MLH1 deficiency enhances tumor cell sensitivity to ganciclovir.

Suicide gene therapy with herpes simplex virus thymidine kinase (HSV-TK) and ganciclovir (GCV) is notable for producing multi-log cytotoxicity in a unique pattern of delayed cytotoxicity in S-phase. As hydroxyurea, a ribonucleotide reductase inhibitor that activates mismatch repair, can increase sensitivity to GCV, we evaluated the role of MLH1, an essential mismatch repair protein, in GCV cytotoxicity. Using HCT116TK (HSV-TK-expressing) colon carcinoma cells that express or lack MLH1, cell-survival studies demonstrated greater GCV sensitivity in the MLH1-deficient cells, primarily at high concentrations. This could not be explained by differences in GCV metabolism, as the less sensitive MLH1-expresssing cells accumulated more GCV triphosphate and incorporated more of the analog into DNA. SiRNA suppression of MLH1 in U251 glioblastoma or SW480 colon carcinoma cells also enhanced sensitivity to high concentrations of GCV. Studies in a pa nel of yeast deletion mutants confirmed the results with MLH1, and further suggested a role for homologous recombination repair and several cell-cycle checkpoint proteins in GCV cytotoxicity. These data suggest that MLH1 can prevent cytotoxicity with GCV. Targeting mismatch repair-deficient tumors may increase efficacy of this suicide gene therapy approach to cancer treatment.

GCV phosphates are transferred between HeLa cells despite lack of bystander cytotoxicity.

The role of gap junctional intercellular communication (GJIC) in bystander killing with herpes simplex virus thymidine kinase (HSV-TK) and ganciclovir (GCV) was evaluated in U251 cells expressing a dominant-negative connexin 43 cDNA (DN14), and in HeLa cells, reportedly devoid of connexin protein. These cell lines both exhibited 0% GJIC when assayed by Lucifer Yellow fluorescent dye microinjection. Bystander cytotoxicity was still apparent in 50:50 cocultures of DN14 and HSV-TK-expressing U251 cells, but not in 50:50 cocultures of HeLa cells. However, the sensitivity of HeLa HSV-TK-expressing cells to GCV decreased nearly 100-fold (IC90=109 microM) when cocultured with bystander cells compared to results in 100% cultures of HSV-TK-expressing cells (IC90=1.2 microM). A more sensitive flow cytometry technique to measure GJIC over 24 h revealed that the DN14 and HeLa cells exhibited detectable levels of communication (29 and 23%, respectively). Transfer of phosphorylated GCV to HeLa bystander cells occurred within 4 h after drug addition, and GCV triphosphate (GCVTP) accumulated to 213+/-84 pmol/10(6) cells after 24 h. In addition, GCVTP levels were decreased in HSV-TK-expressing cells in coculture (867+/-33 pmol/10(6) cells) compared to 100% cultures of HSV-TK-expressing cells (1773+/-188 pmol/10(6) cells). The half-life of GCVTP in the HSV-TK-expressing cells was approximately four times that measured in the bystander cells (12.3 and 3.1 h, respectively). These data suggest that the lack of bystander cytotoxicity in HeLa cocultures is due to low transfer of phosphorylated GCV and a rapid half-life of GCVTP in the bystander cells. Thus, GCV phosphate transfer to non-HSV-TK-expressing bystander cells may mediate either bystander cell killing or sparing of HSV-TK-positive cells, depending upon the cell specific drug metabolism.

Hydroxyurea significantly enhances tumor growth delay in vivo with herpes simplex virus thymidine kinase/ganciclovir gene therapy.

We have previously demonstrated with several cell lines in vitro that hydroxyurea (HU) synergistically enhances ganciclovir (GCV)-mediated cytotoxicity in bystander cells. In this study, we evaluated the role of DNA synthesis inhibition on enhanced bystander killing and assessed whether addition of HU would improve the efficacy of the HSV-TK/GCV system in vivo. Compared with GCV treatment alone, addition of HU resulted in increased DNA synthesis inhibition and delayed progression through S phase following removal of drug. In a xenograft tumor model, 1:10 and 1:1 mixtures of HSVtk- and LacZ-expressing SW620 cells were injected s.c. in the flanks of nude mice and treated i.p. (100 mg/kg GCV, 1500 mg/kg HU) daily for 5 days. Tumors from mice treated with GCV alone grew rapidly and increased to 10 times their initial size in 15.7 +/- 1.8 and 16.0 +/- 0.9 days for 1:10 and 1:1 mixtures, respectively. However, when both GCV and HU were administered in combination, a single complete tumor regression was observed in both the 1:10 and 1:1 groups. In the remaining mice treated with GCV/HU, it took 23.2 +/- 2.1 (1:10) and 26.4 +/- 3.8 days (1:1) to obtain a similar 10-fold increase in tumor size.

Synergistic enhancement of herpes simplex virus thymidine kinase/ganciclovir-mediated cytoxicity by hydroxyurea.

We have previously demonstrated (L. Z. Rubsam et al., Cancer Res., 59: 669-675, 1999) that low ganciclovir (GCV) triphosphate (TP) levels similar to cellular deoxynucleotide concentrations can induce multilog killing in cells stably expressing herpes simplex virus thymidine kinase (HSV-TK). In this study, we evaluated whether reducing the endogenous competitor of GCV-TP, dGTP, enhanced GCV-mediated cytotoxicity. In SW620 human colon carcinoma cells stably expressing HSV-TK, the addition of the ribonucleotide reductase inhibitor, hydroxyurea (HU), decreased cellular dGTP pools and simultaneously increased the accumulation of GCV-TP levels. The amount of GCV nucleotide transfer from HSV-TK-expressing to nonexpressing (bystander) cells was quantitated in physically separated pHook-expressing bystander cells. Elevation of the GCV-TP:dGTP ratio by HU resulted in increased levels of GCV nucleotides transferred from HSV-TK-expressing to bystander cells during a 24 h drug incubation and enhanced GCV monophosphate incorporation into DNA after drug removal. Isobologram analysis demonstrated that the combination of GCV and HU was additive in 100% HSV-TK cultures and synergistic in HSV-TK/bystander mixtures. IC50 values for GCV in 1:1 cocultures of HSV-TK-expressing and nonexpressing SW620 cells were reduced from 1.5 microM to 0.07 microM with 2 mM HU. A similar reduction was also observed with HT-29 cells and U251 cells. With 2 mM HU, IC50 values for GCV in 10:90, 5:95, and 1:99 SW620 HSV-TK-expressing and nonexpressing cocultures were reduced from 55 microM to 0.3 microM, 71 microM to 0.8 microM, and 118 microM to 7 microM, respectively. These results demonstrate the ability to pharmacologically enhance HSV-TK/GCV-mediated bystander killing and may have an important therapeutic impact.

Cytotoxicity and accumulation of ganciclovir triphosphate in bystander cells cocultured with herpes simplex virus type 1 thymidine kinase-expressing human glioblastoma cells.

The ability of herpes simplex virus type 1 thymidine kinase (HSV-TK)-expressing cells incubated with ganciclovir (GCV) to induce cytotoxicity in neighboring HSV-TK-negative (bystander) cells has been well documented. Although it has been suggested that this bystander cell killing occurs through the transfer of phosphorylated GCV, there is little direct proof that bystander cells can accumulate GCV nucleotides. We have studied the ability of U251 human glioblastoma cells expressing HSV-TK (U251tk cells) to induce cytotoxicity in neighboring U251 bystander cells that lack the viral kinase (U251beta gal cells) and evaluated whether this bystander cell killing is mediated by GCV nucleotides. The cytotoxicity studies demonstrated that the ratio of HSV-TK-expressing cells:bystander cells was important in determining the sensitivity of both cell types to GCV. U251tk cells cocultured with an equal number of U251beta gal cells (a 50:50 ratio) exhibited a sensitivity to GCV similar to that observed in the absence of bystander cells, with >99.8% cell kill at 1 microm GCV. However, in cultures with 10% U251tk cells and 90% bystander cells (a 10:90 ratio), 1 microM GCV decreased the survival of U251tk cells by only 54%. Strong bystander cell killing was observed at both ratios. In a 50:50 coculture of U251tk and U251beta gal cells, the survival of bystander cells was decreased by >99.5% with 3 microM GCV, whereas 30 microM GCV was required to effect a similar decrease in bystander cell survival when 90% of the culture consisted of U251beta gal cells. To determine whether this bystander cell killing may be mediated by GCV nucleotides, we developed a technique to separate the two cell populations after coculture. A U251 bystander cell line was developed from the parental cell line by transfection with the cDNA coding for green fluorescent protein (U251gfp cells), which permitted the separation of U251gfp cells from nonfluorescing U251tk cells by flow cytometry with cell sorting. With this technique, bystander cells were isolated in a viable state with >97% purity within 1 h after harvest, permitting analysis of the nucleotide pools for the presence of phosphorylated GCV. The results demonstrated that significant levels of the triphosphate of GCV (GCVTP) accumulated in bystander cells within 4 h of coculture, and this accumulation was dependent upon the percentage of HSV-TK-expressing cells as well as the concentration of GCV and the length of incubation. The proportion of GCVTP in bystander cells was consistently 50-80% of that in HSV-TK-expressing cells in the 50:50 or 10:90 cocultures, suggesting a facile transfer of phosphorylated GCV. However, the actual amount of GCVTP was as much as 8-fold lower in both the U251tk and U251beta gal cells cocultured at a ratio of 10:90 compared to those cocultured at a ratio of 50:50, which is consistent with the lesser effect on cell survival. When U251tk and U251gfp cells were cultured with 1-beta-D-arabinofuranosylthymine (araT), the 5'-triphosphate of araT accumulated in the bystander cells, demonstrating that the transfer of phosphorylated compounds between these cell types is not restricted to GCV nucleotides. However, the proportion of araT-5'-triphosphate in bystander cells compared to that in HSV-TK-expressing cells was lower than that for GCVTP, and the amount was not sufficient to decrease survival in the bystander population.

Differential ganciclovir-mediated cytotoxicity and bystander killing in human colon carcinoma cell lines expressing herpes simplex virus thymidine kinase.

The two human colon carcinoma cell lines HT-29 and SW620, which stably express herpes simplex virus thymidine kinase (HSV-TK), are sensitized to the cytotoxic effects of the antiviral drug ganciclovir (GCV). Compared with HT-29 cells, SW620 cells were more sensitive to lower GCV concentrations (<1 microM), accumulated GCV triphosphate more rapidly, and incorporated higher levels of GCV into DNA. Following a 24-hr exposure to 10 microM GCV, bystander killing was as much as sixfold greater in SW620 cells than HT-29 cells. This bystander effect was dependent on the level of HSV-TK expression, the number of cells expressing HSV-TK, and the overall confluency of the cells. However, bystander killing did not correlate with gap junctional intercellular communication as determined by microinjection of Lucifer Yellow fluorescent dye. SW620 cells were coupled to <3% adjacent cells (compared with >50% for HT-29 cells), but were still able to transfer phosphorylated GCV to bystander cells as soon as 4 hr after drug was added. These results emphasize the importance of cell-specific metabolism in HSV-TK/GCV-mediated cytotoxicity and may suggest a novel mechanism for bystander killing.

Lysine residues at positions 234 and 236 in yeast porin are involved in its assembly into the mitochondrial outer membrane.

Various point mutations of lysyl residues in yeast mitochondrial porin (283 residues) were tested for their ability to assemble in vitro into the outer membranes of intact yeast mitochondria. Assembly was evaluated by protection from proteinases. The extent of assembly of two of the mutants, K234E and K236E porins, was much less than for wild-type in either post-translational or co-translational assembly assays. Lysine to glutamate mutants at other positions and K234R porin assembled as well as wild-type, but K234Q porin was poorly inserted. When both Lys-234 and Lys-236 were mutated, K234R/K236R porin was inserted better than K234Q/K236Q porin, which was inserted better than K234E/K236E; however, none of these mutants assembled as well as wild-type porin. It was concluded that optimal assembly of yeast porin depended on the presence of positively charged residues at both positions 234 and 236 and a lysine at one of these positions. After undergoing the assembly reaction, mutants that were vulnerable to proteinase K (i.e. K234E, K234Q, and K236E porins) seemed to be incompletely digested and were, to varying degrees, resistant to extraction by Na2CO3 (pH 11.5). These experiments suggested that these mutants were incompletely inserted into the outer membrane. Both Lys-234 and Lys-236 are included in an internal pentapeptide, VKAKV, that is conserved in porins from protists, plants, and animals, and it is possible that, at least, the lysines in this tract are one of the signals for the membrane assembly of these proteins.

Partial characterization of the human CYP1A1 negatively acting transcription factor and mutational analysis of its cognate DNA recognition sequence.

Previous studies in our laboratory identified a negative regulatory domain in the 5'-flanking region of the human CYP1A1 gene containing two negative regulatory elements (NRE). Characterization of one of these elements revealed three nuclear protein binding regions: a 21-bp palindrome with a point of symmetry at -784 and two guanine- and cytosine-rich elements that flank the palindrome. Functional studies suggested the palindrome is critical for transcriptional repression, whereas the guanine- and cytosine-rich sequences play a secondary role. In this study, the interaction between nuclear proteins and the CYP1A1 NRE was further defined. Electrophoretic mobility shift assays (EMSA) indicated that the NRE -784 palindrome alone, but not the guanine- and cytosine-rich sequences minus the palindrome, was capable of specific nuclear protein binding. Competitive cotransfection experiments confirmed this observation in intact cells. Specific residues important for DNA-protein interactions were identified by site-directed mutagenesis and competitive EMSA. The loss of specific protein binding was also correlated with the loss of negative regulatory activity in a transient-expression assay. Finally, competitive EMSA was performed with consensus oligonucleotides for known transcription factors. An NF-Y consensus sequence efficiently competed with the NRE probe for specific nuclear protein binding. EMSA supershift analyses indicate that a protein immunologically related to NF-YB is part of the specific nuclear protein complex binding the human CYP1A1 NRE. These studies have refined our understanding of the sequences critical for the transcriptional repression of human CYP1A1. To our knowledge, this is also the first report implicating a member of the NF-Y transcription factor family in negative gene regulation.

In vitro binding and functional studies comparing the human CYP1A1 negative regulatory element with the orthologous sequences from rodent genes.

In previous studies, we identified a 21 bp palindrome (-794 to -774) located within the negative regulatory element of the human CYP1A1 gene consisting of an 8 bp inverted repeat and 5 bp spacer. This element specifically binds protein(s) present in HepG2 nuclear extract preparations and is capable of down-regulating heterologous promoters and enhancers in transient expression assays. Conserved guanine/cytosine-rich regions which flank the palindrome also were implicated in this activity. In the present study, we examined similar regions from the rat (-881 to -746) and mouse (-822 to -683) CYP1A1 genes for their ability to bind nuclear protein and down-regulate heterologous promoters and enhancers. These rodent DNA fragments contain the conserved guanine/cytosine-rich sequences, as well as half-sites similar to those found in the human CYP1A1 palindrome. However, each half-site is separated by approximately 40 bp. DNase I footprint analyses revealed the presence of rat and mouse nuclear proteins which gave a similar protection pattern as that observed with nuclear proteins from the human cell line, HepG2. Electrophoretic mobility shift assays with the human negative regulatory element demonstrated the formation of specific DNA-protein complexes with rat and mouse nuclear protein(s). Interestingly, two specific DNA-protein complexes were observed with rodent extracts as compared to the single specific complex seen with human extract. Specific binding was not observed with either the orthologous rat or mouse fragments using human or rodent extracts. In transient expression assays, the rat and mouse fragments were unable to down-regulate enhancer/promoter activity. This absence of negative regulatory activity occurred whether transfections were performed in human, rat or mouse hepatoma cell lines. The human negative regulatory element, which was previously shown to down-regulate heterologous enhancers/promoters approximately 70% in human cells, did not exhibit this activity in rodent cell lines. UV cross-linking and southwestern blot analyses indicated a high degree of similarity between human and rodent NRE binding proteins, although some differences also were apparent. The possible implications of these findings with regards to species differences in the regulation of CYP1A1 expression are discussed.

Specific nuclear protein binding to a negative regulatory element on the human CYP1A1 gene.

Employing reporter gene/CYP1A1 chimeric plasmids, we previously identified a 275-base pair (bp) cis-element (-833 to -558; NRE275) that down-regulated the CYP1A1 promoter. In the present study, this negative regulatory activity was further localized to two subfragments of 105 bp (-833 to -728; NRE105) and 170 bp (-728 to -558; NRE170), each of which exhibited activity with a heterologous promoter/enhancer. Co-transfection studies demonstrated a dependence on cellular trans-acting factors present at limiting concentrations. Electrophoretic mobility shift assays revealed the presence of protein(s) that specifically bound to NRE275, NRE105, and NRE170. Based upon competition studies, the protein(s) that bound to NRE105 appeared to recognize sites similar to those recognized by the NRE170-binding proteins. DNase I footprint analysis of NRE105 demonstrated nuclear protein binding to a 21-bp palindrome (-794 to -774). Protection was also observed along conserved guanine/cytosine-rich sequences that flank the palindrome, but in a strand-specific manner. Guanine residues involved in protein binding were identified by methylation interference experiments. Based on transient expression assays with each of the three NRE105 components, all three appear to be necessary for complete negative regulatory activity. However, it is clear the palindrome is the most important sequence with the guanine/cytosine-rich elements playing an ancillary role.