Di Luzio and coworkers first proposed a role for lipid peroxidation in liver injury resulting from acute and chronic alcohol administration (Di Luzio 1963, Di Luzio and Hartman 1967). Lipid peroxidation is a process that involves free radical intermediates, and it has the potential to damage cells through either destruction of membranes or attack of radicals on other critical cellular macromolecules. Many laboratories have confirmed the initial observations that alcohol initiates hepatic lipid peroxidation, and this evidence has been recently reviewed (Nordman et al. 1992). However, because of the reactive nature of most free radicals, it is difficult to detect these intermediates directly.

Spin trapping is a method in which radicals are allowed to add across the double bonds of chemical trapping agents, thereby forming a more stable secondary radical commonly referred to as a spin adduct. This spin adduct then can be subjected to electron paramagnetic resonance (EPR)⁴ spectroscopy, in which the unpaired electron interacts with other atoms in the molecule to produce a characteristic EPR spectrum. Thus, spin trapping provides an opportunity to not only detect radicals but also to partially characterize the types of radicals formed. Other information about the spin trapping method and its applications in biology can be obtained from a recent review (Knecht and Mason 1993).

Our group was the first to use the spin trapping method for detection of free radicals in livers of alcohol-fed rats (Reinke et al. 1987). Furthermore, the EPR signals of the presumed lipid radicals detected were most intense when alcohol was fed in combination with high fat diets. Subsequent investigations have further characterized hepatic radical intermediates formed after acute alcohol administration (Reinke et al. 1991), and more recently we have developed methods to reproducibly detect the free radical metabolite of ethanol, the 1-hydroxyethyl radical (HER), in livers of anesthetized rats (Moore et al. 1995). In the current report, we employed these improved methods to reexamine effects of chronic alcohol administration on HER formation in vivo, as well as effects of dietary fat on alcohol-associated free radical formation.

MATERIALS AND METHODS

Spin trapping experiments were conducted as described in detail previously (Moore et al. 1995). Briefly, rats were anesthetized with isoflurane (2%), and a cannula of PE-10 tubing was placed into the common bile duct. After the rats were given -(4-pyridyl-1-oxide)-N-t-butylnitrone (POBN, 700 mg/kg, intravenously), bile was collected over sequential 10-min intervals into microcentrifuge tubes containing 30 µL of dipyridyl (30 mmol/L) and bathocuproinedisulfonic acid (300 mmol/L) to prevent ex vivo radical formation. After two bile samples were collected, rats were given an intravenous dose of [1-¹³C]ethanol (1 g/kg, diluted in saline), and additional samples were collected. The bile samples were weighed and then frozen on dry ice until the EPR measurements could be made. Thawed bile samples were transferred into flat quartz EPR cells and placed into the cavity of a Bruker 300E EPR spectrometer. Typical EPR operating conditions were as follows: gain, 1 × 10⁶; modulation amplitude, 1.0 G; modulation frequency, 100 kHz; microwave power, 20 mW; conversion time, 92 s; sweep time, 84 s. Ten scans were typically accumulated to intensify the weak signals observed. The relative EPR signal intensities were assessed by measuring the height of the EPR peaks in millimilliters, using standardized spectrometer conditions.

Dipyridyl, bathocuproinedisulfonic acid and xanthan gum were purchased from Sigma Chemical (St. Louis, MO), and all other dietary components were purchased from Bioserve (Frenchtown, NJ). The POBN was obtained from the OMRF Spin Trap Source (Oklahoma City, OK), and [1-¹³C]ethanol was purchased from Isotec (Miamisburg, OH).

RESULTS

The effects of prior ethanol feeding, as well as influences of dietary fat, on the two types of spin adduct signals shown in Figure 1 were investigated through a series of experiments involving pairs of alcohol-fed or control rats. In rats fed the high fat liquid diets, POBN injection resulted in detectable spin adducts with six-line EPR spectra in both bile samples taken prior to alcohol injection (Fig. 2). However, the average signal intensity was two- to fourfold greater in bile from alcohol-fed rats than in bile from the corresponding controls. When [1-¹³C]alcohol was subsequently injected, the signal intensity of this biliary spin adduct increased markedly, but greater increases were observed in the alcohol-fed rats (Fig. 2). Similar results were obtained in bile from rats fed low fat alcohol and control diets. However, the signals in samples from alcohol-fed rats tended to be less intense when the rats had been fed low fat diets than when they had been fed high fat diets (Fig. 2).

Effects of alcohol feeding and dietary fat on HER formation were also compared in these experiments. The POBN-¹³C-HER spin adduct was observed in all four groups of rats, but the most intense signals were observed in bile from rats fed alcohol in the high fat diet (Fig. 3). In rats fed low fat diets, the EPR signal intensity tended to show more variability, and essentially comparable results were observed in both control and alcohol-fed rats (Fig. 3).

DISCUSSION

The current results confirm and extend the observations of Knecht et al. (1990), but using the rat as the experimental model. Rats fed alcohol in combination with a high fat diet excreted more HER adducts in the bile than the corresponding controls or rats fed alcohol in a low fat diet (Fig. 3). These results are similar to those previously reported for liver microsomes, in which induction of HER formation by alcohol was greater with consumption of the high fat diet (Reinke et al. 1987). The permissive effect of dietary fat for induction of cytochrome P-450 by agents other than alcohol has been described elsewhere (Wade et al. 1985).

The data in this report do not allow any conclusions to be drawn regarding mechanisms of HER formation in vivo. This radical is known to be formed in vitro through reactions involving reactive oxygen intermediates and transition metals (Reinke et al. 1990, Knecht et al. 1993, Rashba-Step et al. 1993). Although cytochrome P-450 has been postulated to generate some HER directly from alcohol (Albano et al. 1988), others believe that the role of the ethanol-inducible cytochrome may be in the generation of increased levels of superoxide radical and hydrogen peroxide (Knecht et al. 1993). These reduced oxygen intermediates are likely to initiate a variety of free radical reactions in vivo, and conversion of ethanol to the HER is only one example of this type of reaction. It is likely that endogenous compounds would also be subject to attack by reactive oxygen intermediates.

The other radical adduct(s) detected in these studies cannot be positively identified on the basis of these studies. Although POBN has proven to be a superior spin trapping agent for detection of alcohol-associated free radicals in vivo, different types of radical adducts of this spin trap produce quite similar EPR spectra. There is literature precedent to suggest that two different radicals could be present. The EPR signal of the radical adduct designated as POBN-UNK in Figure 1 was particularly intense in bile from rats fed alcohol in the high fat diet (Fig 2). In our first report of alcohol-associated radical formation, we showed that lipid radicals could be detected in liver extracts from alcohol-fed rats and that the signals were particularly intense when the alcohol had been fed in a high fat diet (Reinke et al. 1987). Although a different spin trapping agent was employed in those studies, it is likely that the radical(s) also react with POBN and that the spin adducts are excreted into bile. The second possibility is that alcohol in the circulation was metabolized to the HER, trapped by POBN, and excreted in the bile. The blood alcohol concentration of rats allowed free access to the alcohol diet was quite variable and ranged between 5 to 10 mmol/L at the time of the experiments. We previously demonstrated that the HER is readily detectable in vivo at blood alcohol concentrations of 10 mmol/L (Moore et al. 1995), so it is quite probable that this radical adduct was also present in bile.

Even though the radical(s) that produce this six-line spectrum cannot be identified, this technical problem should not detract from the point that all free radical signals were greatest in livers from rats that had been fed alcohol in high fat diets. This combination of alcohol and fat in liquid diets has previously been shown to cause greater accumulation of hepatic lipids (Lieber and DeCarli 1982) and greater induction of cytochrome P-450 (Reinke et al. 1987) than were observed with alcohol in low fat diets. Furthermore, in the gastric infusion model of alcohol administration, alcohol combined with high fat (25% of energy as corn oil) resulted in liver fibrosis (Tsukamoto et al. 1986), whereas only steatosis and focal necrosis were observed when the corn oil content was reduced to 4.9% of total energy (Tsukamoto et al. 1985). The close relationship between liver injury and free radical formation associated with high fat diets and alcohol administration suggests that these phenomena may be causally related.