HUMAN GENE THERAPY 14:463–472 (March 20, 2003) © Mary Ann Liebert, Inc.
Liver Failure Caused by Herpes Simplex Virus Thymidine Kinase Plus Ganciclovir Therapy Is Associated with Mitochondrial Dysfunction and Mitochondrial DNA Depletion MAITE HERRAIZ,1 NAIARA BERAZA,1 ABELARDO SOLANO, 2 BRUNO SANGRO, 1 JULIO MONTOYA, 2 CHENG QIAN, 1 JESUS PRIETO, 1 and MATILDE BUSTOS1
ABSTRACT Herpes simplex virus thymidine kinase (HSV-tk) converts ganciclovir (GCV) into an active compound, which can be incorporated into DNA molecules and terminate DNA synthesis. Gene transfer of HSV-tk followed by GCV administration has been used with success to treat experimental cancer and this strategy has entered into clinical trials. Although it is thought that the cytotoxic effect occurs mainly in tumoral dividing cells, where mitotic activity favors integration of the genotoxic compound into nuclear DNA, there are concerns of potential damage to normal nondividing cells. In the present work we have explored the mechanisms of HSVtk/GCV toxicity and in particular whether this therapy may cause lesions of mitochondrial DNA (mtDNA) and mitochondrial dysfunction. We found that the administration of GCV to rats injected with adenovirus encoding HSV-tk induced hepatocellular damage characterized by the presence of apoptotic bodies, ballooning of hepatocytes, and severe hepatic steatosis with mitochondria enlargement and cristae dissolution at the ultrastructural level. Remarkably, Southern blot analysis showed substantial reduction in the amount of mtDNA in the liver. Using radiolabeled GCV we could demonstrate incorporation of this compound into both nuclear and mtDNA in HSV-tk–transduced rat hepatocytic cell line MCA-RH7777 and subsequent alteration of mitochondrial function. Our observations confirm that GCV can damage both nuclear and mtDNA in cells transduced with HSV-tk and that this effect could be responsible for severe mitochondrial dysfunction and toxicity in normal nondividing cells. These data are relevant for the design of clinical trials using adenoviral vectors encoding HSV-tk.
OVERWIEW SUMMARY We have explored the mechanism of liver toxicity induced by the transfer of the herpes simplex virus thymidine kinase (HSV-tk) gene followed by the administration of ganciclovir (GCV). We report that adenovirus–mediated HSVtk transfer to rats followed by GCV administration caused biochemical and histopathologic alterations consistent with mitochondrial dysfunction of liver cells. Cellular damage was characterized by the presence of apoptotic bodies and steatosis and marked mitochondrial changes at the ultrastructural level. A substantial reduction in the amount of mitonchondrial DNA (mtDNA) was demonstrated by Southern blot. In a hepatocytic cell line transduced with HSV-tk, radiolabeled GCV was incorporated into both genomic and
1Division
mtDNA with alteration of mitochondrial function. These observations show that cancer gene therapy using the suicide gene HSV-tk plus GCV causes hepatotoxicity in normal hepatocytes and emphasize the need to adjust the dose and route in the administration of the adenoviral vector in order to avoid undue toxicity.
INTRODUCTION
S
is an approach to the treatment of cancer whereby cells become susceptible to an otherwise nontoxic prodrug by genetic alteration. One of the most widely enzyme prodrug system used is the herpes simplex virus thymidine kinase/ganciclovir (HSV-tk/GCV) (Moolten, 1986). UICIDE GENE THERAPY
of Hepatology and Gene Therapy, University of Navarra, Medical School Pamplona, 31008 Pamplona, Spain. of Biochemistry and Molecular and Cellular Biology, University of Zaragoza, 50013 Zaragoza, Spain.
2Department
463
464
HERRAIZ ET AL.
This is based on the selective phosphorylation of GCV by HVtk to the level of the monophosphate (GCV-MP). Cellular kinases further phosphorylate GCV-MP to the triphosphate metabolite (GCV-TP), which is the active form of GCV (Matthews and Boehme, 1988). It is thought that the mode of action of GCV at the molecular level relies on the inhibition of nuclear DNA synthesis leading to DNA chain termination and cell death (Isley et al., 1995). From this concept it was considered that GCV-TP selectively kills tumoral dividing cells with little harm to quiescent cells such as hepatocytes. However, in a previous study we found that intraportal injection of an adenoviral vector expressing HSV-tk (Ad.tk) followed by GCV administration to rats with diethylnitrosamine-induced multifocal hepatocellular carcinoma resulted in marked reduction of the tumor mass, but this effect was accompanied by high mortality caused by severe hepatotoxicity (Quian et al, 1997). Similar hepatotoxic reactions generated by the HSV-tk/GCV system have been described by other authors (Brand et al., 1997; Van der Eb et al., 1998; Bustos et al., 2000). Adenoviral vectors have liver tropism, which easily explains why tissue damage is mainly concentrated in this organ; however, the mechanisms of liver damage induced by the Ad.tk/GCV administration has not been clearly elucidated. Although it is assumed that incorporation of phosphorylated GCV into nuclear DNA leading to DNA strand breaks (Wutzler and Thust, 2001) and apoptosis (Thust et al., 2000) is the main form of molecular lesion of the cell, no studies have addressed the possible damaging effects of phosphorylated GCV derivatives on mitochondrial DNA (mtDNA). Because other nucleoside analogues produce intense hepatotoxicity by mitochondrial dysfunction and also can cause mtDNA depletion (Martin et al., 1994), the purpose of this study has been to analyze whether the Ad.tk/GCV system may affect liver cells by altering mitochondrial function and mtDNA. Our data show that administration of high doses of Ad.tk followed by GCV results in incorporation of GCV into mtDNA, severe mitochondrial dysfunction with macro and microvesicular steatosis associated and mtDNA depletion leading to liver failure. Furthermore, our data emphasize the need to adjust the dose of adenoviral vectors encoding HSV-tk in order to avoid undue liver toxicity.
MATERIALS AND METHODS Adenoviral vectors The recombinant adenoviral vectors Ad.tk and Ad.lacZ expressing the HSV-tk gene and the reporter gene lacZ under the control of the cytomegalovirus enhancer/promoter were constructed as previously described (Qian et al., 1997). These vectors were expanded in 293 cells and purified by double cesium chloride ultracentrifugation. The purified virus was extensively dialyzed against 10 mmol/L Tris per 1 mmol/L of MgCl2 and stored in aliquots at 280°C. The virus titer was determined by spectrophotometry and plaque assay as described (Qian et al., 1997). A single preparation was used for all experiments.
Animals and treatment protocols Male Wistar rats weighting 250 to 275 g (Harlan Interfauna Iberica, S.A. Barcelona, Spain) were maintained in cages at room
temperature under a 12-hr light/dark cycle in an approved biohazard animal room. Standard pelleted chow and drinking water were available ad libitum throughout all experimental protocols. All animal experimentation described in this study was conducted according to an institutionally approved protocol. In order to explore the mechanisms of HSV-tk/GCV liver toxicity and in particular whether this therapy may cause lesions of mtDNA and mitochondrial dysfunction, two separate experiments were done. The aim of experiment 1 was to determine the early appearance of liver toxicity and evolution of the liver changes. For this purpose, rats in the experimental group (n 5 12) received 2 ml of a suspension containing 4 3 1010 plaque-forming units per kilogram (pfu/kg) of Ad.tk, which was slowly injected intravenously into the tail vein. At this moment, an osmotic pump (2ML2 ALZET Criffa S.A. Barcelona, Spain) with GCV solution (Roche S.A., Madrid, Spain) was placed intraperitoneally to deliver the drug at the dose of 30 mg/kg per day. The rats were sacrificed under anesthesia at 3, 7, and 10 days after injection of the vector (n 5 4 in each group). Experiment 2 was designed specifically to assess the mechanism of mitochondrial dysfunction in the liver failure. Rats in the experimental group (n 5 5) received 2 ml of a suspension containing a higher dose of Ad.tk (8 3 1010 pfu/kg) injected intravenously and GCV was given at the same dose and by the same route as in experiment 1. Rats were sacrificed at day 7 after administration of the vector. As control groups we used rats treated with Ad.tk plus saline (administered using an intraperitoneal osmotic pump) or with Ad.lacZ plus GCV (using an intraperitoneal osmotic pump). Adenoviral vectors were given intravenously at the same dose as the corresponding experimental group. We also analyzed rats treated with saline alone (both intravenously and intraperitoneally) or with saline (intravenously) plus GCV (intraperotoneally). Because the latter two groups did not show any biochemical or morphologic changes, we used as controls only the two former groups of animals. To test the efficiency of gene transfer to liver cells, a group of three rats received intravenously Ad.lacZ at 4 3 1010 pfu/kg and three other rats received Ad.lacZ at 8 3 1010 pfu/kg. After 48 hr, livers were removed and X-gal staining was used to detect lacZ gene expression as previously described (Qian et al., 1997).
Biochemistry, light microscopy, and electron microscopy Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were determined in serum, at the time of sacrifice, by automatic colorimetric assays (Technicon RA-1000, Bayer, Tarrytown, NY). Lactic acid was determined using a Boehringer lactic acid kit (Boehringer Mannheim, Mannheim, Germany). At the time of death, liver samples were removed rapidly. The liver was weighed and the tissues were divided and either fixed in 10% buffered neutral formalin or placed in OCT compound, frozen in cold 2-methylbutane (Merck, Darmstadt, Germany) and stored at 280°C. Frozen tissues in liquid nitrogen were used for molecular studies. All staining procedures for light microscopy were performed on 4-mm thick, paraffinembedded sections or on 6–10-mm thick frozen sections. Routine histologic examinations were made for all tissue samples on sections (paraffin and frozen) stained with hematoxylin and
465
HSV-tk/GCV MITOCHONDRIAL LIVER TOXICITY eosin. For oil-red O staining, frozen samples were cut at 10 mm thickness, placed on gelatin-coated slides, and briefly air-dried. Staining was performed on all sections for 20 min using oil-red O in absolute isopropanol diluted 3:2 v/v with distilled water. Sections were rinsed with 60% isopropanol, washed in running distilled water, counterstained with hematoxylin and mounted in gelatin mounting medium. Liver samples were obtained for electron microscopy. Samples were fixed in 0.1 mM phosphate buffer (pH 7.2) containing 2.5% glutaraldehyde for 2 hr, and then postfixed in 0.1 mM phosphate buffer (pH 7.2) containing 2 % osmium tetroxide for a further 2 hr. After embedding the sample blocks in Epon ultrathin sections were cut with the ultramicrotome Ultratome III (Ultratome, Uppsala, Sweden) and then contrasted with uranyl acetate and lead citrate. Tissues were examined with a Philips 300 electron microscopy (Phillips, Amsterdam, Holland).
Measurement of cytochrome C oxidase activity and cytochrome C oxidase subunit levels Cytochrome C oxidase (COX) activity was measured spectrophotometrically as reported elsewhere (Smith et al., 1955). Rat livers were homogenized and sonicated in 10 mM potassium phosphate buffer, pH 7.0. The protein concentration of homogenates was determined by the Bradford assay (Bio-Rad, München, Germany) and used to normalize COX activity. COX was solubilized by adding dodecyl maltoside to the sonicated extracts to a final concentration of 1.5%. Reaction mixtures contained 22.5 mM COX (Sigma Chemical Company, St. Louis, MO), previously reduced with dithiothreitol, in 75 mM potassium phosphate buffer, pH 7.0 and various concentrations of the cell extracts, in a final volume of 1 ml. The rate of oxidation of COX was measured by following the decrease in absorbance at 550 nm. The relative level of the mitochondrially encoded COXII subunit and the nuclear-encoded COX-IV subunit were determined by Western blots. The monoclonal antibodies against COX-II and COX-IV were from Molecular Probes (Eugene, OR) and were devoid of cross-reactivity. Anti-mouse immunoglobulin labeled with peroxidase was purchased from Sigma. Frozen liver tissue was homogenized and sonicated in RIPA buffer (9.1 mmol/L Na2HPO4, 1.7 mmol/L NaH2PO4, 150 mmol/L NaCl, 1.0% Nonidet P-40, and 0.1% sodium dodecyl sulfate (SDS), 0.05% sodium deoxycholate, and 1.1 mmol/L phenylmethylsulfonyl fluoride) with protease inhibitors cocktail tablets (Boehringer Mannheim). A total of 60 mg of protein was dissolved in the sample buffer and analyzed by 12% SDS-polyacrylamide gel electrophoresis. Equal loading of proteins always were confirmed by Ponceau S staining and with the monoclonal antibody glyceraldehyde-3-phosphatedehydrogenase (GAPDH) (Biogenesis Ltd, Poole, England). Immunoreactivity was visualized by incubation with Supersignal CL-HRP chemiluminescence substrate (Pierce Chemical, Rockford, IL).
Isolation of RNA and Northern blot analysis Total RNAs were isolated by acid/guanidine thiocyanate/ phenol/chloroform method (Chomczynski and Sacchi, 1987). RNA concentrations were determined spectrophotometrically. For Northern blotting total RNA (20 mg) was electrophoresed under denaturing conditions, transferred onto positively charged
nylon membranes (Boehringer Mannheim), and fixed by UV cross-linking. The comparative quantity was estimated on gels by ethidium bromide staining. A cDNA insert encoding 230 base pairs of the rat COX-II mitochondrial gene was isolated from rat liver and sequenced. This cDNA probe was used to measure COX-IV mRNA levels. As an internal control the Northern blots were rehybridized with 28S. The probes were labeled by random priming (Amersham Pharmacia Biotech Inc., Buckinghamshire, UK). Membranes were prehybridized for 1 hr and hybridized 2 hr with QuikHyb (Stratagene, La Jolla, CA) at 68°C. The membranes were stringently washed and exposed to Kodak XAR film. The autoradiogram was scanned using a LKB Ultroscan XL laser densitometer and a Gel Scan XL software.
Isolation of total DNA and Southern blot hybridization Total DNA was prepared from approximately 500 mg of liver tissue by standard procedures. In order to quantify mtDNA levels, 5 mm of DNA was digested with PvuII to linearize the mtDNA, electrophoresed in a 0.7% agarose gel, and transferred to a nylon membrane. The membrane was then hybridized simultaneously with two probes, one for the mtDNA (rRNA area) and the other corresponding to nuclear-encoded 18S rRNA. Both probes were synthesized by polymerase chain reaction (PCR) amplification and nonradioactive labeled by the random primer method using digoxigenin-UTP as labeled precursor. Hybridizations and washings were performed at 68°C. Detection of the hybrids was done by using antidigoxigenin antibody conjugated to alkaline phosphatase, which catalyzes a chemiluminescent reaction, following the manufacturer’s instructions (Boehringer Mannheim). To quantify the mtDNA and rDNA signals, the autoradiogram was scanned using a LKB Ultroscan XL laser densitometer and a Gel Scan XL software. The ratio of mtDNA to rDNA was then calculated.
Cell cultures Rat hepatocarcinoma MCA-RH7777 cell line was grown in Dulbecco9s modified Eagle9s medium (DMEM, Gibco BRL, Carlsbad, CA) supplemented with 20% heat-inactivated horse serum (ICN, Aurora, OH) and 5% heat-inactivated fetal calf serum (Gibco). Cultures were performed in a humidified atmosphere of 5% CO2 and 95% air at 37°C.
Incorporation of GCV into mitochondrial and genomic DNA and analysis of mitochondrial function The efficiency of gene transfer to MCA-RH7777 cells using adenovirus vector was tested using a control adenovirus Ad.lacZ at different multiplicities of infection (MOI; 10, 40, and 100) and X-gal staining as previously described (Qian et al., 1997). A total of 2 3 106 MCA-RH7777 cells were plated on 15cm dishes, grown for 36 hr and infected with Ad.tk (40 MOI) and Ad.lacZ as a control (40 MOI). After 24 hr the cells were subsequently treated with 3[H]-GCV (Moravek Biochemical, Brea, CA: Specific activity of 20 Ci/mM) at 0.5 mM/L. Cells were harvested by trypsinization after 24 hr. Mitochondria were prepared according to the protocol published by Enriquez and Attardi (1996). Briefly, cells were suspended in cold isolation
466
HERRAIZ ET AL.
medium (0.32 M sucrose, 1 mM potassium ethylenediaminetetraacetic acid [EDTA], 10 mM Tris-HCl, pH 7.4) and homogenized in a loose-fitting Potter-Elvejhem homogenizer by using three to four up-and-down strokes. The homogenate was centrifuged at 1.000g for 5 min to sediment unbroken cells and nuclei. This cycle was repeated 3 times. The supernatants were divided into microcentrifuge tubes and centrifuged at full speed (13,000g) for 2 min in a microcentrifuge. The pellets were resuspended in 500 ml of isolation medium and were washed 3 times. The number of washes used in this preparation completely eliminates completely any trace of cytoplasmic RNA (estimated by the absence of 28S and 18S rRNA in the electrophoretic patterns of the ethidium bromide-stained gels), making the micrococcal nuclease treatment of the mitochondrial fraction to eliminate contaminating extramitochondrial nucleic acids unnecessary (Enriquez and Attardi, 1996). Mitochondrial and genomic DNA were isolated using the QIAamp Blood Kit (Qiagen, Hilden, Germany) according to the manufacture’s protocol. The DNA was measured by ultraviolet spectrophotometry. The amount of GCV incorporated into mitochondrial and genomic DNA was calculated from the specific radioactivity and specific activity of the GCV. To evaluate the effect of Ad.tk/GCV on mitochondrial function, MCA-H7777 cells were infected with Ad.tk or Ad.lacZ (40 MOI). After 24 hr the cells were subsequently exposed to various concentrations of GCV for 4 days. Cells incubated in drug-free medium were used as a control. The culture medium and drugs were replaced everyday. The level of lactic acid in
the medium after 4 days of treatment with GCV was determined with the Boehringer lactic acid kit.
Statistical analysis Mann-Whitney U test was used to compare variables between groups. A value of p , 0.05 was considered significant.
RESULTS In vivo experiments Biochemical and structural changes in the liver depend on Ad.tk/GCV dose. After an intravenous injection of 4 3 1010 pfu/kg and 8 3 1010 pfu/kg of Ad.lacZ, approximately 50% and 70% of hepatocytes were stained positively with X-gal, respectively. No staining was found in organs other than the liver except for a faint signal in renal proximal tubules in one of the rats tested (data not shown). Rats treated with Ad.tk/GCV at doses of 4 3 1010 pfu/kg (experiment 1) showed a progressive increase in serum transaminases with a parallel reduction in liver weight. Differences in transaminase values with the control group were statistically significant at day 10 (Table 1). Histologic analysis revealed no changes at day 3 and the presence of infiltrating immune cells and acidophilic hepatocytes at day 7. On day 10, some apoptotic bodies were found and all liver sections exhibited moderate microvesicular steatosis (data not shown). Rats treated with
TABLE 1. SERUM ACTIVITIES OF TRANSAMINASES AND LIVER WEIGHT AT DIFFERENT TIME POINTS AFTER ADENOVIRAL ADMINISTRATION IN THE PRESENCE OR ABSENCE OF GANCICLOV IR AST (UI/L) basal values (71 6 11)
Ad.tk (4 3 1010 pfu/kg) 3 days after adenoviral administration 7 days after adenoviral administration 10 days after adenoviral administration Ad.lacZ (4 3 1010 pfu/kg) 3 days after adenoviral administration 7 days after adenoviral administration 10 days after adenoviral administration Ad.tk (8 3 1010 pfu/kg) 7 days after adenoviral administration Ad.lacZ (8 3 1010 pfu/kg) 7 days after adenoviral administration
ALT (UI/L) basal values (35 6 8)
Liver weight (% total body weight)
GCV
Saline
GCV
Saline
GCV
Saline
72 6 10
74 6 12
36 6 60
39 6 40
3.10 6 0.72
3.05 6 0.52
129 6 490
90 6 90
89 6 27
42 6 60
2.92 6 0.70
2.87 6 0.63
*263 6 150*
75 6 11
142 6 85*
56 6 16
2.08 6 0.22
2.93 6 0.41
74 6 80
—
29 6 60
—
3.10 6 0.60
—
0.110 6 36.68
—
0.38 6 6.13
—
3.00 6 0.50
—
00.90 6 10.78
—
54.5 6 11.9
—
3.10 6 0.40
—
*752 6 336*
112 6 200
0414 6 234*
86 6 32
*1.12 6 0.13*
2.68 6 0.23
121 6 300
—
054.5 6 17.80
—
3.00 6 0.30
—
Data are means 6 SD. Experimental group is represented by rats treated with Ad.tk plus GCV. Control groups are represented by rats treated with Ad.tk plus saline and rats treated with Ad.lacZ plus GCV. *p , 0.05 compared with control groups at the same time point. AST, aspartate aminotransferase; ALT, alanine aminotransferase; GCV, ganciclovir; pfu, plaque-forming units; SD, standard deviation.
467
HSV-tk/GCV MITOCHONDRIAL LIVER TOXICITY the same dose of Ad.tk plus saline or with Ad.lacZ plus GCV showed increased levels of transaminases, however there were not statistically differences with respect basal values (Table 1). Histologic studies demonstrated normal liver parenchyma with mononuclear infiltration in portal tracts (data not shown). Rats receiving 8 3 1010 pfu/kg of Ad.tk plus GCV (experiment 2) suffered rapid weight loss and fast deterioration and were sacrificed at day 7. Serum levels of AST and ALT increased dramatically and the livers appeared atrophic and yellowish. Differences with the control group were statistically significant (Table 1). Histologically these livers showed an intense microvesicular and macrovesicular steatosis (confirmed by oil red-O staining), apoptotic bodies and ballooning of hepatocytes (Fig. 1A). At the ultrastructural level we found prominent lipid droplets in the cytoplasm and mitochondrial distortion with swelling and decreased density of the matrix (Fig. 1C and 1D). In control rats treated with 8 3 1010 pfu/kg of Ad.tk plus saline or with the same dose of Ad.lacZ plus GCV, increased levels of transaminases without statistically significant differences with the basal values were observed. Histologic studies revealed a mononuclear infiltrate in the portal tracts. Lactic acid levels, a reflection of mitochondrial function, were markedly elevated in rats treated with high dose of Ad.tk plus GCV (experiment 2) but were only minimally and not significantly increased in those receiving low dose of Ad.tk plus GCV (experiment 1). Lactic acid showed no changes in rats treated with saline instead of GCV (Table 2). Because the two control groups, Ad.tk/saline and Ad.lacZ/GCV, exhibited the same changes in the liver, in subsequent studies we used as controls the Ad.tk/saline-treated rats. Changes in mitochondrial COX by treatment with Ad.tk/GCV. Activity of COX has received particular attention because it reflects abnormalities in mitochondrial enzyme activity. The COX enzyme is a multisubunit complex composed of three large catalytic subunits encoded in the mitochondrial genome and up to 10 smaller regulatory subunits encoded in the nuclear genome. An efficient functioning of the COX requires a tight coordination of the level production of their nuclear and mitochondrial components as well as their assembly. The activity of COX in the liver extracts showed a marked decrease at day 10 in rats from experiment 1 (receiving 4 3 1010 pfu/kg of Ad.tk plus GCV) and even more in rats from experiment 2, which received a higher dose of the adenoviral vector (40% and 70%, respectively, compared to normal livers). Those differences were statistically significant compared with rats from the control groups sacrificed at the same days (Fig. 2A). Livers from Ad.tk/salinetreated rats (Fig. 2A) or Ad.lacZ/GCV or saline/GCV-treated rats (data not presented) showed no differences as compared to normal healthy livers. To investigate whether the decline of COX activity depends on the relative level of the mitochondrially encoded COX-II subunit and/or nuclear-encoded COX-IV subunit we analyzed these two subunits by Western blot. A significant reduction of COX-II subunit content was observed in rats from experiment 1 treated with 4 3 1010 pfu/kg of Ad.tk plus GCV at day 10 after vector administration and in rats from experiment 2 treated with 8 3 1010 pfu/kg of Ad.tk plus GCV. No changes in the levels of nuclear-encoded COX-IV subunit were observed in any of the Ad.tk/GCV-treated rats (Fig. 2B). All rats from the control groups showed no alterations in the levels of COX subunits (data not shown).
A
B
C
D
FIG. 1. Histologic analysis of the livers after high doses of Ad.tk (8 3 1010 plaque-forming units per kilogram [pfu/kg]) plus ganciclovir (GCV). A: Representative liver section showing a severe microsteatosis and macrosteatosis, presence of apoptotic bodies, and ballooning of hepatocytes (hematoxylin and eosin [H&E] 3 200). B: Electron micrograph of control liver tissue. C and D: Electron micrograph of liver from Adtk/GCV–treated rats showing the presence of lipids, variable sized mitochondria, and distortion with decreased density and absence of dense matrical granules.
468
HERRAIZ ET AL. TABLE 2. LACTIC ACID LEVELS AT DIFFERENT TIME POINTS AFTER Ad.tk ADMINISTRATION IN THE PRESENCE OR ABSENCE OF GANCICLOV IR Lactic acid (mg/dl) basal values (253 6 24)
Ad.tk (4 3 1010 pfu/kg) 3 days after adenoviral administration 7 days after adenoviral administration 10 days after adenoviral administration Ad.tk (8 3 1010 pfu/kg) 7 days after adenoviral administration
GCV
Saline
253 6 28*
247 6 32
268 6 11*
259 6 21
279 6 22*
257 6 18
368 6 11*
258 6 19
Data are means 6 SD. Experimental group is represented by rats treated with Ad.tk plus GCV. Control group is represented by Ad.tk/saline-treated rats. *p , 0.05 compared with control group at the same time point. GCV, ganciclovir.
Changes in mitochondrial mRNA and DNA levels induced by Ad.tk/GCV. mtDNA and COX-II mRNA levels were measured in order to determine if the observed changes in the activity of proteins encoded by mtDNA was because of changes in mitochondrial transcription. We found no differences in the relative level of COX-IV mRNA between rats treated with Ad.tk/GCV at different time points with low and high vector doses compared with control rats (Ad.tk/saline) sacrificed at the same days (data not shown). To examine the quantitative changes of mtDNA in rats receiving Ad.tk/GCV, we performed Southern blot analysis of DNA isolated from the liver tissue. We found that the relative mtDNA content in the liver from rats receiving the highest dose of Ad.tk (experiment 2) was significantly decreased compared to healthy livers (Fig. 3A and 3B). A slight reduction of mtDNA was observed at day 10 in rats from experiment 1 receiving a lower dose of the vector (Fig. 3A and 3B). All rats from the control groups (treated with the same Ad.tk dose as the experimental groups but saline instead of GCV) sacrificed at the same days as rats from experimental groups showed identical relative mtDNA content as found in healthy normal livers (data not shown).
In vitro experiments Incorporation of GCV into mtDNA and nuclear DNA and analysis of mitochondrial function. Ad.lacZ was able to infect MCA-RH 7777 cell line. The ratio of infected cells at 100, 40, and 10 MOI was 90%, 40%, and 10%, respectively. To evaluate the interaction of GCV with nuclear and mtDNA, a confluent rat hepatocarcinoma cell line was infected with Ad.tk at 40 MOI and exposed for 1 day to radiolabeled 3[H]-GCV at the final concentration of 0.5 mM. Cells infected with Ad.lacZ at 40 MOI and incubated with the same amount of 3[H]-GCV were used as a control. Both nuclear and mtDNA were isolated, purified, and analyzed as described in Materials and Methods. In the cells treated with Ad.tk, we found a significant incorporation of GCV into both nuclear and mtDNA, although the incorporation
to the mitochondrial genome was 100-fold lower compared with the incorporation into the nuclear genome (Fig. 4A). To observe the effect of GCV on mitochondrial function, the lactic acid content in hepatoma cell culture medium was determined. Figure 4B demonstrates that GCV induced increased lactic acid production in Ad.tk-infected cells in a dose-dependent manner. After 4 days of incubation, enhancement in lactic acid production was detected at concentrations of 0.1 mM and predominately 10 mM. In contrast, under similar conditions, the addition of GCV to Ad.lacZ-infected cells at concentrations comprised between 0.1 and 10 mM had no substantial effect on lactic acid production.
DISCUSSION This study shows that the HSV-tk/GCV system, which has been proposed as a potential gene therapy procedure to treat cancer, may cause injury not only to dividing malignant cells but also to normal quiescent cells such as hepatocytes. Here we provide data indicating that this hepatotoxic reaction is at least partly the result of severe mitochondrial damage. We found that microvesicular steatosis was a distinctive histopathologic feature in all animals treated with Ad.tk plus GCV, particularly in rats treated with high doses of the vector. This finding, associated with lactic acidosis, was suggestive of mitochondrial derangement and was verified at the ultrastructural level by the presence of marked changes in the shape, cristae, and matrix of mitochondria. All these features have been shown to occur with long-term use of nucleoside analogues such as zidovudine (Chariot et al., 1999), didanosine (Bissuel et al., 1994), stavudine (Saboulard et al., 1999), and fialuridine (Lewis et al., 1997). The activity of the COX complex is dependent on the coordinated synthesis of COX subunits encoded by the nuclear genome (COX-IV subunit) and by mtDNA (COX-II subunit). We found a significant dose-dependent reduction of COX ac-
469
HSV-tk/GCV MITOCHONDRIAL LIVER TOXICITY
(% of healthy livers)
COX activity
A) 120 100
*
80 60
*
40 20 0
GCV
_
_
+
_
_
+
4x1010
Ad.tk doses (pfu/Kg) Time after administration of Ad.tk (days)
+
3
7
+
8x1010
10
7
B) Ad.tk doses (pfu/Kg)
4x1010
8x1010
+ GCV Time after administration of Ad.tk (days)
3
7
10
7
COXII subunit GAPDH COXIV subunit GAPDH FIG. 2. Cytochrome C oxidase (COX) activity and steady-state levels of the mitochondrially-encoded subunit 2 (COX-II) and the nuclear-encoded subunit 4 (COX-IV) of COX in livers from Ad.tk/GCV–treated rats. A: Data are presented as the percent of healthy liver activity 1 standard deviation (SD). Samples obtained at day 10 after administration of Ad.tk 4 3 1010 plaque-forming units per kilogram (pfu/kg) plus ganciclovir (GCV) and at day 7 with higher doses showed a significant decrease compared to control group (Ad.tk at the same doses plus saline). * p , 0.05. B: Representative Western blots from rats at different time points (3, 7, and 10 days) after the treatment with Ad.tk/GCV at doses 4 3 1010 pfu/kg and 7 days after higher doses (8 3 1010 pfu/kg). tivity (at day 10 in experiment 1 and at day 7 in experiment 2) in animals that received Ad.tk plus GCV. This decrease was associated with a selective depletion of COX-II protein while the levels of COX-IV remained essentially unchanged. This interesting finding appears to indicate that the toxic effects of the phosphorylated GCV derivatives affect more readily the ex-
pression of mitochondrial genes than the products encoded by nuclear DNA. Similar differential effects on nuclear and mitochondrial-encoded COX subunits have been previously described for a typical mitochondrial toxin such as ethanol (Coleman et al., 1994), and also it has been described in treatment with zidovudine (McCurdy and Kennedy, 1996). In our study
470
HERRAIZ ET AL.
A) Ad.tk doses (pfu/Kg) + GCV Time after administration of Ad.tk (days)
4x1010 _
3
8x1010
7
10
7 mtDNA
nDNA
B) (% of control)
mt/DNA/nDNA
400 300 200 100
*
0
Time after administration of Ad.tk (days) Ad.tk doses (pfu/Kg) + GCV
_
3
7 4x1010
10
7 8x1010
FIG. 3. Southern blot hybridization analysis of total DNA extracted from Ad.tk/GCV–treated rat livers. A: Representative Southern blot from animals treated with different doses of Ad.tk. The intensity of hybridization to the mitochondrial DNA (mtDNA) probe (top bands) corresponds to the quantity of mtDNA present in each sample. The nuclear DNA (nDNA) signals (bottom bands) serve as an internal control for the amount of total DNA loaded in each lane. B: Blot intensities were quantified by densitometry analysis. The mtDNA/nDNA hybridization ratio was expressed as the percentage with respect to healthy livers (2) set at 100%. Data are the mean 6 standard deviation (SD) from all the rats analyzed (n 5 4 in experiment 1 and n 5 5 in experiment 2; see Material and Methods). * Significant difference (p , 0.01) between experimental group and healthy livers.
the fact that the decreased COX-II protein appears in the presence of unmodified COX-II mRNA levels suggests a derangement of mitochondrial transductional processes rather than alteration of mitochondrial transcription. Again, our data are in agreement with previous reports on embryonic chick hearts exposed to ethanol in ova (Kennedy et al., 1993) and with observations of rat livers after ethanol administration (Coleman et al., 1994). In both cases the reduction of the mitochondrial-encoded COX subunits was present without changes in their corresponding mRNAs.
Interestingly, when we analyzed the ratio of mtDNA to nuclear DNA in the livers of rats treated with Ad.tk/GCV we found that those that received 4 3 1010 pfu of adenoviral vector showed an increase of relative mtDNA content at day 3 with reversion to normal values at day 7 after therapy. The mechanism responsible for the early increase of mtDNA copy number after Ad.tk/GCV administration is not known and has not been explored in the present paper. This phenomenon could be the result of stimulation of mtDNA polymerase g or of reduced degradation of mtDNA at this early step. The equilibrium be-
471
HSV-tk/GCV MITOCHONDRIAL LIVER TOXICITY
B) 500 Ganciclovir (mM/L)
fmol / mgr DNA
A)
60 20
10 Ad.tk/GCV
1
Ad.LacZ/GCV
0.1 0 100
Ad.tk
Ad.lacZ Ad.tk Ad.lacZ
Nuclear DNA
Mitochondrial DNA
200
300
Lactic acid levels (% of control)
FIG. 4. Effect of ganciclovir on rat hepatocytic cell line (MCA-RH7777). A: Incorporation of radiolabeled ganciclovir (GCV) into genomic and mitochondrial DNA (mtDNA). Cells were infected with Ad.tk and Ad.lacZ and exposed to tritium-labeled GCV (0.5 mM) for 24 hr. Genomic and mtDNA fractions were collected and quantified. B: Levels of lactic acid in the medium of hepatocytic cells infected with Ad.tk or Ad.lacZ and treated with various concentrations of GCV for 4 days. The lactic acid levels were compared with those in noninfected cells incubated in GCV-free medium. Data represent the means and standard deviation from three experiments.
tween mtDNA synthesis and degradation may change in the following days, explaining return to normal levels or to values slightly below normal at day 10. Because mtDNA replication has been shown to occur during the initial phase of liver regeneration (Koyama et al., 1998) it could be speculated that the injury inflicted by Ad.tk/GCV can activate mtDNA synthesis causing an increase of mtDNA as a preparation for cell proliferation. Our obseration is in keeping with other published reports showing that toxic cell injury of moderate intensity stimulates mitochondrial DNA polymerase g as in the case of ethanol, which enhances this enzyme in myocardial cells (Kennedy, 1998). When hepatocellular damage is too intense, mitochondrial integrity is severely affected, and in fact, one of the most relevant findings of our study was the presence of a marked depletion of mtDNA in rats treated with Ad.tk plus GCV. These changes reflect the profound genotoxic effect exerted by phosphorylated GCV on mitochondrial genome. To analyze this effect further we performed in vitro experiments to determine whether GCV could be incorporated into mtDNA. The use of primary rat hepatocytes were considered; however, the time required in this experimental approach is longer than 48 hr. It has been reported that primary hepatocytes maintained in monolayer culture during the first 24–48 hr are highly dedifferentiated and they exhibit a substantial loss of carriers, transporters, and different enzyme activities (Van Dyke et al., 1982), making experimentation with primary hepatocytes unfeasible. The fact that hepatoma cell lines retain a number of biochemical synthetic pathways typical of hepatocytes has made these cells very useful in toxicology, especially to study the incorporation of nucleoside analogues in the mitochondria of liver cells (Cui et al., 1995). Moreover, we found that GCV increased levels of lactic acid in hepatoma cells expressing HSV-tk, the same effect than GCV in vivo. Thus, hepatoma cells reproduce in vitro the toxicologic phenomena observed in rats and therefore provide useful data regarding the mechanisms underlying the toxicologic effects. Thus, we used rat hepatocarcinoma cell line, and showed for the first time that GCV is incorporated into mtDNA. In absolute terms the incorporation
is 10 times less in mtDNA than in nuclear DNA but taking into consideration the relative proportions of nuclear versus mtDNA, it seems clear that phosphorylated GCV is incorporated with high efficiency into mtDNA, thus explaining the sensitivity of mitochondria to the toxic effect of this compound. The potential role of GCV incorporation into mtDNA in mediating toxicity is further suggested by its effects on mitochondrial function as it has been reported with other nucleoside analogues in liver cells (Cui et al., 1995). The increased lactic acid content in the Ad.tk-infected cell medium after 4 days exposure to GCV was a phenomenon consistent with the in vivo observation in the Ad.tk/GCV-treated rats. Both the insertion of GCV into mtDNA and the described inhibition of mitochondrial DNA polymerase (Martin et al., 1994) might account for the marked depletion of mtDNA and the mitochondrial dysfunction observed in rats treated with high doses of Ad.tk/GCV. Toxicities associated with drugs that induce selective depletion of mtDNA typically have a delayed onset (Lewis and Dalakas, 1995; Brinkman et al., 1998). It has been proposed that this delay is because of the time required for mtDNA copy number to drop below some critical threshold level where mitochondrial function is compromised (Brinkman et al., 1998). In our model, livers treated with Ad.tk plus GCV exhibited mtDNA depletion very early (as soon as day 7 in experiment 2) after the administration of the vector. This fast deleterious effect indicates a great sensitivity of mitochondria to the toxic effects of phosphorylated GCV. The fact that mitochondria exhibit enhanced respiratory function and an increase in their endogenous thymidine kinase activity during the early phases of liver regeneration (Koyama et al., 1998; Elholm et al., 2001) might explain the high sensitivity to GCV of mitochondria in damaged livers. In summary our data show that gene transfer of HSV-tk followed by GCV administration causes damage to normal quiescent hepatocytes and that mitochondrial genotoxicity is an important contributory factor to this adverse effect. These data are of relevance for the design of clinical trials using this gene therapy strategy to combat cancer.
472
HERRAIZ ET AL.
ACKNOWLEDGMENTS This work was supported by grants from Ramón Areces Foundation, Spanish Ministerio de Ciencia y Tecnología BCM 2001-2421, and Diputación General de Aragón (P032-2000). A.S. is recipient of a fellowship from CONACYT (México). N.B. is supported by FIS (grant 01/0723 to M.B.), Red Nacional de Investigacion en Gastroenterologia y Hepatologia.
REFERENCES BISSUEL, F., BRUNEEL, F., HABERSETZER, F., CHASSARD, D., COTTE, L., CHEVALLIER, M., BERNUAU, J., LUCET, J.C., and TREPO, C. (1994). Fulminant hepatitis with severe lactate acidosis in HIV-infected patients on didanosine therapy. J. Intern. Med. 235, 367–371. BRAND, K., ARNOLD, W., BARTELS, T., LIEBER, A., KAY, M.A., STRAUSS, M., and DORKEN, B. (1997). Liver-associated toxicity of the HSV-tk/GCV approach and adenoviral vectors. Cancer Gene Ther. 4, 9–16. BRINKMAN, K., TER HOFSTEDE, H.J., BURGER, D.M., SMEITINK, J.A., and KOOPMANS P.P. (1998). Adverse effects of reverse transcriptase inhibitors: mitochondrial toxicity as common pathway. AIDS 12, 1735–1744. BUSTOS, M., SANGRO, B., ALZUGUREN, P., GIL, A.G., RUIZ, J., BERAZA, N., QIAN, C., GARCIA PARDO, A., and PRIETO J. (2000). Liver damage using suicide genes. A model for oval cell activation. Am. J. Pathol. 157, 549–559. CHARIOT, P., DROGOU, I., DE LACROIX SZMANIA, I., ELIEZER VANEROT, M.C., CHAZAUD, B., LOMBES, A., SCHAEFFER, A., and ZAFRANI, E.S. (1999). Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis, and mitochondrial DNA depletion. J. Hepatol. 30, 156–160. CHOMCZYNSKI, P., and SACCHI, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156–159. COLEMAN, W.B., CAHILL, A., IVESTER, P., and CUNNINGHAM C.C. (1994). Differential effects of ethanol consumption on synthesis of cytoplasmic and mitochondrial encoded subunits of the ATP synthase. Alcohol Clin. Exp. Res. 18, 947–950. CUI, L., YOON, S., SCHINAZI, R.F., and SOMMADOSSI, J.P. (1995). Cellular and molecular events leading to mitochondrial toxicity of 1-(2-deoxy-2-fluoro-1-b-D-arabinofuranosyl)-5-iodouracil in human liver cells. J. Clin. Invest. 95, 555–563. ELHOLM, M., HOLLAS, H., ISSALENE, C., BARROSO, J.F. BERGE, R.K., and FLATMARK T. (2001). Transient up-regulation of liver mitochondrial thymidine kinase activity in proliferating mitochondria. IUBMB Life. 51, 99–104. ENRIQUEZ, J.A., and ATTARDI, G. (1996). Analysis of aminoacylation of human mitochondrial tRNAs. Methods Enzymol. 264, 183–196. ILSLEY, D.D., LEE, S.S., MILLER, W.H., and KUCHTA, R.D. (1995). Acyclic guanosine analogs inhibit DNA polymerases alpha, delta, and epsilon with very different potencies and have unique mechanisms of action. Biochemistry 34, 2504–2510. KENNEDY, J.M. (1998). Mitochondrial gene expression is impaired by ethanol exposure in cultured chick cardiac myocytes. Cardiovasc. Res. 37, 141–150. KENNEDY, J.M., KELLEY, S.W., and MEEHAN J.M. (1993). Ventricular mitochondrial gene expression during development and following embryonic ethanol exposure. J. Mol. Cell. Cardiol. 25, 117–131. KOYAMA, H., KUROKAWA, T., NONAMI, T., NAKAO, T., SUGIYAMA, S., and MURAKAMI, T. (1998). Increases in the mitochondrial DNA replication and transcription in the remnant liver of rats. Biochem. Biophys. Res. Commun. 243, 858–861.
LEWIS, W., and DALAKAS, M.C. (1995). Mitochondrial toxicity of antiviral drugs. Nat. Med. 1, 417–422. LEWIS, W., GRINIUVIENE, B., TANKERSLEY, K.O., LEVINE, E.S., MONTIONE, R., ENGELMAN, L., DE COURTEN MYERS, G., ASCENZI, M.A., HORNBUCKLE, W.E., GERIN, J.L., and TENNANT, B.C. (1997). Depletion of mitochondrial DNA, destruction of mitochondria, and accumulation of lipid droplets result from fialuridine treatment in woodchucks (Marmota monax). Lab. Invest. 76, 77–87. MARTIN, J.L., BROWN, C.E., MATTHEWS DAVIS, N., and REARDON J.E. (1994). Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis. Antimicrob. Agents Chemother. 38, 2743–2749. MATTHEWS, T., and BOEHME, R. (1988). Antiviral activity and mechanism of action of ganciclovir. Rev. Infect. Dis. 10(Suppl 3), S490–S494. MCCURDY, D.T., and KENNEDY, J.M. (1996). Skeletal muscle mitochondria from AZT-treated rats have a diminished response to chronic electrical stimulation. J. Appl. Physiol. 81, 326–334. MOOLTEN, F.L. (1986). Tumor chemosensitivity conferred by inserted herpes thymidine kinase genes: paradigm for a prospective cancer control strategy. Cancer Res. 46, 5276–5281. QIAN, C., IDOATE, M., BILBAO, R., SANGRO, B., BRUNA, O., VAZQUEZ, J. and PRIETO, J. (1997). Gene transfer and therapy with adenoviral vector in rats with diethylnitrosamine-induced hepatocellular carcinoma. Hum. Gene Ther. 8, 349–358. SABOULARD, D., NAESENS, L., CAHARD, D., SALGADO, A., PATHIRANA, R., VELAZQUEZ, S., MCGUIGAN, C., DE CLERCQ, E., and BALZARINI, J. (1999). Characterization of the activation pathway of phosphoramidate triester prodrugs of stavudine and zidovudine. Mol. Pharmacol. 56, 693–704. SMITH, L. (1955). Spectophotometric assay of cytochrome C oxidase. Methods Biochem. Anal. 2, 427–434. THUST, R., TOMICIC, M., KLOCKING, R., VOUTILAINEN, N., WUTZLER, P., and KAINA B. (2000). Comparison of the genotoxic and apoptosis-inducing properties of ganciclovir and penciclovir in Chinese hamster ovary cells transfected with the thymidine kinase gene of herpes simplex virus-1: Implications for gene therapeutic approaches. Cancer Gene Ther. 7, 107–117. VAN DER EB, M.M., CRAMER, S.J., VERGOUWE, Y., SCHAGEN, F.H., VAN KRIEKEN, J.H., VAN DER EB, A.J., RINKES, I.H., VAN DE VELDE, C.J. and HOEBEN, R.C. (1998). Severe hepatic dysfunction after adenovirus-mediated transfer of the herpes simplex virus thymidine kinase gene and ganciclovir administration. Gene Ther. 5, 451–458. VAN DYKE, RW., STEPHENS, JE., and SCHARSCHMIDT, BF. (1982). Bile acid transport in cultured rat hepatocytes. Am. J. Physiol. 243, G484–G492. WUTZLER, P., and THUST, R. (2001). Genetic risks of antiviral nucleoside analogues: A survey. Antiviral Res. 49, 55–74.
Address reprint requests to: Matilde Bustos Internal Medicine Medical School University of Navarra Pamplona, 31008 (Spain) E-mail:
[email protected] Received for publication March 20, 2002; accepted after revision February 26, 2003. Published online: March 14, 2003.
