David, Chella S.

Abstract Optimum function of HLA-DR molecules in transgenic mice requires efficient interaction between the class II molecules on APCs and CD4 on T cells. Residues 110 and 139 of the second domain of class II molecules are considered to be critical for recognition of CD4. We generated an HLA-DR4β(NT) transgene construct in which positions 110 and 139 were altered to resemble endogenous mouse H2 Aβ molecules. This construct was introduced into (B10 × SWR) embryos, and DR4β(NT) transgenic mice were produced. The transgene was transferred into B10.RFB3 (Eβ0 Eαp) mice. The transgene-encoded DR4β molecules paired with endogenous Eα chains to form stable DR4β/Eα dimers expressed on the cell surface. The hybrid dimers showed similar Ag-binding specificity to HLA-DR4 molecules and positively selected CD4+ T cells in vivo. Immunization of HLA-DR4β(NT) transgenic mice with DR4-restricted peptides induced T cell proliferation in vitro. While the purified T cells from DR4β(NT) transgenic mice responded strongly to the HA(307–319) presented by M12C3 transfectants expressing altered DR4β/Eα heterodimers, the response to the same peptides presented by transfectants expressing wild-type DR4β/Eα molecules was substantially reduced. Taken together, these data confirmed in vitro studies on the importance of these residues in CD4-MHC class II interaction. The altered HLA-DR4β transgenic mice were able to overcome the species barrier and generate efficient HLA-DR4-restricted CD4-specific immune responses. Thus, residues 110 and 139 were critical for the interaction of class II with CD4 T cells during thymic selection as well as peripheral immune responses.

Abstract DQw6b transgenic mice have been generated by microinjecting a linearized cosmid clone containing 34-kb DQb genomic DNA, isolated from HLA-homozygous B cell line AKIBA (DR2, Dw12, DQw6), into embryos of (CBA x B10.M)F2 or (SWR x SJL)F2. Among 85 mice screened, eight mice were transgene-positive. The transgene in seven of eight founders was germline-transmitted. FACS analysis and immunohistochemical studies with DQ beta-specific mAb demonstrated that DQ beta molecules in association with mouse A alpha f molecules are expressed on peripheral mononuclear cells, spleen cells, and in thymic medulla. More interestingly, V beta 11-, V beta 5.1-, and V beta 5.2-bearing T cells, but not V beta 8.2-bearing T cells, were clonally deleted in the H-2E-negative but DQ beta+ progeny of two selected founders (260-23 and 258-10). The deletion was found to take place intrathymically during the transition stage from immature to mature thymocyte development. We postulate that although human DQ genes are more homologous to mouse H-2A genes, A alpha f/DQ beta hybrid molecules may possess the same self-peptide- (or superantigen)-presenting epitope as E alpha/E beta molecules critical for deletion of V beta 11-, V beta 5.1-, and V beta 5.2-bearing T cells in thymus. Our results also confirm the previous findings that accessory molecules on thymocytes such as CD4 may be involved in thymic selection, and further suggest that an interaction of mousE CD4 and mouse A alpha chain is required for the clonal deletion.

Abstract Mice of strain B6, but not AKR/J, respond to immunization with Torpedo acetylcholine receptor (AChR) by manifesting in vitro an Ag-specific T lymphocyte proliferative response. Our analysis of (AKR x B6)F1 mice reveals that the T cell unresponsiveness of AKR/J is inherited as a dominant trait, possibly associated with expression of the Mls-1a allele. Mice derived from backcrossing (AKR x B6)F1 x B6 were selected for H-2b homozygosity and were classified as Mls-1a or Mls-1b according to the relative numbers of peripheral blood T cells that expressed the TCR Vβ6 gene product. After challenge by injection with AChR in CFA, lymph node cells from mice classified as having less than 2% of Vβ6+ peripheral T cells had low responsiveness to AChR, whereas mice with greater than 7% Vβ6+ peripheral T cells had high T cell responsiveness to AChR. These results are consistent with the notion that regulation of the T cell repertoire by Mls loci may be a determinant of susceptibility to autoimmunity.

Abstract Interactions between MHC class II genes and minor lymphocyte stimulating (Mls) associated products are responsible for clonally deleting self-reactive T cells in mice. Here we demonstrate the role of the intact I-A and I-E molecules as well as the individual A alpha and E alpha chains in the deletion of cells bearing the V beta 6 TCR. DBA/1 (H-2q, Mls-1a) mice were crossed with various inbred congenic, recombinant, and transgenic strains and the F1's were screened for V beta 6 expression. All I-E+ strains were fully permissive in deleting V beta 6+ T cells. I-E- strains expressing I-A b,f,s,k,p permitted only partial deletion, while I-Aq strains showed no deletion. Recombinant I-Aq and I-Af strains which expressed E kappa alpha chain in the absence of E beta chain showed a decrease in V beta 6+ T cells as compared to their H-2q and H-2f counterparts. Furthermore, transgenic mice expressing E kappa alpha Aq beta gene in an H-2q haplotype (E kappa alpha Aq beta?) gave similar results to that of the recombinants in deleting V beta 6 T-cells. The role of the 1-A molecule was also shown by the partial deletion of V beta 6+ T cells in H-2q mice expressing transgenic I-Ak molecules. These results demonstrate that the E alpha chain is important in the deletion of V beta 6 T-cells in Mls-1a mice. The role of A alpha chain is also implied by the permissiveness of E kappa alpha Aq beta but not Aq alpha Aq beta molecules in the deletion of V beta 6+ T cells.

Abstract Hemopoietic histocompatibility (Hh) Ag are noncodominantly expressed on bone marrow stem cells and other normal and neoplastic cells of hemopoietic origin. H-2/Hh-1 allogeneic or parental-strain bone marrow grafts are eliminated in a determinant specific manner by NK cells. In inbred mouse strains, seven Hh-1 alleles representing combinations of five different Hh-1 antigenic determinants are described. Each Hh-1 allele maps in the vicinity of H-2D, and the genes that map to Hh-1 are transacting regulatory genes. The expression of a particular determinant depends on the absence of the regulatory gene and the presence of the appropriate structural gene. The primary focus of this study is to ascertain whether the Hh-1 phenotype and the serologic H-2DL typing are always correlated or whether recombinant can separate the two. To achieve this, we used a panel of irradiated hosts that are able to recognize the different Hh-1 determinants on the bone marrow cells of congenic intra-H-2 recombinant donors. We report: 1) the majority of strains show a correlation between Hh-1 and H-2DL: 2) B10.RQDB and B10.WB strains dissociate Hh-1 from Lb: 3) nine H-2S/D interval recombinant strains exhibit no correlation between the H-2DL type and Hh-1 phenotype; and 4) in two strains from this group, B10.D2 (R106) and B10.RSF5, H-2S/D crossovers occurred within Hh-1r, (Hh-1 regulatory). We conclude that Hh-1r is a distinct regulatory locus mapping telomeric of H-2S and centromeric of, although probably closer to, H-2D.

Abstract B10.Ek alpha transgenic mice were mated with H2-E B10.Q and SWR mice. F1 and F1 x parental strain backcross progeny were tested for arthritis and autoimmune reactivity to mouse type II collagen (MII) after immunization with bovine, chick, deer, or human type II collagen. The results were correlated with the H-2 haplotype (b/q vs q/q) and the TCR V beta profile of peripheral blood T cells in each mouse. Hybrid progeny expressed TCR profiles different from either parent because of the TCR V beta genomic deletions of SWR mice (V beta a), the wild-type TCR allele of C57Bl/10 (B10) mice (V beta b), and the intrathymic negative selection processes resulting from cell surface expression of Ek alpha-A q beta or Eb beta-Ek alpha, together with the integrated retroviral genes Mtv-9 originating in B10 mice and Mtv-7 (Mls-1a) from SWR mice. (B10.Ek alpha x SWR)F1 mice developed higher IgG anti-MII Ab titers, but much milder arthritis than (B10.E x B10.Q)F1 mice. Expression of Ek alpha did not change the level of IgG anti-MII Ab nor the degree of susceptibility to collagen-induced arthritis (CIA) in the H-2q/q and H-2b/q progeny of (B10.Ek alpha x B10.Q)F1 x B10.Q matings, indicating that the Mtv-9-reactive, TCR V beta 5+, and V beta 11+ T cells are not critical to CIA. Among bovine type II collagen-immunized (B10.Ek alpha x SWR)F1 x SWR backcross mice: 1) arthritis severity is associated with the presence of V beta b (p < or = 0.01) and expression of Ek alpha (p < or = 0.05), but not with the MHC haplotype (b/q vs q/q); 2) regression analysis showed a significant association (R = 0.99) between IgG anti-MII Ab titers and the level of Mtv-7-reactive V beta 6+ T cells that was detectable in the IgG1, but not the IgG2a subclass. The data prompt the speculation that Mtv-7-reactive V beta 6+ (or V beta 7+) T cells in (B10.EK alpha x SWR)F1 x SWR mice express Th2-type properties, and thus contribute to the combination of mild arthritis but high anti-MII Ab titers that characterize mice of SWR heritage.

Abstract Trophoblast, the only fetal tissue in direct contact with maternal cells, fails to express the polymorphic HLA class I molecules HLA-A and -B, but does express the nonpolymorphic class I molecule HLA-G. It is thought that HLA-G may provide some of the functions of a class I molecule without stimulating maternal immune rejection of the fetal semiallograft. As a first step in identifying the cis-acting DNA regulatory elements involved in the control of class I expression by extraembryonic tissue, several types of transgenic mice were produced. Two HLA-G genomic fragments were used, 5.7 and 6.0 kb in length. These included the entire HLA-G coding region, 1 kb of 3' flanking sequence, and 1.2 or 1.4 kb of 5' flanking sequence, respectively. A hybrid transgene, HLA-A2/G, was produced by replacing the 5' flanking sequence, first exon, and early first intron of HLA-G with the corresponding elements of HLA-A. Comparison of transgene mRNA expression patterns seen in HLA-A2/G and HLA-G transgenic mice suggests that 5' flanking sequences are largely responsible for the differing patterns of expression typical of the classical class I and HLA-G genes. Studies comparing the extraembryonic HLA-G expression levels of founder embryos transgenic for either the 5.7- or 6.0-kb HLA-G transgene showed that the 6.0-kb transgene directed HLA-G expression far more efficiently than did the 5.7-kb HLA-G transgene, producing extraembryonic HLA-G mRNA levels similar to those seen in human extraembryonic tissues. The results of these studies suggest that the 250-bp fragment present at the extreme 5' end of the 6.0-kb HLA-G transgene and absent from the 5.7-kb HLA-G transgene contains an important positive regulatory element. This 250-bp fragment lies further upstream than any of the previously documented class I regulatory regions and may function as a locus control region.

Abstract Demyelination induced by Theiler's virus was examined in nonrecombinant H-2 congeneic strains of mice on common backgrounds expressing independent haplotypes. Light and electron microscopy of spinal cord sections from mice with s, f, p, r, v, or q haplotypes on a C57BL/10 background showed perivascular inflammation and multifocal areas of demyelination in the white matter. The demyelination in these mice was usually associated with clinical neurologic deficits. In contrast, mice with identical genetic backgrounds but b, k, or d haplotypes showed no pathologic or clinical abnormalities. Qa or Tla haplotypes did not appear to influence demyelination. The observations support the hypothesis that susceptibility to Theiler's murine encephalomyelitis virus-induced demyelination is influenced by genes that are linked to the H-2 complex. The findings suggest that the demyelination may be mediated by immune cells rather than being a direct cytolytic effect of virus on oligodendrocytes.

Abstract The I-E and I-C subregion products have been identified specifically in haplotypes H-2k and H-2d. Haplotypes H-2p and H-2r are assumed to express products of I-E/I-C subregion on the basis of cross-reactions with anti-I-Ek sera. Anti-I-Ek or I-Ed sera do not cross-react with haplotypes H-2s, H-2b, H-2q, and H-2f. Specific immunizations to produce anti-I-Eb and anti-I-Es alloantisera have been unsuccessful. We turned to xenoimmunization to serologically identify the gene products of this chromosomal segment in the above haplotypes. NP-40 detergent solubilized antigenic material mixed in Freund's complete adjuvant were immunized into rabbits. Two weeks later a booster was given and the rabbits bled one week later for antisera analyses. The rabbits were then boosted once a month and bled periodically. The rabbit antiserum reacted with all mouse strains. Absorptions with appropriate recombinants showed the presence of specific antibodies against the I region products. For example, rabbit anti-B10 serum after absorption with B10.A (3R) left residual activity against B10.A (18R), suggesting antigen(s) mapping between I-J and H-2D. The antiserum was further absorbed with B10.F (13R) to make it specific for I-E/IC subregions. Rabbit anti-B10.M sera, after absorption in ATFR.1 reacts with ATFR.5, suggesting activity in the K and I regions. Since absorption with platelets did not remove activity for spleen cells, presence of anti-If antibodies is suggested. Sequential immunoprecipitation analysis will be performed to screen for anti-I-E/I-C antibodies. Rabbit anti-B10.G and Rabbit anti-B10.S sera are currently being analyzed for I region activity.

Abstract Previously we have reported the precise determination of the entire antigenic structure of sperm-whale myoglobin. This protein which is made up on a single polypeptide chain (153 residues, m.w. 17,816) has five antigenic sites located in exposed conformationally sensitive continuous portions of its polypeptide chain. The five antigenic sites occupy the following regions: Site 1: sequence 15–22; Site 2: sequence 56–62; Site 3: sequence 94–99; Site 4: sequence 113–119; Site 5: sequence 145–151. We investigated the genetic control of T lymphocyte proliferative response to myoglobin and its synthetic antigenic sites in mice (PETLES Assay). Two parts of the myoglobin molecule (sequences 1–6 and 121–127) that were not within antigenic sites were also synthesized and employed as controls. Also used as control was a synthetic ‘nonsense’ heptapeptide that did not resemble any part of myoglobin. Haplotypes H-2d, H-2f and H-2s gave substantial T lymphocyte proliferative response to whole myoglobin whereas haplotypes H-2d, H-2k, H-2p, H-2q and H-2r were low responders. Studies with the recombinant strains suggested that expression of a high responder allele in the I-A subregion was enough to give a good proliferative response. On the other hand expression of I-Cd gave a good response while I-Cs gave a poor response. In preliminary studies, mice were primed with whole myoglobin and then challenged in vitro with the synthetic peptides containing the antigenic sites. Among the high responder strains, H-2d haplotype responded to Sites 1, 2, 3 and 5, whereas H-2s and H-2f haplotypes responded only to Sites 1, 2 and 3. We were surprised to find that the low responder strains, haplotypes H-2b, H-2k and H-2r responded to Site 4 but failed to respond to Sites 1, 2, 3 and 5. Mice did not respond to the control peptides. Further studies are currently underway in recombinant strains and in F1 crosses and will be reported at the Ir Gene Workshop. We are also investigating the ability of synthetic antigenic sites in the priming process.