Phase I Dose-Escalation Study of the Safety and Pharmacokinetics of Atrasentan: An Endothelin Receptor Antagonist for Refractory Prostate Cancer¹

The ETs,³ a family of three 21-amino acid isopeptides (ET-1, ET-2, and ET-3), were first identified in 1988 as secretory products of vascular endothelial cells with potent, sustained vasoconstrictor activity (1). ET-1, the main isoform, is produced by endothelial and epithelial cells and has been recognized as a mediator of growth and survival signal in both normal physiology and tumor growth (2, 3, 4). ET-1 is also produced by many epithelial-derived human tumors (3, 5) and is considered to be a mediator and an autocrine and a paracrine growth factor for local cancer and stromal cells (6, 7).

The action of ET-1 is thought to be mediated via two G-protein-coupled receptors, ET_A and ET_B, which are distinguished by different binding affinities for the ETs. The ET_B receptor binds the three isotypes with equal affinity and functions primarily as both a clearance receptor and a modulator of ET-1 expression (8). In contrast, the ET_A receptor binds ET-1 with a higher affinity than the other isoforms. Binding of ET-1 to the ET_A receptor stimulates vasoconstriction in normal vascular endothelium, stimulates proliferative responses in normal and neoplastic cells, and potentiates other growth factors in malignant growth (2, 8, 9, 10). Recently, ET-1 has also been found to prevent apoptosis (11).

Secreted by normal prostate epithelial cells (12), ET-1 is found in human semen in the highest concentrations thus far observed in any body fluid (13). A dysregulation of the ET axis leading to increased ET-1 production was first reported in prostate cancer; ET-1 concentrations were found to be significantly elevated in patients with metastatic hormone-refractory disease (14). The tumor-promoting effects of ET-1 in prostate cancer were observed to occur via ET_A receptors (15), with ET_A receptors predominating (10). Tumor growth was shown to be promoted by increased production of ET-1, decreased ET-1 clearance, and ET_A overexpression (10, 16). ET-1 expression correlates well with the stage and grade of human prostatic cancer (10, 14, 17).

A dysregulation of the ET-1/ET_A receptor system similar to that in prostatic cancer has been reported for ovarian and human papilloma virus-associated cervical carcinomas (18, 19, 20), in which functional ET_A receptors are overexpressed. Selective antagonism of ET_A receptors, but not ET_B receptors, has been found to inhibit ET-1-mediated growth effects in various tumor models (18, 19, 20). Additional evidence in support of ET-1 and/or ET_A receptor expression in human cancers suggests a wider participation of the ET axis in both human tumor cells and normal stromal cells (21, 22).

Selective blockade of the ET-1 receptor represents a targeted approach to abrogating the pathophysiological effects of ET in cancer (14). Atrasentan, a p.o. bioavailable drug, is a potent (Ki = 0.034 nm) and selective (Ki ET_B = 63.3 nm) ET_A receptor antagonist (23, 24). This Phase I dose-escalation study was designed to evaluate the safety and tolerability of atrasentan (ABT-627), estimate the MTD, and characterize the PK of the drug after once-daily oral administration in patients with refractory adenocarcinomas, particularly those with hormone-refractory prostate cancer requiring pain relief with opioids.

Patients selected for this single-center open-label Phase I study had an adenocarcinoma (documented cytologically or histologically) refractory to standard therapy, with preferred enrollment offered to patients with hormone-refractory prostate cancer requiring pain relief with opioid analgesics. Participants had to be at least 18 years old, have an Eastern Cooperative Oncology Group performance status of 0–2, and have a life expectancy of at least 3 months. Women with child-bearing potential were required to have a negative pregnancy test. Patients were excluded from the study if they had brain metastases, severe left ventricular dysfunction (assessed by ECG or multigated acquisition), clinically significant ECG abnormalities, history of migraines, total bilirubin ≥ 25.6 mm, aspartate transaminase or alanine transaminase ≥ 1.5× the upper limit of the reference range, calculated creatinine clearance < 50 ml/min, WBC count ≤ 2,000/mm³, absolute neutrophil count ≤ 1,000/mm³, platelet count ≤ 100,000/mm³, or hemoglobin ≤ 1.4 nm.

Before enrollment, patients must have discontinued strontium or suramin therapy for 3 months, rhenium-186-etidronate for 8 weeks, antiandrogen treatment with biclutamide or nilutamide for 8 weeks, antiandrogen treatment with flutamide for 4 weeks, and corticosteroids, radiotherapy, or chemotherapy for 4 weeks. Due to potential blood pressure response with atrasentan, antihypertensive agents such as beta blockers, angiotensin-converting enzyme inhibitors, calcium channel blockers, or alpha blockers were discontinued at least 7 days before the first atrasentan dose, and diuretic agents were tapered off. To begin treatment in the study, patients’ blood pressure had to be under adequate control.

The study was approved by the hospital review board for studies on human subjects at The University Hospital Utrecht (Utrecht, the Netherlands) and conformed to the guidelines of the Declaration of Helsinki. All patients were required to give written informed consent before undergoing any study-related procedures.

This dose-escalation study was designed to enroll successive cohorts of patients (3 patients/cohort), each to be started on a fixed dose of atrasentan. The initial protocol specified 10 mg/day for the first cohort. Planned dose levels for subsequent cohorts were 20, 30, 45, 60, 75, and 95 mg/day. The protocol was later amended to include 2.5- and 5-mg doses so that more safety and pharmacokinetic data could be collected at these lower doses. Atrasentan was administered once daily on day 1 and days 3–28; the drug was withheld on day 2 to allow for pharmacokinetic analyses over a 48-h period. The drug was supplied as capsules (2.5, 5, 10, and 25 mg) for oral administration. The patients took the capsules under fasting conditions on study days 1, 7, 14, and 28.

Dose escalation was to be halted when the MTD was reached; MTD was defined as one dose level below that at which DLT was observed in one-third or more of the patients. If one of the three patients in a dose cohort experienced a DLT, three more patients were added to the cohort. If no further DLTs were observed in the group after 28 days, an additional cohort of three patients was enrolled at the next higher dose.

Toxicities were graded using the NCI Common Toxicity Criteria (version 2). If a NCI grade did not apply, the adverse event was graded as mild, moderate, or severe. DLT was defined as any atrasentan-related adverse event qualifying as NCI grade 3 or 4.

Health status assessments, conducted weekly, included physical examination, 12-lead ECG (days 7 and 28), blood chemistry, hematology, and urinalysis.

Blood samples were collected for pharmacokinetic analysis of atrasentan concentrations before the initial dose on days 1 and 28 and at the following intervals thereafter: 15, 30, and 45 min and 1, 1.5, 2, 4, 6, 8, 12, 16, 24, 30, 36, and 48 h. Blood samples were collected over a dosing interval on days 1 and 28 for determination of the plasma concentrations of iET. In addition, a predose sample was collected on day 14. Patients were required to fast for 8 h before dosing and, on days 1 and 28, for at least 2 h afterward.

Atrasentan plasma concentrations were determined using a validated liquid chromatography method with fluorescence detection (25). iET plasma concentrations were measured using a validated enzyme-linked immunoassay (26). Measurements included C_max, C_min, T_max, and AUC. The AUC of atrasentan was determined from zero to infinite time (AUC_∞) after dosing on day 1, with extrapolation after the last measurable concentration. The terminal-phase elimination rate constant (β) was obtained using a least squares linear regression analysis of the terminal log-linear portion of the plasma concentration-time profile. At least three concentration-time data points were used to determine β. t_1/2 was calculated using ln(2)/β. Apparent oral clearance (CL/F, where F = bioavailability) was calculated by dividing the dose by the corresponding AUC. Volume of distribution (V_β/F) was calculated by dividing CL/F by β. For iET, AUC was determined over 24 h on both days 1 and 28 (AUC_0–24).

To test for linearity and dose-proportionality of pharmacokinetic parameters of β, T_max, and dose-normalized C_max, C_min, and AUC, analyses of covariance were performed with body weight as the covariate. In these analyses, the primary test was a linear contrast orthogonal to the dose. To test for time independence, analyses of variance with dose as the only factor were performed on the difference between day 1 and day 28 pharmacokinetic parameters. The same analyses were performed for the iET parameters. All analyses used the procedures of the General Linear Model of SAS/STAT version 6.12 (SAS Institute, Gary, NC). P < 0.05 was considered to be statistically significant.

This clinical trial (excluding the extension study) was conducted between April 1997 and April 1999. The demographic characteristics of patients, types of cancer, and assignment to cohorts are shown in Table 1. Of the 39 patients enrolled, 3 were female (8%), and 36 were male (92%); all were Caucasian. Their ages ranged from 54 to 84 years (mean age, 69.2 years), and their mean baseline Eastern Cooperative Oncology Group score was 0.54; the median score was 0.

Of the 39 patients, 36 completed the study; 3 patients were withdrawn before day 28 due to disease progression (1 patient each from the 45-, 60-, and 95-mg cohorts). A 56-year-old woman with ovarian cancer discontinued the study drug (45 mg) after 7 days due to NCI grade 3 nausea and vomiting that the investigator attributed to disease progression. A 59-year-old man with prostate cancer discontinued 60 mg of atrasentan on day 26 because of hematuria attributed to disease progression. A 73-year-old man with prostate cancer discontinued 95 mg of atrasentan after 18 days due to increased pain from spinal cord compression secondary to tumor progression.

Atrasentan was generally well tolerated in all patients; there were no deaths during the study. A 62-year-old man receiving 75 mg of atrasentan experienced supraventricular tachycardia 28 h after his last dose of the study drug; his pulse was 212 beats/min, and blood pressure was 95/45 mm Hg. After carotid massage and administration of i.v. verapamil and a plasma volume expander, the patient made a complete recovery and had no further episodes.

The most commonly reported adverse events were rhinitis or nasal congestion (29 patients, 74%), headache (24 patients, 62%), peripheral edema (22 patients, 56%), nausea (12 patients, 31%), and dyspnea (12 patients, 31%; Table 2). The majority of adverse events were rated as mild or as NCI grade 2 or lower. The 29 cases of rhinitis and nasal congestion were mild in 28 cases and moderate in 1 case; symptoms were tolerated without treatment or relieved with over-the-counter medications such as antihistamines or oxymetazoline nasal spray. Headaches were mild in 17 cases and moderate in 7 cases and were tolerated without treatment or controlled with paracetamol analgesic therapy. Six of the seven patients with moderate headaches received atrasentan doses of 60 mg or higher. No patient discontinued treatment as a result of rhinitis or headaches. Peripheral edema was mild in 14 cases, moderate in 7 cases, and severe in 1 case. Six of the seven patients with grade 1 edema received atrasentan at doses of 45 mg or higher. A 77-year-old patient with grade 2 edema of the lower legs (rated as severe) had his dose reduced from 45 to 30 mg and went on to complete the study at this dose level.

Patients exhibited an overall mean decrease in blood pressure during the study. Systolic pressure decreased by −2.9 mm Hg during week 1 (P = 0.233), reaching an overall reduction of −9.7 mm Hg at week 4 (P = 0.002). Diastolic pressure decreased −6.44 mm Hg during week 1 (P = 0.002) and −6.0 mm Hg by week 4 (P = 0.006). Pressures returned to baseline levels within 1 week after dosing.

Within 1 week after atrasentan administration was begun, a mean decrease in hemoglobin of 1 g/dl (0.67 mm) was observed across all dose groups. This was accompanied by a 3% mean decrease in hematocrit and a 0.35 × 10¹²/liter mean decrease in RBC count across all dose groups. The changes did not appear to be progressive over time. Many patients had pretreatment anemia due to their underlying disease. Four patients received transfusions for anemia. In all cases, the anemia was considered by the investigator to be unrelated to study drug administration. There was also a decrease in WBC count from baseline over the 4 weeks of the study. The mean decrease in total protein for all dose groups combined ranged from −0.22 g/dl at week 2 to −0.48 g/dl at week 4. Recovery to the baseline values was observed for all laboratory tests as early as 1 week after dosing. Except for the blood samples collected for pharmacokinetic analyses, there was no evidence of blood loss, i.e., no evidence of change in bilirubin level, platelet number, peripheral blood smears, or urinalysis values.

In general, the PK of atrasentan were linear (dose-proportional) and time independent (Table 3). Dose-normalized AUC did not vary with dose on either day 1 or day 28 (P > 0.67). Dose-normalized AUC_0–24 on day 28 did not differ from dose-normalized AUC_∞ on day 1 (P = 0.66). Atrasentan PK were characterized by a global mean ± SD CL/F and V_β/F of 24 ± 15 liters/h and 726 ± 477 liters, respectively. PK varied considerably among individuals. On day 1, percent coefficient of radiation values for CL/F and V_β/F were 46% and 45%, respectively; on day 28, these values were 73% and 83%, respectively.

Atrasentan was rapidly absorbed, with mean T_max values ranging from 0.3 to 1.6 h. No consistent trend with dose was observed with T_max, which increased with increasing dose on day 1 (P = 0.02), but not on day 28 (P = 0.54). There was no difference in T_max between days 1 and day 28 (P = 0.89). A statistically significant increase in dose-normalized atrasentan C_max with increasing dose was observed on both days 1 and 28, in part due to relatively low C_max values for the lowest dose (2.5 mg). Mean C_min values on day 28 increased roughly dose-proportionally for the dose range 2.5–45 mg (slope = 1.0; R² = 0.94) and somewhat less than dose-proportionally at higher doses. Across all dose groups, a statistically significant trend with dose in dose-normalized C_min was observed (P = 0.02).

After peak levels of atrasentan were reached, plasma concentrations declined biexponentially, with a terminal phase half-life averaging 21 h. No consistent trend with dose was observed for the elimination rate constant (β), which increased with increasing dose on day 1 (P < 0.01), but not on day 28 (P = 0.40). There was no difference between days 1 and 28 (P = 0.84). Atrasentan concentrations accumulated with multiple dosing; on day 28, they were about double the levels observed on day 1, which is the expected accumulation for a compound with a 21-h half-life dosed once daily (Figs. 1 and 2). Steady-state plasma concentrations of atrasentan would be expected after 5 days of dosing (five half-lives). Predose concentrations of atrasentan on days 14 and 28 were similar (P = 0.08). Body weight was not a statistically significant covariate in the analysis of any parameter except T_max on day 1 (P = 0.04).

iET concentrations increased over time with atrasentan dosing and were about 40% greater on day 28 than on day 1 (Figs. 3 and 4). This increase appeared to be consistent across dose groups. iET C_max, C_min, and AUC_0–24 were all greater on day 28 than on day 1 (P < 0.01; Table 3). Predose iET concentrations on days 14 and 28 were similar (P = 0.79). On either day 1 or day 28, iET AUC was independent of atrasentan dose (P > 0.67), as was C_min on day 28 (P = 0.96). As expected, iET C_min was also independent of atrasentan dose on day 1 (P = 0.21). iET C_max increased somewhat with increasing dose on day 28 (P < 0.01), but not on day 1 (P = 0.25). iET T_max decreased with increasing dose on both days 1 and 28 (P ≤ 0.03). Body weight was not a statistically significant covariate in the analysis of any iET parameter (P > 0.14). The iET results from this study were generally similar to those reported previously for atrasentan doses of 10 to 75 mg once daily administered to adults with refractory adenocarcinoma (27).

Atrasentan was well tolerated in patients with refractory adenocarcinomas at doses of up to 95 mg/day. The protocol-defined MTD was not achieved in the study, and none of the adverse effects was dose-limiting. The most frequently reported adverse effects of headache, rhinitis, and peripheral edema were attributable to the vasoactive pharmacology of this class of compound. These effects were mild to moderate in intensity, reversible, and, when necessary, readily controlled with symptomatic treatment.

Most cases of headache and peripheral edema occurred in patients taking doses of 60 mg of atrasentan or higher. In another Phase I study of atrasentan in oncology patients, headaches of moderate intensity occurred at the 75-mg dose; whereas they were not considered dose-limiting events, they were of sufficient severity and duration to halt dose escalation at 75 mg (27). Headaches were the DLT in healthy subjects receiving atrasentan; healthy male volunteers were unable to tolerate headaches that occurred at the 40-mg dose (26, 28). In healthy volunteers, the headaches were typically migrainous and were severe enough to necessitate bed rest and cause photophobia, nausea, and vomiting. Headaches of this severity were not observed in this study. Headaches have also been reported with administration of other ET antagonists (29, 30). Changes in laboratory values were consistent with a mild, self-limited hemodilution effect of atrasentan associated with increased body weight, a phenomenon that has also been noted in healthy volunteers (28). No differences in atrasentan PK were found between healthy volunteers and oncology patients to explain the contrast in tolerability.

The observed effects of atrasentan are consistent with the pharmacological activity of ET_A receptor antagonists. The vasodilatory activity of atrasentan could have caused the headaches, rhinitis, hemodilution effect, peripheral edema, and blood pressure changes seen in our patients. In two recent studies of nonselective ET antagonists in patients with heart failure, high doses of these agents sometimes exacerbated cardiac symptoms when administered without an appropriate dose titration schedule (31, 32). Until more data are available on the administration of ET antagonists in patients with underlying cardiovascular conditions, these agents should be used with the same caution observed for calcium channel blockers in patients with a history of heart failure.

Pharmacokinetic parameters in oncology patients in this study were linear, dose-proportional, and independent of time over the range of 2.5–95 mg/day. Steady-state plasma concentrations consistent with biological activity in preclinical models were achieved; for example, the mean unbound C_min for the 10-mg regimen (accounting for 98.8% binding to plasma proteins) was approximately 8-fold greater than the K_i for the ET_A receptor (23). For this reason, dosing above 95 mg/day appeared to be unwarranted, especially because the severity of the headaches seemed to increase at the higher doses. Across atrasentan dose groups, plasma iET concentrations increased modestly (about 40%) during dosing. Whether the increases in iET concentrations over baseline were due to ligand displacement or antagonism of an inhibitory feedback pathway is not known.

ET-1, a peptide expressed by neoplastic cells and acting via the ET_A receptor, has been identified in prostate and other cancers (33). Atrasentan, a selective ET_A receptor antagonist, has the potential to attenuate the progression and morbidity of numerous neoplastic diseases through its action on the ET system. With generally mild to moderate side effects for doses of up to 95 mg/day, predictable PK, and the advantage of once-daily oral dosing, atrasentan is a targeted cytostatic agent with significant therapeutic potential in patients with prostate and other cancers.

We thank Annelies Epping and Rolien Kronemeijer, data managers and research nurses from the Pharmacology Unit, University Hospital Utrecht, for important contributions to this research. We also thank the patients and healthcare providers who participated in this clinical trial.