Bone marrow transplant (BMT) recipients were evaluated for subclinical varicella-zoster virus (VZV) viremia and symptoms of herpes zoster after transplantation. Viremia was demonstrated by testing peripheral blood mononuclear cells using polymerase chain reaction and was documented in 19% of 37 patients. When reactivation was defined as herpes zoster and/or subclinical VZV viremia, 41% of VZV-seropositive BMT recipients experienced VZV reactivation. None of 12 patients tested before VZV reactivation had T lymphocyte proliferation to VZV antigen (mean stimulation index, 1.0 +/- 0.42 [SD] at less than 100 days; 12.0 +/- 6.03 at greater than 100 days [P = .003]). Among patients tested at greater than 100 days, 5 (63%) of 8 with detectable T cell proliferation had subclinical or clinical VZV reactivation compared with none of 6 who lacked VZV T cell responses. Recovery of VZV-specific cytotoxic T lymphocyte function was observed in 50% of BMT patients, but BMT recipients had significantly fewer circulating cytotoxic T lymphocytes that recognized VZV immediate early protein (P = .03) or glycoprotein I (P = .004) than did healthy VZV immune subjects. In vivo reexposure to VZV antigens due to subclinical VZV viremia or symptomatic VZV reactivation may explain the recovery of virus-specific T cell immunity after BMT.

The cytotoxic T lymphocyte (CTL) response was evaluated in adults given live attenuated varicella vaccine, using target cells expressing varicella-zoster virus (VZV) immediate-early protein (IE62) or VZV glycoproteins gpI, gpIV, or gpV to determine viral protein specificity. The frequency of CTL that recognized IE62 was 1:171,000 +/- 46,000 SE in subjects tested 10 days to 8 weeks after the initial vaccine dose; the induction of CTL specific for gpI was equivalent. CTL recognition of VZV proteins was mediated by CD4+ or CD8+ cells. CTL recognition of IE62 and gpIV persisted in vaccinees (tested approximately 4 years later) and was comparable to that in the naturally immune. The mean frequency of CTL specific for gpV was lower (but not significantly) in vaccinees than in naturally immune subjects. Assay of responder cell frequencies showed persistence of equivalent numbers of T lymphocytes that recognized IE62 and gpI in vaccinees and naturally immune subjects. Immunization with this vaccine elicited memory T lymphocyte responses to VZV comparable to those induced by natural infection.

Each of the two mutants isolated from a fliC (= hag, flagellin-deficient) Escherichia coli strain made motile by a plasmid carrying the fliC gene of Salmonella muenchen by selection for motility in the presence of anti-d (Salmonella flagellar antigen) serum had both lost and gained one or more subfactors of the wild-type antigen. In one mutant codon 246 was GAC (alanine) instead of GCC (asparagine); the other had a deletion of 105 base pairs, explicable by a 10bp direct repeat, starting at bases 782 and 887. The in vitro removal of a 48bp EcoRV(631)/EcoRV(679) fragment produced plasmid pLS408, which was found to lack a subfactor of wild-type antigen d but able to confer motility on flagellin-negative Salmonella sp. (and used for insertion of epitope-specifying oligonucleotides at its EcoRV site). Immunoblotting with absorbed and unabsorbed sera from rabbits immunized with E. coli with wild-type or mutated antigen d showed that the fusion proteins specified by lambda gt11 with the N-terminal part of gene lacZ joined to a restriction fragment coding for residues 145-391 of flagellin gave the same pattern of parent-specific and mutant-specific reactions as the flagellate bacteria. Four out of five similarly selected mutants had the same 105 bp deletion as the first-isolated mutant; the fifth had a 72 bp deletion made possible by a 7-base pair direct repeat, starting at positions 649 and 721. All these changes in serological character without loss of function affected segment IV, specifying residues 182 to 308 of the total of 505, where there is little homology between different flagellar-antigen alleles.