Phase I Trial of a Combination of the Multikinase Inhibitor Sorafenib and the Farnesyltransferase Inhibitor Tipifarnib in Advanced Malignancies

Abstract

PURPOSE

We evaluated the safety, maximum tolerated dose, pharmacokinetics, and biologic effects of the combination of the Raf-1, RET, KIT, platelet-derived growth factor receptor (PDGFR) and VEGFR2 kinase inhibitor sorafenib and the farnesyltransferase inhibitor tipifarnib.

EXPERIMENTAL DESIGN

A standard 3+3 phase I dose-escalation design was used with a 28-day cycle (sorafenib daily and tipifarnib for 21 days, by mouth).

RESULTS

Fifty patients were treated; 43 reached restaging evaluation after cycle 2. The most common side effects were grade 1–2 rash, hyperglycemia and diarrhea. Dose-limiting toxicity was rash, and the recommended phase II dose is sorafenib 400 mg po qam/200 mg po qpm and tipifarnib po 100 mg BID. Despite the low doses of tipifarnib, one-quarter of patients had ≥50% reduction in farnesyltransferase (FTase) levels. Interestingly, 6 of 8 patients with medullary thyroid cancer (MTC) had durable stable disease (N=3) or partial remissions (N=3), lasting 12 to 26+ months. Five of the six responders had available tissue, and RET gene mutations were identified in them. Prolonged (≥ 6 months) stable disease was also seen in nine patients as follows: papillary thyroid cancer (N=4; 18+ to 27+ months); adrenocortical cancer (N=2; 7 and 11 months); and one each of melanoma (PDGFR mutation-positive) (14 months), renal (6 months) and pancreatic cancer (6 months).

CONCLUSION

Our study shows that the combination of tipifarnib and sorafenib is well tolerated. Activity was seen, especially in patients with medullary thyroid cancer, a tumor characterized by RET mutations.

STATEMENT OF TRANSLATIONAL RELEVANCE.

We are submitting an original article entitled “Phase I Trial of a Combination of the Multikinase Inhibitor Sorafenib and the Farnesyltransferase Inhibitor Tipifarnib in Advanced Malignancies”. We believe the findings in this phase I combination study represent an important contribution to the phase I, combination targeted therapy, and thyroid cancer literature. We have reported here the safety, pharmacokinetic, pharmacodynamics, responses and MTD of the unique combination of sorafenib, a mulitkinase inhibitor (Raf, RET) and the farnesyletranseferase (Ras) inhibitor.

Our data support that this combination is well tolerated and has shown significant activity in thyroid malignancies, particularly medullary thyroid cancer patients with RET kinase mutations. This is the first report of significant responses of medullary thyroid cancer with targeted therapies such as sorafenib and tipifarnib. Moreover, the unique combination adds to the growing body of literature of the tolerability, toxicities and activity of these rationally targeted combinations.

INTRODUCTION

The transition from normal cells to the cancer phenotype is mediated by processes such as uncontrolled cell division, apoptosis evasion, angiogenesis, and activation of abnormal signal transduction pathways. In multiple tumor types, multiple pathways implicated in these processes involve receptor tyrosine kinases and intracellular signals such as those transduced through Ras and Raf oncogenic proteins(1, 2). Thus, combination therapy with targeted agents is thought to be needed for producing optimal responses in cancer patients.

Sorafenib (Nexavar), an orally potent multikinase inhibitor of Raf-1, PDGFR, RET, KIT and VEGFR2, exerts in vivo antitumor efficacy against diverse human tumor xenografts and cell lines, and was approved by the US Food and Drug Administration (FDA) for treating renal cell and hepatocellular carcinoma(3, 4). Tipifarnib, a potent, selective farnesyltransferase (FTase) inhibitor, induces antiproliferative effects against many human tumor cell lines, and has clinical activity in a number of malignancies(1, 5–10). Farnesylation of Ras is the rate-limiting step in its posttranslational modification and is required for its oncogenic activity(1),(11),(12).

The development of FTase and Raf kinase inhibitors such as tipifarnib and sorafenib, respectively, provides a unique opportunity to test the hypothesis that by combining these agents, a synergistic or additive effect on the Ras/Raf/MEK/ERK and related pathways might be achieved clinically in advanced cancer. As a first step, we report a phase I study of the combination that describes its safety, toxicities, maximum tolerated dose (MTD), pharmacokinetics, pharmacodynamic effects and preliminary indications of efficacy.

PATIENTS, MATERIALS, AND METHODS

Patient eligibility and selection

Inclusion criteria: ≥18 years; histologically confirmed advanced cancer with ≤ 4 prior cytotoxic chemotherapies or no standard therapy that could increase survival by three months; Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2; Response Evaluation Criteria in Solid Tumors (RECIST)⁽13) measurable disease that was biopsiable, although biopsies were optional; leukocytes (≥3000/mcL); absolute neutrophil count (≥1500/mcL); platelets (≥1000,000/mcL); total bilirubin (≤1.5); aspartate aminotransferase (AST) (SGOT)/ alanine aminotransferase (ALT) (SGPT) ≤2.5 X the upper limit of normal (ULN); creatinine within ULN (≥60m/L/min/1.73m² for patients with creatinine levels above ULN); discontinuation of therapies ≥4 weeks prior to study entry.

Exclusion criteria: continuing grade 3 adverse events resulting from therapy administered ≥4 weeks earlier; central nervous system metastases except patients having prior radiation; allergies to imidazoles or compounds similar to sorafenib or tipifarnib; uncontrolled hypertension (systolic pressure >140 mm Hg, diastolic >90 mm Hg); current bleeding diathesis; ≥grade 2 peripheral neuropathy; uncontrolled intercurrent illness; New York Heart Association (NYHA ) classification >2; impaired swallowing; therapeutic anticoagulation; human immunodeficiency virus (HIV)-positive; pregnancy; childbearing potential individuals unwilling to use adequate contraception.

Study design

Study Design

The Division of Cancer Treatment and Diagnosis of the National Cancer Institute (NCI-CTEP) supplied both sorafenib (BAY43–9006, Nexavar) and tipifarnib (Zarnestra, R115777).

All patients signed a written, informed consent meeting M. D. Anderson Cancer Center Institutional Review Board (IRB) policy and NCI requirements. A standard 3+3 dose-escalation design was used (Table 1). Each cycle consisted of 28 days of sorafenib, and 21 days of tipifarnib (3 weeks on, 1 week off per 28-day cycle). Toxicity was graded according to the Cancer Therapy Evaluation Program Common Toxicity Criteria, version 3.0. Dose-limiting toxicity (DLT) was defined as any grade 4 hematologic toxicity delaying the next course for over 2 weeks, accompanied by a lasting infection or bleeding requiring hospitalization. A clinically significant non-hematologic DLT was defined as ≥ grade 3 adverse event possibly attributable to drug. Exceptions were alopecia, insomnia, weight gain, amenorrhea, and galatactorrhea. Grading for nausea, vomiting, and diarrhea was based upon toxicity despite maximal symptomatic treatment. The DLT window encompassed the first 28 days of treatment. The MTD was defined as dose level in which ≤1 of 6 patients experienced a DLT.

Table 1.

Dose escalation schedule^*

Dose Level	BAY 43-9006 (Sorafenib)	R115777 (Tipifarnib)	Patients (N)	Patients with dose reduction (N)	DLT's (No. and type)
Level −1	400 mg PO QOD	100 mg po QD	0	0	0
Level 1	400 mg PO QD	100 mg PO QD	5	0	0
Level 2	400 mg PO QD	100 mg PO BID	4	0	0
Level 3 (Phase II recommended dose or MTD)	400 mg PO qam, 200 mg qpm	100 mg PO BID	31	5	3 (Elevated Lipase and rash)
Level 4	400 mg PO qam, 200 mg PO qpm	200 mg PO BID	4	1	2 (Rash)
Level 5	400 mg PO qam, 200 mg PO qpm	300mg PO BID	0	0	0
Level 6*	400 mg PO BID	100 mg PO BID	6	2	2 (Rash)

^*Current modified dose levels. Trial was initially started at Level 6 but 2/6 patients experienced grade 3 rash and dosing was modified.

Baseline evaluations were conducted within one week prior to protocol start. Physical exams were conducted every three weeks, with weekly hematologic/biochemical laboratories. Scans were required within four weeks prior to therapy. RECIST response evaluation occurred every eight weeks and evidence of a partial or complete response, confirmed within four weeks.

Patients continued on treatment until disease progression, unacceptable adverse events, intercurrent illness preventing further administration, or patient withdrawal. Dose reduction occurred if grade 2–3 non-hematologic adverse events observed despite symptomatic treatment, excluding non-clinically significant metabolic or laboratory abnormalities.

Laboratory correlative studies

FTase enzyme analysis in peripheral blood mononuclear cells (PBMCs)

Blood samples (30 mL in ethylenediaminetetraacetic acid (EDTA) tubes) were drawn within 2 weeks before starting the drug combination and during day 21 of cycle 1. PBMCs were snap frozen at –80C° until analysis for FTase activity using previously described methods(14, 15). The FTase value for the pretreatment sample (baseline) was set at 100% and the FTase activity on day 21 presented as a percent of baseline.

Pharmacokinetics

Plasma drug levels were assessed during course 1 (weeks 1, 2, 3). Plasma samples (5 mL) from 24 patients were obtained: immediately prior to the combination, at 30 minutes, 2, 4, 6, and 24 h (before the following second daily dose). Plasma was removed, frozen, and stored at –80°C until analysis.

Levels of tipifarnib were assessed using a validated HPLC-UV method. Plasma was alkalinized with Na hydroxide (0.01 mol/L) and then extracted with heptane:isoamyl alcohol (90:10, v/v). An internal standard (IS; R121550) was used to correct for extraction efficiency. The chromatographic peaks of both tipifarnib and R121550 were detected at 240 nm. Based on quality control samples assayed along with patient samples, intra-day and inter-day precision ranged from 3.9% to 11.2%, and accuracy ranged from 92.0% to 1.5.5%. Identity of peaks measured as tipifarnib was confirmed by LC/MS/MS. Analysis of sample content of sorafenib was determined by a validated LC/MS/MS assay (b) with a lower limit of quantitation of 1 ng/mL. Based on quality control samples, intraday and interday precision ranged from 2.4% to 14.3%, and accuracy ranged from 89.8% to 111.5%. Tolfonate (Sigma Chemical, St. Louis, MO) was used as an internal standard. Following protein precipitation, drug and IS were extracted with diethyl ether which was then dried under N_2; residue was reconstituted in MeOH prior to analyses.

RET, KIT and PDGFR sequencing

Hereditary medullary thyroid cancer (MTC) patients have germline RET mutations, and a subset of patients with sporadic MTC also harbor RET mutations in tumor. To test for these mutations, DNA was extracted from paraffin-embedded tumor using the DNeasy Tissue kit (Qiagen, Maryland, USA). Polymerase chain reaction (PCR) was performed to amplify exons 10,11,13,14,15,16 of the RET gene. Similarly, exons 9, 11, 13, and 17 of the KIT gene, and exons 12 and 18 of the PDGFR gene(16) were examined in a melanoma patient with prolonged stable disease. PCR was done using LA-Taq (TaKaRa, Otsu, Shiga, Japan) and was carried out in a PTC-100 thermocycler (MJ Research, Watertown, MA). After the Exonuclease I – Shrimp Alkaline Phosphatase purifying method (Roche, Indianapolis, IN), the products were directly sequenced in an ABI PRISM 3730 Genetic Analyzer (Applied Biosystems, Foster City, CA). Mutations were cross-referenced with the Human Gene Mutation Database (http://www.hgmd.cf.ac.uk/ac/index.php), Entrez SNP (http://www.ncbi.nlm.nih.gov/sites/entrez) and PubMed (http://www.pubmed.org).

RESULTS

Patient demographics

From November, 2005 to December, 2007 fifty patients were enrolled. Their median age was 56 years (range, 18–81 years) (Table 2). Forty-three patients reached their first restaging evaluation. Seven patients came off study early due to toxicities (rash N=2; elevated lipase N=2) or co-morbidities unrelated to drug (atrial fibrillation N=1, secondary malignancy N=1, thrombus N=1) necessitating early withdrawal.

Table 2.

Dose escalation schedule^*

(n=50)
Characteristics
Gender, n(%)
Male	25	(50)
Female	25	(50)
Race, n (%)
Caucasian	44	(88)
Others	6	(12)
Median Age, (y)	56
Age Range	18 – 81
ECOG, n (%)
0	21	(42)
1	26	(52)
2	3	(6)
Diagnosis, n (%)
Thyroid	15	(30)
Medullary	8	(16)
Papillary	5	(10)
Follicular	1	(2)
Anaplastic	1	(2)
Melanoma	7	(14)
Breast	6	(12)
RCC	3	(6)
Sarcoma	3	(6)
Colorectal	3	(6)
Head & Neck	3	(6)
Adrenal Cortical	2	(4)
Pancreatic	4	(8)
HCC	1	(2)
Thymoma	1	(2)
SCC of the Skin	1	(2)
Lung cancer	1	(2)
Sites of disease
1–3	43	(86)
≥ 4	7	(14)
Prior anticancer therapy regimens (incl immunotherapy), n (%)
0	7	(14)
1	11	(22)
2	10	(20)
3	8	(16)
≥ 4	14	(28)

Dose escalation and maximum tolerated dose

No patients received therapy at dose levels −1 or 5 (Table 1). The initial protocol started at dose level 6 (sorafenib 400 mg BID; tipifarnib 100 mg BID) but two of six patients developed a grade 3 maculopapular erythematous rash requiring hospitalization, and dosing was modified. Five were treated at new dose level 1 (sorafenib 400 mg QD; tipifarnib 100 mg QD) with one patient coming off early after discovery of a second malignancy (gastric cancer) and a second for early progression. Four were treated at dose level 2, with one coming off before completion of the DLT window due to progression. Three were treated at dose level 3, with one patient developing a grade 3 rash. Therefore, an additional three were enrolled with no further DLTs. At dose level 4, two out of four patients developed a grade 3 rash, therefore, the MTD was defined as dose level 3 (sorafenib 400 mg QAM, 200 mg QPM; tipifarnib 100 mg BID). An additional 25 patients were treated at the MTD to further delineate the pharmacokinetics, pharmacodynamics, and activity.

Safety

The most common toxicities for all cycles were grade 1–2 rash (48% of patients), hyperglycemia-nonfasting (46%), and diarrhea (38%) (Table 3). The most common dose-limiting toxicity was grade 3 rash. Two patients at expansion experienced grade 3 lipase elevation without symptoms but were taken off after dose reduction and continued lipase elevation.

Table 3.

Incidence of all drug-related toxicities

Category	Grade	Soraf 400 bid, Tipi 100 bid (n=6), n (%)		Soraf 400 qd, Tipi 100 qd (n=5), n (%)		Soraf 400 qd, Tipi 100 bid (n=4), n (%)		Soraf 400 qam & 200 qpm, Tipi 100 bid (n=31), n (%)		Soraf 400 qam & 200 qpm, Tipi 200 bid (n=4), n (%)		Total (n=50), n (%)
Blood/Bone marrow
Anemia	1 or 2	1	(17)	0	(0)	1	(25)	1	(3)	0	(0)	3	(6)
	3	0	(0)	0	(0)	0	(0)	1	(3)	0	(0)	1	(2)
Lymphopenia	1 or 2	3	(50)	0	(0)	1	(25)	7	(23)	2	(50)	13	(26)
	3	2	(33)	1	(20)	1	(25)	5	(16)	1	(25)	10	(20)
Thrombocytopenia	1 or 2	0	(0)	0	(0)	0	(0)	0	(0)	1	(25)	1	(2)
	3	0	(0)	0	(0)	0	(0)	1	(3)	0	(0)	1	(2)
Cardiac
Hypertension	1 or 2	2	(33)	2	(40)	0	(0)	0	(0)	0	(0)	4	(8)
Constitutional Symptoms
Fatigue (asthenia, lethargy, malaise)	1 or 2	2	(33)	0	(0)	1	(25)	11	(35)	2	(50)	16	(32)
	3	0	(0)	1	(20)	0	(0)	1	(3)	1	(25)	3	(6)
Dermatology
Hand/foot	3	1	(17)	0	(0)	0	(0)	2	(6)	0	(0)	3	(6)
Rash	1 or 2	2	(33)	3	(60)	1	(25)	16	(52)	2	(50)	24	(48)
	3	2	(33)	0	(0)	0	(0)	1	(3)	1	(25)	4	(8)
Gastrointestinal
Anorexia	1 or 2	1	(17)	1	(20)	1	(25)	6	(19)	1	(25)	10	(20)
Nausea/Vomiting	1 or 2	3	(50)	0	(0)	2	(50)	3	(10)	2	(50)	10	(20)
Diarrhea	1 or 2	2	(33)	1	(20)	2	(50)	12	(39)	2	(50)	19	(38)
Mucositis	1 or 2	1	(17)	0	(0)	0	(0)	4	(13)	1	(25)	6	(12)
Metabolic/Lab
Elevated Bilirubin	1 or 2	0	(0)	0	(0)	0	(0)	3	(10)	2	(50)	5	(10)
Elevated AST/ALT	1 or 2	2	(33)	2	(40)	2	(50)	9	(29)	1	(25)	16	(32)
Elevated Lipase	1 or 2	0	(0)	0	(0)	0	(0)	1	(3)	0	(0)	1	(2)
	3	0	(0)	0	(0)	0	(0)	4	(13)	0	(0)	4	(8)
Hyperglycemia	1 or 2	5	(83)	3	(60)	2	(50)	10	(32)	3	(75)	23	(46)
	3	0	(0)	0	(0)	0	(0)	1	(3)	0	(0)	1	(2)
Neurology
Neuropathy	1 or 2	0	(0)	0	(0)	0	(0)	5	(16)	0	(0)	5	(10)
Renal
Elevated creatinine	1 or 2	1	(17)	0	(0)	0	(0)	0	(0)	0	(0)	1	(2)
Other
Sq. Skin cancer	3	1	(17)	0	(0)	0	(0)	0	(0)	0	(0)	1	(2)

Pharmacology

Plasma levels of tipifarnib and sorafenib were assessed during course 1 (weeks 1, 2, 3) (Figure1). Complete plasma sample collections were obtained from 24 patients. Plasma levels of tipifarnib (100 mg BID) reached equilibrium within 6 hours and were maintained at approximately 100 ng/mL during the 21-day treatment cycle (Figure 1B). In contrast, steady-state plasma levels of sorafenib were reached after 7 days with only slight additional accumulation (Figure 1A). Neither sorafenib’s frequency nor its dose had an apparent effect on tipifarnib plasma levels or vice versa. Plasma levels of sorafenib (2.5 – 4 µg/mL) were similar to those reported(17). Tipifarnib steady-state levels, however, averaged 100 ng/mL and were lower than those previously reported, but the small number of patients in previous reports at a 100 mg BID dose level precluded determining whether these differences were significant (18, 19).

Figure 1.

Steady-state plasma levels of A. sorafenib (ng/mL), and B. tipifarnib (ng/mL). Error bars are standard-error of the mean.

Analysis of patient PBMCs for FTase activity

Of 50 enrolled patients, 31 had matching basal and on-study PBMC samples were analyzed for FTase activity summarized as percent of baseline activity of FTase (Table 4). All patients, except for patient 24 (200 mg BID) received 100 mg BID of tipifarnib. There was no correlation between the dose of sorafenib administered and inhibition of FTase activity.

Table 4.

FTase activity level in PBMCs

Cancer Type	Patient#	Tipifarnib Dose (mg)	Sorafenib Dose (mg)	FTase PBMC D21 (% of baseline)
Colorectal	4	100 mg BID	400 mg BID	44.1
Renal cell	6	100 mg BID	400 mg BID	112.1
Thymic	7	100 mg BID	400 mg QD	96
Melanoma	9	100 mg BID	400 mg QD	69.6
Sarcoma	11	100 mg BID	400 mg QD	87.3
Melanoma	12	100 mg BID	400mg QD	64.5
Breast	13	100 mg BID	400mg QD	66.7
Non-small cell lung	14	100 mg BID	400mg QD	54.4
Adrenal Cortical	15	100 mg BID	400 mg QD	54.8
Melanoma	16	100 mg BID	400 mg QAM/200 mg QPM	67.2
Colorectal	17	100 mg BID	400 mg QAM/200 mg QPM	101.1
Squamous cell skin	18	100 mg BID	400 mg QAM/200 mg QPM	187.1
Head & Neck	22	100 mg BID	400 mg QAM/200 mg QPM	39
Medullary Thyroid	24	200 mg BID	400 mg QAM/200 mg QPM	100.6
Breast	28	100 mg BID	400 mg QAM/200 mg QPM	68.5
Head & Neck	30	100 mg BID	400 mg QAM/200 mg QPM	56.3
Follicular Thyroid	31	100 mg BID	400 mg QAM/200 mg QPM	40
Adrenal Cortical	32	100 mg BID	400 mg QAM/200 mg QPM	57
Breast	36	100 mg BID	400 mg QAM/200 mg QPM	48.8*
Medullary Thyroid	37	100 mg BID	400 mg QAM/200 mg QPM reduced to 400 mg QD	209.3*
Breast	38	100 mg BID	400 mg QAM/200 mg QPM reduced to 400mg QD	6.8*
Pancreatic	39	100 mg BID	400mg QAM/200 mg QPM reduced to 400 mg QD	76.6*
Anaplastic Thyroid	40	100 mg BID	400 mg QAM /200 mg QPM	532*
Hepatocellular	41	100 mg BID	400 mg QAM /200 mg C reduced to 400 mg QD	72.3*
Papillary Thyroid	42	100 mg BID	400 mg QAM /200 mg QPM	37.3*
Pancreatic	43	100 mg BID	400 mg QAM /200 mg QPM	24.9*
Pancreatic	45	100 mg BID	400 mg QAM /200 mg QPM	39.3*
Medullary Thyroid	47	100 mg BID	400 mg QAM /200 mg QPM	50.3*
Medullary Thyroid	48	100 mg BID	400 mg QAM /200 mg QPM reduced to 400 mg QD	84*
Medullary Thyroid	49	100 mg BID	400 mg QAM /200 mg QPM	55.8*
Papillary Thyroid	50	100 mg BID	400mg QAM /200 mg QPM	48.9*

^*some PBMC samples contained enough protein for 2 experiments and the number is an average of the 2 experiments.

FTase activity showed less inhibition than previous reports even after taking tipifarnib for three weeks, probably because of the low doses of tipifarnib(14, 15), but in spite of the low dose, 8/31 analyzable samples showed inhibited FTase activity ≥50%. Three patients had increased FTase activation on day 21 of cycle 1 (187%, 209.3%, and 532%) compared to basal FTase activity.

Tumor responses

Overall, 43 of 50 patients reached first restaging; the other seven came off early due to toxicities (N=4) or unrelated co-morbidities (N=3). Twenty (47%) of the 43 patients who reached their first restaging had progression as their best response.

Twelve of 15 patients with thyroid cancer enrolled reached first restaging. Figure 2a depicts their best response. Three of the six patients with medullary thyroid cancer who reached first restaging had a partial response, lasting 14, 16+ and 26+ months, while three had minor regressions (stable disease) lasting 12 to 16 months. None of the six had a family history of MEN2 or familial thyroid cancer. RET mutation analysis from tumor was positive in all five for whom tissue was available (see RET mutation analysis). All six patients also had marked reductions in calcitonin, and five of the six had a decrease in CEA (Figure 2b).

Figure 2.

Figure 2a. Waterfall plot showing best response (by RECIST) of patients with thyroid cancer who reached their first restaging evaluation. All five of the MTC patients with available paraffin blocks had RET kinase mutations of exon 11. These mutations locate in the extracellular cysteine rich domain, which can cause ligand independent homodimerization and constitutive activation of the RET kinase.

Figure 2b. Best percent change in calcitonin and CEA levels from baseline in the six patients with medullary thyroid cancer who reached their first restaging. All six patients had a rapid decline in their calcitonin levels, and all patients except patient #2 had a significant decrease in CEA levels.

All four patients with papillary thyroid cancer who reached a first restaging had durable tumor regressions (16%–20%) lasting 18+, 19+, 20, and 27+ months. The patient with follicular thyroid cancer had short-lived tumor stability (4 months) and the patient with anaplastic thyroid cancer showed rapid tumor progression (1 month).

Ten additional patients had stable disease for 4 to14 months. These patients included three of the seven patients with melanoma (4, 4, and 14 months), both patients with adrenocortical cancer (7 and 11 months), two of three patients with renal cancer (5 and 6 months), one of four patients with pancreatic cancer (6 months), one of six patients with breast cancer (5 months), and one of three patients with colorectal cancer (4 months). Of interest, the patient with melanoma who remained stable for 14 months showed a mutation of PDGFR-alpha but not of KIT. The tumor was found to have a heterozygous mutation (GTC to CTC at codon 824) that resulted in an amino acid substitution from valine to leucine. This mutation was not evident in normal tissue sequenced from the same patient.

In addition, we reviewed the prior history of our thyroid cancer patients, including assessment of their scans immediately prior to treatment. Because patients had a variety of treatments immediately prior to tipifarnib and sorafenib, including radioactive iodine, surgery, and systemic agents, and their scans were performed over a variety of intervals. Over a median of 3 months (range, 2 to 5 months), the patients with thyroid cancer (excluding anaplastic cancer) showed a median of 39% (range, 14 to 110%) increase in disease by RECIST. This reflects the fact that patients with the most aggressive disease were generally referred to the phase I clinic.

RET mutational analysis

All five of the patients with medullary thyroid cancer who reached their first restaging and who had available blood had no identifiable germline RET mutation in blood as per standard screening for MEN2/FMTC (familial medullary thyroid carcinoma) and no family history of MEN2/FMTC (Figure 2a). One patient declined testing and had no family history of MEN2/FMTC. All five patients who reached first restaging and who had available paraffin tissue had an exon 11 RET mutation (Figure 2a). Of interest, one patient had a novel RET mutation of exon 11, a 6-bp deletion (TGTGCG) seen as double peaks after codon 628. This deletion alters Leu, Cys and Asp at codons 629 to 631 and has been previously reported by us(20). All the mutations are located in the extracellular cysteine-rich domain, which can cause ligand independent homodimerization and RET kinase constitutive activation.

DISCUSSION

This phase I study of combined sorafenib and tipifarnib demonstrated safety and excellent tolerance, albeit with each drug given at lower doses than the individual recommended doses. The recommended phase II dose for the combination is sorafenib 400 qam/200 mg qpm and tipifarnib 100 mg BID. For the most part, toxicities of the combination were similar to those seen with tipifarnib and sorafenib as single agents, and mainly included diarrhea and rash. Rash was the dose-limiting toxicity, and this was not surprising given that rash develops in a significant minority of patients treated with either sorafenib or tipifarnib. Of note, 46% of patients developed grade 1–2 hyperglycemia (non-fasting), which is high relative to what might be expected from either agent alone(3, 21, 22). The reason for the hyperglycemia is unclear, but it has recently been demonstrated that side effect similarities may point to unexpected drug targets.²⁵ It is therefore conceivable that the hyperglycemia reflects an off-target effect on a signal such as mTOR, given that mTOR inhibitors result in frequent hyperglycemia²⁶. One patient developed multiple squamous cell skin cancers within three months of starting treatment²⁷. The treatment was stopped and the tumors excised and did not recur. This effect is most likely due to sorafenib, as it has been reported with this agent²⁸.

Our pharmacokinetic studies show that steady-state plasma levels of sorafenib are similar to those reported elsewhere¹⁹. Steady-state levels of tipifarnib are low, but the small number of patients treated at the 100 mg bid dose level in previous studies precludes a comparison^20,21. Of interest, despite the low doses of tipifarnib, 8 of 31 patients evaluated showed >50% inhibition of FTase activity. Regardless, studies have demonstrated that even these low doses of tipifarnib can induce responses, including complete remissions, in myelodysplastic syndrome or acute myelogenous leukemia patients(10, 14, 15).

Recently, investigators have reported significant responses and increased median progression-free survival with sorafenib in patients with differentiated (papillary and follicular) thyroid cancers(23, 24) . Our study also demonstrates significant activity for tipifarnib/sorafenib in patients with papillary thyroid cancer, with all four treated showing regressions lasting 18+ to 27+ months. These patients had shown clear progression by RECIST (median progression=39%) over a median of 3 months before starting the study. In addition, we report significant responses in patients with medullary thyroid cancer with all six patients who reached their first restaging achieving prolonged stable disease or a partial response (duration = 12 to 26+ months) (Figure 2a). In all cases, calcitonin dropped significantly, as did carcinoembryonic antigen (CEA) in five of the patients (Figure 2b) with none showing any significant new side effects after long-term administration.

Medullary thyroid cancer can be hereditary or sporadic, with the molecular hallmark of the hereditary form being germ-line mutations in the RET kinase gene. A subset of patients with sporadic medullary thyroid cancer, especially those with more aggressive disease, will demonstrate mutations in RET kinase in their tumors³¹. In our study, of the five responders with medullary thyroid cancer who had paraffin tissue available for RET mutation analysis, each had an activating mutation. It is unclear if the activity of tipifarnib/sorafenib in medullary thyroid cancer was due entirely to sorafenib’s inhibition of RET as the RET kinase pathway is complex. Ligand activation of the RET kinase activates a cascade of signaling pathways, e.g., JAK, MAPK, c-Jun, NH2-terminal kinase, Ras/Raf/MAPK, NF-κB, and the PI3K/AKT pathways³¹. Because FTase inhibitors can also inhibit AKT and MEK activation, it is possible, but not known, whether tipifarnib inhibited these pathways, thereby increasing sorafenib’s activity against RET. Moreover, it is conceivable that sorafenib’s suppression of VEGFR kinase also contributed to the medullary thyroid cancer responses.

In conclusion, we demonstrate that combining sorafenib and tipifarnib is well tolerated at doses up to and including sorafenib 400 mg po qam/200 mg po qpm and tipifarnib 100 mg po BID. The most clinically significant side effect was rash. Of interest, patients with medullary thyroid cancer bearing RET kinase mutations had durable partial responses (N=3) or stable disease (N=3), lasting 12 to 26+ months(20), and four patients with papillary thyroid cancer have been stable for 18+ to 27+ months. Prolonged stable disease (14 months) was also seen in a patient with melanoma harboring a PDGFR-alpha mutation; responses to sorafenib have also been previously reported anecdotally in patients with melanoma bearing a KIT mutation(25). In addition, ≥ 6 months stable disease was observed in patients with adrenocortical (7,11 months), renal (6 months) and pancreatic cancer (6 months), suggesting that the activity and mechanism of action of tipifarnib/sorafenib in these malignancies warrants additional exploration (21). Finally, larger studies are planned in patients with medullary thyroid cancer to better assess response rates, and determine whether response is based on the activity of one drug, i.e., sorafenib, or if the suppression of multiple pathways by combining the two drugs, contributed to the salutary effects seen.