Differential Mechanism of Cytostatic Effect of (E)-5-(2-Bromovinyl)-2’-deoxyuridine, 9-( 1,3-Dihydroxy-2-propoxymethyl)guanine, and Other Antiherpetic Drugs on Tumor Cells Transfected by the Thymidine Kinase Gene of Herpes Simplex Virus Type 1 or Type 2”

Abstract After they have been transfected with the herpes simplex virus type 1 (HSV-1) or type 2 (HSV-2) thymidine kinase (TK) gene murine mammary carcinoma (FM3A) cells become highly sensitive to the growth inhibitory properties of the antiherpetic agents (E)-5-(2-bromovinyl)-2'-deoxyuridine (BVDU), 9(-)[(2-hydroxyethoxy)methyl]guanine (acyclovir, ACV), 9(-)[(1,3-dihydroxy-2-propoxy)methyl]guanine (DHPG, ganciclovir), and 1-(2-deoxy-2-fluoro-beta-D-arabinofuranosyl)-5-methyluracil (FMAU). BVDU was 100-fold more potent an inhibitor of HSV TK gene-transfected tumor cell growth (50% inhibitory concentration (IC50), 0.0020-0.0047 microM) than FMAU or DHPG (IC50, 0.051-0.277 microM) and 1000-fold more potent than ACV (IC50, 0.42-4.9 microM). As a rule, the test compounds were more cytostatic to HSV-2 TK than HSV-1 TK gene-transfected FM3A cells. This may be ascribed to the higher phosphorylating capacity (Vmax/Km) of HSV-2 TK than HSV-1 TK and/or to the higher TK enzyme levels of the HSV-2 TK gene-transfected FM3A cells than the HSV-1 TK gene-transfected FM3A cells. Thymidylate synthase of the HSV TK gene-transfected FM3A cells appears to be the target enzyme for the cytostatic action of BVDU, but not FMAU, DHPG, or ACV. Instead, the cytostatic activity of DHPG seems to be correlated with its conversion to the triphosphate form and subsequent incorporation into the DNA of HSV TK gene-transfected FM3A cells.

Several nucleoside analogues have been reported to selectively inhibit the replication of herpes simplex virus type 1 (HSV-1)' and type 2 (HSV-2) both in vitro and in uiuo. Foremost among the antiherpetic compounds that have demonstrated efficacy in the treatment of herpesvirus infections in animals and/or humans are 9-[ (2-hydroxyethoxy)methyl] guanine (acyclovir, ACV), 9-[(1,3-dihydroxy-2-propoxy)-* This investigation was supported by Krediet 3.0026.91 from the Belgian Fonds voor Geneeskundig Wetenschappelijk Onderzoek, Krediet 7.0049.90 from the Belgian Nationaal Fonds voor Wetenschappelijk Onderzoek, and Krediet 91/94-2 from the Belgian Geconcerteerde Onderzoeksacties. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Inhibition of Cell Proliferation-The methods for evaluating the cytostatic effects of the test compounds against tumor cells have been described previously (26). Briefly, 5 X 10' cells were suspended in growth medium and added to microplate wells in the presence of varying concentrations of the test compounds. In those assays that were aimed at reversing the inhibitory effects of the test compounds, dUrd (125 pg/ml) or dThd (5 pg/ml) were added to the cells, together with varying amounts of the test compound. The concentrations employed for dUrd and dThd correspond to the maximal concentrations of dUrd and dThd that are themselves not inhibitory to tumor cell proliferation. The cells were allowed to proliferate for 48 h (FM3A, L1210) or 72 h (CEM, C8166, Molt 4lclone 8) at 37 "C in a humidified C0,-controlled atmosphere. At the end of the incubation period, the cells were counted in a Coulter counter. The inhibitory concentrations of the test compounds for the different cell lines listed in Table I are the mean of at least three to five independent determinations.
FM3A TK-/HSV-1 TK+/TS-Cell Growth Assay-The FM3A TK-/HSV-1 TK+/TS-cells are triple mutant cells that are deficient in cytosol TK and thymidylate synthase (TS) but transfected with the HSV-1 TK gene. These cells grow only in the presence of exogenous dThd. Before initiating the cell growth experiments, the cells were thoroughly washed with dThd-free medium before they were seeded in each well of a microtest plate at 5 X lo' FM3A TK-/ HSV-1 TK+/TS-cells in the presence of a given amount of test compound (1000,400,100,40,10,4, and 1 p~) .
The cells were allowed to proliferate for 48 h at 37 "C in a humidified, COz-controlled atmosphere. At the end of the incubation period, the cells were counted in a Coulter counter.
Determination of Tritium Release from [5-3H]dCyd-Activity of TS in intact FM3A cells was measured by estimation of tritium release from [5-3H]deoxyuridylate (formed in the cells from [5-3H] dCyd) in the reaction catalyzed by TS. This method has been described previously (27) but was modified as follows. FM3A cells were collected by centrifugation at 200 x g for 8 min and resuspended in fresh medium; 300 p1 of this cell suspension (0.75 X lo6 cells) were added to 60 p1 of medium containing an appropriate amount of test compound and 40 p1 of 4 pCi [5-3H]dCyd. At 0, 10,20, and 30 min, 100 pl of the reaction mixture were withdrawn and mixed with 250 pl of a cold suspension of carbon black (160 mg/ml) in 5% trichloroacetic acid. After centrifugation at 1100 X g for 10 min, 200-p1 samples of the supernatants were analyzed for radioactivity.
Phosphorylation of PHIDHPG in FM3A Cells-The metabolism of [3H]DHPG was monitoreM as follows. FM3A/O, FM3A TK-/HSV-l TK', or FM3A TK-/HSV-2 TK' cells were seeded at 4 X 10' cells/ ml in 5-ml culture bottles and incubated with 0.05 pM [3H]DHPG (5 pCi/bottle). At 6 or 24 h, cells were centrifuged at 200 X g, washed twice with cold medium, and precipitated with cold methanol 66%. After centrifugation at 10,000 X g for 3 min, the supernatants were subjected to high performance liquid chromatography analysis using a Partisphere-SAX column. A linear gradient of 5 mM (NH4)HzPOI, pH 5.0 (Buffer A) to 500 mM (NH4)H2P04, pH 5.0 (Buffer B) was used to separate the metabolites as follows: 5 min of 100% Buffer A, 15 min of linear gradient to 10% Buffer B, 15 min of linear gradient to 100% Buffer B, 10 min of 100% Buffer B, 5 min of linear gradient to 100% Buffer A, and 5 min of equilibration with Buffer A. The different fractions of the eluate were assayed for radioactivity in a toluene-based scintillant. The time of elution of DHPG and its 5'monophosphate, 5'-diphosphate, and 5"triphosphate derivatives were 1, 14, 28, and 36 min, respectively. The precipitates of the methanol extractions (containing FM3A nucleic acids) were further washed three times with cold methanol 66% in HzO and assayed for radioactivity.
Quuntitation of Thymidine Kinase Activity in FM3A Cells and Determination of the Inhibitory Effects of BVDU, DHPG, and ACV against the Enzyme-FMBA/O, FM3A TK-/HSV-1 TK+, and FM3A TK-/HSV-2 TK' cell pellets (-500 X lo6 cells) were first washed with cold phosphate-buffered saline and then washed twice with suspension buffer (50 mM potassium phosphate, pH 7.6, containing 2 mM dithiothreitol, 2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 5 mM benzamidine). The cell pellet could either be stored at -70 "C or suspended in -4 ml of suspension buffer. The cell suspension was then sonicated three times for 10 s, supplemented with sucrose to obtain a final sucrose concentration of 0.25 M, and cleared by centrifugation at 43,000 X g for 20 min. The 70% (NHASO, precipitate of the cell homogenate was resuspended in -500 pl of suspension buffer and dialyzed against at least 100 volumes of 50 mM Tris-HC1, pH 7.6, containing 5 mM dithiothreitol, 5 mM MgClz, and 20% glycerol for 2-3 h at 4 "C and stored in aliquots at 70 "C. The cell extracts were assayed for dThd kinase activity in a standard reaction mixture containing 2.5 mM MgCIz, 10 mM dithiothreitol, 1 mg/ml bovine serum albumin, 2.5 mM ATP, 10 mM NaF, 1 pM [meth~l-~HIdThd (specific radioactivity, 900 cpm/pmol), an FM3A/O protein or 2 pg of FM3A TK-/HSV TK+ protein) in a total appropriate amount of inhibitor, and 5 p1 of the cell extract (1 pg of volume of 100 p1 of 50 mM Tris-HC1, pH 8.0. The reaction mixture was incubated at 37 "C for 15 and 30 min, and the reaction was terminated by spotting an aliquot (25 ~1 ) onto DE-81 discs that were instantly immersed in ethanol (70%) and thoroughly washed in ethanol. The filters were then dried and assayed for radioactivity in a toluene-based scintillant.
Phosphorylation of PHIDHPG by Thymidine Kinase from FM3A

RESULTS
Cytostatic Activity of Antiherpetic Compounds against HSV TK Gene-transfected FM3A Celkr-BVDU, DHPG, ACV, and FMAU were evaluated for their inhibitory effects on the proliferation of a number of tumor cell lines, including the human T lymphocytes CEM/O, CEM/TK-, C8166, and Molt 4/clone 8 cells, and the murine L1210, FM3A/O, FM3A TK-/ HSV-1 TK+, and FM3A TK-/HSV-2 TK+ cells. None of the human tumor cells and the murine leukemia L1210 and mammary carcinoma FM3A cells were markedly inhibited by BVDU, DHPG, and ACV (ICm, 50 to >500 p~) ( Table I). FMAU proved inhibitory to FM3A cells at -20 PM (Tables I  and 11). In contrast, the growth of FM3A TK-/HSV-l TK+ and FM3A TK-/HSV-2 TK+ cells was inhibited by BVDU at concentrations of 4.7 and 2.0 nM, that is at concentrations that are 10,000-25,000-fold lower than the concentrations required to inhibit the growth of the corresponding wild-type FM3A/O cells. Also, DHPG and FMAU were inhibitory to the HSV-2 TK gene-transfected FM3A cells at 0.051 and 0.073 PM and to HSV-1 TK gene-transfected cells at 0.190 and 0.277 PM, respectively. These concentrations were 750-3,000fold (DHPG) or 70-300-fold (FMAU), lower than those found inhibitory to the wild-type FM3A/O cells (Tables I and 11). Acyclovir inhibited the proliferation of HSV-TK gene-transfected FM3A cells at a concentration that was 50-600-fold lower than that required to inhibit the proliferation of the wild-type tumor cells. As a rule, the test compounds were markedly more inhibitory to FM3A TK-/HSV-2 TK+ than FM3A TK-/HSV-1 TK+ cells (2-fold for BVDU, 4-fold for DHPG and FMAU, and 11-fold for ACV) (Table I). Thus, BVDU proved 70-100-fold more effective than DHPG and FMAU in its cytostatic activity against HSV-2 TK genetransfected FM3A tumor cells (Table I).
Effect of Thymidine, BVDU, DHPG, and ACV on the Growth of FM3A TK-/HSV-1 TK+/TS-Celkr"FM3A TK-/ HSV-1 TK+/TS-cells were incubated in the presence of different concentrations of thymidine (dThd), BVDU, DHPG, and ACV for 48 h, and then the cell number was determined and compared with the initial cell number. In the absence of dThd, no marked cell growth could be recorded. None of the antiherpetic compounds were able to sustain FM3A TK-/ HSV-1 TK+/TS-cell growth at a concentration ranging from 1 to 1000 p~ (data not shown). Cell number was equal if not lower after the 48-h incubation period than the initial cell number. In contrast, dThd was able to sustain cell growth at all concentrations tested (1-1000 pM); 20-100 pM was the optimal. At these thymidine concentrations, FM3A TK-/ HSV-1 TK+/TS-cells did not differ in their growth rate, as compared with the wild-type FM3A/O cells cultured in the absence of dThd (generation time, 12-14 h) (data not shown).
However, DHPG had a K , value of 45 and 162 p~ for the TK from FM3A TK-/HSV-l TK+ and FM3A TK-/HSV-2 TK+ cells, respectively, and the V,, values for the HSV-1 and HSV-2 TK, derived from the HSV TK gene-transfected FM3A cells, were as high as 133 and 964 pmol/mg/min (Table VII). Consequently, the phosphorylative capacity of HSV-2 TK for DHPG was 2-fold higher than that of HSV-1 TK.
The fact that DHPG exhibits a stronger cytostatic activity than acyclovir may also be ascribed to the greater affinity of DHPG than of ACV for HSV-1 TK and HSV-2 TK; V J K , of HSV-1 TK for DHPG is 30-244-fold higher than for ACV (5,301 and V,,/K,,, of HSV-2 TK for DHPG is 60-fold higher than for ACV (30). Also, DHPG monophosphate has a greater affinity for GMP kinase than ACV monophosphate; V,,/K, of GMP kinase for DHPG monophosphate is 56-492-fold higher than for ACV monophosphate (5,29). A factor that may play a role in the higher cytostatic effects of DHPG and ACV in HSV-2 TK gene-transfected cells than in HSV-1 gene-transfected cells are the higher levels of HSV TK activity expressed in FM3A TK-/HSV-2 TK' cells, as compared with FM3A TK-/HSV-l TK* cells (Table VI). In addition, HSV-2 TK has a higher phosphorylating capacity for DHPG than HSV-1 TK (Table VII).
Our antimetabolic data indicate that BVDU, but not the other antiherpetic drugs, inhibits [1',2'-3H]dUrd incorporation into DNA at concentrations that are 10,000-to >150,000fold lower than the concentrations required to inhibit [methyl-3H]dThd incorporation (Table 111). This discriminative behavior is compatible with an antimetabolic action targeted at thymidylate synthase. Also, the ability of dThd, but not dUrd, to almost completely reverse the cytostatic activity of BVDU against the HSV TK gene-transfected cells (Table 11), as well as the extremely potent inhibitory effect of BVDU on tritium release from [5-3H]dCyd in HSV-TK gene-transfected FM3A cells (4-7 nM) (Table IV), are in full agreement with thymidylate synthase being the molecular target for the antiproliferative effect of BVDU. In fact, BVDU 5'-monophosphate is a potent inhibitor of cell-free thymidylate synthase (KiIK,,,, 0.66) (31).
The lack of marked cytostatic activity of BVDU against FM3A/O (and other cell lines) may obviously be explained by the inability of the cellular TK of normal cells to phosphorylate BVDU to its 5'-monophosphate. It is well established that BVDU is not a substrate for cytosol TK (31).
In contrast with the cytostatic activity of BVDU, the cytostatic activity of the other antiherpetic drugs DHPG, ACV, and FMAU is almost equally well reversed by dUrd and dThd. Furthermore, DHPG, ACV, and FMAU do not inhibit tritium release from [5-3H]dCyd at relatively high concentrations (20-100 p~) . This means that, whereas the cytostatic activity of BVDU against FM3A TK-/HSV TK' cells can be readily explained by an action targeted at thymidylate synthase, the other antiherpetic drugs must exert their cytostatic activity by an entirely different mechanism of action.
DHPG is 30-50-fold better phosphorylated by FM3A TK-/ HSV-1 TK' and 100-200-fold better phosphorylated by FM3A TK-/HSV-2 TK' cells than by wild-type FM3A/O cells to its 5'-triphosphate form. Also, DHPG is incorporated to a 4-and 10-fold greater extent into nucleic acids of FM3A TK-/HSV-l TK' and FM3A TK-/HSV-2 TK' cells (as compared with FM3A/O cells). These findings point to the necessity of DHPG (i) to be phosphorylated to its 5"triphosphate and (ii) to be subsequently incorporated into DNA to achieve its cytostatic effect on HSV TK gene-transfected FM3A cells. Moreover, DHPG is unable to sustain cell growth upon incorporation into FM3A TK-/HSV-l TK+/TS-cells, which means that the incorporation of DHPG into the cells' nucleic acid must have a deleterious effect on cell growth. The different mode of cytostatic action of DHPG (and ACV) versus BVDU in HSV TK gene-transfected FM3A cells corresponds well with the inhibitory effects of the triphosphate derivatives of the test compounds against DNA polymerase a. Indeed, the triphosphate derivatives of ACV and DHPG were reported to show a much greater binding affinity for, and preferential inhibition of, DNA polymerase a (Ki, 0.096 and 0.146 p~, respectively) (32) than did the triphosphate derivative of BVDU (Ki, 3.6 p~) (33). In fact, the observations of Mar et al. (32) that the triphosphate derivative of DHPG is able to (i) substitute for dGTP, (ii) act as a substrate for DNA polymerase, and (iii) incorporate into DNA, resulting in a drastic suppression of DNA chain elongation, are in agreement with our findings that DHPG is incorporated into cellular DNA (Table V), does not allow FM3A TK-/HSV-1 TK'/TS-cell growth, and principally acts at the level of DNA polymerization.
All antiherpetic compounds proved more inhibitory to FM3A TK-/HSV-2 TK+ than FM3A TK-/HSV-l TK+ cells. The greater inhibitory effect of BVDU for FM3A TK-/HSV-2 TK+ cells can be expected from an action targeted at thymidylate synthase. In HSV-2 TK gene-transfected FM3A cells, BVDU is readily metabolized to its 5'-monophosphate but not further onto the 5'-triphosphate, whereas in FM3A TK-/HSV-l TK' cells, high levels of BVDU 5"triphosphate are found (14,15). The reason for this differential phosphorylation pattern is that HSV-1 TK is able to convert dThd (and dThd analogues such as BVDU) to their 5"diphosphate form (which is then further phosphorylated to the 5"triphosphate forms by cellular enzymes), whereas HSV-2 TK converts dThd (and related analogues) to the 5'-monophosphate but not further onto the 5"diphosphate (34).
In conclusion, we have demonstrated that BVDU is by far more effective than DHPG, ACV, or FMAU in inhibiting the proliferation of HSV-1 TK and HSV-2 TK gene-transformed FM3A tumor cells. All compounds effected a relatively greater cytostatic activity against HSV-2 TK gene-transfected than HSV-1 TK gene-transfected cells due to higher TK levels in FM3A TK-/HSV-2 TK' cells than FM3A TK-/HSV-1 TK' cells and/or the higher phosphorylating capacity of HSV-2 TK than HSV-1 TK for these compounds. The mechanism of cytostatic activity of BVDU is clearly different from that of DHPG and other antiherpetic drugs. If HSV TK gene therapy would become an amenable modality for the selective treatment of tumors, HSV-2 TK gene-transfected cells should be preferred over HSV-1 TK gene-transfected cells, and BVDU should be considered as the prime candidate drug for the treatment of tumors transfected with the HSV-2 TK gene.