CD3 antibodies are proven immunosuppressants capable of reversing transplant rejection episodes. Their general application has been limited both by their immunogenicity and, in particular, by the "first-dose" cytokine-release syndrome experienced by patients after the initial administration of antibody. We have produced a set of variants of the humanized YTH 12.5 CD3 monoclonal antibody (mAb) (Routledge et al., Eur. J. Immunol. 1991. 21: 2717) bearing different human heavy (H) chain constant regions, with the intention of finding a form of the antibody that is not able to activate T cells. Comparison of the variants having gamma 1, gamma 2, gamma 3 and gamma 4 H chains in a competitive binding assay showed that antibody avidity was not affected by IgG subclass. Using a sensitive indicator of FcR binding activity (the capacity of the CD3 mAb to redirect cytotoxic T cells to kill the monocytic cell line U-937) we demonstrated a functional hierarchy of gamma 1 = gamma 4 > alpha 2 =/> gamma 3 mb >> gamma 2. An aglycosyl version of the gamma 1 CD3 mAb, produced by site-directed mutagenesis (Asn297 to Ala), still had considerable activity in this assay (intermediate to the gamma 1 and alpha 2 CD3 mAb), albeit at a level approximately 10-fold lower than that of the parental gamma 1 form. When we tested their ability to stimulate T cell proliferation in vitro in the presence of 5% human serum, all of the wild-type immunoglobulin isotypes were found to be active, although there were T cell donor-dependent variations in the extent of the responses. The aglycosyl gamma 1 mAb was, however, completely non-mitogenic in all of ten donors tested, unless the assay was performed in IgG-free medium. Despite being non-stimulatory, this mAb was also able to inhibit the mixed lymphocyte reaction responses of both naive and primed T cells. Comparison of the gamma 1 and aglycosyl gamma 1 mAb in an experimental mouse model for CD3 mAb-induced cytokine release indicated that removal of the carbohydrate moiety from the gamma 1 constant region reduced the in vivo tumor necrosis factor-alpha response by a factor of at least 16-fold. These data suggest that the aglycosyl gamma 1 CD3 mAb is a promising candidate for immunosuppressive therapy without "first dose" side effects.

Two isolates of the murine hepatitis virus (MHV) strain JHM, which differed in their ability to express the nonstructural gene product ns2, were characterized. The MHV Wb3 isolate encodes a 30,000-molecular-weight ns2 protein that can be readily detected in infected cells by using a specific monoclonal antibody, MAb 2A. The MHV Wbl isolate is a deletion mutant that lacks a functional ns2 gene and the transcriptional signals required for the synthesis of an ns2 mRNA. However, there are no obviously significant differences in the growth of the MHV Wbl and MHV Wb3 isolates in continuous cell lines or in the synthesis of viral mRNAs or proteins in infected cells. These results demonstrate that the ns2 gene product is not essential for MHV replication in transformed murine cells and suggests that the function of the ns2 gene may only be manifest in vivo.

The murine coronavirus surface glycoprotein gene was expressed as a fusion protein in bacteria, and the expressed protein was used to generate S protein-specific monoclonal antibodies (MAbs). Three of the MAbs, 11F, 30B, and 10G, were able to neutralize virus infectivity, and two of them, 11F and 10G, were able to block virus-induced, cell-to-cell fusion. The binding sites of the 11F, 30B, and lOG MAbs were determined by Western immunoblotting and epitope mapping. The 11F and 30B MAbs bound to sites located, respectively, between amino acids 33 to 40 and 395 to 406 in the amino-terminal (Sj) subunit of the S protein, and the lOG MAb bound to a site located between amino acids 1123 and 1137 in the carboxy-terminal (S2) subunit. These data define more precisely the interactions between the SI and S2 subunits of the murine coronavirus S protein and provide further insights into its structure and function.

Twelve clones of murine hybridoma cells secreting antibody specific for respiratory syncytial (RS) virus were classified into four groups on the basis of their pattern of staining of unfixed RS virus-infected HEp-2 cells in an indirect immunofluorescence test. Three of the groups reacted with virus antigens present on the membrane of the cells, whilst the fourth group failed to stain most live cells, suggesting specificity for an antigen expressed internally. Representative monoclonals from the membrane antigen staining groups immunoprecipitated the 86K glycoprotein (G), 50K plus 19K glycoprotein (F1,2) and a 23K non-glycosylated protein (VP23). A representative monoclonal from the fourth group that appeared to stain an internally expressed protein immunoprecipitated the virion 34K phospho-protein (P). All four monoclonals stained acetone-fixed tissue culture cells infected with either the Long strain of RS virus or with strains isolated in Newcastle during the 1965, 1972, and 1983 winter epidemics. The anti-fusion protein antibody stained acetone-fixed cells from all of 26 nasopharyngeal secretions from infants with RS virus infection. The anti-G glycoprotein antibody and the anti-VP23 antibody stained cells from secretions poorly or not at all, whilst the anti-P protein antibody stained cells in half the secretions tested but reacted with only a small proportion of cells in comparison with the anti-F or polyclonal antibodies. A pool of all four monoclonals produced more intense staining than the anti-F monoclonal alone and gave a more clearly defined staining reaction than the polyclonal antiserum used for routine diagnosis in over half the secretions. These results indicate that monoclonal antibodies will be of value in the diagnosis of RS virus by indirect immunofluorescence if care is taken in the selection of a suitable pool.

No abstract