(Received for publication, July 17, 1996, and in revised form, January 31, 1997)
From the Instituto de Parasitología y
Biomedicina, Consejo Superior de Investigaciones Científicas,
18001 Granada, Spain and the § Institute of Cancer Research,
Chester Beatty Laboratories, London SW3 6JB, United Kingdom
Perturbing deoxyribonucleoside triphosphate (dNTP) metabolism with inhibitors of the de novo synthesis of dNTP causes apoptosis in the interleukin-3 (IL-3)-dependent pre-B cell line BAF3. Under these conditions apoptosis is prevented when deoxyribonucleosides for dNTP synthesis are supplied in the culture medium. On the other hand, removal of IL-3 from cultures of BAF3 cells resulted in down-regulation of thymidine kinase activity, rapid imbalance in dNTP levels, and apoptosis. In this study we show that overexpression of a heterologous thymidine kinase, herpes simplex virus thymidine kinase (TK), in BAF3 cells protects these cells from apoptosis induced by either inhibitors of dNTP synthesis or IL-3 deprivation. This protection against apoptosis is abrogated by 9-(4-hydroxybutyl)-N2-phenylguanine, a specific inhibitor of herpes simplex virus-1 TK. These results suggest that deoxyribonucleoside kinases, particularly TK, may be important in the regulation of apoptosis in hemopoietic cells.
Cell population dynamics depend upon changes in the balance between cell proliferation and death. Transformed cells may be those that either proliferate in the absence of growth factors or fail to undergo apoptosis upon factor removal (1). Tumor cells are susceptible to apoptosis, and certain therapeutic strategies for cancer have been developed to induce apoptosis. Indeed, a number of antineoplastic drugs and treatments exert their cytotoxic effect by inducing apoptosis (2). Furthermore, the drug and radiation resistance of many tumors can be ascribed to the failure of cancer cells to enter apoptosis due, for example, to mutations in p53 or deregulation of the expression of proteins of the bcl-2 family (1, 3).
The maintenance of balanced dNTP1 pools is
critical for DNA replication and repair; under normal culture
conditions, it is mainly achieved by regulation of the activity of
enzymes of the de novo pathway of synthesis of dNTPs (4).
Whereas moderate perturbation of dNTP pools affects genetic stability
of cells (5), with the appearance of frequent mutations (6) and strand breaks (7), severe imbalance of dNTP pools causes cell death (8).
Several antineoplastic agents that inhibit DNA precursor synthesis have
been reported to kill lymphoid cells by induction of apoptosis (3), and
inherited deficiencies in enzymes such as adenosine deaminase and
purine nucleoside phosphorylase, which produce imbalanced accumulation
of dNTPs, result in lymphoid cell death (9). Previous results from our
laboratory have shown that inhibition of thymidylate synthase (an
enzyme of the de novo synthesis of dNTP) with
5-fluoro-2-deoxyuridine, which perturbs deoxyribonucleotide
metabolism, drives the IL-3-dependent cell line BAF3 to
enter apoptosis even in the presence of IL-3 (10). Furthermore, removal
of IL-3 from BAF3 cell cultures leads to an early imbalance in dNTP
pools during apoptosis (11).
In addition to the de novo pathway for dNTP synthesis, mammalian cells, in particular those of the immune system, contain kinase activities for deoxyribonucleosides (4). These salvage enzymes function in the reutilization of degradation products from nucleic acids or precursors from the extracellular medium. Thymidine kinase is perhaps the best characterized of these enzymes; its substrates are thymidine and deoxyuridine, and it is inhibited by dTTP (4). It has been used as a malignancy marker for a variety of tumors (12, 13), and tumor cells deficient in this enzyme show a lower oncogenic potential than the wild-type cells from which they derive (14). Moreover, a recent report has also demonstrated that TK is a major radioresponse determinant in rat glioma cells (15).
It has been suggested that this reutilization pathway, acting in a concerted manner with other salvage enzymes and the de novo pathway, could also regulate intracellular levels of dNTPs (4) and protect cells from apoptosis (16-18). Here, we show that the potentiation of this salvage pathway either by the supply of dNTP precursors or by overexpression of a heterologous TK (HSV-1 TK) delays programmed cell death induced by IL-3 deprivation and drugs that inhibit dNTP metabolism in BAF3 cells.
RPMI 1640 medium and fetal bovine serum were
obtained from Life Technologies, Inc. Methotrexate was from Cyanamid
Iberica, Division Lederle. Deoxy[8-3H]ATP (24 Ci/mmol),
deoxy[8-3H]GTP (16.9 Ci/mmol),
deoxy[5-3H]CTP (18 Ci/mmol), and
[methyl-3H]thymidine (49 Ci/mmol) were from Amersham.
[methyl-3H]TTP (40 Ci/mmol) was purchased from ICN
Biochemicals. Synthetic DNA templates, 5-fluoro-2-deoxyuridine,
hydroxyurea, and other reagents of the purest grade available were
obtained from Sigma and Boehringer Mannheim. The HSV-1 TK inhibitor
9-(4-hydroxybutyl)-N2-phenylguanine
(HBPG) was kindly provided by Dr. George E. Wright (University of
Massachusetts Medical School, Worcester, MA).
Murine IL-3-dependent BAF3 cells (19) were maintained in RPMI medium containing 10% fetal bovine serum, 1 mM glutamine, and 10% conditioned medium from the IL-3-producing cell line Wehi-3B. For transient expression experiments, BAF3 cells were transfected with HSV1-TK cDNA cloned into MFG-S plasmid (Somatix) or with a control cDNA (the neomycin-resistance gene, neo) by electroporation of 10-20 µg of DNA at 350 V or alternatively by lipofection with 5 µg of Lipofectin (Life Technologies, Inc.) and 2.5 µg of DNA. Cells were cultured in the presence of IL-3 for 24-48 h, and thymidine kinase activity was measured. For stable expression of HSV thymidine kinase activity, cells were cotransfected by electroporation with the previous TK cDNA and a plasmid conferring resistance to puromycin, pBabe Puro (20). Puromycin was added at 4 µg/ml to the cultures to select cells expressing resistance to this marker.
Analysis of DNA fragmentation and cell cycle were performed according to published procedures (21).
Deoxyribonucleoside Triphosphate Pool AssayPreparation of cell extracts was essentially as described (22). dNTPs were determined by the DNA polymerase assay (23) using a synthetic DNA template and the Klenow fragment of DNA polymerase. The intracellular concentrations of dNTP were estimated from calibration curves obtained using pure standards.
Thymidine Kinase ActivityCells were pelleted and washed twice with ice-cold PBS. The pellet was then resuspended in 20 mM Tris-HCl, pH 7.0, 2 mM MgCl2, 20 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride, subjected to Dounce homogenization, and sonicated to completely disrupt the cells. The extract was cleared of cellular debris by centrifugation, and the remaining supernatant containing a minimum of 2 mg of protein/ml was assayed for TK activity essentially as described (24).
Statistical AnalysisThe statistical significance of the data was determined after applying Student's t test.
Dihydrofolate reductase and ribonucleotide reductase
catalyze key steps in the de novo production of dNTPs.
Inhibitors of these enzymes, such as methotrexate (MTX) and hydroxyurea
(HU), are commonly used in antineoplastic treatment, and they also
induce apoptosis in hemopoietic BAF3 cells even in the presence of IL-3 (10). After 8 h of treatment with either drug, loss of cell viability, as assessed by cell membrane integrity, was not observed (Fig. 1A); however, at this time DNA was
digested into oligonucleosome-sized fragments (Fig. 1B).
Cell death could be clearly observed after 15 h of treatment with
either inhibitor. Addition of deoxyribonucleosides to the medium
facilitates dNTP synthesis; these precursors are transported by a
nonspecific permease across the cell membrane and modified by the
action of deoxyribonucleosides kinases to generate dNTP pools (4), thus
bypassing the de novo synthesis pathway. The presence of 50 µM thymidine in the culture medium completely prevented
the appearance of the sub-G1 peak of apoptotic cells in
cultures of BAF3 cells treated with MTX for 15 h (Fig. 1C) and inhibited DNA fragmentation and cell death (data not
shown). Inhibition of HU-induced apoptosis was achieved by incubating the cells with 1 µM deoxyadenosine and 100 µM deoxyguanosine (Fig. 1C). This precursor
combination was used in the study of Lagergren and Reichard (25), in
which it reversed the inhibition of DNA synthesis by HU. These data
suggest that salvage enzymes involved in the phosphorylation of these
precursors are probably important in the maintenance of dNTP balance
and the inhibition of cell death (17, 18). A similar role of this
salvage pathway has been demonstrated in erythroblasts from mice with
experimental folate deficiency anemia, in which the addition of
thymidine is sufficient to protect cells from apoptosis (16).
Overexpression of HSV-1 Thymidine Kinase Suppresses Apoptosis Induced by IL-3 Removal and Inhibitors of dNTP Synthesis
To
determine the role of TK activity in the regulation of apoptosis, we
have examined the effect of overexpressing heterologous HSV-1 TK (26)
on the entry of cells into apoptosis after IL-3 withdrawal or drug
treatment. Initially, we transiently expressed HSV-1 TK in BAF3 cells
and determined both the level of TK activity and apoptosis in
bulk-transfected cells. Bulk-transfected cell populations were deprived
of IL-3 for 24 h, and viability was measured at this time. The
results from eight independent transfection experiments are shown in
Fig. 2A; a correlation was found between the
level of expression of TK activity and protection from cell death.
Whereas cell viability decreased to 20% in control
(vector-transfected) cells, transfection of HSV-1 TK resulted in up to
83% viable cells after IL-3 deprivation. These trypan blue-excluding
cells in the HSV-1 TK-transfected cultures were viable because they
were able to grow in response to IL-3, with a generation time similar
to that of cells not subjected to IL-3 withdrawal (data not shown).
We next determined whether stable HSV-1 TK expression could inhibit
apoptosis. By transfecting BAF3 cells with the HSV-1 TK cDNA,
several clones were generated that expressed high levels of TK
activity. From these clones, clone TK2 was chosen because it did not
release any soluble factor to the culture media to allow cell survival
in the absence of IL-3, which would mask the effect of TK
overexpression. Cell viability in a control clone expressing only
resistance to puromycin (puro3) decreased to less than 30% after
24 h in the absence of IL-3, and there was a complete loss of cell
viability after 48 h of cytokine deprivation (Fig. 2B).
In contrast, viability of TK2 cells was maintained at values higher
than 90 and 30% after 24 and 48 h of IL-3 withdrawal,
respectively (Fig. 2B). Furthermore, the clone
overexpressing HSV-1 TK was able to maintain TK activity after
deprivation of IL-3, in contrast to what was observed in puro3 cells
(data not shown) and parental BAF3 cells (11). We have also determined
the effect of HSV-1 TK overexpression on the viability of cells treated
with inhibitors of dNTP metabolism such as MTX and FdUrd, which rapidly
deplete the cells of dTTP. Results in Fig. 3A
and B show that cells overexpressing HSV-1 TK are more
resistant to cell death induced by increasing doses of FdUrd or
MTX.
We also measured intracellular dNTP levels after MTX treatment or IL-3
withdrawal in control and HSV-1 TK-transfected cells. The dNTP pools in
untreated cultures from both cell types were not significantly
different (Fig. 4, legend), but the cell response to an
antimetabolite drug or IL-3 deprivation was markedly different (Fig.
4). Control cells treated for 3 h with 5 µM MTX
exhibited a 66% decrease in dTTP and dGTP levels, whereas dATP and
dCTP remained unchanged. In IL-3 deprivation experiments, the levels of
dATP, dGTP, and dTTP after 8 h in the absence of IL-3 decreased to
about 40-50% of the initial levels, whereas the dCTP level was only
slightly changed, as reported previously (11). Interestingly, in cells
overexpressing HSV-1 TK, treatment with MTX had no effect on the
intracellular dATP, dGTP, and dTTP levels and had little effect on the
dCTP pool (67% of the level found in untreated cells). Furthermore,
removal of IL-3 from these cells did not induce a decrease in the
values of all four dNTPs, showing a general maintenance of dNTP levels
as a result of TK overexpression. Although one would have expected only
the dTTP level to be maintained in HSV TK-overexpressing cells,
allosteric activation of ribonucleotide reductase by dTTP might be
responsible for the elevated pool of dGTP, which in turn can stimulate
the reduction of ADP to dADP (27).
Earlier reports have shown that introduction of herpes virus DNA into
TK-deficient cells was able per se to alter the sensitivity to -interferon, irrespective of the acquisition of TK activity (28,
29). To rule out this possibility in our experiments, we took advantage
of a recently characterized specific inhibitor of HSV-1 TK (30). This
compound, HBPG, did not inhibit cellular TK activity as assayed
in vitro on cell extracts from control cells and had no
effect on [3H]thymidine incorporation in growing cells
(data not shown). Fig. 5A shows that HBPG did
not affect the cell viability of control cells in either the presence
or absence of IL-3. However, in clones overexpressing HSV-1 TK, 10 µM HBPG produced a 67% inhibition of the TK activity
present in cell extracts from these cells (data not shown). In these
cells, the inhibitor did not induce apoptosis in the presence of IL-3
(Fig. 5B). However, in IL-3-deprived TK-transfected cells,
HBPG significantly reduced cell viability (Fig. 5B) and induced a sub-G1 population of apoptotic cells (Fig.
5D) to values similar to those of control cells (Fig. 5,
A and D). The inhibition of viral TK by HBPG also
increased the sensitivity of TK2 cells to FdUrd and MTX as determined
by measuring cell viability (Fig. 5C) or by the generation
of sub-G1 cells in cell cycle analysis experiments (Fig.
5D). In summary, these results indicated that the protection
from apoptosis in cells overexpressing HSV-1 TK was the direct outcome
of the presence of the viral TK activity and not a consequence of the
introduction of viral DNA.
Whereas many intracellular signals are known to decline after removal of IL-3 from dependent cells (31-33), only overexpression of oncogenes such as myc (34) or bcl-2 (21), activators of the ras pathway (35, 36), or activated abl kinase (37) has previously been shown to modulate apoptosis. Our data demonstrate that a single enzyme involved in nucleotide metabolism can exert a similar effect, which suggests that regulation of the dNTP supply may be a control point in apoptosis.
Although mammalian TK expression is cell cycle-regulated, with peak gene expression occurring immediately before entry into S-phase (38), we have demonstrated that IL-3 also regulates TK activity posttranslationally (11). Serum has also been shown to regulate the phosphorylation state and activity of TK in HL-60 cells (24). The vector used in our transfection experiments constitutively expresses HSV TK throughout the cell cycle, and the viral TK does not seem to be down-regulated after withdrawal of IL-3, resulting in increased TK activity even in the absence of IL-3.
Under normal culture conditions, the viability and growth of cells do not seem to require the uptake of exogenous deoxynucleosides unless the endogenous synthesis of dNTP has been inhibited by treatment with drugs (17, 18). This is because most culture media contain high levels of folic acid, which allows the continued synthesis of tetrahydrofolate derivatives required for the synthesis of nucleotides and particularly of thymidylate. However, in culture media containing concentrations of folic acid comparable to the serum levels, the growth of cells is dependent on the supply of exogenous thymidine, and therefore, under these conditions, TK activity should play an important role in the pathway leading to dTTP formation for DNA synthesis and repair (39). Therefore, in cells such as hemopoietic BAF3 cells that are strictly dependent on growth factors for cell viability and proliferation, regulation of TK activity by IL-3 (11) could be an important event in the maintenance of balanced dNTP pools when the extracellular concentration of folic acid is comparable with that of serum.
Due to the complexity of dNTP metabolism and the multiple allosteric mechanisms involved in this metabolism, at present it is difficult to identify the critical variable that induces apoptosis during dNTP pool imbalances. The fact that inhibition of one specific pathway such as that leading to de novo dTTP synthesis might influence the formation of other dNTPs (27) makes it difficult to propose a more specific model. The mechanism by which viral TK overexpression might protect cells from IL-3 removal or dNTP synthesis inhibitors is probably related to the increased capacity of the cell to maintain dNTP pools for DNA repair and synthesis (40). dNTP pool balance has been shown to be essential for fidelity in DNA synthesis in dividing cells (4). In BAF3 cells, IL-3 withdrawal does not result in cell cycle arrest (21), thus DNA synthesis is probably taking place in cells with imbalanced dNTP pools. In other cells, promotion of entry into S-phase by disregulated myc (41) or E2F (42) expression in the absence of growth factors also induces apoptosis. It has been previously shown that nucleotide misincorporation during DNA replication and repair under limiting metabolic conditions (43) may serve as a signal to initiate a death program (8) to avoid the appearance of genetic variants with chromosomal abnormalities. Although the dNTP pool imbalances observed in BAF3 cells may not be sufficiently large to induce an important increase in replication error rates, it is possible to speculate that small changes in DNA replication accuracy or misincorporation of uracil residues into DNA might activate genes that are important for the execution of apoptosis (8) or induce the accumulation of proteins such as p53 that trigger the apoptotic program (44).
Cells lacking TK activity have been shown to have a lower rate of DNA
synthesis, a longer generation time (45), and an increased sensitivity
to DNA alkylating agents (46) that is probably related to their
inability to maintain dTTP pools. These agents are also potent inducers
of apoptosis (47). Our results raise the hypothesis that regulation of
TK activity by growth factors, and therefore the cellular dNTP levels,
could be a checkpoint to signal proliferating cells whether to continue
the progression through the cell cycle or to enter a program of cell
death by apoptosis. Deregulated expression of cellular TK activity has
been reported to be correlated with tumorigenicity (12, 13) and cell
resistance to radiation (15). On the other hand, viral TK could be an
apoptosis-inhibiting factor during productive or latent infection of
cells with herpes simplex virus, acting perhaps in a concerted manner
with the viral proteins 34.5 and ICP4, recently described as
inhibitors of apoptosis (48, 49).