Phase I Study of Lovastatin, an Inhibitor of the Mevalonate Pathway, in Patients with Cancer

Vol. 2, 483-491, March 1996 Clinical Cancer Research 483

Alain Thibault,” 2 Dvorit Samid,2 7 consecutive days is well tolerated. The occurrence of my- Anne C. Tompkins, William D. Figg, opathy, the dose-limiting toxicity, can be prevented by ubiquinone supplementation. To improve on the transient Michael R. Cooper,2 Raymond J. Hohl, inhibitory activity of this dosing regimen on the mevalonate Jane Trepel, Bertrand Liang, Nicholas Patronas, pathway, alternative schedules based on uninterrupted ad- David J. Venzon, Eddie Reed, ministration of bovastatin should also be studied. and Charles E Myers2 Clinical Pharmacology Branch IA. T.. D. S., A. C. T.. M. R. C.. INTRODUCTION W. D. F., J. T., B. L.. E. R., C. E. M.] and Biometrics Section The role of lipid metabolism in cancer was initially inves- [D. J. V.. National Cancer Institute and Neuro-Radiology Department. Warren Magnussen Clinical Center IN. P.]. tigated in the 1950s by Fumagalli et al. (1) who observed that NIH, Bethesda, Maryland 20892-1576. and Divisions of neoplastic cells synthesize large quantities of cholesterol from Hematology-Oncology and Clinical Pharmacology, University of precursors such as acetate and mevalonate. They concluded that Iowa College of Medicine, Iowa City, Iowa 52242 [R. J. H.] this ‘ ‘high rate of cholesterol synthesis may provide a useful basis for testing whether tumor growth can be inhibited by

impairing its sterol synthesis’ ‘ ( 1 ). Twenty years later, Maltese ABSTRACT (2) demonstrated that the activity of HMG-CoA3 reductase, the Lovastatin, an inhibitor of the enzyme 3-hydroxy-3- major regulatory enzyme of de novo cholesterol synthesis, was methylglutaryl-coenzyme A reductase (the major regulatory increased in neoplastic tissue. This enzyme catalyzes the for- enzyme of the mevalonate pathway of cholesterol synthesis), mation of mevalonate, which is also the precursor of isoprenoid displays antitumor activity in experimental models. We moieties that are incorporated into or linked to several mole- therefore conducted a Phase I trial to characterize the tol- cules essential for cell growth and replication. The latter include erability of lovastatin administered at progressively higher ubiquinone (an isoprenylated benzoquinone involved in the mi- doses to cancer patients. From January 1992 to July 1994, 88 tochondrial electron transfer chain), dolichol, haem A, isopentyl patients with solid tumors (median age, 57 ± 14 years) were transfer RNA, and several proteins involved in signal transduc- treated p.o. with 7-day courses of lovastatin given monthly tion (Fig. 1 ; for review, see Ref. 3). The scope of these obser- at doses ranging from 2 to 45 mg/kg/day. The inhibitory vations was later broadened when it was shown that inhibition effects of lovastatin were monitored through serum concen- of HMG-CoA reductase by lovastatin selectively inhibited tu- trations of cholesterol and ubiquinone, two end products of mor growth in vitro and in animal models of hepatocellular. the mevalonate pathway. Concentrations of lovastatin and pancreatic, and central nervous system tumors (4-7). These its active metabolites were also determined, by bioassay, in studies demonstrated that growth arrest was associated with the serum of selected patients. Cyclical treatment with by- marked inhibition of isoprenoid synthesis and could be achieved astatin markedly inhibited the mevabonate pathway, evi- with minimal toxicity to the tumor-bearing animals, including denced by reductions in both cholesterol and ubiquinone the absence of myelosuppression. concentrations, by up to 43 and 49% of pretreatment values, These findings suggested that inhibition of the mevalonate respectively. The effect was transient, however, and its mag- pathway by lovastatin, a fungal antibiotic used in the treatment nitude appeared to be dose independent. Drug concentra- of hypercholesterolemia (8), may offer a novel approach to the tions reached up to 3.9 LM and were in the range associated treatment of cancer. We therefore designed a Phase I study to with antiproliferative activity in vitro. Myopathy was the determine the maximum tolerated dose of bovastatin when ad- dose-limiting toxicity. Other toxicities included nausea, di- ministered at progressively higher doses to patients with cancer. arrhea, and fatigue. Treatment with ubiquinone was associ- The study rationale was to attempt to achieve in patients drug ated with reversal of bovastatin-induced myopathy, and its concentrations associated with the experimental antiprolifera- prophylactic administration prevented the development of tive activity. It was supported by animal toxicology studies this toxicity in a cohort of 56 patients. One minor response which indicated that much higher doses of lovastatin than are was documented in a patient with recurrent high-grade currently recommended for the treatment of hypercholesterol- glioma. Lovastatin given p.o. at a dose of 25 mg/kg daily for emia (up to 80 mg/day, or 1 mg/kg/day) could be administered for short periods of time and be well tolerated (9). This information led us to administer lovastatin in cycles, to allow for Received S/I 8/95; revised 9/25/95; accepted I 1/21/95. recovery from acute drug-induced toxicity, while preserving the

I To whom requests for reprints should be addressed. at University of Virginia Health Sciences Center, Division of Hematology-Oncology, Jordan Hall. P. 0. Box 513, Charlottesville, VA 22908.

2 Present address: Cancer Center, University of Virginia Health Sci- 3 The abbreviation used is: HMG-CoA, 3-hydroxy-3-methylglutaryl co- ences Center. Charlottesville. VA 22908. enzyme A.

Acetyl CoA I Acetoacetyl CoA I 3 Hydroxy-3I Methyl Glutaryl CoA HMG-CoA reductase 1- LOVASTATIN

Mevalonate I Mevabonate-PP I Isopentenyl-PP . isopentenyl t-RNA I Geranyl-PP I Farnesyl-PP . haem A . farnesylated proteins

Squalene I Geranylgeranyl-PP . side chain of ubiquinone I . geranylgeranylated proteins Desmosterol . dolichol I Cholesterol

Fig. 1 Outline of the mevalonate pathway. The rate-limiting step is catalyzed by HMG-CoA reductase and inhibited by lovastatin. End products include cholesterol, a cell membrane component; ubiquinone, an electron carrier of the respiratory chain; and isoprenoid moieties, required for the posttranslational processing of proteins involved in intracellular signaling.

potential for drug activity. The secondary objectives of the trial partial thromboplastin time, serum alanine transferase concen- were to quantitate the pharmacological effects of high-dose trations less than twice the upper limit of normal, total bilirubin lovastatin in humans and to evaluate the clinical efficacy of within the normal range, and and serum creatinine concentration ubiquinone supplementation in the prevention of lovastatin- of 2.0 mg/dl or less. Patients were ineligible if they had exten- induced rhabdomyolysis, since there is evidence that ubiquinone sive (>50%) liver replacement by tumor or if they had not depletion may play a role in the pathophysiology of this side recovered from the toxicities of previous radiation or chemo- effect(lO, 11). therapy, or if they had received such therapy within a 4-week period. Patients had to have failed standard therapy for their PATIENTS AND METHODS disease or harbor a disease for which no acceptable therapy is Patient Population known. For example, patients with prostate cancer were re- Adults with a histological diagnosis of cancer confirmed by quired to have disease progression despite total androgen block- the Pathology Department of the NIH Clinical Center were ade therapy (hormone-independent prostate cancer) and subse- eligible for this Phase I trial. Other inclusion criteria included: quent cessation of flutarnide therapy. Patients with primary Eastern Cooperative Oncology Group performance status of 2 or central nervous system tumors had to have undergone maxi- better, ability to take p.o. medication, hemoglobin concentration mally tolerated surgery followed by radiation therapy. Previous >9.0 mg/dl, platelet count > 100,000/mm3, absolute granulo- treatment with adjuvant or palliative chemotherapy was not cyte count > 1 ,500/mm3, normal prothrombin time and activated required, nor did it constitute an exclusion criterion. No other

form of antitumor therapy was allowed during the study period. 45 mg/kg/day prompted us to return to the 25-mg/kg/day dose Patients taking anticonvulsants and corticosteroids were main- level and subsequently characterize the 30- and 35-mg/kg/day tamed on the same therapy. The dosage of anticonvulsants was dose levels. In a second part of the trial, lovastatin was given at modified according to symptoms and plasma concentrations. four dose levels (30, 35, 40, and 45 mg/kg/day) while coadmin- The dose of corticosteroids was kept identical to preprotocol istering ubiquinone (Vitaline Corporation, Ashland, OR) p.o. conditions or decreased as a function of clinical improvement. (240 mg daily, in four divided doses, given at the same time as All patients or their proxy signed an informed consent document lovastatin) to prevent lovastatin-induced myotoxicity. Supple- in compliance with the rules of the National Cancer Institute’s mentation started 7 days before the initiation of bovastatin and Institutional Review Board. continued for as long as the patient remained on protocol. Other hypolipidemic drugs (niacin, fibric acid derivatives, and other Initial Clinical Evaluation HMG-CoA reductase inhibitors) were not concurrently admin- Each patient underwent a complete history and physical istered. examination along with an assessment of the performance status. Pretreatment laboratory investigations included a hemogram Biochemical Measurements with leukocyte differential and platelet count, standard serum Frequency of Laboratory Evaluation. Initially, the co- biochemistry and coagulation studies, urinalysis, tumor markers hort of patients treated with lovastatin alone had blood drawn on relevant to the tumor type, total and high-density lipoprotein days 0, 1 . 3, 5, 8, and 28 of each cycle of therapy. Additional cholesterol, triglycerides, ubiquinone, chest X-ray, and electro- blood was obtained from most patients on days 15 and 21 . The cardiogram. Radiological studies appropriate to the disease type cohort treated with lovastatin in combination with ubiquinone and location were obtained within 4 weeks of entering protocol. was monitored on the day ubiquinone supplementation began (day -7), on the day treatment with lovastatin began (day 1), Treatment and subsequently on days 3, 5, 8, 15, 21, and 28. As a rule, blood Preclinical toxicology and pharmacology, available in the samples were obtained prior to the administration of the morn- mouse, rat, rabbit, and dog (9), demonstrated linear pharmaco- ing dose of lovastatin. kinetics in every species and indicated that doses up to 200 Measurement of Pharmacological Parameters. The mg/kg/day would yield drug concentrations in the range of 2-20 following biochemical effects of lovastatin therapy were mon- p.M. Although the maximum tolerated dose and organ toxicity itored: (a) inhibition of cholesterol synthesis, by measuring differed among species, progressive anorexia and death devel- serum total and high-density lipoprotein cholesterol; (b) inhibi- oped after 9-14 days of uninterrupted drug administration to tion of the synthesis of isoprenylated end products of the me- rabbits (the most sensitive species) and was associated with valonate pathway, by measuring serum ubiquinone concentra- sustained drug concentrations of 20-25 JiM. In contrast, circu- tions; and (c) circulating concentrations of lovastatin and its lating concentrations of 2-4 i.M were well tolerated for months metabolites. in all animal models. For the purposes of this Phase I study, Measurements of Ubiquinone Concentrations in Blood. intermittent drug administration, starting at a dose of 2 mg/kg/ Serum ubiquinone concentrations were assayed by normal phase day, was therefore predicted to be well tolerated. high-performance liquid chromatography according to a method Lovastatin (Mevacor; Merck, West Point, PA) was admin- previously described by Abe et a!. ( 12). The assay was linear istered p.o. following a four times a day schedule, for 7 con- between 0.08 and 10.67 .g/ml (r = 0.995 + 0.009, mean ± secutive days, in monthly cycles. The trial was initially con- SD; n = 12 standard curves), with a coefficient of variation ducted over seven dose levels, ranging from 2 to 45 mg/kg/day < 17%. The lower limit of quantification was 0.08 p.g/ml. (2, 4, 6, 8, 10, 25, and 45 mg). The initial dose level was twice Measurements of Lovastatin Concentrations in Blood. the dose currently recommended for prolonged administration in Serum concentrations of lovastatin and its metabolites, which humans and was chosen based on the preclinical information are also responsible for inhibition of HMG-CoA reductase in and unpublished safety data made available by the manufac- vivo, were measured indirectly and retrospectively from serum turer. The large dose increments of the last two dose levels were samples collected for other monitoring purposes. The method arbitrarily chosen based on the paucity of toxicity episodes at used was a standardized bioassay which quantitates the total the first five dose levels. At least three new patients were treated inhibitory activity of a patient’s serum (referred to as drug at each dose level. The first cycle of therapy was administered concentrations, in this report) against a microsomal suspension in the outpatient clinic of the Clinical Pharmacology Branch, of HMG-CoA reductase (13). The lower limit of quantification National Cancer Institute. Patients then returned home and were was 0.03 M. subsequently seen monthly. Compliance with the p.o. regimen was monitored through pill count and weekly telephone inter- Assessment of Toxicity views. In the absence of disease progression or severe (grade 3 The laboratory evaluation of toxicity was implemented by or greater) drug-induced toxicity, treatment cycles were re- weekly hemograms with leukocyte differential and platelet peated every 4 weeks. The dose of lovastatin could be increased count, electrolytes, blood urea nitrogen, creatinine, albumin, and for a given patient from one cycle to the next, according to the total protein. Since myopathy and elevations of hepatic dose escalation schedule, provided the preceding cycle had been transaminases have been recognized in 0.2-2% of patients tak- well tolerated in that patient and shown to be safe in at least ing lovastatin for the treatment of hypercholesterolemia (14), three new patients. The occurrence ofdose-limiting myopathy at patients were specifically monitored for the possible occurrence

of these events with measurements of hepatic aminotransferases, Table 1 Demographic and treatment characteristic s of the study bilirubin (total and direct), alkaline phosphatase, lactate dehy- population drogenase, creatine phosphokinase with isoenzyme fraction, and Characteristic n

aldobase. Urinalysis and urine dipstick for myoglobin (positive Median age, 57 ± 14 hemoglobin reaction) were also performed. Toxicity was graded Performance status (Eastern Cooperative according to the National Cancer Institute Common Toxicity Oncology Group scale) Grade 0 21 Criteria. Since no standard criterion exists for the grading of Grade I 53 musculoskeletal toxicity, the following scale was used: myalgias Grade 2 14 for <2 days without elevations of serum creatine phosphokinase Previous therapy represented grade 1 toxicity; myalgias of >2 days duration or Surgery 44 Radiation 54 elevations of serum creatine phosphokinase to < 10 times the Chemotherapy 67 upper limit of normal defined grade 2 toxicity; and muscle pain, Hormonotherapy 46 falls, or weakness sufficient to prevent the performance of daily Tumor type activities and/or elevations of serum creatine phosphokinase Prostate (hormone independent) 38 Primary central nervous system 24 concentrations to 10 times the upper limit of normal or above Astrocytoma 12 were defined as grade 3 toxicity. If, at any dose level, one Glioblastoma 8 patient developed grade 3 toxicity or above, additional patients Anaplastic oligodendroglioma 2 were entered at that dose level until at least six patients had been Others 2 Breast 7 treated. Dose escalation was stopped once grade 3 or higher Colorectal 4 toxicity developed in two of the six patients. The next lower Ovary 4 dose of lovastatin was defined as the maximum tolerated dose Sarcoma 3 and at least three additional patients were treated at this lower Lung 2 Others 6 dose to confirm its safety.

Assessment of Response The response status of malignancies was determined monthly, prior to each cycle of therapy, using conventional nadirs while receiving lovastatin were assessed with the Wil- anatomical criteria (15). For patients with prostate cancer, the coxon signed rank test (19). Cholesterol declines in the two criteria from the National Prostate Cancer Project ( 16) and patient groups (one supplemented with ubiquinone, the other published criteria for decline in prostate-specific antigen were not) were compared using the Wilcoxon rank sum test (19). used ( 1 7, 1 8). A technetium bone scan was obtained every 3 Least-squares linear regression was used to test for trends in months if initially positive or in the presence of new bone circulating drug concentrations, which were logarithmically symptoms. The assessment of patients with gliomas is compli- transformed before analysis. The distributions of toxicities cated by the variability in tumor-associated edema and its re- across grade and dose level were compared using the Cochran- sponse to steroid therapy, and technical factors which preclude Armitage test (20). SAS (Version 6.04; SAS Institute, Cary, using the intensity of gadolinium enhancement on magnetic NC) and StatXact (Version 2.04a; Cytel Software Corp., Cam- resonance imaging to determine tumor response. In these pa- bridge, MA) statistical packages were used for all analyses. tients, special attention was therefore paid to changes in performance status and steroid requirements, which were assessed at RESULTS each visit. Complete response was defined as complete disap- Patient Characteristics. Eighty-eight Caucasian pa- pearance of lesions on magnetic resonance imaging (assessment tients (63 men and 25 women) entered the trial from January performed in two different planes) and weaning from steroids. 1992 to July 1994. The demographic and treatment character-

Partial and minor responses were defined by conventional ana- istics of the study population are shown in Table 1 . All patients tomical criteria, absence of deterioration in performance status, (other than prostate and brain cancer patients) had failed at least and stable or decreased corticosteroids requirements. Progres- one course of chemotherapy with doxorubicin and/or cyclophos- sive disease was defined by either anatomical criteria, deterio- phamide (breast cancer and sarcoma), cisplatin or Taxol (ovar- ration in performance status, or the need to increase steroid ian cancer), and 5-fluorouracil (colon cancer). Patients with doses to maintain function. Disease stabilization was defined as primary brain tumors and hormone-independent prostate cancer the absence of a significant (more than 25%) increase or de- made up the two largest cohorts of the trial (39 and 24 patients, crease in tumor size while the patient maintained or improved respectively). performance status compared to pretreatment level. Disease Pharmacological Parameters. A total of 200 cycles of stabilization in these patients had to be maintained for at least 3 therapy was administered over 13 dose levels (Table 2). Forty- months to be considered significant. five patients received only one cycle of lovastatin, and 43 were given more than one (occasionally administered at different Statistical Methods doses since dose escalation was allowed in 16 patients). To Trend analysis of cholesterol change and lovastatin dose simplify the analysis, therefore, only data from the first cycle of was performed using the Spearman rank correlation method therapy for each treated patient has been used to characterize the ( I 9). Changes from pretherapy ubiquinone concentrations to pharmacological effects of lovastatin. All cycles of therapy were

Table 2 Distribution of patien ts and number of cy des per dose level 80

Lovastatin dose Patients” Cycles 70 (mg/kg/day) (n) (n) . 60 2 4 8

4 5(1) 7 . 50 6 4(1) 7 8 3 3 - 40 10 4 7 30 25 13(2) 23 30 8 18 20 35 12(1) 22 45 6 6 10 30 with ubiquinone 3 6

35 with ubiquinone 1 1 (2) 20 2 4 6 8 10 25 30 35 45 40 with ubiquinone 20 (8) 30 Lovastatin dose 45 with ubiquinone 22 (10) 43 (mg/kg/day)

“ Numbers in parentheses. number of patients escalating from lower dose level. Fig. 2 Pharmacological effects of lovastatin. Percentage of decline in serum concentrations of cholesterol () and ubiquinone () as a function of bovastatin dose (mean). Ubiquinone concentrations were determined only in patients treated at the 25-mg/kg/day dose level or higher. included, however, in the analysis of toxicity and antitumor Bars, SE. activity. Over the range of lovastatin doses administered without ubiquinone supplementation, cholesterol concentrations de- dined by 23-43% (nadir, mean ± SE: 143 ± 38 mg/dl; base- Serum HMG-CoA Reductase Inhibitory Activity. The line: 220 ± 56 mg/dl; normal range: 180-230 mg/dl; Fig. 2). bioactivity of lovastatin and its metabolites was retrospectively There was no direct correlation between the dose of lovastatin measured in 40 patients treated at the 4-mg/kg dose level and and the magnitude of the declines. Cholesterol concentrations above, with lovastatin alone or in combination with ubiquinone. rapidly declined during lovastatin administration. The time to A total of 149 samples was analyzed. The range of HMG-CoA reach the nadir ranged from 7 to 1 1 days after starting therapy reductase bioactivity achieved during the 7-day period of by- and also appeared to be dose independent. In all patients, the astatin administration was assessed in 1 19 samples. Thirty ad- declines were not sustained once treatment was discontinued ditional samples obtained from patients treated with more than and were completely resolved prior to the initiation of the next one cycle of therapy were analyzed to document the presence or cycle of therapy (Fig. 3). absence of drug activity at various time intervals after the Serum concentrations of ubiquinone were measured in 24 cessation of lovastatin. Peak bioactivity was reached within 4 h patients treated at the 25-mg/kg dose level and above. The in all patients and ranged from 0. 10 to 3.92 fiM (mean ± SD, declines in this isoprenylated end product of the mevalonate 2.32 ± 1 .27 .LM). Marked interpatient variability and no direct pathway ranged from 36 to 49% and were independent of the relationship to the dose administered were noted (by least- dose administered (P = 0.72, Spearman’s rank correlation test). squares linear regression). Trough activity at the 25-mg/kg dose The mean baseline and nadir ubiquinone concentrations mea- level and above averaged 0.28 ± 0.09 p.M. Drug activity in sured before and after 7 days of lovastatin therapy were 0.94 ± patients treated with lovastatin and ubiquinone did not differ 0.47 .tg/ml and 0.55 ± 0.28 p.g/ml, respectively (Table 3). from those measured in patients treated with lovastatin alone. In Similar to the changes noted in cholesterol concentration, the four patients treated with more than six cycles of therapy, no declines in ubiquinone were not sustained after stopping lovas- drug activity was detected 1 week after completing lovastatin tatin. administration. In the second part of the trial, ubiquinone supplementation Toxicity. Sixty patients (68%) experienced a total of 128 was started 7 days before the beginning of lovastatin and con- episodes of clinical toxicity. As can be seen from Table 4, the tinued until the patient was taken off protocol. The administra- incidence and severity of toxicity increased markedly once the tion of ubiquinone over 1 week had no effect on the circulating 25-mg/kg dose level was reached. In the cohort of patients concentrations of cholesterol. The magnitude of the decline in treated with lovastatin alone (n 32 patients; 104 cycles of cholesterol concentrations following the administration of by- therapy), grade 1 and 2 toxicity encompassed 92% of the epi- astatin was not different from those achieved in the first part of sodes. Gastrointestinal dysfunction was the most commonly the trial (P = 0.79, Wilcoxon’s rank sum test) and remained recognized toxicity, comprising 56% of all episodes. The most dose independent. severe clinical toxicity was related to the musculoskeletal sys- Oral ubiquinone supplementation for a week resulted in a tern and manifested primarily as myalgias and muscle weakness. 3-fold increase in serum concentrations (Table 3) from baseline No musculoskeletal toxicity was recognized at doses <25 mg/ concentrations of 1 .23 ± 0.78 p.g/ml to 4.58 ± 3.20 i.g/ml kg/day. At higher doses, however, no direct correlation could be (mean ± SD, n = 27). Following the administration of lovas- established between the incidence of myotoxicity and the dose tatin for 7 days, ubiquinone concentrations decreased on average of lovastatin administered (P = 0.24, Cochran-Armitage test). by 49% (to 1.88 ± 0.97 p.g/rnl, P = 0.001, Wilcoxon’s signed In one patient with high-grade glioma treated with lovas- rank test) but still exceeded baseline measurements. tatin alone at a dose of 35-mg/kg/day and whose course is

350’

ill C 0 300’

C 0 0 250’ -0--- patient 1 C . patient 2 OE00i -- a- patient 3

-“-,“--‘..“,‘ patient 4 0 200’ : patient 5 0 U) U) 0 150 -C C-)

100 -U- I 0 5 10 15 20 25 30 Cycle Days

Fig. 3 Time course of cholesterol decline after bovastatin treatment. Lovastatin induced declines in serum cholesterol concentrations in five representative patients treated for 7 days at the 25-45-mg/kg/day dose levels (horizontal bar). Notice the recovery in cholesterol concentrations by the end of each treatment cycle.

Table 3 Serum ubiqu inone concentration s in patients treated wit h lovastatin alone or in combination with ubiquinone

Ubiquinone concentration (pg/mi) Lovastatin dose (mg/kg/day) n Baseline Postubiquinone Postbovastatin

25 5 0.87 ± 0.36 0.49 ± 0.30 30 6 0.89 ± 0.50 0.64 ± 0.49 35 10 1.01 ± 0.58 0.53 ± 0.15 45 3 0.94 ± 0.60 0.49 ± 0.21 30 with ubiquinone 3 1.31 ± 1.21 3.50 ± 1.34 1.70 ± 0.60 35 with ubiquinone 7 0.83 ± 0.16 5.16 ± 4.21 2.03 ± 1.47 40 with ubiquinone 9 1.46 ± 0.57 4.98 ± 2.88 1.93 ± 0.67 4swithubiquinone 8 1.28 ± 1.1 1 2.83 ± 1.34 1.68 ± 0.91

summarized in Fig. 4, the onset of myopathy occurred after the any correlation between occurrence and cumulative lovastatin sixth treatment cycle, simulating disease progression. The max- dose. imum severity of this patient’s symptoms was associated with a Elevations in serum hepatic aminotransferases and creatine low serum concentration of ubiquinone (0.33 pg/ml). Magnetic phosphokinase concentrations above the upper limit of the nor- resonance imaging of the brain at that time revealed no increase mal range (alanine aminotransferase, 45 units/liter; aspartate in tumor volume. The patient’s symptoms resolved almost corn- aminotransferase, 42 units/liter) were the most common labora- pletely 48 h after p.o. supplementation of ubiquinone (60 mg, tory toxicities recognized in 23% of the evaluable cycles. Of p.o., four times daily), at which time serum concentrations of note, no elevation of grade 3 severity or higher was noted. The ubiquinone had increased 9-fold to 3.0 pg/mb. The patient was peak concentrations of these enzymes correlated with the max- subsequently maintained on p.o. ubiquinone and tolerated addi- imum intensity of symptoms and occurred between days 7 and tional cycles of lovastatin with no recurrence of symptoms nor 10 in most patients. radiological evidence of disease progression for 3 additional To prevent rnyotoxicity and improve the tolerability of months. Whether intracellular accumulation of lovastatin may lovastatin, ubiquinone was prophylactically administered to a have underlied the occurrence of this toxicity remains specula- second cohort of patients. Fifty-six patients were treated with tive at this time. In this patient with advanced brain cancer, bovastatin at doses of 30 mg/kg/day or more and received p.o. however, the administration of dexamethasone, a synthetic glu- ubiquinone prophylaxis; 27 (98%) patients experienced grade 1 cocorticoid with known myopathic effects, is a confounding or 2 toxicity and 1 patient experienced grade 3 nausea (2%). One factor that must be taken into consideration. With the exception death occurred secondary to disease progression and was unre- of this case, analysis of the toxicity patterns failed to disclose bated to treatment. Nausea and diarrhea were the most common

Table 4 Toxicity: Major types and frequency per dose level Frequency grade Lovastatin dose (mg/kg/day) n Toxicity I 2 3 4 A. Lovastatin alone 0 0 2 4 Nausea 0 Diarrhea 0 0 0 4 5 Abdominal pain 0 0 0 6 4 Constipation 0 0 0 8 3 Constipation 0 0 0 10 4 0 0 0 0 25 13 Myalgias 2 0 0 Muscle weakness 0 0 Nausea/vomiting 4 0 0 0 Diarrhea 2 0 0 0 Skin rash 0 0 Others 4 0 0 0 30 8 Diarrhea 0 0 0 Fatigue 0 0 0 35 12 Myalgias 4 0 0 Muscle weakness 0 0 Nausea/vomiting 4 0 0 Diarrhea 2 0 0 Fatigue 2 0 0 0 Others 4 0 0 45 6 Myalgias 0 2 0 0 Muscle weakness 0 0 Nausea/vomiting 5 0 0 Diarrhea 2 0 0 0 Fatigue 0 0 0

44 10 4 B. Ubiquinone supplementation

30 3 Myalgias 0 0 0 Trush 0 0 0 35 11 Myalgias 2 0 0 0 Muscle weakness 0 0 0 Nausea/vomiting 4 0 0 Diarrhea 0 0 0 Others 4 0 0 40 20 Myalgias 0 0 0 Muscle weakness 0 0 Nausea/vomiting 9 0 0 Diarrhea 6 0 0 Constipation 4 0 0 0 Others 3 0 0 45 22 Myalgias 7 0 0 0 Muscle weakness 0 0 0 Nausea/vomiting 6 0 0 0 Diarrhea 3 0 0 Constipation 3 0 0 0 Others 4 0 0 0

62 6 0

gastrointestinal toxicities. The administration of ubiquinone lovastatin at the 30- and 35-mg/kg/day dose levels and achieved alone for a week was not associated with toxicity. Ubiquinone a minor response (45% reduction in tumor size) that was main- prophylaxis did not decrease the incidence of musculoskeletal tamed for 8 months. No activity was documented in the cohort toxicity, but significantly reduced its severity (P = 0.01 1, Co- of patients with hormone-independent prostate cancer. chran Armitage test), which was limited to grade 1 ( 15/17, 88%) and 2 (2/17, 12%). The maximum tolerated dose for the com- DISCUSSION bination regimen was not reached. Response to Treatment. Evidence of antitumor activity Lovastatin is a prodrug which yields several metabolites was documented in one patient with an anaplastic astrocytoma that are responsible for the inhibition of HMG-CoA reductase in that was progressing after surgical resection, radiation therapy, vivo. Incompletely absorbed from the gastrointestinal tract, lo- and two cycles of carmustine. The patient was treated with vastatin undergoes extensive first-pass metabolism in the liver,

Several cases point to the depletion of ubiquinone as an important pathophysiobogical mechanism responsible for lovas-

tatin-induced muscle damage (1 1 ). Our experience establishes

C that this type of myopathy can be treated and prevented with p.o. 0 ubiquinone supplementation. Moreover, we could not identify a

C an antagonistic effect from ubiquinone in the patients who had e

C appeared to benefit from lovastatin, which is consistent with 0- oE C2 C3 C4 CS C6 laboratory data (5). Since it did not interfere with the cholester- a,; ol-lowering effect of lovastatin, ubiquinone may also prove 0 * * * + * useful in the management of hypercholesterolemic patients un- 0 . able to tolerate lovastatin’s most serious side effect.

E The drug concentrations measured in this Phase I trial a (0. 1-3.9 pM) were comparable to those found to be active (I) against glioma cells in vitro (0.2-2.0 prvi; Refs. 5 and 7) and provide a rationale to further study this approach in patients with high-grade gliomas. Several limiting factors, however, must be 25 50 75 100 125 150 175 taken into consideration. Although it is not known at present whether potential antitumor activity would result from the sys- Time (days) temic inhibitory effect of lovastatin on the mevalonate pathway or by the actual drug concentrations in the cerebrospinal fluid or Fig. 4 Lovastatin-induced myopathy: changes in serum concentrations brain parenchyma, the presence of a relatively intact blood-brain of ubiquinone and response to p.o. supplementation. Time course of barrier at the periphery of malignant gliomas represents a for- bovastatin-induced changes in serum ubiquinone concentrations in a patient who developed myopathy after several cycles (C) of lovastatin midable obstacle to delivering a highly protein-bound drug such therapy. Notice the transient inhibitory effect of lovastatin (documented as lovastatin to the actively dividing tumor cells. In addition, the after cycles 2 and 3) and the rapid increase in ubiquinone concentrations reversible inhibitory effect of lovastatin and its impact on p0- after p.o. supplementation (horizontal bar). tential antitumor activity must not be overlooked. Intermittent administration of the drug is associated with normalization of cholesterol and ubiquinone concentrations after treatment is stopped, and by the absence of detectable lovastatin concentra- the main site of its therapeutic activity as a hypocholesterolemic tions as early as I week after drug administration. These find- drug. Several studies have characterized the human pharmaco- ings suggest that a temporary blockade of HMG-CoA reductase kinetics of lovastatin given at doses not exceeding 2 mg/kg activity is taking place, which would not be expected to result in (21-24). Plasma concentrations of active lovastatin inhibitors sustained biological activity. In this perspective, it is possible increase linearly following single doses ranging from 60 to 120 that sustained inhibition of the mevalonate pathway by uninter- mg. Prolonged administration at the maximum recommended rupted administration of lovastatin may yield improved clinical dose (80 mg/day) results in steady-state concentrations of active results. lovastatin inhibitors ranging from 0. 15 to 0.3 pM. The pharma- In the context of isoprenylation inhibition as an anticancer cological effects of low-dose lovastatin in hypercholesterolemic approach, peptidic inhibitors of farnesyl transferase, the enzyme patients are better described (24). In these patients, maximum responsible for the isoprenylation of numerous proteins in mam- reduction appears to be dose-related (in the order of 30-40% in malian cells (25), must be mentioned. The therapeutic goal has patients treated with 80 mg/day) and occurs within 4-6 weeks been to improve upon the specificity of isoprenylation inhibition of initiating therapy. by targeting the ras oncogene in particular. Currently, however, That high doses of lovastatin may be more effective at technical limitations related to the intracellular delivery of these inhibiting the mevalonate pathway than currently recommended compounds remain to be solved (26), while their antiprolifera- doses is suggested by the observation that comparable declines tive activity may actually be independent of the ras prenylation in cholesterol concentrations occurred more rapidly in our study status of the tumor (27). (7-10 days) than has been reported with conventional doses. On We conclude that the administration of lovastatin at a dose the other hand, although marked interpatient variability was of 25 mg/kg daily for 7 consecutive days is well tolerated by noted, neither the magnitude of the inhibitory effect nor the cancer patients, and that high-grade gliomas represent a reason- circulating drug concentrations correlated directly with the dose able target for Phase II clinical trials. Alternative treatment administered. Similarly, the incidence of myotoxicity was inde- schedules aimed at achieving sustained inhibition of mevalonate pendent of the dose once 25 mg/kg lovastatin or more were synthesis should be investigated. administered. These findings, which led us to interrupt dose escalation at the 45-mg/kg level, are possibly accounted for by ACKNOWLEDGMENTS saturation of drug absorption at the higher dose levels or by the We thank Natalie McCall and Anne Schleifer for their laboratory short duration of drug administration, which may have pre- assistance, as well as Nancy Chen for her help in preparing the manu- vented larger, dose-dependent effects from taking place. script.

REFERENCES The Pharmacological Basis of Therapeutics. 8th ed, pp. 874-897. New York: Pergamon Press, 1990. 1. Fumagalli, R., Grossi, E., Paoletti, P., and Paoletti, R. Studies on lipids in brain tumors. 1 . Occurrence and significance of sterol precur- 15. Hoogstraten. B. Reporting treatment results in solid tumours. In: sors ofcholesterol in human brain tumors. J. Neurochem., 11: 561-565, M. E. Buyse (ed), Cancer clinical trials, pp. 139-156. Oxford: Oxford 1964. Medical Publications, 1990. 2. Maltese, W. A. 3-Hydroxy-3-methylglutaryl-coenzyme A reductase 16. Schmidt, J. D., Scott, W. W., Gibbons, R., Johnson, D. E., Prout, in human brain tumors. Neurology, 33: 1294-1299, 1983. G. R., Loening, S., Soloway, M., deKernion. J., Pontes, J. E.. Slack. 3. Goldstein, J. L., and Brown, M. S. Regulation of the mevabonate N. E., and Murphy, G. P. Chemotherapy programs of the National pathway. Nature (Lond.), 343: 425-430, 1990. Prostatic Cancer Project (NPCP). Cancer (Phila.), 45: 1937-1946, 1980. 4. Maltese, W. A., Refendini, R., Green, R. A., Sheridan, K. M., and 17. Thibault, A., Sartor, A. 0.. Cooper, M. R., Figg, W. D.. and Myers, Donley, D. K. Suppression of murine neuroblastoma growth in vivo by C. E. A 75% decline in prostate specific antigen (PSA) predicts survival mevinolin, a competitive inhibitor of 3-hydroxy-3-methylglutaryl-coen- in hormone refractory prostate cancer. Proc. Am. Assoc. Cancer Res., zyme A reductase. J. Clin. Invest., 76: 1748-1754, 1985. 34: 192, 1993. 5. Prasanna, P., Thibault, A., Liu, L., and Samid, D. Lipid metabolism 18. Kelly, W. K., Scher, H. I., Mazumdar, M., Vlamis, V., Schwartz, as a target for brain cancer therapy: synergistic activity of bovastatin and M., and Fossa, S. D. Prostate-specific antigen as a measure of disease phenylacetate against glioma cells. J. Neurochem., 66: 710-716, 1996. outcome in metastatic hormone-refractory prostate cancer. J. Clin. On- 6. Littman, M. L., Taguchi, T., and Mosbach, E. H. Effect of choles- col., 11: 607-615, 1993. terol-free, fat-free diet and hypocholestrolemic agents on growth of 19. Hollander, M., and Wolfe, D. A. Nonparametric Statistical Meth- transplantable animal tumors. Cancer Chem. Rep. 50: 25-45, 1966. ods. New York: John Wiley & Sons, 1973. 7. Shack, S., Prasanna, P., Hudgins, W. R., and Samid, D. Experimental therapies for malignant gliomas: targeting the mevabonate pathway of 20. Armitage, P. Test for linear trend in proportions and frequencies. cholesterol synthesis. Proc. Am. Assoc. Cancer Res., 35: 409, 1994. Biometrics, ii: 375-386, 1955. 8. Brown, M. D., and Goldstein, J. L. The hyperlipoproteinemias and 2 1. Alberts, A. W., MacDonald, J. S., Till. A. E., and Tobert, J. A. other disorders of lipid metabolism. ln: J. D. Wilson (ed), Harrison’s Lovastatin. Cardiovasc. Drug Rev., 7: 89-109, 1989. Principles of Internal Medicine, 12th ed., pp. 1814-1825. New York: 22. Stein, E. A., Till, A. E., Dobrinska, M. R., Gregg, M. H., Entwistle, McGraw-Hill, 1991. L. A., Hessey, J. Lovastatin multiple-dose pharmacokinetics. Pharm. 9. Kornbrust, D. J., MacDonald, J. S., Chennekatu, P. P., Duchai, Res., 5 (Suppl.): S-l50, 1988. D. M., Stubbs, R. J., Germerhausen, J. I., and Alberts, A. W. Toxicity of 23. Dobrinska, M. R., Stubbs, R. J., and Gregg, M. H. Effects of dose the HMG-coenzyme A reductase inhibitor, lovastatin, to rabbits. J. and food on HMG-CoA reductase inhibitor profiles after bovastatin Pharmacol. Exp. Ther., 248: 498-505, 1989. (MEVACOR). Pharmacol. Res., 5 (Suppl.): 5-182, 1988. 10. Folkers, K., Langsjoen, P., Willis, R., Richardson, P., Xia, L. J., Ye, 24. Querin, S., Lambert, R., Cusson, J. R., Gregoire, S., Vickers, S., Q. C., and Tamagawa, H. Lovastatin decreases coenzyme Q levels in Stubbs, R. J., Sweany, A. E., and Larochelle, P. Single dose pharma- humans. Proc. Natl. Acad. Sci. USA, 87: 8931-8934, 1990. cokinetics of ‘4C-lovastatin in chronic renal failure. Clin. Pharmacol. I 1. Chariot, P., Abadia, R., Agnus, D.. Dana, C., and Charpentier, C. Ther., 50: 437-441, 1991. Simvastatin-induced rhabdomyolysis followed by a MELAS syndrome. Am. J. Med.. 94: 109-1 10, 1993. 25. Kohl. N. E., Mosser, S. D., deSolms, S. J., Giuliani, E. A., Ponpli- ano, D. L., Graham, S. L., Smith, R. L., Scolnick, E. M., Oliff, A., and 12. Abe, K., Ishibashi, K., Ohmae, M.. Kawabe, K., and Katsui, G. Gibbs, J. B. Selective inhibition of ras-dependent transformation by a Determination of ubiquinone in serum and liver by high-speed liquid farnesyltransferase inhibitor. Science (Washington DC), 260: 1934- chromtography. J. Nutr. Sci. Vitaminol., 24: 555-567, 1978. 1937, 1993. 13. Alberts, A. W., Chen, J., Kuron, G., Hunt, V., Huff, J., Hoffman. C.. Rothrock, J., Lopez, M., Joshua, H., Harris, E., Patchett, A., Monahan, 26. Vogt, A., Qian. Y., Blaskovich, M. A., Fossum, R. D., Halmiton. R., Currie, S., Stapley, E., Albers-Shonberg, G., Hensens, 0., Hirshfield, A. D., and Sebti, S. M. A non-peptide mimetic of ras-CAAX: selective J., Hoogsteen, K., Liesch, J., and Springer, J. Mevinolin: a highly potent inhibition of farnesyltransferase and ras processing. J. Biol. Chem.. 270: competitive inhibitor of hydroxy methylglutaryl coenzyme A reductase 660-664, 1995. and a cholesterol-lowering agent. Proc. Nat. Acad. Sci. USA, 77: 27. Sepp-Lorenzino. L., Oliff, A., Gibbs, J. B., Kohl, N. E., Ma, Z., 3957-3961, 1980. Lebwhol, D. E., and Rosen, N. Farnesyl:protein transferase inhibitors 14. Brown, M. D., and Goldstein, J. L. Drugs used in the treatment of block the anchorage dependent and -independent growth of human hyperlipoproteinemias. In: A. G. Gilman (ed). Goodman and Gilman’s tumor cell lines. Proc. Am. Assoc. Cancer Res., 36: 431, 1995.

Downloaded from clincancerres.aacrjournals.org on October 1, 2021. © 1996 American Association for Cancer Research. Phase I study of lovastatin, an inhibitor of the mevalonate pathway, in patients with cancer.

A Thibault, D Samid, A C Tompkins, et al.

Clin Cancer Res 1996;2:483-491.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/2/3/483

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/2/3/483. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.