Frequent hereditable shutdown of murine retrovirus gene expression in murine cell lines

Genetic determinants of feline leukemia virus-induced lymphoid tumors: patterns of proviral insertion and gene rearrangement

The genetic basis of feline leukemia virus (FeLV)-induced lymphoma was investigated in a series of 63 lymphoid tumors and tumor cell lines of presumptive T-cell origin. These were examined for virus-induced rearrangements of the c-myc, flvi-2 (bmi-1), fit-1, and pim-1 loci, for T-cell receptor (TCR) gene rearrangements, and for the presence of env recombinant FeLV (FeLV-B). The myc locus was most frequently affected in naturally occurring lymphomas (32%; n = 38) either by transduction (21%) or by proviral insertion (11%). Proviral insertions were also common at flvi-2 (24%). The two other loci were occupied in a smaller number of the naturally occurring tumors (fit-1, 8%; pim-1, 5%). Examination of the entire set of tumors showed that significant numbers were affected at two (19%) or three (5%) of the loci. Occupation of the fit-1 locus was observed most frequently in tumors induced by FeLV-myc strains, while flvi-2 insertions occurred with similar frequency in the presence or absence of obvious c-myc activation. These results suggest a hierarchy of mutational events in the genesis of feline T-cell lymphomas by FeLV and implicate insertion at fit-1 as a late progression step. The strongest links observed were with T-cell development, as monitored by rearrangement status of the TCR beta-chain gene, which was positively associated with activation of myc (P < 0.001), and with proviral insertion at flvi-2 (P = 0.02). This analysis also revealed a genetically distinct subset of thymic lymphomas with unrearranged TCR beta-chain genes in which the known target loci were involved very infrequently. The presence of env recombinant FeLV (FeLV-B) showed a negative correlation with proviral insertion at fit-1, possibly due to the rapid onset of these tumors. These results shed further light on the multistep process of FeLV leukemogenesis and the relationships between lymphoid cell maturation and susceptibility to FeLV transformation.

Ping-pong amplification of a retroviral vector achieves high-level gene expression: human growth hormone production

Retroviral vectors offer major advantages for gene transfer studies but have not been useful for producing proteins in large quantities. This deficiency has resulted in part from interference to superinfection, which limits the numbers of active proviruses in cells. Recently, we found that these vectors amplify when they are added as calcium phosphate precipitates to cocultures of cells that package retroviruses into ecotropic and amphotropic host range envelopes. Helper-free virions from either cell type can infect the other without interference, resulting in theoretically limitless back-and-forth (ping-pong) vector replication. In initial studies, however, amplifications of a vector that contained the human growth hormone gene ceased when the hormone produced was 0.3% or less of cellular protein synthesis. This limit was caused by two factors. First, recombinant shutoff viruses that are replication defective and encode envelope glycoproteins form at a low probability during any round of the vector replication cycle and these spread in cocultures, thereby establishing interference. Single cells in shutoff cocultures therefore synthesize both ecotropic and amphotropic envelope glycoproteins, and they release promiscuous (presumably hybrid) virions. The probability of forming shutoff viruses before the vector had amplified to a high multiplicity was reduced by using small cocultures. Second, cells with large numbers of proviruses are unhealthy and their proviral expression can be unstable. Stable expresser cell clones were obtained by selection. Thereby, cell lines were readily obtained that stably produce human growth hormone as 4 to 6% of the total protein synthesis. A ping-pong retroviral vector can be used for high-level protein production in vertebrate cells.

Macrophage-resistant murine simian virus 40 tumors express a retroviral type-specific gp70

Prototype macrophage-resistant and -sensitive cells were subcloned. Among several subclones of the resistant line, one subclone showed partial reversion to a sensitive phenotype. Analysis with monoclonal antibodies specific for different serotypes of endogenous murine leukemia virus revealed that expression of only one such gp70 (gp70a) correlated with the macrophage-resistant phenotype.

Variable stability of a selectable provirus after retroviral vector gene transfer into human cells

Human lymphoblasts deficient in the enzyme hypoxanthine-guanine phosphoribosyltransferase (HPRT) were infected with an amphotropic helper-free retroviral vector expressing human HPRT cDNA. The stability and expression of the HPRT provirus in five cell lines with different proviral integration sites were examined by determining HPRT mutation and reversion frequencies and by blot hybridization studies. Mutation to the HPRT-negative phenotype occurred at frequencies of approximately 4 X 10(-5) to 3 X 10(-6) per generation. Most mutations in each of the five cell lines were associated with partial or complete deletions or rearrangements of the provirus. Several mutants retained a grossly intact HPRT provirus, and in one such mutant HPRT shutdown resulted from a revertible epigenetic mechanism that was not associated with global changes in proviral methylation. Therefore, mutation and shutdown of the HPRT provirus in human lymphoblasts result from mechanisms similar to those reported for several other avian and mammalian replication-competent retroviruses.

Variable stability of a selectable provirus after retroviral vector gene transfer into human cells

Human lymphoblasts deficient in the enzyme hypoxanthine-guanine phosphoribosyltransferase (HPRT) were infected with an amphotropic helper-free retroviral vector expressing human HPRT cDNA. The stability and expression of the HPRT provirus in five cell lines with different proviral integration sites were examined by determining HPRT mutation and reversion frequencies and by blot hybridization studies. Mutation to the HPRT-negative phenotype occurred at frequencies of approximately 4 X 10(-5) to 3 X 10(-6) per generation. Most mutations in each of the five cell lines were associated with partial or complete deletions or rearrangements of the provirus. Several mutants retained a grossly intact HPRT provirus, and in one such mutant HPRT shutdown resulted from a revertible epigenetic mechanism that was not associated with global changes in proviral methylation. Therefore, mutation and shutdown of the HPRT provirus in human lymphoblasts result from mechanisms similar to those reported for several other avian and mammalian replication-competent retroviruses.