Investigational New Drugs

, Volume 25, Issue 5, pp 445–451

Phase I dose escalation and pharmacokinetic study of AZD2171, an inhibitor of the vascular endothelial growth factor receptor tyrosine kinase, in patients with hormone refractory prostate cancer (HRPC)


    • Urologic Oncology ProgramUCSF Comprehensive Cancer Center
  • Walter M. Stadler
    • University of Chicago
  • Bruce Roth
    • Vanderbilt University Medical Center
  • Douglass Hutcheon
    • Urologic Oncology ProgramUCSF Comprehensive Cancer Center
  • Shauna Conry
    • Urologic Oncology ProgramUCSF Comprehensive Cancer Center
  • Thomas Puchalski
    • AstraZeneca
  • Charles Morris
    • AstraZeneca
  • Eric J. Small
    • Urologic Oncology ProgramUCSF Comprehensive Cancer Center

DOI: 10.1007/s10637-007-9050-y

Cite this article as:
Ryan, C.J., Stadler, W.M., Roth, B. et al. Invest New Drugs (2007) 25: 445. doi:10.1007/s10637-007-9050-y


To explore the pharmacokinetics and tolerability of AZD2171, an inhibitor of vascular endothelial growth factor receptors 1 and 2, in patients with hormone refractory prostate cancer. Twenty-six patients received oral daily dosing of AZD2171 at 1, 2.5, 5, 10, 20, 30 mg. The maximum tolerated dose (MTD) was defined as the dose below that at which ≥33% of patients experienced a dose-limiting toxicity (DLT) within 21 days of initiating therapy. Pharmacokinetic analysis was performed. DLTs occurred at the 30 mg dose and included grade 3 events in three patients: fatigue (n = 3) and muscle weakness (n = 2). The pharmacokinetic profile revealed an effective half-life of approximately 27 h. At steady state, the unbound drug concentration was 4.4 times above the concentration required to inhibit endothelial cell proliferation in vitro. Four patients experienced PSA reductions within 30 days following drug discontinuation (one on 2.5 mg, two on 20 mg and 1 on 30 mg). In two patients treated with 20 mg, post therapy PSA declines persisted for >17 months, despite a PSA increase on therapy. Resolution of adenopathy occurred in one patient persisting for >17 months. Plasma concentrations were maximum 2–8 h post dosing with an overall median value of 2 h. The dose of 20 mg daily was declared as the MTD. One objective response and several PSA declines following the discontinuation of therapy for toxicity suggest that evidence of clinical efficacy may be delayed. While further study is indicated, careful attention must be paid to the novel toxicities of this agent with prolonged dosing.


Prostate cancerHormone refractory prostate cancerAngiogenesis inhibitorsPharmacokinetics


Hormone refractory prostate cancer (HRPC) poses a significant clinical challenge. Although secondary hormonal therapies and chemotherapy are capable of inducing responses and, in the case of docetaxel based chemotherapy, prolonging life, there is a critical need to develop newer therapies for this common malignancy.

A significant body of work supports the hypothesis that tumor growth is dependent on the development of a neovasculature, mediated predominantly by tumor-associated vascular endothelial growth factor (VEGF) and its family of receptors on endothelial cells. Further, recently developed hypotheses suggest that potent inhibition of the vascular permeability function of VEGF may lead to an improved delivery of systemic chemotherapy to the tumor compartment, thus raising the possibility that a VEGF receptor inhibitor can be added to standard therapies [1].

The contribution of angiogenesis to the morbidity of prostate cancer is established. Microvessel density (MVD) is higher in the primary tumors of patients with concurrent metastatic disease as compared to those with localized disease and no concomitant metastases [2]. High urine levels of VEGF [3] and plasma [4] are associated with poor survival in patients with HRPC. The expression of VEGF receptor type 2 (VEGFR2 or kinase insert domain-containing receptor, KDR) in transgenic tumors has been implicated in the development of high grade, poorly differentiated tumors [5]. These data cumulatively support the hypothesis that angiogenesis is a major mediator of tumor progression and death, and justify the study of angiogenesis inhibitors in HRPC.

Clinical evaluation of angiogenesis inhibitors is complicated by a number of properties observed in preclinical models. This includes their possible growth inhibitory as opposed to cytotoxic effects, their possible utility in combination despite little to no single agent activity, and the possible disassociation between anti-tumor activity and PSA secretion. To begin to explore these issues early in drug development, and because prostate cancer patients are typically older [6] than the usual phase I enrolee and have specific comorbidities as a result of long term androgen ablation, a phase I study confined to prostate cancer patients only was initiated to evaluate the VEGF receptor tyrosine kinase inhibitor AZD2171.

AZD2171 inhibits both VEGFR1 (IC50 = 0.005 μm) and VEGFR2 (IC50<0.002 μm) receptor subtypes [7]. The primary objective of this Phase I multicenter, dose-escalation trial was to assess the safety and tolerability of AZD2171 in HRPC administered via once daily oral dosing. Secondary objectives included determining the single and multiple dose pharmacokinetic (PK) profile of AZD2171, and exploratory assessments of efficacy.

Patients and methods

The study was approved by the investigational review boards (IRB) at all three participating institutions. Eligible patients included men with histologically proven prostate cancer that had progressed despite androgen deprivation therapy (ADT) consisting of a leutinizing hormone releasing hormone (LHRH) agonist, combined with an oral antiandrogen. All patients were required to experience progression of disease following antiandrogen withdrawal, and as defined by consensus criteria [8]. Exclusion criteria included prior anti-angiogenic therapy and greater than one prior chemotherapy. Pretherapy staging consisted of a bone scan, a chest x-ray and abdomen/ pelvis computed tomography. Treatment consisted of continuous oral daily dosing according to a planned dose escalation of 1, 2.5, 5, 10, 20, 30, 45 and 60 mg. Each dose cohort initially consisted of three patients. The maximum tolerated dose (MTD) was defined as the dose level immediately below the dose at which ≥33% of patients in a cohort experienced a dose-limiting toxicity (DLT—defined as any CTCAE grade ≥3 not clearly related to underlying disease but considered by the investigator to be related to AZD2171 treatment) within the initial 21 days of therapy. The following dose escalation rules were utilized: If <33% of subjects in a given cohort experienced a DLT, then dose escalation will continue, if DLT is observed in ≥33 to ≤50% of subjects, then the cohort will be expanded to include at least three further subjects. If DLT is observed in >50% of subjects then the dose will be considered above the MTD and dose escalation will stop. Serial pharmacokinetics and blood pressure measurements were performed on day 1 and 21. Dosing beyond 28 days was allowed provided withdrawal criteria were not met and, in the investigator’s opinion, they were receiving benefit from the drug.

Exploratory efficacy assessments

Prostate-specific antigen (PSA) was assessed at Screening, on days 1 and 21 of all cycles (prior to dosing), and at withdrawal. The percentage of patients with a PSA decline of ≥50% from baseline confirmed by a second PSA value at least 4 weeks later was calculated. Initial staging evaluation consisted of PSA, bone scan and CT scan of the abdomen and pelvis at the start of therapy, following cycle 1 and every 12 weeks thereafter. Objective tumor response was assessed according to RECIST criteria.


Specimen collection and quantitation

Plasma pharmacokinetics (PK) of AZD2171 were characterized following a single dose with a 24 h washout and after 21 day of once-daily consecutive multiple doses. Blood samples were obtained at pre-dose and 1, 2, 3, 4, 6, 8 and 24 h post-dose on day 1 and 21. Plasma concentrations of AZD2171 were determined by solid-phase extraction followed by reverse-phase liquid chromatography and detection by tandem mass spectrometry. The method was validated to cover the concentration range 0.1–500 ng/ml, with 0.1 ng/ml adopted as the lower limit of quantification.

PK data analysis

Single and multiple dose PK parameters of AZD2171 were determined by non-compartmental methods using WinNonlin (Professional Version; Scientific Consulting, Apex, NC). Cmax, Cmax ssCmin ss and tmax were determined by visual inspection of the plasma concentration time profile. The area under the plasma concentration-time curve from zero to 24 h following a single dose (AUC(0–24)) and the area under the plasma concentration-time curve at steady state (AUCss) were estimated by a linear trapezoidal algorithm to the last data point of the 24 h dosing interval. Attainment of steady state was assessed by visual inspection of the pre-dose plasma concentration time samples obtained on study days 8, 15, and 21. To estimate the degree of accumulation following repeated doses the accumulation ratio (Rac) was calculated as the ratio of AUCss–AUC(0–24).

Pharmacokinetic/pharmacodynamic relationships

Correlative studies included serial measurement of markers of bone resorption alkaline phosphatase, N-terminal propeptide of type I procollagen (PINP), N-telopeptide (NTx), and the NTx/Creatinine ratio as well as serum VEGF as well as serum thyroxine and thyroid stimulating hormone levels. Linear regression models were used to explore the relationship between pharmacokinetic parameters (AUCss, Css,max and Css,min at Cycle 1, Day 21) and PSA, plasma VEGF, and supine mean arterial pressure.

Statistical methods

Safety analysis was performed on all patients who received at least one dose of study medication. Patients in the safety analysis set were included in the efficacy and pharmacokinetic analyses if they were evaluable for each specific efficacy or pharmacokinetic parameter. There was no formal statistical analyses for safety, tolerability and efficacy. The treatment groups were compared descriptively using summary statistics and percentage counts.


Patient characteristics

A total of 26 patients were treated between March, 2004 and October, 2005. All had received prior ADT, ten patients (38.5%) had received prior chemotherapy. Patient characteristics are summarized in Table 1. AZD2171 was administered at daily doses of 1, 2.5, 5, 10 (all n = 3), 20 (n = 10) and 30 mg (n = 4).
Table 1

Summary of demographic and baseline characteristics of all patients treated with AZD2171



Age, median (range)

70 years (50.1–89.7)

PSA, median (range)

54.92 ng/ml (6.2–1051.3)

Alkaline phosphatase, median (range)

96 U/L (25.2–182.4)

Hemoglobin, median (range)

12.4 g/dl (9.8–14.5)

Gleason score, number (%)













Race, number (%)







Extent of disease, number (%)

 Locally advanced




 Not recorded


 Prior chemotherapy, number (%)


 Performance status (ECOG 0/1), number (%)

0 = 16(61.5%)

1 = 10(38.5%)

Table 2

Pharmacokinetic parameters of AZD2171 following a single dose (cycle 1, day 1)


Summary statistics

AZD2171 dose (mg)







(n = 3)

(n = 3)

(n = 3)

(n = 3)

(n = 9)

(n = 4)

Cmax (ng/ml)















tmax (h)























(ng h/ml)








AUC(0–24), area under the plasma concentration-time curve from zero to 24 h following a single dose; Cmax, maximum plasma (peak) drug concentration after single dose administration; CV, coefficient of variation; gmean, geometric mean; tmax, time to reach peak or maximum concentration or maximum response following drug administration.

Dose limiting toxicities (DLT) and maximum tolerated dose (MTD)

The maximum duration of therapy was 190 days. The median duration on treatment for all patients was 43 days (range 1–190; two cycles) and the median number of days of exposure for doses 1, 2.5. 5. 10. 20 and 30 mg were 22 (range 22–43), 43 (range 43–70), 22 (range 21–63), 22(range 22–84), 63(range 1–190), 30 (range 7–53), respectively. Dysphonia (N = 8 [31%]), fatigue (N = 8 8 [31%]), headache (N = 8 8 [31%]), hypertension (N = 8 8 [31%]), and diarrhea (N = 8 6 [23%]) were the most commonly reported AEs considered causally related to treatment. In total, 12 AEs leading to discontinuation occurred in seven patients (27%). Three patients treated at the 30 mg dose experienced DLTs consisting of grade 3 fatigue (N = 3), muscle weakness (N = 2), hypertension (N = 1) and dehydration (N = 1). The muscle weakness was confined to proximal musculature not associated with elevations of creatine phophokinase. Based on this toxicity profile, it was determined that long-term dosing at 30 mg would not be tolerable and dose of 20 mg daily was therefore declared as the MTD.

Because the DLT of muscle weakness observed did not reflect the safety profile observed from all other studies to date involving approximately 200 patients treated with AZD2171, IRB permission was granted to accrue additional patients at the 20 mg dose to a total of ten patients at this dose. In total, ten patients were treated at the 20 mg dose for a median of 63 days (range 1–190 days). Drug related adverse events of any grade occurred in 9 (90%) patients and discontinuation occurred in 4 (40%) due to hypertension, hematuria, transient ischemic attack and the combination of fatigue and anorexia all of which occurred in one patient each. The CTCAE grade 3 adverse events observed in the 20 mg dose that were deemed to be possibly drug related included three hypertensive episodes, one muscle weakness similar to that observed in the two patients treated at the 30 mg dose and one transient ischemic attack). One patient treated at 20 mg (10%) developed an asymptomatic decline in serum thyroxine (grade 1—no replacement therapy indicated). No deaths occurred while on study therapy.

Pharmacokinetic data

Single dose pharmacokinetics are summarised in Table 2. Plasma concentrations reached their maximal value from 2 to 8 h post dosing with an overall median value of 2 h. Day 21 pharmacokinetics are summarised in Table 3. Plasma concentrations reached their maximal value from 1 to 6 h post dosing with an overall median value of 3 h. The arithmetic mean accumulation ratio for the 20 mg dose level of 2.15 is equal to an effective half-life of approximately 27 h. Pre-dose plasma concentrations are supportive of steady-state being achieved on cycle 1, day 8 or after 7 days of repeated once-daily dosing.
Table 3

Pharmacokinetic parameters of AZD2171 following 21 days of once-daily dosing (cycle 1, day 21)


Summary statistics

AZD2171 dose (mg)a






(n = 3)

(n = 1)

(n = 3)

(n = 2)

(n = 6)

Css,min (ng/ml)













Css,max (ng/ml)













tmax (h)













AUCss (ng h/ml)


























AUCss, area under plasma concentration-time curve during any dosing interval at steady state; Cmax ss, maximum steady-state drug concentration in plasma during dosing interval; Cmin ss, minimum steady-state drug concentration in plasma during dosing interval; CV, coefficient of variation; gmean, geometric mean; NC, non calculable; Rac, accumulation ratio; SD, standard deviation; tmax, time to reach peak or maximum concentration or maximum response following drug administration.

aMultiple-dose pharmacokinetic parameter data were not available for the AZD2171 30 mg cohort.

bCalculated as cycle 1, day 21 AUCss/cycle 1, day 1 AUC(0–24)

cArithmetic mean

Pharmacodynamic testing

No significant trends were observed for markers of bone resorption: alkaline phosphatase, N-terminal propeptide of type I procollagen (PINP), N-telopeptide (NTx), and the NTx/Creatinine ratio. Plasma VEGF levels increased from baseline within 8 h following initial dosing in most patients receiving 5 mg or higher dose and remained elevated over baseline on day 21.

No significant PK/pharmacodynamic (PD) relationships were observed between the change from baseline cycle 1, day 21 plasma VEGF and cycle 1, day 21 PSA with steady-state pharmacokinetic parameters of AUCss, Cmax ss or Cmin ss. A significant relationship was observed between the change from baseline supine mean arterial pressure (MAP) and the steady-state pharmacokinetic parameter of Cmin ss (p = 0.05, R2 = 0.48).

Clinical outcomes

Nineteen patients completed at least 28 days of therapy and were thus eligible for an analysis of response. Seven patients did not complete 28 days of therapy due to toxicity. Of the 19 evaluable patients, none achieved a PSA decline of ≥50% during protocol therapy. Four patients (15.4%) experienced reductions in PSA between 10 and 50% while on study. No objective responses were observed during the study period in the 14 patients with bi-dimensionally measurable disease. Four (28.6%) patients had RECIST defined best response of SD. Three of these four (75%) were in the 20 mg cohort.

Clinical responses following discontinuation of study therapy

Reductions in PSA and one objective response in extensive retroperitoneal nodal disease were observed in patients following the discontinuation of study therapy, without intervening subsequent therapy. In total, four patients treated with AZD2171 experienced PSA reductions within 30 days following discontinuation of the drug (one patient on 2.5 mg, two on 20 mg and 1 on 30 mg). In two of the patients (2.5 and 30 mg) the PSA had risen on study therapy (by +173 and +68%, respectively), and declined at the time of a 30 day post study measurement to a level that remained above the baseline (Pre-AZD2171) level to +44 and +33%, respectively—(the serial PSA levels of the patient treated at 30 mg is shown in Fig. 1c).

Two patients have experienced long term reductions in PSA following the discontinuation of therapy, despite experiencing dose limiting toxicity while on therapy. The first such patient was treated in the 30 mg cohort and developed a 27% reduction in PSA during 15 days on study therapy before discontinuing due to muscle weakness, fatigue and hypertension. Subsequently the PSA declined to 58% of baseline 30 days after discontinuing therapy. The patient has had a PSA that has been maintained within a narrow range (53–92.6 ng/ml with a PSA doubling time of 99 months) for the past 15 months (Fig. 1a) with no subsequent therapy during which time the CT scan and Bone scan remained stable.

The second patient was treated on the 20 mg dose and received 28 days of therapy during which the PSA rose from 68 to 73.89 ng/ml prior to discontinuing due to grade 3 muscle weakness and hypertension. Thirty days following therapy discontinuation the PSA had declined to 4.92 ng/ml, a 93% reduction that was confirmed with serial measurements and correlated with a resolution of retroperitoneal adenopathy. The patient has received no further therapy for the past 17 months, and the PSA has not risen above 5.38. (Fig. 1b) All toxicity has resolved.


In the present trial of AZD2171 in patients with HRPC the MTD, and recommended phase II dose, was 20 mg/day. Dose limiting toxicities of hypertension and muscle weakness were observed in three of four patients treated at 30 mg/day. This dose recommendation differs from prior phase I studies of AZD 2171 in other tumor types which have resulted in higher MTDs (up to 45 mg daily). Dosing at 20 mg daily was associated with acceptable toxicity in the majority of patients and resulted in a plasma concentration of drug that is well above the concentration required to inhibit endothelial cell growth.

Although hypertension was observed in the 20 mg cohort, this was an expected toxicity and not generally dose limiting, as no patient developed HTN despite maximal medication or required withdrawal from the study as a result of HTN. The presence of muscle weakness in five patients, however, is a heretofore undescribed toxicity of this drug and of VEGF receptor tyrosine kinase inhibitors in general. The clinical syndrome observed in patients typically involved proximal weakness which developed within 15–21 days of therapy, resulted in an inability to perform activities of daily living, and readily resolved (within 3–7 days) upon discontinuation of the therapy. Markers of muscle inflammation such as creatine phosphokinase were not elevated. No association with prior or current exposure to agents such as statins was observed. Relatively little is known regarding the interface of angiogenesis, androgen deprivation and skeletal muscle. In animal models of spinal and bulbar muscular atrophy (SBMA), a neurodegenerative disorder characterized by muscle weakness and lower motor neuron degeneration, an inverse association has been made between AR mutation and VEGF signalling, i.e. motor neuron cell death is accelerated in the absence of VEGF signalling [9]. These data suggest that a link may exist between the dual insults of VEGF receptor inhibition and the long-term muscle wasting effects of androgen deprivation.

Notable clinical activity was observed in this study. Interestingly, the significant PSA changes and bidimensional objective response in patients on this study occurred after discontinuation of the drug. In three patients, PSA levels climbed during the course of study therapy and declined following discontinuation of the drug (Fig. 1a–c). In one case, this was associated with a resolution of retroperitoneal adenopathy, and a drop in PSA by >80%, a response which was maintained for >17 months. A similar PSA effect was observed in two other patients who did not have measurable soft tissue lesions (Fig. 1).
Fig. 1

Delayed PSA and objective responses in patients following Discontinuation of Therapy with AZD2171. a (20 mg), PSA and scans have remained stable for 17 months following discontinuation. b (20 mg) therapy was discontinued due to muscle weakness, retroperitoneal adenopathy has resolved and no further therapy has been given for 17 months. c (30 mg), AZD2171 was discontinued due to myalgias and hypertension, PSA remained at or below baseline values for 5 months prior to the development of disease progression

Prior studies of the VEGF receptor tyrosine kinase inhibitors have suggested similar discordance between PSA and objective findings. In one report with sunitinib, HRPC patient(s) experienced a decline in PSA with an improvement in bone scan [10]. The present study, to our knowledge, is the first to demonstrate a prolonged objective and serologic response that developed after the discontinuation of a drug of this class. While this phenomenon may reflect a withdrawal response similar to that described with non-steroidal and steroidal antiandrogens [11], it is also possible that an antitumor activity of AZD2171 did not become apparent until AZD2171 was discontinued was masked by a concomitant increase in PSA secretion. These mechanisms require further study. These observations further suggest that the utility of changes in PSA in response to therapy are agent specific, and may not be an adequate intermediate marker of outcome for studies of angiogenesis inhibitors. Given these factors, subsequent evaluation of AZD2171 in HRPC would best be accomplished utilizing a time to progression endpoint, rather than utilizing changes in PSA as the primary endpoint.

Pharmacokinetic data show that AZD2171 is orally available following both single and multiple dose administration. The tmax for both single and multiple doses was attained in a majority of patients between 2 and 4 h post dosing. Steady-state plasma concentrations were attained after approximately 7 days of repeated once-daily dosing. The observed accumulation ratio supports an effective half-life for 20 mg AZD2171 of approximately 26.6 h. Following multiple oral doses of 20 mg, the unbound Cmin ss was 4.44 times above the HUVEC proliferation IC50. This suggests that doses of 20 mg AZD2171 and above will produce plasma concentrations throughout the entire 24 h dosing interval at the steady-state levels needed for inhibition of a target in vitro enzyme. The lack of available data from the 30 mg dose preclude any conclusions regarding the change in drug level from day 1 to 21 or the Rac in this cohort of patients. The PK findings in this study are similar to those observed in a single agent solid tumor Phase I study.

In an attempt to better understand the effect of AZD2171 on BP, VEGF and PSA changes, PK/PD modelling of the appropriate data was conducted. No significant PK/PD relationships were observed between the change from baseline Cycle 1, Day 21 plasma VEGF and Cycle 1, Day 21 PSA with any steady-state pharmacokinetic parameters. Although a significant relationship was observed between the change from baseline supine MAP and the steady-state pharmacokinetic parameter of Cmin ss, there was no significant relationship with AUCss or Cmax ss. These data fail to definitively define the relationship between AZD2171 plasma exposure and change in BP.

Further development of oral VEGF receptor tyrosine kinase inhibitors in prostate cancer will need to address the unique toxicity profile in elderly men treated with prior androgen deprivation. Strategies to circumvent the toxicities may include intermittent dosing or selective dose reductions at the early sign of drug related toxicity may be required to ensure maximum tolerability with chronic dosing. Further evaluation of the interface between muscle toxicity, androgen deprivation and VEGF inhibition (particularly VEGFR2) may also be useful.


The antiangiogenic agent AZD2171 was considered to be well tolerated at doses ≤20 mg/day in this study of patients with HRPC. PSA changes noted and the off-study objective response suggest that further investigation of the treatment of prostate cancer with AZD2171 is warranted, as is a more thorough evaluation of unique toxicities in this patient population. Muscle weakness is appears to be a novel toxicity that is relatively specific for this agent in patients with HRPC and requires further study to clarify its mechanism of action.

Copyright information

© Springer Science+Business Media, LLC 2007