No abstract

In human immunodeficiency virus (HIV)-infected individuals, the proportion of circulating mononuclear cells (PBMCs) which carry HIV provirus and the number of HIV proviral sequences per infected PBMC have been matters for conjecture. Using a double polymerase chain reaction which allows the detection of single molecules of provirus and a method of quantifying the provirus molecules, we have measured provirus frequencies in infected individuals down to a level of one molecule per 10' PBMCs. As a general rule, only a small proportion of PBMCs contain provirus (median value of samples from 12 patients, one per 8,000 cells), and most if not all of the infected cells carry a single provirus molecule. The frequency of provirus-carrying cells correlated positively both with the progression of the disease and with the success with which virus could be isolated from the same patients by cocultivation methods. Of seven asymptomatic (Centers for Disease Control stage H) patients, all but one contained one provirus molecule per 6,000 to 80,000 cells; of five Centers for Disease Control stage IV patients, all but one contained one provirus molecule per 700 to 3,300 cells. When considered in conjunction with estimates of the frequency of PBMCs that express viral RNA, our results suggest that either (i) the majority of provirus-containing cells are monocytes or (ii) most provirus-containing lymphocytes are transcriptionally inactive. We also present nucleotide sequence data derived directly from provirus present in vivo which we show is not marred by the in vitro selection of potential virus variants or by errors introduced by Taq polymerase. We argue from these data that, of the provirus present in infected individuals, the proportion which is defective is not high in the regions sequenced.

Brain tissue was examined for evidence of human immunodeficiency virus (HIV) infection in 23 intravenous drug users who died suddenly some years after seroconversion but while still in presymptomatic stages of infection. None showed giant cell encephalitis, but 14 showed T cell lymphocytic leptomeningitis and 3 showed other significant neuropathologic features. Quantitative polymerase chain reaction for HIV was applied to 13 of the 23 with negative results in 6 and very low positive results in the other 7, a finding consistent with contamination by residual infected blood in the brain tissue. This contrasted with findings in AIDS-infected tissue, in which substantial amounts of provirus were found. It is concluded that significant infection in brain tissue does not occur in presymptomatic stages of HIV infection and that invasion of the central nervous system may be delayed until the transition to symptomatic AIDS.

No abstract

The frequency and dynamics of infection with different genotypes of hepatitis C virus were investigated in a cohort of hemophiliacs repeatedly exposed to non-virus-inactivated clotting factor. Among 63 infected hemophiliacs, genotype 1 (n = 38, subtypes 1a [27] and 1b [11]) was predominant; genotypes 2a (n = 1), 2b (n = 3), 3a (n = 20), and 5a (n = 1) accounted for the remainder. This distribution was similar to that found in Scottish blood donors from whom the infected blood products were manufactured. Hemophiliacs with severe disease were more likely to be polymerase chain reaction-positive than those with moderate or mild disease. Over 10 years, changes in the circulating major genotype and serotype were observed in 9 of 29 hemophiliacs and from one subtype to another in 3, although there was no clear trend toward replacement with any particular variant. Replacement occurred after the introduction of inactivated clotting factor in 4 subjects, implicating reactivation rather than reinfection. Those coinfected with human immunodeficiency virus were more likely to show a change in genotype.