Phase I and Pharmacokinetic Study of Triapine, a Potent Ribonucleotide Reductase Inhibitor, Administered Daily for Five Days in Patients with Advanced Solid Tumors¹

Triapine [3-aminopyridine-2-carboxaldehyde thiosemicarbazone (Fig. 1)], is a new potent inhibitor of RR,³ an enzyme that mediates the conversion of ribonucleotides to deoxyribonucleotides (1). Treatment of tumor cell lines with Triapine causes a reduction in one or more of the deoxyribonucleotide triphosphates (dNTPs), resulting in the arrest or slowing of DNA synthesis and cellular proliferation. Triapine was screened for antitumor effect in the National Cancer Institute panel of 60 tumor cell lines and demonstrated an average GI₅₀ (concentration required to inhibit growth by 50%) in the range of 1.6 μm. Triapine also showed antitumor activity in several murine models (2). When administered twice daily for five days at doses of 6–10 mg/kg, Triapine inhibited the growth of s.c. implanted A2780 human ovarian and M109 murine lung cancer cell lines, and the growth of L1210 leukemia implanted in the peritoneum. The antitumor effects of Triapine in the M109 tumor model were comparable with those of paclitaxel.

The mechanism by which Triapine inhibits RR is similar to that of hydroxyurea, an agent with clinical antitumor activity against both solid tumors and hematological malignancies (3, 4). Both agents inhibit the M2 subunit of RR by quenching a tyrosyl-free radical (5, 6, 7). However, Triapine is ∼100–1000-fold more potent than hydroxyurea in both enzyme and tumor cell-growth inhibition assays. Two tumor cell lines selected for resistance to hydroxyurea had unchanged sensitivity to Triapine compared with the parent cell lines, indicating that Triapine can overcome resistance to hydroxyurea attributed to overexpression of the M2 RR subunit (1, 2, 8).

In the toxicology studies of Triapine, short 15-min i.v. infusions caused emesis in dogs, which was reduced by extending the infusion to 2 h (9). Therefore, single- and multiple-dose Phase I trials of Triapine were initiated using a 2-h i.v. infusion schedule. The results of the Phase I trial examining an every-4-week schedule were recently published and demonstrated that doses up to 105 mg/m² could be administered without DLT (10). Dose escalation in that study was terminated without reaching a MTD, because preclinical studies indicated that multiple-dose schedules were required for antitumor activity. Peak plasma concentrations of Triapine at the highest dose were ∼5 μm and the half-life of elimination was in the range of 30–120 min. Herein, we describe the results of the Phase I trial of a 2-h i.v. infusion administered daily for 5 days.

The study was approved by the Human Subject Committees at each of the participating centers. All of the patients were required to sign informed consent before study participation. Patients were eligible if they were 18 years or older with pathologically documented malignant disease and no standard therapeutic options. The last dose of chemotherapy or radiotherapy must have been given >3 weeks before study entry (6 weeks for nitrosoureas or mitomycin C), and patients must have recovered from all acute toxicities of prior therapy. Eligibility criteria included the following: Karnofsky performance status >70%; body weight >50 kg; serum creatinine ≤2.0 mg/dl; total bilirubin ≤2.0 mg/dl; ALT and AST no more than three times the upper limit of normal range; alkaline phosphatase no more than three times the upper limit of normal range, or if the alkaline phosphatase was determined to be of nonhepatic origin, no more than five times the upper limit of normal; WBC >3,000/mm³; absolute neutrophil count ≥1500/mm³; platelet count >100,000/mm³; and hemoglobin ≥10 g/dl. Males and females were required to practice adequate contraception or abstinence, and pregnant or lactating women were excluded. Patients were also excluded if they had evidence of active heart disease including myocardial infarction within the previous 3 months; symptomatic coronary insufficiency or heart block; uncontrolled congestive heart failure; moderate or severe pulmonary dysfunction; active infectious process; known, active central nervous system metastases; or prior radiotherapy administered to more than 30% of marrow-bearing bone mass.

The study was originally designed as a dose-escalation study, based on a modified Fibonacci scheme, in which cohorts of three to six patients would be entered at each dose level until two patients developed DLT. Dose escalation to the next higher level was permitted in patients without clinical disease progression if drug-related adverse events from the previous course were less than grade 2; thus, a single patient could be evaluated for toxicity at more than one dose level. The MTD was defined as the highest dose level at which less than two patients of six experience first-course DLT. The starting dose level was 5 mg/m², daily for 5 days, repeated every 4 weeks. This starting dose was ∼10% of a conservatively estimated LD₁₀ (dose lethal to 10% of animals) for the daily-for-5-days schedule in a rat toxicology study (9). When the first patient developed an unexpected severe hepatic adverse event, the starting dose was reduced to 2 mg/m²/day. After the first five dose levels were evaluated (2, 3, 5, 8, and 12 mg/m²/day) and no additional toxicities of more than grade 1 were observed, the study was amended to escalate the dose in successive cohorts based on an accelerated titration design as described by Simon et al. (11). Beginning with a dose of 24 mg/m²/day, one to two patients were assigned to a cohort, and doses were escalated by 100% for each new cohort, until the first instance of drug-related grade 2 toxicity or DLT. Intrapatient dose escalation was permitted if drug-related toxicity in a previous course was less than grade 2, and patients in whom the dose was escalated were considered evaluable for toxicity at the higher dose level. At a dose of 96 mg/m²/day, grade 2 or higher toxicities were observed. The adverse events at the 96 mg/m²/day were not dose limiting but were of sufficient severity (grade 4 neutropenia) to indicate that further dose escalation would cause excessive myelosuppression. Because adverse events recovered to no more than grade 1 by day 15, the protocol was amended to evaluate the safety of the 96-mg/m² dose, daily for 5 days every other week, in an expanded cohort of 10–15 patients. For this latter cohort, dose escalation to 105 mg/m²/day was permitted if toxicity in the previous course was less than grade 2.

The National Cancer Institute Common Toxicity Criteria version 2.0 was used to grade toxicity. DLT was defined as nonhematological toxicity of at least grade 3, or grade 4 neutropenia or thrombocytopenia lasting more than 3 days, or grade 4 neutropenia of any duration associated with the development of a life-threatening infection.

Before study entry, patients were assessed with a complete history and physical exam, vital signs, electrocardiogram, pregnancy test in women of child-bearing potential, tumor staging studies including computed tomography scans as appropriate to the disease, CBC with differential, coagulation studies, serum chemistries including electrolytes and liver function tests, chest X ray, urinalysis, and iron studies (serum iron, total iron binding capacity, ferritin). The CBC and serum chemistries were repeated before the fourth dose, and serum iron studies were repeated before the fifth dose of each course. CBC with differential was obtained twice weekly, and serum chemistries were obtained once weekly between courses. Toxicities were monitored and recorded continuously. Patients were evaluated for tumor response with full staging studies every 2 months. Patients demonstrating stable disease or partial response were eligible to continue treatment for up to 1 year, and patients with complete response would receive two to four additional courses of treatment. Complete remission was defined as the disappearance of all clinical evidence of active tumor and symptoms for a minimum of 4 weeks. Partial remission was defined as a decrease (by >50%) in the sum of the product of the perpendicular diameters of all of the measured lesions for at least 4 weeks. No simultaneous increase in the size of any lesion or the appearance of new lesions could occur. The disease was considered stable when the response was less than partial and did not meet evidence for progression for a minimum of 8 weeks. Progressive disease was defined as an unequivocal increase of at least 50% in the size of any measured lesion, or the appearance of significant new lesions.

Triapine (3-aminopyridine-2-carboxaldehyde thiosemicarbazone) was supplied by Vion Pharmaceuticals, Inc. Studies conducted by Vion indicate that Triapine is stable when diluted to a final concentration of 0.01–2 mg/ml for up to 96 h. Triapine was diluted into 500 ml of normal saline or 5% dextrose in water and was administered by i.v. infusion over 2 h. Dilutions of Triapine were performed in glass bottles and not in polyvinyl chloride plastic containers to avoid extraction of the plasticizer di(ethylhexyl)phthalate by the nonaqueous solvents in the Triapine formulation. For the same reason, polyethylene-lined administration sets were not used.

Antiemetics were administered only as required for patients developing nausea or vomiting with a previous dose. Administration of hematopoietic growth factors was permitted after the second or subsequent courses if patients developed grade 4 neutropenia not resolving to grade 3 or less in <48 h in the previous course, or if neutropenia had not resolved to less than grade 2 by the scheduled date of the next course.

Blood samples for determination of Triapine pharmacokinetics were collected from the arm opposite the infusion on days 1 and 5 of the first course immediately before the start of infusion, during the infusion at approximately 15, 30, 60, 90, and 120 min (at the end of infusion), and 15, 30, 45, 60, 240, and 480 min after the end of the infusion. Five ml of blood were collected in a Vacutainer red top tube without anticoagulant. The blood was allowed to clot at room temperature for 15–20 min, and then was centrifuged at 3,000 rpm for 10–20 min. Serum was separated, transferred to two separate labeled glass vials, immediately frozen at −20°C, and stored frozen until analysis. Urine was collected before the infusion, and on days 1 and 5 was collected and pooled from 0–4 and 4–8 h after the start of the infusion. An aliquot of 100 ml of the pooled urine from each period was saved in a labeled plastic storage vial and was frozen at −20°C until analysis.

HPLC with UV detection was used to analyze the serum and urine samples for Triapine concentration. An Agilent Technologies 1100 series HPLC system was used. Chromatographic separation was achieved using a Supelco Discovery C18 column (5 μm, 250 mm × 4.6 mm; Supelco, St. Louis, MO) and detection at 400 nm. Agilent ChemStation software was used for data acquisition and processing. Serum or urine samples (0.5 ml) were extracted with 1.0 ml of methanol (containing 4 mm EDTA). After centrifugation, the extract was concentrated to dryness and was reconstituted with 0.25 ml of a solvent consisting of 10% acetonitrile and 90% mobile Phase A [20 mm potassium phosphate buffer, 15 mm 1-heptanesulfonic acid, and 1 mm EDTA (pH 3.0)]. The reconstituted solution sample (30 μl) was then injected into the HPLC system. External calibration standards were prepared in pooled control human serum or urine and were processed identically to test samples. The validated assay has a nominal curve range of 0.02–10 μg/ml for serum, and 0.05–10 μg/ml for urine. Quality control samples of both serum and urine were prepared at various concentration levels, stored with test samples in a freezer, and analyzed with each sample batch.

Pharmacokinetic modeling and pharmacokinetic parameter calculations were conducted using WinNonlin software program (Pharsight Corporation, Mountain View, CA) with compartmental as well as noncompartmental methods. The following pharmacokinetic parameters were computed: area under the serum concentration-time curve (AUC) from time zero to the last data point, peak serum concentration (C_max), elimination half-life (T_1/2), volume of distribution at steady state (V_d,ss), and total body clearance (Cl_tot). Descriptive statistics (mean and SD) were calculated and were used to characterize the pharmacokinetic parameters at each dose level. For urine samples, cumulative urinary recovery of unchanged drug (Triapine) was determined for the time period of 0 (start of the infusion) to 8 h.

Thirty-two patients were treated on the study between August 1998 and April 2001. Seventeen patients were treated with Triapine administered on the every-4-week schedule. Fifteen patients were treated on the every-other-week schedule. The characteristics of the patients are listed in Table 1 for each schedule, respectively. Overall, the median age was 67 years. There were 23 (72%) men and 9 women enrolled. Karnofsky performance status was 90–100 in 38% and 70–80 in 62%. Twenty-nine (91%) patients had received prior chemotherapy. Fifty % of patients had metastatic gastrointestinal tumors.

Table 2 provides the total number of patients, new patients, and total number of courses delivered by dose level. Among the 15 patients treated with 96 mg/m² daily for 5 days every other week, one patient received only one dose of treatment, five required dose reduction of 12.5–50%, and one patient’s dose was escalated to 105 mg/m². Of the total 49 courses (second and subsequent courses) administered on the every-other-week schedule, 63% were delivered on time.

Adverse events that were considered possibly, probably, or definitely related to Triapine administration by the treating physician are shown in Table 3. The first patient entered on the trial, a 52-year-old male with colon cancer metastatic to liver, received 5 mg/m²/day and developed grade 3 alkaline phosphatase, AST, ALT, and bilirubin by day 5 of the first course. Tumor-related distal common bile duct obstruction was suspected and was ultimately confirmed 2 months later by endoscopic retrograde cholangiopancreatography. However, early diagnostic procedures were inconclusive, and the hepatic adverse events were initially considered possibly related to Triapine administration. The trial was temporarily halted and restarted at a dose of 2 mg/m²/day. In one patient, the first dose of a third course of treatment given at 3 mg/m²/day was complicated by elevation of blood pressure to 190/110 mm Hg that required hospital admission and treatment with a diuretic. Hypertension was first noted ∼4 h after completing the infusion; because of the patient’s long-standing history of hypertension, the event was judged unlikely to be related to treatment.

Through a dose of 12 mg/m²/day, Triapine produced minimal nonhematological and hematological toxicity. Therefore, the protocol was amended to permit 100% dose escalation in single-patient cohorts until the first episode of drug-related grade 2 adverse events or DLT. Among the four patients treated at a dose of 96 mg/m²/day every 4 weeks, two developed grade 4, and one patient developed grade 3 leukopenia. The leukopenia occurred in the week after treatment, was reversible to grade 1 or less by day 15, and was not dose limiting. Rather than continuing dose escalation on the every-4-week schedule, which would likely have produced excessive myelosuppression, the protocol was amended to determine the safety and tolerability of the 96-mg/m² dosage daily for 5 days every other week. Fifteen patients began treatment on the latter schedule, and 14 completed at least one 5-day course of treatment. The one patient who withdrew early from the trial developed burning of the hands and heartburn transiently during the first infusion, followed several hours later by grade 2 fever and flushing, grade 1 nausea, vomiting, and hypertension, and transient hypoxia. The events resolved by the following morning, but the patient elected not to continue protocol treatment.

In the 14 patients completing at least one course of treatment on the every-other-week schedule, most nonhematological toxicity was grade 1–2 and rapidly reversible. The most common events were asthenia, fever, nausea and vomiting, mucositis, decreased serum bicarbonate, and hyperbilirubinemia. The highest bilirubin levels were observed on days 8–10 of the treatment course; it was grade 2 in five patients and grade 3 in two patients. Routine serum chemistries obtained on day 4 (before the fourth dose in a course) detected decreases in serum bicarbonate in nine patients. The decreases were mild (grade 1 in eight patients, grade 2 in one patient), and generally returned to baseline by day 8 or 9. Because serum chemistries were not obtained on day 5 or 6, the possibility exists that some patients had greater changes in serum bicarbonate. The low serum bicarbonate was not associated with an increase in the anion gap or changes in serum potassium concentrations.

Hematological toxicity on the every-other-week schedule consisted of leukopenia (grade 4 in 93% of the patients in at least one course) and anemia (grade 2 in 71%, grade 3 in 22%). Thrombocytopenia was less common and was grade 3–4 in severity in only 22%. Four patients (29%) developed febrile neutropenia.

Pharmacokinetic evaluation was performed in a total of 29 patients, who received a total of 50 treatment courses. Triapine was quantifiable in serum for up to 6–8 h after dosing (the 8-hour sample was not collected for all of the patients). For most patients, the pharmacokinetic parameters were derived from a one-compartment model. However, because this model did not fit well in all of the patients for all of the courses, a two-compartment model was used for two patients, and a noncompartmental analysis was performed for one patient. Mean AUC and C_max versus dose on days 1 and 5 are plotted in Fig. 2, and the mean serum concentration versus time on day 1 for five of the dose levels is shown in Fig. 3. Triapine showed linear pharmacokinetic behavior although interpatient variability was relatively high. AUC and C_max values showed no significant difference between the first dose (day 1) and the last dose (day 5), suggesting no drug accumulation under the multiple-dosing schedule.

Table 4 summarizes the mean pharmacokinetic parameters at each dose level. These parameters are presented as the mean of day 1 and day 5 values. Peak concentrations at the 96-mg/m²/day dose averaged 8 μm. The elimination T_1/2 ranged from 35 min to 3 h, with a median value of ∼1 h. In urine, Triapine was detected during the 8 h after drug administration. Cumulative urinary recovery averaged 1–3% of the administered dose, suggesting that the elimination of Triapine was primarily through metabolism.

No partial or complete responses were observed. Three patients treated on the every-other-week schedule had evidence of antitumor effect. The first patient, a 73-year-old female with breast cancer metastatic to the peritoneum and progressive after multiple chemotherapy and hormonal treatments, demonstrated a 53% decrease of an elevated, baseline CA27.29 serum level after two courses of treatment. Her intra-abdominal disease was not measurable. The patient died of a bowel perforation after a fifth course. A 39-year-old male with an adenocarcinoma of the esophageal-gastric junction metastatic to supraclavicular and neck nodes, and progressive after two prior chemotherapy regimens, had a 51% decrease in carcinoembryonic antigen over four courses associated with reduction in neck adenopathy and improvement in pain. The patient developed progressive disease after an additional three courses of treatment. A 56-year-old male with esophageal cancer and continued involvement of the distal esophagus and proximal stomach had improvement in tumor visualized by endoscopy after four courses of treatment. The disease progressed after an additional four courses of treatment. Two additional patients treated on the every-other-week schedule (metastatic breast cancer and colon cancer), both with substantial disease involvement and multiple prior treatments, were felt to have evidence of disease stabilization, and received 10 and 6 courses of treatment, respectively.

Ribonucleotide reductase is the only known pathway for reduction of ribonucleotides to deoxyribonucleotides necessary for DNA synthesis and cell replication. Expression of this enzyme is controlled by the retinoblastoma protein, which is inactivated in many cancers, suggesting that expression of RR is dysregulated in cancer cells (12). The only currently available RR inhibitor acting at the same site within the enzyme (M2 subunit) as Triapine is hydroxyurea. However, hydroxyurea has low potency against the enzyme, and resistance can develop readily by overexpression of the M2 subunit (8, 13). Other available RR inhibitors are nucleoside analogues, e.g., gemcitabine and fludarabine, which require uptake and intracellular activation and then interact with the M1 subunit of the enzyme (14, 15). Although nucleoside analogues are potent RR inhibitors, they appear to mediate cytotoxicity primarily through DNA incorporation (16). Interestingly, in some cell lines, resistance to a nucleoside analogue has been attributed to overexpression of the M2 subunit of RR (17, 18). Recently, the activity of gemcitabine in patients with non-small cell lung cancer was reported to be reduced in tumors that overexpress RR (19).

Triapine is 100–1000-fold more potent than hydroxyurea and was shown to be active in cell lines selected for resistance to hydroxyurea. The increased activity of Triapine compared with hydroxyurea is thought to be a consequence of its ability to chelate iron, which is essential to regenerate the tyrosyl-free radical in the M2 subunit that initiates reduction of ribonucleotides (20, 21, 22). Recently published data indicate that Triapine may have a second mechanism contributing to its antitumor activity (22). A Triapine-iron complex was shown to produce DNA damage in vitro through a redox cycling mechanism at clinically relevant concentrations. Of interest, the authors also conducted experiments showing preferential cytotoxic effects of Triapine on tumor cell lines compared with a fibroblast cell line.

In the present study, we initially sought to establish the MTD for a daily-for-5-days regimen administered every 4 weeks. At a dose of 96 mg/m²/day, the most prominent toxicity of Triapine was grade 4 leukopenia of short duration. Mild to moderate anemia and thrombocytopenia were also observed, and nonhematological toxicities were limited and mild. Rather than escalating the dose, which was expected to produce even more severe myelosuppression and possibly greater nonhematological toxicity, we elected instead to administer 96 mg/m²/day in a condensed schedule of daily-for-5-days every other week. The every-other-week regimen was generally well tolerated but produced grade 4 leukopenia in one or more courses in 13 of the 14 patients. Grade 2–3 anemia was also observed in 93% of patients, but thrombocytopenia was less common and was grade 3–4 in only 22% of patients. To reduce the incidence of severe leukopenia, a daily-for-4-days every-other-week schedule will be used in subsequent Phase II studies. The Phase II studies will also include intrapatient dose escalation to minimize underdosing of some patients.

Although not dose limiting, hyperbilirubinemia and decreased serum bicarbonate were noted in 50% and 64% of patients, respectively, treated on the every-other-week schedule. Hyperbilirubinemia was initially suspected to be caused by hemolysis because the time of occurrence of the highest levels often coincided with decreases in hemoglobin. Furthermore, 5-hydroxy-2-formylpyridine thiosemicarbazone, another agent in the class of α-(N)-heterocyclic thiosemicarbazones to which Triapine belongs, had been studied briefly in clinical trials in the early 1970s and was shown to cause hemolysis (23). Unfortunately, appropriate studies to confirm hemolysis, including fractionation of bilirubin and serum haptoglobin, were not obtained, and the cause of the elevated bilirubin remains in question; alternative possibilities include Triapine-induced ineffective erythropoiesis secondary to its iron-chelating effects, or interference with bilirubin uptake, conjugation, or excretion by the liver. The mechanism by which Triapine induces decreased serum bicarbonate also remains unclear. Triapine may be affecting kidney function directly and could possibly be producing a mild drug-induced renal tubular acidosis.

Because Triapine is known to chelate iron, an attempt was made to measure serum iron, iron-binding capacity, and ferritin, serially in patients entered on the trial. The studies were collected inconsistently, and no clear trend in any direction was observed (data not shown). Only small amounts of Triapine are excreted in the urine. It is possible that, during metabolism of Triapine, any bound iron is salvaged. The higher serum concentrations of Triapine achieved at the proposed Phase II dose could affect the function of iron-containing enzymes and perhaps iron-containing proteins such as hemoglobin, in the latter case interfering with oxygen binding and delivery. The duration of adverse events related to interference with iron-containing proteins is expected to be dependent on the serum and intracellular half-life of Triapine. On the basis of the findings in this trial, such adverse events are subclinical in most patients and rapidly reversible.

The 2-hour infusion of Triapine at a dose of 96 mg/m² produced peak serum concentrations in the range of ∼8 μm, substantially higher than concentrations required to inhibit RR in vitro. On the basis of the elimination half-life of ∼1 h, Triapine serum concentrations exceeded the mean GI₅₀ for cancer cell lines (∼1.6 μm in a 48-h exposure assay) for ∼3–5 h/day. In vitro studies conducted at Vion indicate that Triapine effects on RR begin to reverse in some tumor cell lines within 4–6 h. Thus the 2-hour infusion would be predicted to maintain a pharmacological effect for 6–10 h/day. The duration of effect is sufficient to produce myelosuppression, and some evidence of antitumor effect was observed in a very heavily pretreated patient population.

The schedule of administration may be an important determinant of Triapine antitumor activity. In vitro, antitumor activity requires both a threshold concentration and minimum duration of exposure. In one murine tumor model, twice-daily administration was required to produce growth inhibition; once-daily administration was inactive (2). In the murine model, maximum single doses were limited by the toxicity of the formulation. Humans appear to tolerate higher doses than do mice based on surface area; thus, a 2-hour infusion administered for 4–5 days may be sufficient to observe antitumor activity in patients. It may be possible to further increase antitumor effects by developing multiple-day i.v. continuous infusion or twice-daily administration schedules. A Phase I 96-h i.v. continuous infusion study has been initiated, and an oral formulation is being developed to permit twice-daily administration schedules.

Although Triapine exerts an antitumor effect in vitro and in vivo in preclinical studies, it remains to be determined whether it will have sufficient activity as a single agent, in any schedule, to cause tumor regression or to inhibit tumor progression in a meaningful proportion of patients with advanced cancer. The most effective use of Triapine may be in combination with other agents. For example, Triapine-induced depletion of deoxyribonucleotide triphosphates is predicted to inhibit the repair of DNA damage created by other cytotoxic agents. Indeed, Triapine demonstrates additive or synergistic activity in vitro and in vivo with various DNA-damaging agents (2). In vitro, Triapine also increases uptake and DNA incorporation of several nucleoside analogues in a schedule-dependent manner, resulting in enhanced cytotoxicity (24). The toxicity profile demonstrated in the present trial indicates that combinations of Triapine with various standard agents are likely to be feasible and warrant clinical study.

We thank Drs. John Mao and Samson Tom for their technical assistance and advice in conducting and analyzing the pharmacokinetic studies.