Invest New Drugs (2007) 25:139–146DOI 10.1007/s10637-006-9019-2

PHASE I STUDIES

A phase I and pharmacokinetic study of silybin-phytosomein prostate cancer patientsThomas W. Flaig · Daniel L. Gustafson · Lih-Jen Su · Joseph A. Zirrolli ·Frances Crighton · Gail S. Harrison · A. Scott Pierson · Rajesh Agarwal ·L. Michael Glode

Received: 21 August 2006 / Accepted: 28 September 2006 / Published online: 1 November 2006C© Springer Science + Business Media, LLC 2006

Summary Silibinin is a polyphenolic flavonoid isolatedfrom milk thistle with anti-neoplastic activity in several invitro and in vivo models of cancer, including prostate cancer.Silybin-phytosome is a commercially available formulationcontaining silibinin. This trial was designed to assess thetoxicity of high-dose silybin-phytosome and recommend aphase II dose. Silybin-phytosome was administered orallyto prostate cancer patients, giving 2.5–20 g daily, in threedivided doses. Each course was 4 weeks in duration. Thirteenpatients received a total of 91 courses of silybin-phytosome.Baseline patient characteristics included: median age of 70years, median baseline prostate specific antigen (PSA) of4.3 ng/ml, and a median ECOG performance status of 0.The most prominent adverse event was hyperbilirubinemia,with grade 1–2 bilirubin elevations in 9 of the 13 patients.The only grade 3 toxicity observed was elevation of alanineaminotransferase (ALT) in one patient; no grade 4 toxicitywas noted. No objective PSA responses were observed. Weconclude that 13 g of oral silybin-phytosome daily, in 3divided doses, appears to be well tolerated in patients withadvanced prostate cancer and is the recommended phase II

T. W. Flaig (�) · L.-J. Su · G. S. Harrison · A. S. Pierson ·L. M. GlodeDepartment of Medicine, Division of Medical Oncology,University of Colorado at Denver and Health Sciences Center,Denver, Colorado, USA

D. L. Gustafson · J. A. Zirrolli · R. AgarwalDepartment of Pharmaceutical Sciences, School of Pharmacy,University of Colorado at Denver and Health Sciences Center,Denver, Colorado, USA

F. CrightonUrologic Oncology, University of Colorado Hospital,Denver, Colorado, USA

dose. Asymptomatic liver toxicity is the most commonlyseen adverse event.

Keywords Silibinin . Phase I . Prostate cancer

Introduction

Silibinin is a naturally occurring polyphenolic flavonoid, iso-lated from milk thistle (Silybum marianium). The crude ex-tract of milk thistle, known as silymarin, has a long history ofhuman use dating back to ancient Greece, with contemporarystudies investigating its use in liver diseases [1–3]. Silibinin,also know as silybin or silibin, is the main component ofsilymarin and has a known chemical structure (Fig. 1).

Silibinin has shown anti-neoplastic activity in several can-cer models, including prostate cancer. Using a DU145 humanprostate cancer cell xenograft model in athymic mice, dietarysilibinin significantly reduced the xenograft tumor growth,without any apparent adverse effects [4]. Free plasma silib-inin was measured, with mean plasma levels ranging from14 to 27 µM in the low and high silibinin-diet groups, re-spectively. Kohno et al. used silymarin in a prostate cancermodel, injecting rats with 3,2′-dimethyl-4-aminobiphenyl toinduce cancer [5]. After 60 weeks, dietary silymarin treat-ment significantly reduced the appearance of prostate cancer,with 50.0% of the control and 17.6% of the silymarin-treatedgroup developing cancer. Milk thistle derivatives have alsoshown in vivo anti-neoplastic or cancer prevention effects incolon [6], bladder [7], lung [8], and skin [9], cancer models.

There appear to be several important components in sili-binin’s anti-cancer activity. Using the human prostate can-cer cells LNCaP, silibinin treatment decreases cell growthwith selective G1 arrest [10]. Consistent with this finding,silibinin also was shown to induce the cyclin-dependent

Springer

140 Invest New Drugs (2007) 25:139–146

O

O

O

O

HO

OHOH

OH

H

H

H

H

OCH3

CH2OH

O

O

O

O

HO

OHOH

OH

H

H

OCH3

CH2OH

Fig. 1 The structure of silibinin

inhibitors Cip1/p21 and Kip1/p27, while decreasing levelsof cyclin D, cyclin-dependent kinase (CDK) 4 and CDK 6.In DU145 cells, silymarin decreases the activation of theepidermal growth factor receptor and its downstream targetSHC, while also directing G1 arrest [11]. The insulin-likegrowth factor (IGF) pathway appears to be an importantactor in prostate cancer, with epidemiologic evidence asso-ciating an increased IGF level with increased prostate cancerrisk [12]. Dietary silibinin significantly increases both theplasma [4] and prostate tissue [13] levels of IGF bindingprotein-3 (IGFBP-3), a modulator of IGF and positive prog-nostic factor, in nude mice with DU145 xenograft.

The oral bioavailability of silibinin and silymarin is poor,which has led to the development of silibinin complexesfor improved absorption. The pharmacokinetics of low-doseIdB 1016 (silipideTM), a silibinin-phosphatidylcholine for-mulation similar to silybin-phytosome, have been reported[14]. Nine healthy volunteers consumed a single, oral dose ofsilipide (equivalent to 360 mg of silibinin), achieving a peakconcentration of 298 + / − 96 ng/ml, with peak plasma lev-els achieved at 1.6 + / − 0.3 h. IdB 1016 showed improvedoral availability compared with silymarin, although inter-patient variability of plasma levels was high. In a separatecohort, 9 healthy individuals were given IdB 1016 (equiva-lent to 120 mg silibinin) twice daily for 8 days, with bloodsampling on days 1 and 8. The terminal half-life of IdB1016was short (2.6 + / − 1.0 h) when measured on day 1 andshowed little change at day 8. Schandalik and Perucca havealso reported the pharmacokinetics of silipide in patientswith extrahepatic biliary obstruction [15]. Fourteen hospi-talized patients with malignant and non-malignant causes ofextrahepatic biliary obstruction were given one oral dose ofsilipide (120 mg silibinin equivalents). Compared to the pre-vious studies in healthy volunteers, the free silibinin plasmalevels were similar, suggesting that extrahepatic cholestasishas little effect on free silibinin kinetics. In 2006, IdB 1016was studied giving standard daily doses (360–1440 mg) topatients before surgery for adenocarcinoma of the colon [16].This work revealed silibinin levels of 0.3–4 µmol/L in theplasma and 20–141 nmol/L in colorectal tissue, with no effecton circulating IGFBP-3 levels noted at these doses. Takentogether, these data indicate that silibinin is better absorbedwhen joined to phosphatidylcholine, has a rapid peak, promi-

nent interpatient variability and a short half-life when takenat standard doses.

This phase I study examines the use of silybin-phytosome (SiliphosR), a commercially available silibinin-phosphatidylcholine complex, to determine a phase II doserecommendation and examine the toxicity of high-dose ther-apy. While silibin-phosphatidylcholine formulations havebeen used in clinical investigations, the typical daily dosehas been less than 2 g and set without clear foundation [14–16]. With strong preclinical data supporting silibinin’s in vivoantineoplastic activity, the pharmacokinetic and toxicity dataprovided here will support the design of future human stud-ies.

Patients and methods

Patient selection

Patients participating in this trial provided both verbaland written consent according to federal and institutionalguidelines, with an IRB-approved consent form. Eligibil-ity criteria included: Age ≥ 18 years, histologically con-firmed adenocarcinoma of the prostate, life expectancy ≥ 3months, ECOG performance status of ≤ 2, adequate organfunction (absolute neutrophil count (ANC) > 1500 mm3,platelets > 100 × 10 /mm3, hemoglobin > 9 g/dL, biliru-bin < 1.5 mg/dL, aspartate transaminase (AST) and ala-nine transaminase (ALT) < 2.5 × upper limits of normal,creatinine < 1.5 mg/dL or creatinine clearance of at least50 ml/min), with progressive disease defined by a rising PSAor measurable disease by radiological assessment. Exclusioncriteria included radiation therapy within the 4 weeks preced-ing study enrollment, uncontrolled brain metastasis, use ofan investigational drug or device within one month of startingthe trial, or any other serious medical condition which couldinterfere with the conduct of the trial. Patients were allowedto use luteinizing hormone-releasing hormone analogs (i.e.leuprolide and goserelin) during the trial.

Dosage and drug administration

Silybin-phytosome (SiliphosR) is a commercial formulationof silibinin and phosphatidylcholine (approximately 30%silibinin by weight), purchased from the Indena Corpora-tion (Seattle, WA). It was obtained as a powder and mixedwith applesauce at the ratio of 1/4 teaspoon of silybin-phytosome to 1 Tablespoon of applesauce. Patients weredirected to take silybin-phytosome 3 times a day, at least30 min before meals. The first daily dose-level was 2.5 gof silybin-phytosome, then 5 g and then increased by incre-ments of 5 g (i.e. 10, 15, 20 g daily); due to the toxicityobserved with chronic administration of 15 and 20 g daily,

Springer

Invest New Drugs (2007) 25:139–146 141

the dose level was reduced to 13 g daily. Once a singlepatient was treated for 1 month at an indicated dose level,2 patients from a lower dose-level were escalated to thenew dose-level. Once they had been treated for a monthat this level with no observed toxicities, all remaining pa-tients were escalated to that dose, with the next new patientenrolled at the next dose-level. Intra-patient dose escala-tion was allowed in this manner. This escalation approachwas altered for the 2 lowest dose-levels, with only 1 patientstarted at each level. Toxicities were graded with the Na-tional Cancer Institute’s Common Toxicity Criteria (version2.0). No dose adjustments were made for bilirubin eleva-tion, although patients did withdrawal from the study forthis reason. Dose limiting toxicity was defined as grade 3 or4 non-hematologic toxicity, or grade 4 hematologic toxicity(thrombocytopenia, neutropenia > 5 days or complicated byfever).

Pretreatment assessment and follow-up studies

Patients were carefully screened and followed with monthlyvisits during the study period. Prior to starting treatment, allpatients had a physical examination, medical history, electro-cardiogram (ECG), performance status assessment, and labo-ratory assessment with a complete blood count (CBC), com-prehensive metabolic panel (CMP), uric acid, phosphorus,lactate dehydrogenase (LDH) within 14 days of starting plusradiographic assessment of all know disease sites, testos-terone, and prostate specific antigen (PSA) within 28 daysof starting. A course was defined as 4 weeks, with patientshaving a history and physical examination, performance sta-tus assessment, and laboratory evaluation with CBC, CMP,uric acid, phosphorus, LDH and PSA every course. Follow-up radiographic assessment was left to the discretion of thetreating physician. Patients were removed from the study forprogression of disease (by established PSA or radiographiccriteria), unacceptable toxicity, patient non-compliance, adelay in treatment of greater than 21 days for any reason,and at the discretion of the Principal Investigator.

Pharmacokinetic sampling

To characterize the pharmacokinetics of high-dose silybin-phytosome, both blood and urine were collected for all pa-tients during the first month of treatment on days 1, 8, 15,and 22. Blood was collected at time 0, 30 min, 1 h, 2 h, and4 h after the administration of silybin-phytosome on eachof theses days. The urine was collected, as available duringthe patients’ time in the clinic (no catheters were used fortimed collections). Samples were obtained in EDTA tubes,centrifuged at room temperature to separate the plasma andstored at − 80◦C until analysis.

Silibinin and silibinin glucoronide analysis in plasmaand urine

Naringenin, naringin, alamethicin, saccharolactone, uridine5′-diphosphoglucuronic acid (UDP-GA) and silibinin wereobtained from Sigma (St. Louis, MO). All other chemicalsand solvents used in the analysis were of reagent or higherquality and obtained from Fisher Scientific (Pittsburgh, PA).Human microsomes (20 mg/mL in 250 mM sucrose) wereobtained from XenoTech, LLC (Lenexa, KS).

Silibinin glucuronide was prepared by the method ofGhosal [17] with minor modifications. Briefly, to a 1.5 cen-trifugation tube the following were added: 560 uL of Trisbuffer (100 mM, pH 7.4), 200 uL of human microsomes(4 mg), 20 uL of alamethicin (50 ug/mg of microsome),100 uL of saccharolactone (50 mM in 1 M Tris, pH 7.4),100 uL of MgCl2 (100 mM in 1 M Tris, pH 7.4 and 20 uLof silibinin (50 mM in ethanol). The solution was warmedfor 5 min at 37◦C and the incubation then started with theaddition of 10 uL of UDP-GA (200 mM in 100 mM Tris, pH7.4). After 4 h the solution was chilled for 5 min and the in-cubation stopped with the addition of 100 uL of formic acid(88%). The solution was centrifuged for 10 min @ 10,000RCF, the supernatant removed and analyzed by LC/MS/MSmethod described below.

Silibinin glucuronide was isolated from the incubationsupernatant by solid phase extraction (SPE) using 3 mL,200 mg, BondElut-C18 columns (Varian, Harbor City, CA).SPE columns were conditioned first with 15 mL of methanolfollowed with 15 mL of distilled, deionized water. Super-natant (1 mL) was loaded on to the SPE column, washedwith 3 mL of 0.1% formic acid and eluted sequentially with1 × 2 mL of 10% acetonitrile in 0.1% formic acid, 3 × 2 mLof 20% acetonitrile in 0.1% formic acid, 1 × 2 mL of 30%acetonitrile in 0.1% formic acid and 1 × 2 mL of 40% ace-tonitrile in 0.1% formic acid. Eluates were analyzed for freesilibinin and silibinin glucuronide by LC/MS/MS (describedbelow). Only silibinin glucuronide was detected in the firsttwo 20% acetonitrile eluates, then a mixture of silibinin glu-curonide and free silibinin were detected in the 3rd 20% and30% acetontrile eluates and finally only free silibinin wasdetected in the 40% acetontrile eluate. Silibinin and silib-inin glucuronide have identical UV spectra (200–350 nm)with a maximum at 285 nm. Purified silibinin glucuronidewas quantified by UV absorbance (A285) using a silibininreference standard curve.

Standard dilutions of silibinin and silibinin glu-curonide were prepared in acetonitrile, control humanurine (10–20,000 ng/mL) and control human plasma (10–20,000 ng/ml). Each dilution contained 250 ng of naringeninand naringin (25 uL of 10 ug/mL stock), the internal stan-dards, in a final volume of 1 ml. Naringenin and narigininwere prepared at 10 ug/mL stock solutions in methanol.

Springer

142 Invest New Drugs (2007) 25:139–146

Urine and plasma samples were prepared by similar meth-ods. To 200 µL of either plasma or urine, 250 ng (25 µL of10 µg/mL) of naringenin and naringin were added and thenextracted with 1 mL of acidified ethyl acetate (0.1% formicacid) for 10 min (vortex). Organic and aqueous layers wereseparated by centrifugation (10 min, 12000 RCF) and the or-ganic layer removed, evaporated and reconstituted in 1 mLof 20% acetonitrile for LC/MS/MS analysis.

Negative ion electrospray ionization (ESI) mass spectrawere obtained with a PE Sciex API-3000 triple quadrupolemass spectrometer (Foster City, CA) with a turbo ionspraysource interfaced to a PE Sciex 200 HPLC system. Sampleswere chromatographed with a Prodigy 5 u Phenyl-3, 5 µm,100 A◦, 150 × 2 mm column (Phenomenex, Torrance, CA).The LC elution was isocratic with 50% acetonitrile contain-ing 10 mM ammonium acetate and 0.1% acetic acid at a flowrate of 200 µl/min and sample injection volume of 20 µl.The analysis time was 5 min.

The mass spectrometer settings were: turbo ionspray tem-perature, 400◦C, spray needle, − 3500 V, declustering po-tential (DP), − 30 V, focus plate (FP), − 130 V, collisionenergy (CE) − 40 V, collision gas, N2, (CAD) 10 units.Samples were quantified by the internal standard referencemethod in the MRM mode by monitoring the transition m/z481 to m/z 125 for silibinin, m/z 271 to m/z 119 for narin-genin (internal standard), m/z 561 to m/z 481 for silibininglucuronide and m/z 579 to m/z 271 for naringin (internalstandard). Each ion transition was integrated for 250 ms.

Quantitation of silibinin was based on standard curvesin spiked matrix using the ratio of silibinin peak area tonarigenin peak area using 1/x weighting. Pharmacokineticparameters were determined by non-compartmental analysis.

Results

Patient characteristics

Thirteen prostate cancer patients were enrolled fromOctober of 2003 through November of 2005. Data was an-alyzed through March 23, 2006, with 91 courses of silybin-phytosome administered. Table 1 describes the characteris-tics of the participants. They had a good performance status(median ECOG = 0) with a median age of 70 years, con-sistent with a typical prostate cancer patient population.

Drug delivery

The dose-escalation scheme is depicted in Table 2. The dailydose of silybin-phytosome was escalated from 2.5 g to 20 g,at which time persistent grade 2 hyperbilirubinemia wasnoted in 3 of the 6 patients at the 15 and 20 g dose level. Al-though dose limiting toxicity (DLT) was not seen in the first

Table 1 Patient characteristics

Characteristics No.

No. of patients 13Total no. of courses 91No. courses per patient

Median 6Range 3–13

Age (years)Median 70Range 57–84

ECOG performance status0 91 4

Median starting PSA (ng/mL) 4.3Gleason score total∗

5 16 07 58 39 3

Previous treatmentsRadical prostatectomy 7Radiation therapy 8Androgen deprivation therapy 8

Note: Eastern Cooperative Oncology Group (ECOG), Prostate SpecificAntigen (PSA).∗1 patient’s original Gleason score could not be confirmed (12 assess-able and included).

course of treatment at any dose level, this persistent grade 2toxicity was deemed unacceptable for phase II testing and anadditional 6 patients were subsequently accrued at a lower(13 g a day) dose level, per the protocol guidelines for doseadjustment. All patients were available for evaluation withgood follow-up compliance.

Toxicity

The most notable toxicity observed was gastrointestinal,with grade 1 or 2 unconjugated hyperbilirubinemia observedcommonly. The only grade 3 or 4 toxicity noted was onepatient with transient grade 3 elevation of ALT.

Table 2 Dose-escalation scheme

Silybin-phytosomedaily dose No.of patients∗ No. of courses

2.5 g 1 15 g 2 1010 g 5 2415 g 5 1920 g 3 513 g 6 32Total 91

∗Intra-patient dose escalation was allowed.

Springer

Invest New Drugs (2007) 25:139–146 143

Table 3 Gastrointestinaltoxicities of silybin-phytosome Silybin-

phytosome No. of Bilirubinemia ALT elevation AST elevationdose courses G1 G2 G3 G1 G2 G3 G1 G2 G3

2.5 g 1 0 0 0 0 0 0 0 0 05 g 10 0 0 0 0 0 0 0 0 010 g 24 6 1 0 0 0 0 0 0 013 g 32 16 2 0 4 1 1 2 1 015 g 19 5 7 0 3 0 0 0 0 020 g 5 0 3 0 0 0 0 0 0 0

Note: Alanine transaminase(ALT), Aspartate transaminase(AST), Grade 1 (G1), Grade 2(G2), and Grade 3 (G3).

Gastrointestinal toxicity

Grade 1 or 2 bilirubin elevation was seen in 40 of 91 (44%)courses (see Table 3). Despite this common elevation in totalbilirubin, no abnormal direct bilirubin levels were observed,thus characterizing this as an unconjugated bilirubinemia.There was no compelling evidence for hemolysis as a sourceof this elevation, with only 3 patients exhibiting grade 1anemia, 3 other patients with grade 1 LDH elevation and 1patient with grade 1 elevation of both measures. The notedgrade 1 and 2 total bilirubin elevation resolved to normal inall but one patient, who continued to have a mildly elevatedbilirubin (grade 1) of 1.3 mg/dL approximately 6 months af-ter stopping the silybin-phytosome. Increased transaminaselevels were also seen, but with less frequency. One patientat the 13 g dose-level was noted to have grade 3 elevationin ALT and grade 2 elevation of AST after 2 courses. Areevaluation of these laboratory tests showed significant im-provement, with only grade 2 and grade 1 elevation of theALT and AST, respectively, without any dose reduction. Atthe last assessment, 4 months after the grade 3 elevation, thepatient’s transaminase levels had normalized with an ALTof 43 U/L and an AST of 35 U/L, without dose adjustment.Mild diarrhea was also commonly seen, but was generallywell tolerated. Only one patient cited this as a reason forwithdrawal from the trial. The diarrhea was consistently de-scribed as a mild increase in frequency (e.g. 2 additionalbowel movements daily), without urgency or prominent dis-comfort. Eleven of the patients reported grade 1 diarrhea andone patient had grade 2 diarrhea.

Other toxicity

Other toxicities were generally mild and isolated. Grade 1elevation in creatinine was noted in 6 patients and usuallyresolved spontaneously, without dose change. Six patientsalso had grade 1 hypercalcemia, which as also mild and self-limited. Two patients reported minor halitosis which theyattributed to the silybin-phytosome. One case of grade 1 andone of grade 2 bone pain were seen and attributed to progres-sive disease. One patient had sensory neuropathy (grade 2)in one arm, although this was attributed to thoracic outlet ob-

struction and not believed related to the silybin-phytosometreatment. No other grade 2–4 toxicity was observed.

Pharmacokinetic studies and analysis

Plasma samples were collected to analyze the pharma-cokinetic characteristics of single-agent, high-dose silybin-phytosome. Complete samples (times 0, 30 min, 60 min,120 min, and 240 min after administration) were avail-able for weeks 1–4 in 11 of the patients; 2 patients hadincomplete samples for all or part of one of the 4 weeks.The mean plasma concentration-time plot for weeks 2–4 is seen in Fig. 2. The largest cohort was started at13 g (6 patients) with 3 patients started at the 10 g leveland 1 patient started at each of the remaining dose lev-els. The area under the curve (AUC) and the Cmax ondays 8, 15, and 22 were determined for each dose level(Fig. 3). Prominent interpatient variability is noted without

0 1 2 3 4 50.1

1

10

100

2.5 g5 g10 g13 g15 g20 g

Time After Dose (hr)

[Sili

bin

in] p

lasm

a( µµ

M)

Fig. 2 Silibinin plasma concentration versus time. The largest cohortwas started at 13 g (6 patients) with 3 patients started at the 10 g leveland 1 patient started at each of the remaining dose levels

Springer

144 Invest New Drugs (2007) 25:139–146

2.5 g 5 g

10g

13g

15g

20g

0

50

100

150

200

250

Silibinin Dose

AU

C0

t( µµ

Mh

r)

2.5g

5g

10g

13g

15g

20g

0

25

50

75

100

125

150

Silibinin Dose

Cm

ax( µµ

M)

←

•

Fig. 3 Area under the curve (AUC) and Cmax. The AUC and Cmaxwere calculated for each patient and plotted with relation to the doselevel

a clear dose response above the 13 g level. The half-lifeof plasma silibinin was short, ranging from 1.79–4.99 h.Thirty-one urine samples were also collected during the 4 hperiod of the pharmacokinetic blood sample collection andassessed for free silibinin and silibinin-glucuronide. Therewas a large amount of interpatient variability in the silibininurine levels and not a clear dose response. The mean silib-inin level in the urine was 6.4 µM (range of undetectable to

28.2 µM); the mean silibinin-glucuronide level was 253.4(range of 1.5–982 µM).

Anti-tumor effect

Although the main objectives of this trial were to assess thetoxicity of high-dose silybin-phytosome and recommend aphase II dose, the PSA was followed in all patients to as-sess for any disease effect. While no patients demonstrateda 50% reduction in PSA, which is required for a formal PSAresponse, several patients demonstrated prolonged stable dis-ease.

Discussion

This study was designed to assess the toxicity of high-dosesilybin-phytosome, gather pharmacokinetic data for high-dose silybin-phytosome, and determine a recommendedphase II dosing schedule. The findings indicate that silybin-phytosome may be given in high doses with acceptable toxic-ity. While mild hyperbilirubinemia was frequently observed,this finding was asymptomatic and improved with cessationof the treatment. At 15 and 20 g of daily silybin-phytosome,grade 2 hyperbilirubinemia was noted in 10 of 24 courses(41.67%); with a dose reduction to 13 g daily, grade 2 hy-perbilirubinemia was only seen in 2 of 32 courses (6.25%),with no grade 3 or 4 bilirubin toxicity noted. One patient diddevelop grade 3 elevation of ALT; however, this improvedto grade 2 on reexamination and resolved to normal levelswith additional follow-up, without reducing the dose. Basedon these findings and practical considerations of consuming15 g or more of silybin-phytosome daily, the recommendedphase II dose of silybin-phytosome is 13 g daily, in 3 divideddoses.

The common finding of asymptomatic hyperbilirubinemiawas unexpected at the time this trial was designed and con-ceptualized; however, recent information provides a rationalexplanation for this event. Sridar et al have described the se-lective in vitro inhibition of Uridinediphosphoglucuronate-glucuronosyltransferase-1A1 (UGT-1A1) by silibinin, withan inhibitory concentration 50% (IC50) of 1.4 µM [18].Clinically, decreased functionality of UGT-1A1 is seen inpatients with Gilbert’s syndrome, a common genetic disor-der which manifests through asymptomatic hyperbilirubine-mia. Since the peak plasma levels seen in the present studygrossly exceed 1.4 µM, silibinin’s selective inhibition ofUGT-1A1 provides a plausible biologic explanation for thehyperbilirubinemia observed.

In addition to its role in the conjugation of bilirubin, UGT-1A1 is important in the normal metabolism of irinotecan,a chemotherapy agent used commonly in the treatment ofcolon cancer. Increased toxicity from irinotecan has been

Springer

Invest New Drugs (2007) 25:139–146 145

noted in patients with Gilbert’s syndrome, believed sec-ondary to the UGT-1A1 deficiency associated with this disor-der [19]. With this consideration, the effect of milk thistle onthe pharmacokinetics of irinotecan has recently been investi-gated [20]. Patients receiving weekly irinotecan (4 weeks outof 6) as treatment for colon cancer were additionally given200 mg of oral milk thistle (80% silymarin), thrice daily,starting before the second dose of irinotecan. Blood sampleswere subsequently collected for pharmacokinetic analysiswith plasma silibinin levels of 0.0249 to 0.257 µmol/l. Theauthors concluded that this formulation and dose of milkthistle was unlikely to alter the metabolism of irinotecan.The present trial demonstrates dramatically higher peak sili-binin blood levels, with large intra- and inter-patient vari-ability, and reports bilirubin elevation consistent with UGT-1A1 inhibition. Therefore, we recommend caution in usingany combination of silibinin and irinotecan, particularly withhigh doses of silibinin, until more is learned about the char-acteristics of the apparent dose-related effect of silibinin onUGT-1A1.

While some peak plasma levels of silibinin were in excessof 100 µM, silibinin is quickly conjugated and excreted intothe urine, demonstrating a short half-life. Silybin-phytosomedemonstrates prominent inter- and intra-patient variabilityat the high doses studied here. Interestingly, there is not astrong dose relationship to peak silibinin blood levels, withthe highest levels noted in patients at the 13 g level andnot the 15 or 20 g level, although there were more patientsin the 13 g cohort with significant interpatient variabilityobserved. Notably, the peak plasma levels obtained in thepresent study exceed the mean plasma levels of free silibininshown to retard xenograft tumor growth in vivo [4].

Defining the maximum tolerated dose (MTD) of silybin-phytosome, as with other non-cytotoxic chemotherapyagents, was challenging. Even though no DLT was observedduring the 1st cycle of treatment at any dose level, persistentgrade 2 gastrointestinal toxicity was seen commonly at the 15and 20 g daily dose levels. Although this didn’t reach the tra-ditional definition of DLT (selected grade 3 and 4 toxicity), itwas judged that this level of toxicity would be unacceptablein chronic administration of this natural product. It was onthis basis that the 13 g daily dose level was studied.

Although not the primary objective of this study, the pa-tient’s response to silybin-phytosome was monitored withmonthly PSA measurements while they were on study. Whileno objective PSA responses (i.e. > 50% reduction from base-line value) were observed in these patients, it is importantto recall that preclinical experiments using prostate cancercells show enhancement or synergistic activity when silib-inin is combined with cisplatin, carboplatin [21] or doxoru-bicin [22]. The recommended phase II dosing regimen fromthis study will be useful for the planning of future single-

agent and combination chemotherapy trials with silybin-phytosome.

In conclusion, this study demonstrates the tolerability ofhigh-dose, oral silybin-phytosome in patients with prostatecancer. The most significant toxicity observed was asymp-tomatic liver abnormalities, most notably unconjugated hy-perbilirubinemia, with chronic administration. Silibinin’sknown in vitro inhibition of UGT-1A1 is the likely cause ofthis finding. The recommended phase II oral dose of silybin-phytosome is 13 g daily, in 3 divided doses.

Acknowledgments The authors wish to acknowledge the invaluableresearch support of Spencer Green and Cynthia Snedden. This work wassupported in part by a grant from the National Center for Complemen-tary and Alternative Medicine (NCCAM), P30 CA046934 supplement.

References

1. Pares A, Planas R, Torres M, Caballeria J, Viver JM, Acero D,Panes J, Rigau J, Santos J, Rodes J (1998) Effects of silymarin inalcoholic patients with cirrhosis of the liver: results of a controlled,double-blind, randomized and multicenter trial. J Hepatol 28:615–621

2. Flora K, Hahn M, Rosen H, Benner K (1998) Milk thistle (Silybummarianum) for the therapy of liver disease. Am J Gastroenterol93:139–143

3. Ferenci P, Dragosics B, Dittrich H, Frank H, Benda L, Lochs H,Meryn S, Base W, Schneider B (1989) Randomized controlledtrial of silymarin treatment in patients with cirrhosis of the liver. JHepatol 9:105–113

4. Singh RP, Dhanalakshmi S, Tyagi AK, Chan DC, Agarwal C,Agarwal R (2002) Dietary feeding of silibinin inhibits advancehuman prostate carcinoma growth in athymic nude mice and in-creases plasma insulin-like growth factor-binding protein-3 levels.Cancer Res 62:3063–3069

5. Kohno H, Suzuki R, Sugie S, Tsuda H, Tanaka T (2005) Di-etary supplementation with silymarin inhibits 3,2′-dimethyl-4-aminobiphenyl-induced prostate carcinogenesis in male F344 rats.Clin Cancer Res 11:4962–4967

6. Kohno H, Tanaka T, Kawabata K, Hirose Y, Sugie S, Tsuda H, MoriH (2002) Silymarin, a naturally occurring polyphenolic antioxidantflavonoid, inhibits azoxymethane-induced colon carcinogenesis inmale F344 rats. Int J Cancer 101:461–468

7. Vinh PQ, Sugie S, Tanaka T, Hara A, Yamada Y, Katayama M,Deguchi T, Mori H (2002) Chemopreventive effects of a flavonoidantioxidant silymarin on N-butyl-N-(4-hydroxybutyl)nitrosamine-induced urinary bladder carcinogenesis in male ICR mice. Jpn JCancer Res 93:42–49

8. Singh RP, Mallikarjuna GU, Sharma G, Dhanalakshmi S, TyagiAK, Chan DC, Agarwal C, Agarwal R (2004) Oral silibinin in-hibits lung tumor growth in athymic nude mice and forms anovel chemocombination with doxorubicin targeting nuclear fac-tor kappaB-mediated inducible chemoresistance. Clin Cancer Res10:8641–8647

9. Lahiri-Chatterjee M, Katiyar SK, Mohan RR, Agarwal R (1999)A flavonoid antioxidant, silymarin, affords exceptionally high pro-tection against tumor promotion in the SENCAR mouse skin tu-morigenesis model. Cancer Res 59:622–632

10. Zi X, Agarwal R (1999) Silibinin decreases prostate-specific anti-gen with cell growth inhibition via G1 arrest, leading to differenti-

Springer

146 Invest New Drugs (2007) 25:139–146

ation of prostate carcinoma cells: implications for prostate cancerintervention. Proc Natl Acad Sci U S A 96:7490–7495

11. Zi X, Grasso AW, Kung HJ, Agarwal R (1998) A flavonoid antiox-idant, silymarin, inhibits activation of erbB1 signaling and inducescyclin-dependent kinase inhibitors, G1 arrest, and anticarcinogeniceffects in human prostate carcinoma DU145 cells. Cancer Res58:1920–1929

12. Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma J,Wilkinson P, Hennekens CH, Pollak M (1998) Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study.Science 279:563–566

13. Singh RP, Sharma G, Dhanalakshmi S, Agarwal C, Agarwal R(2003) Suppression of advanced human prostate tumor growth inathymic mice by silibinin feeding is associated with reduced cellproliferation, increased apoptosis, and inhibition of angiogenesis.Cancer Epidemiol Biomarkers Prev 12:933–939

14. Barzaghi N, Crema F, Gatti G, Pifferi G, Perucca E (1990) Phar-macokinetic studies on IdB 1016, a silybin-phosphatidylcholinecomplex, in healthy human subjects. Eur J Drug Metab Pharma-cokinet 15:333–338

15. Schandalik R, Perucca E (1994) Pharmacokinetics of silybin fol-lowing oral administration of silipide in patients with extrahepaticbiliary obstruction. Drugs Exp Clin Res 20:37–42

16. Hoh C, Boocock D, Marczylo T, Singh R, Berry DP, Dennison AR,Hemingway D, Miller A, West K, Euden S, Garcea G, Farmer PB,Steward WP, Gescher AJ (2006) Pilot study of oral silibinin, a puta-tive chemopreventive agent, in colorectal cancer patients: silibinin

levels in plasma, colorectum, and liver and their pharmacodynamicconsequences. Clin Cancer Res 12:2944–2950

17. Ghosal A, Hapangama N, Yuan Y, Achanfuo-Yeboah J, IannucciR, Chowdhury S, Alton K, Patrick JE, Zbaida S (2004) Identifi-cation of human UDP-glucuronosyltransferase enzyme(s) respon-sible for the glucuronidation of ezetimibe (Zetia). Drug MetabDispos 32:314–320

18. Sridar C, Goosen TC, Kent UM, Williams JA, Hollenberg PF(2004) Silybin inactivates cytochromes P450 3A4 and 2C9 andinhibits major hepatic glucuronosyltransferases. Drug Metab Dis-pos 32:587–594

19. Wasserman E, Myara A, Lokiec F, Goldwasser F, Trivin F,Mahjoubi M, Misset JL, Cvitkovic E (1997) Severe CPT-11 tox-icity in patients with Gilbert’s syndrome: two case reports. AnnOncol 8:1049–1051

20. van Erp NP, Baker SD, Zhao M, Rudek MA, Guchelaar HJ, NortierJW, Sparreboom A, Gelderblom H (2005) Effect of milk thistle(Silybum marianum) on the pharmacokinetics of irinotecan. ClinCancer Res 11:7800–7806

21. Dhanalakshmi S, Agarwal P, Glode LM, Agarwal R (2003) Silib-inin sensitizes human prostate carcinoma DU145 cells to cisplatin-and carboplatin-induced growth inhibition and apoptotic death. IntJ Cancer 106:699–705

22. Chlopcikova S, Psotova J, Miketova P, Simanek V (2004) Chemo-protective effect of plant phenolics against anthracycline-inducedtoxicity on rat cardiomyocytes. Part I. Silymarin and its flavono-lignans. Phytother Res 18:107–110

Springer

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.