Many pathological conditions are induced by chemical exposures, of which an increase in free radical production has been shown to be a primary and ubiquitous event in producing toxicity¹. Oxidative stress or chemical insults that are capable of generating DNA strand breaks cause a remarkable increase in poly(ADP-ribose) polymerase-1 (PARP-1; EC 2.4.2.30) activity², whose activity is barely detectable under normal cellular conditions. When DNA damage is moderate, PARP-1 works for its repair. Upon binding to DNA strand termini, PARP-1 catalyzes the splitting of the ADP-ribose-nicotinamide bond in β-nicotinamide adenine dinucleotide (NAD⁺) with the concomitant attachment of ADP-ribose moiety to protein and then to this protein-bound ADP-ribose, resulting in a long and branched chain of up to 200 units, i.e. poly(ADP-ribose)³,⁴. Various proteins serve as acceptors of ADP-ribose. In addition, PARP-1 catalyzes the transfer of ADP-ribose moieties to acceptor sites on itself, a process called “automodification”³,⁴. This automodification leads to the release of PARP-1 from DNA strand termini, due to the negative charge repulsion between poly(ADP-ribose) and DNA, and allows for other proteins of the DNA excision repair pathways to access and work at sites of damage.

Overactivation of PARP-1 potentiates the deleterious effects of an initial chemical insult by rapidly depleting cellular NAD⁺ and adenosine triphosphate (ATP) and leads to an “energy crisis” that culminates in cell death^5-10. In such situations, PARP-1 inhibitors very often protect cells from death⁸,⁹,^11-15. It seems that PARP-1 is important for both, apoptotic and necrotic, modes of cell death, but in different ways^16-19. DNA-damage-induced PARP-1 overactivation leads to ATP depletion and necrotic cell death¹⁰,²⁰,²¹. Consequently, PARP-1 inhibition attenuates or prevents necrosis in various cell types¹¹,¹²,^21-24. During classical apoptosis (i.e., caspase-dependent programmed cell death), PARP-1 is cleaved by caspases resulting in its inactivation²⁵. PARP-1 cleavage, a hallmark of apoptosis, may prevent its overactivation and, therefore, ATP depletion and necrosis by preserving cellular energy that is required for apoptosis. Thus, intracellular ATP levels can regulate the mode of cell death²⁶,²⁷. A similar effect would be expected from PARP-1 inhibition. In fact, inhibition of PARP-1 preserves cellular ATP levels and in turn minimizes execution of the necrotic death pathway²⁸. Moreover, the maintenance of intracellular ATP levels associated with PARP-1 inhibition shifts cell death from necrosis to apoptosis and from apoptosis to cell survival²⁹. PARP-1 inhibitors (3-aminobenzamide and 4-hydroxyquinazoline) prevent hydrogen peroxide H₂O₂-induced ATP depletion with reversion of the mode of cell death from necrosis back to apoptosis³⁰. 3-Aminobenzamide restores NAD⁺ and ATP levels and decreases both necrosis and apoptosis caused by H₂O₂-injury of PC12 cells³¹. PARP-1 also plays an important role in a caspase-independent programmed cell death that is mediated by the apoptosis-inducing factor, i.e. a mitochondrial flavoprotein that is released and translocated to the nucleus in response to death stimuli³²,³³. PARP-1 activation is necessary for this translocation. Furthermore, during excessive poly(ADP-ribosyl)ation-dependent cell death, there is a rapid cross-talk between the nucleus and mitochondria³⁴,³⁵. Some studies have shown that pharmacological inhibition of PARP-1 is capable of attenuating the adverse effects of selected pharmaceuticals³⁶.

Several lines of evidence have shown that the toxicity of carbon tetrachloride (CCl₄; Chemical Abstracts Service Registry Number: 56-23-5) is dependent on an early efflux of calcium (Ca²⁺) from the endoplasmic reticulum, due to intense lipid peroxidation caused by the trichloromethyl radical³⁷,³⁸. Late preventive effects of nicotinamide and dimethyl sulfoxide (DMSO) against CCl₄- and bromobenzene-induced hepatotoxicity have already been reported³⁹,⁴². These effects were assumed to be the result of the ability of nicotinamide treatment to restore mitochondrial Ca²⁺ transport and the antioxidant effects of DMSO. More recently, we showed that increased ADP-ribosylation of hepatocellular proteins occurred in mice 24 h after the administration of a hepatotoxic dose of CCl₄⁴³. Further, we found that 6(5H)-phenanthridinone (Chemical Abstracts Service Registry Number: 1015-89-0), a potent inhibitor of PARP-1 (IC₅₀ = 300 nM)⁴⁴, offered a protective affect against CCl₄-induced hepatotoxicity by decreasing areas of hepatocellular necrosis and serum transaminase levels⁴⁵. Herein, we report additional parameters from these previous studies that further support the protective effects of PARP-1 inhibition in animals poisoned with CCl₄.

This study was undertaken to further evaluate the protective effect of PARP-1 inhibition on CCl₄-induced hepatotoxicity⁴⁹. The ability of DMSO to intervene in the development of bromobenzene- and chloroform-induced hepatotoxicity has been shown when animals are treated up to 24 h after administration of the toxicant. This effect would be independent of the bioactivation of CCl₄ to the trichloromethyl radical, which occurs between 6 and 10 h after treatment⁴⁹. The in vivo concentration of DMSO utilized in the aforementioned studies was between 2.5 and 3.2% based on the average mouse blood volume of 6–8 mL per 100 g of bodyweight⁴⁶. Interestingly, DMSO has been shown to cause a 20% reduction in PARP-1 activity at a concentration of 4% in vitro⁵⁰. Therefore, it was hypothesized that the late protective effects of DMSO may occur, in part, by preventing the overactivation of PARP-1, thereby allowing partially damaged cells to overcome the initial insult through reparative processes or facilitating apoptosis in excessively damaged cells.

Many of the known PARP-1 inhibitors are water insoluble, and organic solvents must be used as the vehicles for administration. DMSO is the most commonly used⁵⁰. Previous studies have shown that DMSO can increase the acute lethality of CCl₄⁵¹, but decrease its hepatotoxicity⁵². Furthermore, several studies have shown that in human hepatocytes DMSO can alter the activity of various isoforms of cytochrome P450 (CYP), including CYP2E1 that bioactivates CCl₄ to the trichloromethyl radical⁵³; however, an in vitro inhibitory effect was observed only at concentrations greater than 0.1% DMSO. Further, an in vivo concentration of DMSO at 1.66% or lower does not provide any protection against bromobenzene- or chloroform-induced hepatotoxicity, both of which are bioactivated by CYP2E1^54-56. As presented herein, the administered amount of DMSO (<0.65% in vivo) was maintained sufficiently low enough so as to prevent it from altering the hepatotoxicity of CCl₄.

An important consideration in interpreting our previous findings and the results presented herein is which of the inhibitory and antioxidant effect of DMSO or the inhibitory effect of 6(5H)-phenanthridinone on the bioactivation of CCl₄ via CYP2E1 dominated. Since 6(5H)-phenanthridinone is predominantly eliminated as a glucuronide conjugate and is not further metabolized in benzo[a]pyrene-induced animals⁵⁷,⁵⁸, our results suggest that 6(5H)-phenanthridinone protects against CCl₄-induced hepatotoxicity independently of its metabolism.

Recently, Cover et al.⁵⁹ suggested that PARP-1 activation might not contribute to acetaminophen-induced cell death under the experimental conditions used in their study. In the report, two PARP-1 inhibitors were tested and gave mixed results. 3-Aminobenzamide completely protected against acetaminophen hepatotoxicity, whereas 5-aminoisoquinolinone lacked protective effects. The authors hypothesized that 3-aminobenzamide might reduce metabolic activation of acetaminophen or act as an antioxidant, and, in view of the results with the more potent inhibitor 5-aminoisoquinoline (IC₅₀=10 μM), concluded that PARP-1 activation was not a relevant event for acetaminophen-induced oncotic necrosis. Based on their results, it is possible that the 5-aminoisoquinoline used in the study was partially degraded or rapidly metabolized when administered by the intraperitoneal route, as the administered dose did not appear to inhibit PARP-1 to the degree that such a potent inhibitor should inhibit. For instance, the authors report that the staining intensity of poly(ADP-ribose)-positive cells was not significantly reduced in animals treated with acetaminophen and 5-aminoisoquinoline.

Alternatively, the discrepancy between the results of Cover et al.⁵⁹ and the results presented herein may be due to the different pathways by which acetaminophen and CCl₄ cause hepatotoxicity. For instance, the metabolic activation of acetaminophen generates a reactive metabolite, N-acetyl-p-benzoquinone imine, that covalently binds to cellular proteins, whereas the bioactivation of CCl₄ to the trichloromethyl free radical results in lipid peroxidation. Wan et al.⁶⁰ showed that metabolic activation of acetaminophen resulted in an increase in oxidative DNA damage, due, in part, to significantly reduced levels of 8-oxoguanosine DNA glycosylase, a base-excision repair enzyme that removes oxidatively modified DNA bases. In comparison, the types of DNA lesions, commonly called ethenobases, resulting from CCl₄ are predominantly those originating from products of lipid peroxidation (e.g. malondialdehyde and 4-hydroxynonenol) and not an increase in oxo⁸dG, as shown herein. However, the ethenobases are repaired via the base-excision repair pathway, specifically alkyl-N-purine-DNA glycosylase (ANPG). The recent findings of Ogawa et al.⁶¹ indicate that differences exist with regard to the protective effects of PARP-1 inhibition and the type of DNA damage to be repaired via the nucleotide-excision repair pathway or the base-excision repair pathway, with inhibitors conferring a protective effect in the latter case. It is possible that inactivation of Ogg1 during acetaminophen intoxication prevents the activation of PARP-1, given that glycosylases initiate base-excision repair by removing the altered DNA base, followed by excision of the sugar-phosphate backbone, which triggers the activation of PARP-1. Since CCl₄ generates lipid peroxidation with the subsequent formation of ethenobases, the protective effect of PARP-1 inhibition on CCl₄ intoxication may be due to a possible increase in ethenobase removal by ANPG, which, in the absence of PARP-1 inhibition, may lead to PARP-1 overactivation. If the activity of ANPG is inducible as with other glycosylases, the enhanced removal of ethenobases may ultimately lead to overactivation of PARP-1 following excision of the sugar-phosphate backbone and DNA strand break formation⁶². Finally, the degree to which a parent compound and its metabolites may or may not alter the activity of PARP-1 remains an area of investigation that has not been readily explored. In studies where high doses of chemicals are used to treat animals, as performed in this study, saturation of metabolic pathways may end up providing some degree of protection if the parent compound inhibits PARP-1. Some insight into such phenomena was recently reported with several different types of heterocyclic amines, many of which inhibited PARP-1 in their unmetabolized form; however, some compounds caused a significant increase in PARP-1 activity⁶².

In summary, our data demonstrated that 6(5H)-phenanthridinone treatment attenuated CCl₄-induced hepatotoxicity. This effect does not appear to be a result of the use of DMSO as a carrier because the DMSO + CCl₄ produced intense centrilobular necrosis. Since 6(5H)-phenanthridinone is predominantly eliminated as a glucuronide conjugate and is not further metabolized in control or benzo[a]pyrene-induced animals [57, 58], our results suggest that 6(5H)-phenanthridinone protects against CCl₄-induced centrilobular hepatotoxicity without altering CCl₄ bioactivation. Moreover, when viewed in toto with our previous findings, these results show the predominant role of PARP-1 overactivation in chemical-induced hepatotoxicity and strongly suggest the possibility of pharmacological intervention with PARP-1 inhibitors for chemical-induced hepatotoxicity and possibly other cytotoxicities.

This work was supported in part by a research fellowship from the Japan Society for the Promotion of Science (M.B., T.S., and R.P.S.). The author report no conflicts of interest.