Perifosine (NSC 639966) is a synthetic, substituted heterocyclic alkylphosphocholine that acts primarily at the cell membrane targeting signal transduction pathways. Early clinical trials were limited because of dose-limiting gastrointestinal toxicity, and parenteral dosing of this class of agents is not possible because of their hemolytic properties; therefore, related compounds with an improved therapeutic index were developed. Toxicity was minimized and efficacy improved by using a loading dose/maintenance dose schedule, and therefore, this schedule was carried into clinical trials. This phase I trial enrolled 42 patients with incurable solid malignancies. The starting doses were 100 mg p.o. × four doses (every 6 hours) load followed by a 50 mg p.o. once daily maintenance dose with escalation of either component in successive dose levels. No treatment related deaths occurred. The maximum-tolerated dose was determined to be 150 mg p.o. × four doses load and 100 mg p.o. once daily maintenance. Dose-limiting toxicities such as nausea, diarrhea, dehydration, and fatigue were seen early during the loading phase and were surmountable with the use of prophylactic 5-HT3 receptor antagonists, dexamethasone, and loperamide. Toxicities during the chronic phase were difficult to manage and, given that pharmacokinetic data showed biologically active serum concentrations (based on preclinical data), raised the question of less frequent maintenance dosing. Pharmacokinetic data confirmed the maintenance of stable drug levels with chronic dosing and the long half-life. One partial response was seen, as were multiple patients with stable disease beyond course 2. These results suggest perifosine activity in sarcoma and perhaps renal cell carcinoma (stable disease in two patients who continued for 6 and 14 courses), thus justifying additional investigation of this agent in a phase II sarcoma trial.
Perifosine (NSC 639966) is a synthetic, substituted heterocyclic alkylphosphocholine analogue that acts primarily at the cell membrane targeting signal transduction pathways. The first phosphocholine compounds synthesized were analogues of two lysophosphatidylcholine. Subsequent work by Hilgard et al. (1) in 1993 showed that the glycerol moiety was not an essential structural element and that alkylphosphocholines exerted similar antitumor effects. Hexadecyclphosphocholine (miltefosine) emerged as the prototype of this class, showing antitumor activity in vitro against a variety of human cell lines [breast (MDA-MB-231), prostate (PC-3), colon (KM12), lung (HOP-92), and melanoma (M14)] and in vivo against 7,12-dimethylbenz(a)anthracene-induced rat mammary tumors (2) , PC3 xenografts (National Cancer Institute data), and others. Early clinical trials were limited because of dose-limiting gastrointestinal toxicity and parenteral dosing of this class of agents is not possible because of their hemolytic properties (2) ; therefore, related compounds with an improved therapeutic index were developed. Because some of the side effects observed with miltefosine were consistent with parasympathomimetic effects, it was theorized that the metabolite, phosphocholine, might be involved. Thus, synthetic efforts replaced the choline moiety with a heterocyclic nitrogen reducing emetogenic potential (3 , 4) .
Perifosine is a miltefosine derivative with a longer alkyl chain and a piperidine head group, which has improved bioavailability, decreased emetogenic capability, and more activity in preclinical models when compared with miltefosine (ref. 2 ; Fig. 1⇓ ). In vitro activity was seen in various assay systems against human KB (squamous), murine L1210 cell lines (leukemia), LU65A (lung), LNCaP (prostate), PC-1 (lung), hep-2 (larynx), and others (IB +5). In vitro, perifosine caused cell kill at low concentrations in several larynx (IB), colon (IB +2), breast (IB +2), and pancreas cell lines. The majority of in vivo data with perifosine comes from its usage in various doses and schedules in autochthonous tumors in Sprague Dawley rats bearing 7,12-dimethylbenz(a)anthracene-induced breast cancers (2 , 5) . Single oral (loading) dose therapy with high-dose perifosine (68.1 mg/kg) caused growth inhibition for ∼14 days (2) . Daily oral treatments (for 25 days) at lower doses (2.5 to 46.4 mg/kg) caused growth inhibition at all doses tested with the onset of response being dose related. Responses persisted for >20 days after termination of therapy without clear dose-response relationship over this range (2) . Given these results, a combination of loading dose followed by a lower daily maintenance dose schedule was explored and achieved complete regressions over the entire study period without compromising gastrointestinal tolerability. Hilgard et al. (2) note that there was no significant difference between a maintenance dose of 1.47 and 2.15 mg/kg, implying a lower, more tolerable maintenance dose could be used. This data, together with pharmacokinetic modeling done at the National Cancer Institute, prompted the schedules used in both phase I trials conducted in the United States.
Initially, three phase I trials of perifosine were undertaken to assess toxicity, pharmacokinetics, and limited pharmacodynamic parameters on various schedules. A study conducted by Crul et al. (6) at the Netherlands Cancer Institute used a daily oral administration schedule treating 22 patients with solid tumors at doses ranging from 50 to 350 mg/day for 3 weeks on followed by 1 week off. Dose-limiting toxicity (DLT) was not reached, but gastrointestinal complaints (nausea, 52%, grade 1 to 2; vomiting, 38%, grade 1 to 2; and diarrhea, 43%, grade 1 to 2) increasingly caused more patients at higher dose levels to stop therapy prematurely, despite the use of 5-HT3 antagonists. Fatigue occurred in 43% of patients (not graded) and was described as, “persistent in nature and … very inconvenient.” The maximum-tolerated dose (MTD) was established at 200 mg/day. Minor responses were seen in 2 of 16 evaluable patients with stable disease lasting 4 months (liver cell carcinoma, ocular melanoma). Pharmacokinetics showed trough levels of 2595 to 8195 ng/mL and the terminal t1/2 was calculated at 105 hours. Two United States-based trials were undertaken in parallel capitalizing on the preclinical loading dose/maintenance dose schedule—one at the National Cancer Institute and one at the University of Wisconsin Comprehensive Cancer Center.
Here, we present the results of the University of Wisconsin Comprehensive Cancer Center phase I clinical and pharmacokinetic trial. The objectives of the trial were to (a) determine the MTD with the combination of loading dose and maintenance dosing of perifosine, (b) identify the recommended starting dose for continued clinical trials at this schedule, (c) determine the qualitative and quantitative nature of the toxicities encountered, (d) determine the pharmacokinetic properties of perifosine on this schedule, (e) investigate the relationship between pharmacokinetic parameters and toxicity, and (f) to assess any changes in the MTD with prolonged (3 and 6 months) administration of this agent.
PATIENTS AND METHODS
Patients were eligible for this trial if they had histologically proven incurable malignancy (solid tumor or lymphoma) that was unresectable, refractory to standard therapy, or for which no known effective treatment exists. Patients had to be 18 years of age or older, have a Zubrod performance scale ≤ 2 with life expectancy ≥ 12 weeks, and could not have frequent emesis, poor alimentation, or recent weight loss > 10% current body weight. Additionally, patients had to be ≥4 weeks from most recent chemotherapy/radiation (≥6 weeks for mitomycin C) and recovered from toxicities. Adequate end organ function (WBC count ≥ 4000/mm3, platelet count ≥ 100,000/mm3, normal bilirubin, aspartate aminotransferase ≤ 2.5× upper limit of normal, normal serum creatinine, or calculated creatinine clearance ≥ 60 mL/min) was required, and no active central nervous system disease was allowed. Patients with recent major surgery (<21 days) pregnancy/lactation, cataract or retinal diseases, concurrent antitumor therapy (except stable hormonal therapy), or other serious intercurrent medical illness were excluded. All patients signed a written informed consent.
Treatment Plan and Trial Design.
Oral perifosine was provided by ASTA medica via the National Cancer Institute and was formulated in 50 mg white, film-coated tablets. These were stored in closed containers at room temperature and administered with food.
The treatment schedule used an oral loading dose at the start of cycle 1, given in even fractions every 6 hours over a total of 24 to 36 hours depending on dose level. Once daily oral maintenance dose began the next morning and continued until off study criteria were met. Dose escalation alternated loading dose and maintenance dose as seen in Table 1⇓ .
Before initiation of therapy, a complete history and physical (X-ray, electrocardiogram, ophthalmologic exam, laboratory parameters, and tumor assessment/measurement) were done. Patients were seen in follow-up before the start of each course to assess toxicity and to reassess laboratory parameters. Patients completing one or more cycles of therapy were considered evaluable, and tumor reassessments were done every two cycles, unless evidence of interim progression was clinically evident. Clinical response was a secondary objective and was analyzed when available. Responses were characterized as a complete response if all clinical and radiographic evidence of active tumor and symptoms disappeared for at least 1 month. A partial response was defined as a >50% decrease in the sum of the products of the perpendicular tumor diameters of all measurable lesions documented for >4 weeks without increase in any lesion or appearance of new lesion(s). Stable disease was a less than partial response or a <25% increase in the sum of the products as defined above persisting for at least 8 weeks. Progressive disease was defined as ≥25% increase in the sum of the products of measurable lesions and/or the appearance of new lesion(s). A DLT was defined as any grade 4 hematologic toxicity or as any more than or equal to grade 3 nonhematologic toxicity (except nausea and vomiting in the absence of antiemetics) as per the National Cancer Institute Common Toxicity Criteria, version 2.0.
Patients were accrued in cohorts of three, with standard definitions of DLT and MTD. DLT was reached in the loading dose phase of dose level 3. Expansion and escalation continued with the incorporation of prophylactic antiemetics and antidiarrheals because up to that point the maintenance dose had proven tolerable.
Pharmacokinetic Sampling and Methods.
Patients received a loading dose of perifosine, as described under the treatment plan, followed by a once daily maintenance dose for 8 weeks. Blood samples for perifosine assay were collected on day 1 before the first loading dose, at 1, 2, 3, 4, and 6 hours after the first loading dose, and at 1 and 2 hours after the last loading dose. Trough samples to assess steady-state perifosine concentrations were collected weekly during cycle 1, biweekly during cycle 2, on day 1 of each succeeding cycle, and with the occurrence of DLTs. At each time point, heparinized blood was collected into a plastic vacutainer (Becton Dickinson, Franklin Lakes NJ) to minimize adhesion of perifosine. Plasma was separated by centrifugation and stored in polypropylene cryovials at −70°C until assayed.
Perifosine in plasma was measured by a validated reversed-phase liquid chromatography/electrospray mass spectrometry method developed by Woo et al. (7) . Briefly, after addition of the internal standard hexadecylphosphocholine (miltefosine), perifosine was extracted from plasma standards or patient samples with acetonitrile. The extracts were diluted with mobile phase before injection. Separation was achieved on a Develosil RP aqueous 10 × 4-mm column with an isocratic (Phenomenex, Torrance, CA) mobile phase of acetonitrile 95%/9 mmol/L ammonium formate (40:60 v/v). Perifosine was detected by single-ion monitoring in the positive mode at m/z 462.4 and 408.4 for the internal standard, with a mass selective detector (Agilent 1100 LC/MS system, Agilent Technology, Palo Alto, CA). Standard curves were constructed from peak area ratio (perifosine/internal standard) versus concentration. The standard curve was linear from 0.075 to 10.0 μg/mL perifosine (r2 = 0.997 ± 0.003, n = 10 over 12 months). The limit of quantitation was 0.015 μg/mL in plasma. Recovery of perifosine by the method of standard addition was 98.0 ± 5.3% (n = 10). Intra-day variability of triplicate standards was <7% over the concentration range. Repetitive assays of the same control plasma sample over a 4-month period had a variability of 3% (n = 7).
The steady-state plasma concentration (Css) for individual patients was determined by averaging trough concentrations obtained on day 8 and thereafter. The terminal half-life was calculated by simple log linear fit in a cohort of six patients after discontinuation of study drug. Apparent clearance (Cl/F) was calculated by dose/Css.
A total of 42 patients was enrolled onto this trial. Demographic information is listed in Table 2⇓ .
A total number of 94 cycles was completed with the range 1 to 18 (Table 1)⇓ . Six patients were unevaluable: two due to clinical deterioration during cycle 1 (C1), one requiring radiotherapy treatment (new central nervous system disease), and three missed doses within C1.
Toxicities were analyzed for both the loading dose and maintenance dose components (Tables 3⇓ and 4⇓ ). Loading dose toxicities included nausea, vomiting, and diarrhea with one occurrence of dose-limiting diarrhea (level 2), one occurrence of dose-limiting dehydration (level 3), and one occurrence of dose-limiting fatigue (level 4) in a patient subsequently diagnosed with new brain metastasis. Toxicities seen in the maintenance phase again included nausea, vomiting, diarrhea, and fatigue. Both the incidence of and, to some extent, the severity of these toxicities increased with increasing drug dose. Dose-limiting fatigue was seen in level 3, dose-limiting nausea, vomiting, and fatigue (in the same patient) was seen in level 4, along with a second occurrence of dose-limiting fatigue in our expanded cohort for gender equality (2 of 14 DLTs). Dose level 5 (completed before gender equality accrual at level 4) included two DLTs (fatigue and elevated gamma glutamyl transferase) surpassing the MTD. Other toxicities seen in the maintenance phase included joint aches (grade 1), anorexia (grade 1 to 2), leg/foot pain/gout (grade 1 to 2), and gastrointestinal bleeding (fatal; felt related to progressive disease). No significant ophthalmologic or hematologic toxicities were seen.
Given our broadened experience (n = 14, 11 evaluable with two DLTs) at dose level 4 (load 150 mg × 6, maintenance 100 mg once daily) with reasonable tolerance of perifosine, this was our recommended phase II dose.
The loading dose was escalated from a total dose of 400 mg to a total dose of 900 mg, given as divided doses of 100 or 150 mg every 6 hours for four to six doses. All loading dose schedules achieved target concentrations of 2 to 10 μg/mL, ranging from 4.29 +/− 0.90 μg/mL (n = 3) at the lowest dose of 400 mg and 6.83 +/− 0.45 μg/mL (n = 17) at the highest total loading dose of 900 mg (see Table 5⇓ and Fig. 2⇓ ). There were no significant differences between the loading doses administered and the concentrations achieved 2 hours after the final dose of the loading dose by one-way ANOVA (P = 0.07).
After the loading dose, once daily oral maintenance dosing of perifosine was initiated. The maintenance dose was escalated from 50 to 100 mg, with an average Css of 3.40 +/− 0.43 μg/mL in the 24 individuals receiving 50 mg and 6.32 ± 0.52 μg/mL in the 9 individuals receiving 100 mg, which were significantly different by unpaired t test (P = 0.0355). The mean estimated clearance was not different between the two doses, with 0.72 +/− 0.88 liter/h in those receiving 50 mg per day and 0.79 +/− 0.09L/hour in those receiving 100 mg per day (P = 0.896, by unpaired t test; see Table 5⇓ ).
Css was achieved by day 2 (after the loading doses) in all dose levels, except level 2, where the concentration at the end of the loading cycle was significantly higher than the Css (6.05 +/− 1.87 versus 3.45 +/− 1.49, P = 0.026). The mean concentrations versus time for all patients at level 4 are shown in Fig. 3⇓ . Overall, intrapatient variability in Css averaged 16 ± 6% with no significant accumulation over time. There was no significant difference between the day 8 plasma concentration and the last measured perifosine concentration in 12 patients who received three or more cycles (P = 0.122), indicating no accumulation of perifosine over time. Fig. 4⇓ shows all steady-state plasma concentrations for the two patients who remained on study the longest, one at level 4 who received 14 courses of perifosine, and one patient at level 2 who received 18 courses.
To assess the long-term half-life, in six patients one additional plasma sample was obtained ∼8 days after the last dose (mean, 8.8 days; range, 6 to 13 days). In the six patients, off-study plasma concentrations yielded an estimated terminal half-life of 131 +/− 26 hours.
One confirmed partial response (dose level 2; by bi-dimensional measurements, reviewed by an oncologist not affiliated with the study) was observed in a patient with leiomyosarcoma (pelvic; Fig. 5⇓ ). This patient had previously received surgical and radiotherapeutic therapies for her cancer. She completed two cycles of doxorubicin (disease progression), two cycles of dacarbazine (with disease progression), and had been on tamoxifen 10 mg twice per day from October 1996 to March 2000 before receiving perifosine. This patient qualified for partial response after four cycles of perifosine and completed 18 cycles before coming off trial because of patient wishes (hand ache/pain, which did eventually improve months after coming off drug). The patient’s May 2003 visit relates that energy had improved, aches and cramps disappeared over 6 months off drug, and no significant disease progression had occurred off all anticancer therapies. One patient with renal cell carcinoma remained on drug (level 4) for 14 courses, eventually coming off because of disease progression. He had previously received numerous other experimental agents. A second patient with renal cell carcinoma completed six courses (dose level 3) before progression and a third renal cell carcinoma received four cycles (level 4). Both have since expired. Two patients with colon cancer (levels 4 and 5) completed three cycles before documentation of disease progression. Twelve patients completed two cycles, and 17 received less than two cycles before coming off study with either toxicity or disease progression. No perifosine-related deaths occurred on trial.
Perifosine (NSC 639966) is a synthetic substituted heterocyclic alkylphosphocholine that acts primarily at the cell membrane targeting signal transduction pathways. Although perifosine’s true mechanism of action remains to be defined, this class of drugs has shown interference with phosphoinositide metabolism, inhibition of protein kinase C, and inhibition of protein kinase B/Akt phosphorylation (8) . Perifosine has also been shown to block cell cycle progression at G1-S and G2-M by inducing p21 WAF1 (9 , 10) . Thus, perifosine can affect a broad spectrum of cellular functions, including cell proliferation, differentiation and invasion, and metastasis, thereby altering tumor cell growth and function.
Preclinical data confirmed perifosine’s activity in vitro and in vivo in a 7,12-dimethylbenz(a)anthracene-induced breast cancer model. Toxicity was minimized and efficacy improved by the use of a loading dose/maintenance dose schedule (2) . This data carried forward into pharmacokinetic modeling done at the National Cancer Institute and ultimately to the schedule used in our (and that of the National Cancer Institute) phase I trial. This schedule allowed a rapid approach to steady-state drug concentrations and sought to maintain these levels with minimal daily toxicity.
DLT was seen in both loading and maintenance phases and included diarrhea, dehydration (secondary to diarrhea), and fatigue. In the loading phase, nausea, diarrhea, and dehydration were surmountable once prophylactic antiemetics (5-HT3, decadron) were used and antidiarrheals used liberally (loperamide, two tabs after the first loose bowel movement, one tab after each successive loose bowel movement not to exceed 12 tabs in 24 hours). More difficult to ameliorate were the toxicities (dose limiting and compliance limiting) seen in the maintenance phase. DLTs included nausea, vomiting, fatigue, and elevated gamma glutamyl transferase. The incidence of grade 1 to 2 nausea, vomiting, and fatigue increased with increasing maintenance dose. Low-grade chronic nausea and fatigue were difficult to manage. Some patients responded to bedtime dosing, the use of H2 blockers, and/or antiemetics such as prochlorperazine. 5-HT3 antagonists were used occasionally, but, most often, if nausea and/or fatigue persisted, patients withdrew from study because of quality-of-life issues.
All loading schedules were able to achieve clinically relevant concentrations by day 2, and all except level 2 were at Css by day 2. The inability to achieve Css at level 2 was most likely caused by a large loading dose compared with the 50 mg maintenance dose. There were two DLTs with the loading dose, one at level 2 and one at level 4. Taken together, these data suggest that the lowest (900 mg total) loading dose is adequate to achieve clinically relevant serum levels by day 2, with minimal toxicity.
In this trial, 36 evaluable patients (6 patients were unevaluable) received maintenance therapy: 9 patients at 50 mg daily for 33 courses (range, <1 to 18 courses) and 22 patients at 100 mg once daily for 54 courses (range, <1 to 14 courses). Limited experience was gained at 150 mg daily with five patients completing eight courses (range, <1 to 3 courses). Mean steady-state plasma concentrations of perifosine roughly doubled from 50 to 100 mg maintenance dose. Of interest, the one partial responder at the 50 mg maintenance dose had steady-state levels of 5.90 +/− 1.03 μg/mL, closer to most 100-mg dosed patients.
Preclinical data indicate perifosine to be effective at serum concentrations of 2 to 10 μg/mL (2) ; concentrations achieved at both the 50 and 100 mg dose levels. Our trial confirmed the long t1/2 > 100 hours of this agent previously published (6) , and the lack of significant variance in steady-state levels with protracted (1.5 years) maintenance therapy (Fig. 4)⇓ , as well as the lack of significant accumulation beyond day 8. In addition, clearance is unaffected by dose. There is a clear need to balance active levels of the drug with chronically tolerable doses because, as the preclinical data support, chronic dosing of such targeted therapies is likely necessary to maintain efficacy. In our trial, maintenance phase low-grade gastrointestinal symptoms were manageable with nocturnal dosing and/or prochlorperazine but also responded rapidly to withdrawal of perifosine. Fatigue, however, was not as readily reversible. These observations together with the long terminal half-life warrant consideration of a less frequent dosing schedule (i.e., every other day or every third day).
In general, toxicities (primarily gastrointestinal) increased at higher dose levels. Pharmacokinetic studies at steady state show perifosine concentrations doubled in going from a 50 to 100 mg once daily maintenance dose. No additional increase in toxicity was observed in the 150 mg daily dose (Table 4)⇓ . Interestingly, our patient with a partial response had serum levels that approached those of our 100 mg daily cohort, suggesting variations in metabolism. Individual tolerance of the agent varied widely within a given dosing cohort making it difficult to suggest drug concentration alone was the etiology.
One partial response and multiple patients with stable disease beyond course 2 suggest some activity of perifosine in sarcoma (partial response), renal cell (stable disease until 6 and 14 courses), and possibly other tumor types (colon). Our partial response (sarcoma) remains without significant disease progression now at 17 months since discontinuation of study drug. Our recommended phase II maintenance dose is 100 mg once daily, although lower doses may prove to be biologically active and more tolerable. A 150 mg × 4 or × 6 loading dose is adequate to achieve rapid approach to steady state without excessive toxicity. Prophylactic 5-HT3 and steroid antiemetics are recommended. Current plans for continued evaluation of this agent include a phase II trial in refractory sarcoma, which has complete accrual. This trial will use tactics learned in our current phase I trial to balance toxicity and adequate dose intensity.
Grant support: NIH Grant U01 CA62491.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Lynn Van Ummersen, University of Wisconsin Comprehensive Cancer Center, K4/572 Clinical Science Center, 600 Highland Avenue, Madison, WI 53792. Phone: (608) 263-8600; Fax: (608) 263-9132; E-mail:
- Received October 16, 2003.
- Revision received May 18, 2004.
- Accepted July 8, 2004.