Review of the Fialuridine (FIAU) Clinical Trials (1995)

Since the tragic events associated with the (PPPC) clinical trial, Lilly and NIH have conducted and sponsored extensive additional research into the characteristics and mechanism of the FIAU toxicity described in humans and into the development of a possible animal model for this toxicity. The investigators involved gathered in April and December 1994 to provide progress reports to each other and their sponsors, who kindly agreed to allow the IOM committee access (Lilly-NIH FIAU meeting, 1994).

Among several studies utilizing tissue from FIAU-treated patients was work on mitochondrial (mt) enzymology by Douglas Kerr of Case Western Reserve University, which suggested that the liver toxicity from FIAU was related to injury to enzymes of the electron transport chain encoded in part by mt DNA. These liver samples showed relatively normal levels of mt enzymes of pyruvate metabolism, which are encoded largely from nuclear DNA.

Cecil Fox of Molecular Histology Laboratories and Richard Sallie of NIDDK used in situ hybridization with patient tissue to demonstrate widespread DNA injury in patients with severe FIAU toxicity, with associated apoptosis (cell death). Their studies also suggest that the relatively small decrease in total mt DNA measured in homogenized tissue may result in part from no change or even an increase in mt DNA from non-parenchymal cells in the liver.

Frank Richardson of Eli Lilly has been conducting studies in a variety of animal species, looking at FIAU incorporation into DNA, finding concentrations in some rat liver tissue as high as 1 molecule per 90 molecules of thymidine. Timed hydrolysis of rat liver DNA indicated a continual rise in the FIAU concentration, suggesting that FIAU was not chain terminating, that is, that it could continue to be incorporated (Richardson, Engelhardt and Bowsher, in press).

Inhibition of mt DNA replication has been proposed as the mechanism of FIAU clinical toxicity, and William Lewis and colleagues at the University of Cincinnati have shown in vitro that FIAU-triphosphate is a potent inhibitor of DNA polymerase gamma (which is responsible for mt DNA replication) (Lewis, Meyer, Simpson, Colacino and Perrino, in press). Colacino and colleagues (Colacino, Malcolm, and Jaskunas, 1994), however, demonstrated, also in vitro, that FIAU (and ddC) can increase lactic acid levels at doses and durations that do not inhibit mt DNA replication. Therefore, simply screening nucleoside analogs for their abilities to

inhibit mt DNA replication during a short exposure in a cell culture will be inadequate to predict serious problems in a clinical setting.

Despite failing to produce any significant histological or biochemical abnormalities in dogs given FIAU at 3 mg/kg/day for 90 days or in monkeys given 25 mg/kg/day for 30 days, Eli Lilly completed a report in July 1994 describing a study conducted from September 23 through December 2, 1993, in which they set out to produce in rats the toxic syndrome seen in the PPPC clinical trial. Rats (five rats of each sex receiving each dose) were treated with FIAU three times per day, with total daily dosages of 255 or 510 mg/kg/day (a dosage 1,000 times greater than that administered to humans) for a period of 10 weeks. Originally, the study was designed to last only 2 weeks, but since the animals tolerated the high doses better than expected, the study was extended first to 1 month and finally to 10 weeks. The FIAU was administered by gavage as an aqueous acacia suspension.

All rats survived the 10-week period, but clinical signs started to appear during the last 4 weeks. Significant differences in body weight gain and food utilization efficiency were seen in all animals, but the male rats were more severely affected than the females. Decreased body weight gain was attributed primarily to the decreased efficiency of food utilization. Hematologic parameters were assessed on days 27 and 70. Slight to moderate decreases in erythrocytes, lymphocytes, and leukocytes were observed in both male and female rats treated with the higher dose; lymphocytes and erythrocytes were also decreased in male rats treated with the lower dose. Regarding clinical chemistry parameters, blood urea nitrogen values were increased slightly in both dosage groups, but not in a dose-dependent manner; there was a slight increase in total plasma bilirubin concentration in males at both dose levels, but only at the higher dose in females. gamma-glutamyltransferase activity was increased (about 33 percent) in both males and females at the higher dose level. It is noteworthy, however, that no significant increases in ALT or AST were observed in either test group. Dose-related decreases in glucose (both doses, both sexes), total protein (both doses, males; high dose, females), and albumin (both doses, males; high dose, females) were recorded. There were moderate increases in plasma lactate ( versus 16 mg/dl) in animals of both sexes at both dose levels.

Mitochondrial parameters were assessed from liver samples obtained at day 70. Citrate synthase activity was significantly higher than that of control rats in both males ( percent of the control at both doses) and females ( percent of the control at the lower dose; percent of the control at the higher dose) treated with FIAU. On the other hand, cytochrome oxidase activity was significantly lower than that of control rats, in both males ( percent of the control at both doses) and females ( percent of the control at the lower dose; percent of the control at the higher dose) treated with FIAU.

DNA was isolated from liver, heart, spleen, jejunum, and epididymal sperm about 24 hours following the last treatment with FIAU. The FIAU concentration in cellular DNA was assessed by radioimmunoassay and was normalized to the thymidine concentration in each

sample. FIAU was found in nuclear DNA from all tissues; the highest concentration was found in liver (about 1 FIAU molecule/90 thymidine molecules); incorporation in males was twice that seen in females. DNA isolated from hepatic nuclei was reported to contain essentially all of the FIAU-DNA found in the cell. Nuclear FIAU-DNA was also observed in the jejunum and spleen, with lesser concentrations found in the sperm and heart.

Nephropathy was observed in all rats receiving the higher doses. Regenerative myopathy (three of five males, two of five females) was observed in animals receiving the high dose. Changes in hematopoietic tissue were unremarkable in the treated groups.

The liver was affected in all FIAU-treated rats. Histologic evaluation of liver tissue indicated slightly increased apoptosis and increased numbers of multinucleated hepatocytes; mitotic figures were also increased, with some mitoses having atypical morphologies; dose-related nuclear atypia and cytoplasmic hyaline droplet accumulation were also observed. Hepatic lobular architecture was generally normal even in animals most affected by cytologic alterations. By electron microscopy, enlarged mitochondria and increased numbers of secondary lysosomes were observed in rats treated with the higher dose of FIAU. The cytoplasmic droplets seen histologically were attributed to the megamitochondria observed by electron microscopy.

This new study performed by Lilly indicates that hepatic involvement can be observed in laboratory animals under specific treatment conditions. This is an important contribution, since previous animal studies included in the FIAU preclinical toxicology assessment did not show hepatic dysfunction or hepatotoxicity. Although the changes are important, it should be noted that the morphologic and biochemical descriptions do not mimic exactly what was observed in humans during the tragic clinical trials. Elevations in plasma aminotransferase activity were not observed in the rats treated with FIAU for 10 weeks. Histologically, no mention was made of microsteatosis in rat liver at either dose level, although this was a major finding in humans. Finally, the morphologic evidence does not indicate an onset of liver failure. On the other hand, the finding that megamitochondria are observed in FIAU-treated rats is important and relevant. Furthermore, changes in serum lactate levels suggest changes in mitochondrial intermediary metabolism. There are suggestions in the literature, based on in vitro studies, that FIAU may affect mitochondrial function and that this event could be related to some of the toxic effects of the agent. The rat findings in vivo are supportive of the hypothesis; the data indicate that mitochondrial involvement in toxic manifestations of FIAU must be pursued in vivo to test the validity of the hypothesis. The incorporation of FIAU into liver nuclear DNA is also an important finding. Its relationship, however, to undesirable events remains to be established.

An entirely different animal model has been utilized in two studies with FIAU conducted by Bud C. Tennant at Cornell University College of Veterinary Medicine. The Eastern U.S. woodchuck (Marmota monax) harbors an hepadnavirus, woodchuck hepatitis virus (WHV), that is similar to human hepatitis B virus. Tennant provided the IOM committee with a brief written summary of the results of two experiments performed with FIAU-treated woodchucks (Tennant, 1994; Tennant and Gerin, 1994). For both of these studies, laboratory-born and reared woodchucks were inoculated shortly after birth with WHV-positive serum; 95 percent or more became infected and 65-70 percent became chronic carriers with progressively worsening hepatitis. Infected woodchucks generally die of liver cancer during the second or third year of life (the lifetime risk of liver cancer for these animals approaches 100 percent).

The first study was started on December 9, 1992 and was completed on March 30, 1993. Three groups (six animals per group) of woodchucks with chronic WHV infection were treated with daily injections of FIAU (0.0, 0.3, or 1.5 mg/kg) directly into the abdomen (intraperitoneally) for a total of 28 days. On the basis of the metabolic size of woodchucks compared with that of humans, the 1.5-mg/kg/day dose of FIAU was estimated to be equivalent to a human dose of 0.5 mg/kg/day, whereas the 0.3-mg/kg/day dose was estimated to be equivalent to a human dose of 0.1 mg/kg/day. A 12-week observation period following FIAU treatment was included in the design of the experiment. Although the antiviral effect of FIAU at the lower dose (0.3 mg/kg/day) was equivocal, FIAU appeared to have a strong antiviral effect at 1.5 mg/kg/day.

Accumulation of lipid in the liver of one of six animals receiving 1.5 mg/kg/day was observed at the end of the FIAU treatment period and again in a liver biopsy specimen 4 weeks after the end of the treatment. The animal died 3 days after the liver sample was obtained by biopsy, but the death was attributed to complications of the liver biopsy. Another woodchuck receiving FIAU at the same dose level exhibited a milder accumulation of liver lipid at 4 and 12 weeks; this animal also died, but the death was considered to be due to the anesthetic agent. One of six woodchucks treated only with a placebo (no FIAU treatment) also exhibited an accumulation of lipid in the liver at 4 and 12 weeks following treatment. Because lipid accumulation was observed in one placebo-treated animal, the investigators found it impossible to conclude that FIAU was the cause of the liver lipid change in the FIAU-treated woodchucks. On July 2, 1993, the group at Cornell, who had been advised of the adverse effects seen in patients treated with FIAU in the PPPC clinical trial, reinitiated posttreatment observations on the surviving woodchucks; the observations were continued until February 1994, when the remaining animals were euthanized and postmortem evaluations were performed. A significant difference in body weights between the group treated with FIAU at a high dose and the placebo-treated animals was observed; the groups receiving FIAU weighed less (3.4 versus 3.8 kg), but the meaning of this finding is not evident.

The second study in woodchucks was initiated in February 1994 and was completed in May. One objective of the second study was to see if the presence of a chronic HBV infection was a factor of importance in the liver injury observed in humans following treatment with FIAU. Furthermore, the route of administration of FIAU was oral rather than direct abdominal injection. Eighteen woodchucks were infected with virus (WHV) at 3 days of age and were

reared as adult chronic WHV carriers; 18 other woodchucks were reared to be virus-free. Before initiating the daily FIAU treatments (duration, 12 weeks), clinical tests as well as serologic tests for the presence of WHV were performed. There were four separate groups of woodchucks in the 12-week treatment study. Nine animals with chronic WHV infection were given FIAU orally (1.5 mg/kg/day); nine other animals with chronic WHV infection received only the placebo orally. These 18 infected animals were compared with two other groups of uninfected woodchucks (nine animals treated with FIAU at 1.5 mg/kg/day and nine animals receiving only the placebo). All of the data from the study have yet to be completely analyzed. Some preliminary results are available, however. After 4 weeks of treatment, no treatment-related findings were observed in any of the groups. Accumulation of liver lipids was observed at this time. After 8 weeks, the mean body weights of the FIAU-treated woodchucks (both WHV carriers and uninfected animals) were lower than those observed in the placebo groups (receiving no FIAU). Anorexia accompanied the weight loss; the effects persisted in surviving woodchucks during the posttreatment period. The effects were attributed to the FIAU treatment and were similar in WHV carriers and uninfected animals; the toxicity was not dependent on the presence of a persistent WHV infection or underlying hepatitis. Biochemical signs of liver injury were observed in woodchucks treated with FIAU. Furthermore, both WHV infected and uninfected animals receiving FIAU are reported by the investigator as exhibiting elevations in transaminase activity, particularly AST as the animals became moribund. Kidney insufficiency was also observed in FIAU-treated animals; elevations in blood lactic acid concentrations were found in the final stages of FIAU toxicity. Changes were observed in the pancreases of some woodchucks, but the role of FIAU treatment is not yet established. Evidence of liver injury, including increases in liver lipids, was observed in all but one FIAU-treated woodchuck. Similar changes in placebo-treated animals (WHV positive or uninfected) were not observed. Finally, at the dose level of 1.5 mg/kg/day, it was reported that all FIAU-treated woodchucks ''either died or were euthanized because of multiorgan failure 78-111 days after beginning drug treatment.''

The results of the first woodchuck study do not allow one to draw any firm conclusions, since the changes seen in the livers of some of the FIAU-treated animals were also observed in the control placebo-treated woodchucks. Furthermore, the deaths observed could be attributed to the sampling procedures used to obtain tissue specimens rather than to then FIAU treatment. In terms of the chronology of the study, it is important to note that although this study actually preceded the PPPC clinical trial, there really were no indications of remarkable untoward effects caused by FIAU.

The second study was initiated after the PPPC clinical trial. Although the data have not been completely analyzed, the preliminary findings reported by the investigator indicate that with extended treatment, FIAU does exert toxic effects in woodchucks. The study also indicates that the presence of a chronic viral infection is not a necessary factor for the toxic effects observed. In this experiment, there is a much stronger indication that FIAU can effect liver toxicity in woodchucks. One striking point is the presence of toxic signs in this species of

animal at doses of FIAU that are quite small compared with those utilized in previous toxicity testing with mice, rats, dogs, and monkeys. The first doses in fact were just sufficient for an antiviral effect. A burning question is how closely this woodchuck model resembles FIAU toxicity in humans. The incompleteness of the data, particularly the lack of specifics on the numbers of FIAU-attributable deaths (no numbers were supplied to the committee), undermines significantly the interpretation one can make of the results. The situation should undoubtedly be clarified when the final evaluations are available for scientific scrutiny.

It is important to note that the woodchuck is an uncommon animal species for toxicity testing. They were used in the present instance only after the toxicity of FIAU in patients with liver disease was clear. It is highly unlikely that the woodchuck could become a routine species for the preclinical testing of FIAU-like chemicals. To the committee's knowledge these animals are difficult to obtain; the animals used at Cornell University were born and reared in the local facilities, since those trapped in the wild are often rabid. The reproductive capacity of the woodchuck (one breeding cycle annually yielding one or two offspring) unfortunately precludes its use as a routine species for animal toxicity testing, but the data to date suggest that it might be a powerful addition to preclinical testing for anti-HBV drugs.