells. III. Inhibition of antigen-induced T lymphocyte Proliferation with anti-Ia antisera.

The recent development of a reliable murine T lymphocyte proliferation assay has facilitated the study of T lymphocyte function in vitro. In this paper, the effect of anti-histocompatibility antisera on the proliferative response was investigated. The continuous presence of anti-Ia antisera in the cultures was found to inhibit the responses to the antigens poly (Glu58 Lys38 Tyr4) [GLT], poly (Tyr, Glu) ploy D,L Ala-poly Lys [(T,G)-A--L], poly (Phe, Glu)-poly D,L Ala-poly Lys [(phi, G)-A--L], lactate dehydrogenase H4, staphylococcal nuclease, and the IgA myeloma protein, TEPC 15. The T lymphocyte proliferative responses to all of these antigens have previously been shown to be under the genetic control of major histocompatibility-linked immune response genes. The anti-Ia antisera were also capable of inhibiting proliferative responses to antigens such as PPD, to which all strains respond. In contrast, antisera directed solely against H-2K or H-2D antigens did not give significant inhibition. Anti-Ia antisera capable of reacting with antigens coded for by genetically defined subregions of the I locus were capable of completely inhibiting the proliferative response. In the two cases studied, GLT and (T,G)-A--L, an Ir gene controlling the T lymphocyte proliferative response to the antigen had been previously mapped to the same subregion as that which coded for the Ia antigens recognized by the blocking antisera. Finally, in F1 hybrids between responder and nonresponder strains, the anti-Ia antisera showed haplotype-specific inhibition. That is, anti-Ia antisera directed against the responder haplotype could completely block the antigen response controlled by Ir genes of that haplotype; anti-Ia antisera directed against Ia antigens of the nonresponder haplotype gave only partial or no inhibition. Since this selective inhibition was reciprocal depending on which antigen was used, it suggested that the mechanism of anti-Ia antisera inhibition was not cell killing or a nonspecific turning off of the cell but rather a blockade of antigen stimulation at the cell surface. Furthermore, the selective inhibition demonstrates a phenotypic linkage between Ir gene products and Ia antigens at the cell surface. These results, coupled with the known genetic linkage of Ir genes and the genes coding for Ia antigens, suggest that Ia antigens are determinants on Ir gene products.     

Antigen-specific T cell factors: a fine analysis of specificity.

The specificity of T cell factors produced in presence of synthetic polypeptide antigens was studied. Factors prepared with either one of the three antigens: poly(Tyr,Glu)-poly(DLALa)--poly(Lys), (T,G)-A--L, poly(Phe,Glu)-poly(DLALa)--poly(Lys), (Phe,G)-A--L, and poly(His,Glu)-poly(DLALa)--poly(Lys), (H,G)-A--L, successfully cooperated with B cells for antibody production to the homologous as well as to the other two immunogens. Furthermore, the activity of a (T,G)-A--L-specific factor was removed after passage through immunoadsorbents built of Sepharose coupled to: (T,G)A--L, (Phe-G)-A--L and poly(Glu)-poly(DLAa)--poly(Lys), (G)-A--L, but not to poly (DLALa)--poly(LLys),A--L. No cross-reactivity was observed between (T,G)-A--L and poly(Tyr,Glu)-poly(Pro)--poly(Lys), (T,G)-Pro--L, at the level of T cell factors, as shown using the above approaches. These results lead to the conclusion that specificity of T cell factors, although not identical, is similar to that of antibodies.     

Variable regions of antibodies to synthetic polypeptides. II. Analysis of variable region genes encoding antibodies specific for (T,G)-A--L

Monoclonal antibodies specific for the synthetic polypeptide antigen (T,G)-A--L have been produced in two strains of mice, C57BL/10 and C3H.SW. The genes encoding the variable (V) regions of these antibodies have been studied by using the DNA hybridization technique of Southern, as well as by gene cloning and sequencing. Hybridization of DNA from 14 different cell lines with a kappa-chain probe revealed that the different cell lines used one of two different gene rearrangements to encode the recombined V region gene. There was a perfect correlation between light chain rearrangement, idiotype expression, and fine specificity. Hybridization analyses of the heavy chain revealed a more complex pattern. Seven hybridomas had the rearranged heavy chain V region genes on a 4.4 kb EcoRI restriction fragment. Others were found on restriction fragments that differed in length by several hundred base pairs. The recombined heavy chain V region genes were cloned from three different hybridoma cell lines secreting anti-(T,G)-A--L antibodies, all of which express the same idiotype and fine specificity pattern. Restriction mapping and sequencing indicate that all three utilize the same V gene, identified as the 186-2 germline gene. However, different D and J genes are used to encode each of the antibodies. In contrast to the results seen in other antigen systems, heavy chain D and J genes do not have a major influence on idiotype expression and fine specificity of antibodies to the synthetic polypeptide (T,G)-A--L.     

Antigen-specific HLA-restricted human T-cell lines

Human T-cell lines responsive to the polypeptide antigens GAT and (T, G)-A--L were developed. They were specific for the priming antigens and required the participation of accessory cells matched for HLA-linked determinants, as shown in family studies. In panel studies, the ability of accessory cells to present antigen was shown to be associated with HLA-D-region antigens. However, the specificity of these determinants did not fully correspond to any HLA antigens as currently defined. One GAT-specific subline, derived by limiting dilution, utilized a restriction determinant associated with, but distinct from, the DQw3 (MB3) allospecificity. Blocking studies with mouse monoclonal antibodies indicated that this restriction determinant was carried by HLA-DQ molecules. The epitopes recognized in these molecules appear to be distinct from the alloantigenic determinants currently defined by serology.     

Genetic control of the immune response to (T,G)-A--L in C3H in equilibrium C57 tetraparental mice

When 15 C3H ? C57 tetraparental (allophenic) mice were analyzed for coat color, hemoglobin, and immunoglobulin allotype, all but two were shown to be chimeric. These 15 tetraparental mice were immunized with the synthetic polypeptide (T,G)-A--L, and the origin of the (T,G)-A--L-specific antibody produced was determined by using genetic markers (allotypes) on the immunoglobulin heavy chain constant region. Five tetraparental mice were high responders to (T,G)-A--L and had significant amounts of a (low responder) allotype antibody in their total serum. Three of these mice had significant amounts of anti-(T,G)-A--L antibody of the a (low responder) allotype. The antigen binding capacities of the a allotype fractions of these three were 4–5 times higher than the antigen binding capacities of immunized C3H (low responder) control mice. These results are compatible with the hypothesis that the inability of low-responder mice to produce significant amounts of anti-(T,G)-A--L antibody is a function of Ir-1A gene expression at the level of T cells.     

Specificity of H-2-linked Ir gene control in mice: demonstration of T helper cells recognizing branched synthetic polypeptides in low responder mice

This report provides evidence for the presence of T helper cells capable of recognizing the polypeptide antigens T6-A--L and (H,G)-A--L in low responder mice of H-2k and H-2b haplotypes, respectively. Mice were primed in vivo with the T6-A--L-avidin-(H,G)-A--L complex or, in the case of T6-A--L in H-2k mice, with the cross-reactive and permissive antigen T6-S--L. T helper cells cooperating with DNP-primed B cells could be rechallenged in vitro with the DNP-conjugates of T6--A--L or (H,G)-A--L, although the cells were of low responder type with respect to these antigens. This implies that T cell-macrophage interaction required for restimulation is apparently not defective in these low responders. The implications of these results for the concept of Ir gene control are discussed.     

Genetic and other effects on antibody and cell mediated immune response in swine leucocyte antigen (SLA)-defined miniature pigs

Miniature pigs of eight swine leucocyte antigens (SLA) haplotypes were immunized with sheep erythrocytes (SRBC), hen egg white lysozyme (HEWL) and the synthetic peptide (T, G)-A--L to induce antibody. Bacillus Calmette Geurin (BCG) and dinitrochlorobenzene (DNCB) were used to induce cell mediated immune response (CMI). Analysis of variance by least squares was used to assess the effects of SLA haplotype, sire, dam, litter and sex of pig on the magnitude of the primary and secondary antibody response and on dermal delayed type hypersensitivity induced by purified protein derivative of tuberculin (PPD) and DNCB-induced contact hypersensitivity. The statistical model accounted for 43.50-77.30% of the observed variability in antibody and CMI at various times after immunization or challenge. While SLA had a significant effect on both antibody and CMI to some antigens at some, but not all times, sire, dam and litter were more frequently significant and to a greater degree. Haplotypes dd, dg and gg produced more antibody to SRBC and (T, G)-A--L while dg and gg had higher primary, but not secondary antibody response to HEWL. Delayed hypersensitivity to PPD was most marked in pigs of dd, dg and gg haplotypes while contact hypersensitivity to DNCB was expressed least in the dg and gg haplotype pigs. Heritability estimates were high for response to (T, G)-A--L and HEWL indicating feasibility of selective breeding for these traits.     

Immunoregulation of the anti-bovine serum albumin response by polyclonal and monoclonal antibodies

Monoclonal or polyclonal antibodies directed toward determinants on limited structures of bovine serum albumin (BSA) (P505-582) were shown to regulate the entire anti-bovine serum albumin (BSA) immune response when passively administered to mice 24 hr prior to immunization. Regulation was observed as suppression of the humoral IgG immune response toward all BSA determinants except those on fragment P505-582. By Day 21 suppression of humoral response was most pronounced toward determinants present on the carboxy terminal end of the molecule (N 307-582). These observations demonstrate that monoclonal antibodies directed against a single determinant on a protein molecule have the capacity to regulate the immune response to a multiplicity of determinants present on the same protein. The data lend support to concepts of antibody-induced regulation by induction of suppressor cells or idiotype recognition.     

Conformational integrity of myoglobin after immunization with Freund's adjuvant.

The use of Freund's complete adjuvant for immunizing goats with myoglobin produces mainly antibodies directed against antigenic determinants present in the native protein. Only about 9% of the total antibodies produced are directed toward determinants not expressed in tha native molecule. This shows that neither emulsification nor the subsequent in vivo events leading up to the immune response appreciably perturb the conformation of the protein surface.     

Specific antigenic relationships between the RNA-dependent DNA polymerases of avian reticuloendotheliosis viruses and mammalian type C retroviruses

Immunoglobulin G directed against the DNA polymerase of Rauscher murine leukemia virus (R-MuLV) could bind to 125I-labeled DNA polymerase of spleen necrosis virus (SNV), a member of the reticuloendotheliosis virus (REV) species. Competition radioimmunoassays showed the specificity of this cross-reaction. The antigenic determinants common to SNV and R-MuLV DNA polymerases were shared completely by the DNA polymerases of Gross MuLV, Moloney MuLV, RD 114 virus, REV-T, and duck infectious anemia virus. Baboon endogenous virus and chicken syncytial virus competed partially for antibodies directed against the common antigenic determinants of SNV and R-MuLV DNA polymerases. DNA polymerases of avian leukosis viruses, pheasant viruses, and mammalian type B and D retroviruses and particles with RNA-dependent DNA polymerase activity from the allantoic fluid of normal chicken eggs and from the medium of a goose cell culture did not compete for the antibodies directed against the common antigenic determinants of SNV and R-MuLV DNA polymerases. We also present data about a factor in normal mammalian immunoglobulin G that specifically inhibits the DNA polymerases of REV and mammalian type C retrovirus DNA polymerases.     

Antigen-specific T-cell factor in cell cooperation and genetic control of the immune response.

Cell interactions between thymus-derived (T) and bone marrow-derived (B) lymphocytes in the antibody response appear to involve soluble T-cell mediators known as 'factors.' This paper describes the properties of a T-cell factor that has specificity for the inducing antigen, a synthetic polypeptide (T, G)-A--L, and is able to replace T cells in the thymus-dependent antibody response to (T, G)-A--L. Besides antigen specificity, the main features of the molecule are that it is nonimmunoglobulin; it has a molecular weight of about 50,000; and it is a product of the I-A subregion of the H-2 complex (the mouse major histocompatibility complex). These properties suggest that the factor is closely related to the T-cell receptor, which may, by inference, also be a product of the H-2 complex. The factor cooperates well with allogeneic B cells. It can also be absorbed by bone marrow cells and B cells. Studies on the genetic control of the immune response to (T, G)-A--L using the T-cell factor indicate that two immune response genes in the H-2 complex are involved in genetic control, one expressed in T cells and the other in B cells. This two gene hypothesis has been confirmed by showing that an F1 between two low responders to (T, G)-A--L can be a high responder.     

Genetic control of immune response to the L-Glu, L-Lys, L-Phe terpolymer in man.

We have demonstrated that human lymphocytes can respond to the synthetic polypeptide GLPhe upon in vitro challenge by the antigen similar to that of (H,G)-A--L, (T,G)-A--L, (Phe,G)-A--L, and GAT. Family studies further support our postulation that responses to these synthetic polymers are under dual gene control. Three families with intra-HLA-A/B recombinants provided mapping information for Ir-GLPhe genes. The response phenotype of the recombinant of family 21 localized the Ir-GLPhe genes toward the HLA-B of D regions, whereas recombinants of family 24 and 27 placed the Ir-GLPhe genes distal to HLA-B, toward the A region. This discrepant gene assignment can be explained by assuming that recombination occurred at different positions between HLA-A and HLA-B. In family 21, crossover occurred distal to the Ir genes, while for the other two, proximal to them. A second possibility is that as in the mouse the two complementing genes are situated in different regions of the human major histocompatibility complex (MHC) and all three of the crossovers occurred between them with the putative Ir-GLPhe-1 located near the HLA-A region and Ir-GLPhe-2 on the HLA-D region or vice versa. A third possibility is that immune response required interaction between a complete HLA-D-like molecule encoded in the A region and another encoded elsewhere, perhaps in HLA-D.