Inhibition of the Ras/Raf/MEK/ERK and RET Kinase Pathways with the Combination of the Multikinase Inhibitor Sorafenib and the Farnesyltransferase Inhibitor Tipifarnib in Medullary and Differentiated Thyroid Malignancies

Inhibiting the Ras/Raf/MEK/ERK and RET kinase pathways with sorafenib and tipifarnib is well tolerated and active against thyroid cancer.

Abstract

Purpose:

Ras/Raf/MAPK kinase/ERK and rearranged in transformation (RET) kinase pathways are important in thyroid cancer. We tested sorafenib, a B-Raf, RET, and vascular endothelial growth factor receptor kinase inhibitor, combined with tipifarnib, a farnesyltransferase inhibitor that inactivates Ras and other farnesylated proteins.

Patients and Methods:

We treated 35 patients with differentiated thyroid cancer (DTC) and medullary thyroid cancer (MTC) in a phase I trial. Sorafenib and tipifarnib were given for 21 d with 7 d rest in each 28-d cycle.

Results:

We enrolled 22 patients with metastatic DTC (16 papillary, five follicular, and one poorly differentiated) and 13 patients with MTC, of whom 15 with DTC and 10 with MTC reached first restaging. When tissue was available, eight of 15 DTC patients (53%) had B-Raf mutations; eight of 13 MTC (61.5%) patients had RET mutations. MTC partial response rate was 38% (five of 13) (duration = 9+, 12, 13, 16+, and 34+ months), stable disease of at least 6 months was 31% (four of 13). The DTC partial response rate was 4.5% (one of 22), and stable disease of at least 6 months was 36% (eight of 22). Median progression-free survival for all 35 patients was 18 months (95% confidence interval, 14.6 to not reached months). Median overall survival has not been reached, with a median follow-up of 24 months with 80% overall survival. Grade 1–2 toxicities were mainly rash, fatigue, and diarrhea. The most common grade 3–4 toxicities were rash, rise in amylase/lipase, and fatigue.

Conclusions:

Inhibiting the Ras/Raf/MAPK kinase/ERK and RET kinase pathways with sorafenib and tipifarnib is well tolerated and active against thyroid cancer.

The incidence of thyroid cancer in the United States has increased over the last several years with 37,340 new cases reported in 2008 (Surveillance Epidemiology and End Results) (1, 2). Although overall survival (OS) is generally excellent in patients with localized disease, metastatic disease that cannot be treated with surgery or radioactive iodine is associated with short survival rates (3–5). The only Food and Drug Administration (FDA)-approved treatment for advanced differentiated thyroid malignancies (papillary and follicular) is doxorubicin, with a progression-free survival (PFS) of 2 months and median OS of 8 months (6). Although published trials with doxorubicin have been promising (7, 8), no phase III trial showing benefit has ever been published. Furthermore, the addition of doxorubicin to combination chemotherapy has not shown any advantage over single-agent doxorubicin (6, 9). Other thyroid cancers such as medullary thyroid cancer (MTC), which are metastatic and surgically unresectable, also have a poor prognosis, with survival rates of 25% at 5 yr and 10% at 10 yr, and are usually unresponsive to conventional cytotoxic chemotherapy and radiation therapy (9). Recently, several small-molecule kinase inhibitors such as motesanib, axitinib, sunitinib, and sorafenib have been tested in thyroid cancer with some benefit (10–16).

Activating mutations in certain genes reportedly play a critical role in papillary thyroid malignancies (B-Raf), follicular thyroid (Ras), and MTC [rearranged in transformation (RET)] (9, 17). Moreover, papillary thyroid carcinoma (PTC) gene rearrangements of RET tyrosine kinase result in a RET/PTC oncogene that has been shown to be constitutively activated in poorly differentiated thyroid tumors (18).

Sorafenib (BAY 43-9006, Nexavar), a potent oral multikinase inhibitor of Raf-1, platelet-derived growth factor receptor, RET, KIT, and vascular endothelial growth factor receptor-2, is approved by the FDA for the treatment of renal cell and hepatocellular carcinoma (19–22). Tipifarnib (R115777, Zarnestra), a selective oral farnesyltransferase (FTase) inhibitor, induces antiproliferative effects against various human tumor cell lines, with clinical activity in several types of malignancies (23–31). Farnesylation of Ras is the rate-limiting step in its posttranslational modification and is required for its oncogenic activity (21, 24, 25).

Given the importance of the Ras/Raf/MAPK kinase/ERK and RET kinase pathways in thyroid malignancies, the combination of tipifarnib and sorafenib provides an opportunity to test the hypothesis that clinical activity might be achieved by using the combination in advanced thyroid cancer. We conducted a phase I study of tipifarnib and sorafenib given together in patients with advanced solid tumors. In an earlier study, we reported the results of the dose escalation part of the phase I study with the finding that the recommended phase II dose was sorafenib 400 mg in morning (qam) and 200 mg in evening (qpm) and tipifarnib 100 mg by mouth (po) twice daily (bid) and the dose-limiting toxicity was rash (19). Here, we describe the results of 35 differentiated thyroid cancer (DTC) and MTC patients treated on this study, primarily as part of a dose-expansion cohort.

Patients and Methods

Eligibility

Eligibility criteria included age at least 18 yr; histologically confirmed advanced cancer with no more than four previous cytotoxic chemotherapies or no standard therapy that could increase survival by 3 months, Eastern Cooperative Oncology Group performance status no higher than 2, Response Evaluation Criteria in Solid Tumors (RECIST) (32) measurable disease that was able to be biopsied (optional), leukocytes at least 3000/μl, absolute neutrophil count at least 1500/μl, platelets at least 1000,000/μl, total bilirubin no higher than 1.5, aspartate aminotransferase (serum glutamic oxaloacetic transaminase)/alanine aminotransferase (serum glutamic pyruvic transaminase) no higher than 2.5 × the upper limit of normal (ULN), creatinine within ULN (glomerular filtration rate ≥60 m/liter · min · 1.73 m² for patients with creatinine levels above ULN), and discontinuation of therapies at least 4 wk before study entry.

Exclusion criteria included ongoing treatment-related grade 3 adverse events, central nervous system metastases except those treated with previous radiation, allergies to imidazoles or compounds similar to sorafenib or tipifarnib, inadequately controlled hypertension (systolic pressure >140 mm Hg, diastolic >90 mm Hg), current bleeding diathesis, at least grade 2 peripheral neuropathy, uncontrolled intercurrent illness, New York Heart Association classification higher than 2, inability to take oral medications, therapeutic anticoagulation, HIV positive, pregnancy, and individuals with childbearing potential unwilling to use adequate contraception.

Study design

A standard phase I 3 + 3 dose-escalation design was used and described in our previous report (19). Each cycle consisted of 28 d of sorafenib and 21 d of tipifarnib (21 d on, 7 d off per 28-d cycle).

Bayer Inc. and Johnson & Johnson supplied both sorafenib and tipifarnib, respectively, through the Division of Cancer Treatment and Diagnosis of the National Cancer Institute (NCI-Cancer Therapy Evaluation Program). All patients signed a written, informed consent in accordance with the MD Anderson Cancer Center Institutional Review Board (IRB) policy and National Cancer Institute requirements.

Patient monitoring

Baseline evaluations were conducted within 7 d before protocol start, and physical exams were conducted every 21 d, with weekly hematological/biochemical laboratories. Scans of all known disease were required within 28 d before therapy. RECIST response evaluation occurred every 8 wk and evidence of a partial response (PR) or complete response, confirmed within 4 wk.

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

Endpoints and statistical considerations

Patient demographic data were summarized using descriptive statistics. The best tumor responses for individual patients, calculated as a percent change in target lesions compared with baseline, were plotted graphically on a waterfall plot.

The Kaplan-Meier method was used to describe time to treatment failure (TTF), PFS, and OS after therapy and to provide estimates for the median time to event at respective time points. TTF was defined as the length of time from initiation of treatment to discontinuation of treatment due to toxicity, disease progression, or death, whichever came first. PFS was defined as the length of time during and after treatment in which a patient was alive and without progressive disease (PD). If the drug was stopped before PD occurred, and a different therapy initiated, the end date was the date the patient started on the new therapy. All other patients still on treatment or free of progression on their last follow-up visit were censored on that date for TTF and PFS, respectively. OS was defined as the time period from start of treatment to the last treatment. All patients still alive at last follow-up were censored on that date.

Correlative studies

RET, B-Raf, and KRAS sequencing

Hereditary MTC patients have germline RET mutations, and a subset of patients with sporadic MTC also harbor somatic RET mutations in their tumors. To test for RET mutations in MTC, DNA was extracted from paraffin-embedded tumor using the DNeasy tissue kit (QIAGEN, Gaithersburg, MD). PCR was performed to amplify exons 10, 11, 13, 14, 15, and 16 of the RET gene. Similarly, we tested for somatic mutations in the B-Raf (V600E) protooncogene in codons 595–600 from exon 15 and KRAS protooncogene in codons 12, 13, and 61 in available tumor blocks from papillary and follicular thyroid cancer patients, respectively. DNA was extracted from microdissected paraffin-embedded tumor tissue and analyzed by PCR, and a pyrosequencing method was performed, with the sensitivity being approximately one in 10 mutation-bearing cells in the microdissected area. PCR was done using LA-Taq (Takara Bio, Otsu, Shiga, Japan) and was carried out in a PTC-100 thermal cycler (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 October 2009, 76 patients were enrolled in the study, with a total of 35 with MTC or DTC. Thirteen patients had MTC, 22 had metastatic DTC (16 papillary, five follicular thyroid, and one poorly differentiated DTC). The demographics of patients with MTC and DTC are shown in Table 1. All DTC patients were radioactive-iodine refractory. None of the DTC or MTC patients had previously received sorafenib. Three patients received sorafenib (400 mg qam and 200 mg qpm) and tipifarnib (200 mg bid), whereas the remaining patients received the combination at the maximum tolerated dose (sorafenib 400 mg qam and 200 mg qpm and tipifarnib 100 mg bid). Twenty-five of the 35 patients with thyroid cancer reached their first restaging evaluation, whereas 10 patients came off study early: five (14%) withdrew due to toxicities, three patients withdrew consent, and two patients withdrew due to a comorbidity unrelated to study medications.

Table 1.

Baseline characteristics (n = 35)

Characteristics	Baseline value
Gender [n (%)]
Male	20 (57)
Female	15 (43)
Race [n (%)]
White	24 (69)
Others	11 (31)
Median age (yr)	61
Age range (yr)	33–87
ECOG [n (%)]
0	12 (34)
1	23 (66)
Diagnosis [n (%)]
Papillary	16 (46)
Medullary	13 (35)
Follicular	5 (11)
Poorly differentiated	1 (3)
Sites of disease
1–3	24 (69)
≥4	11 (31)
Lymph node	24 (69)
Lung	29 (83)
Liver, kidney, or adrenal	13 (37)
Bone	13 (37)
Previous anticancer therapy regimens (including immunotherapy) [n (%)]
0	7 (20)
1	20 (57)
2	2 (6)
3	4 (11)
≥4	2 (6)
Baseline tumor markers and RECIST measurements [mean/median (range)]
MTC patients
Calcitonin (pg/ml)	8228.7/6369.5 (11.4–18,949)
CEA (ng/ml)	241.7/236.3 (2.4–693.4)
RECIST (cm)	6.9/5.1 (1.2–14.6)
DTC patients
Thyroglobulin (ng/ml)	16,136/559 (0–180,325)
RECIST (cm)	7.8/6.5 (2.7–15.4)

ECOG, Eastern Cooperative Oncology Group; CEA, carcinoembryonic antigen.

Safety and toxicities

The recommended phase II dose was determined to be sorafenib 400 mg qam and 200 mg qpm and tipifarnib 100 mg bid, and this was the dose at which most patients, 32 of 35, with thyroid cancer were initially treated, because they were enrolled in the dose-expansion cohort (19). The most common toxicities in patients with thyroid cancer for all cycles were grade 1–2 rash (40%), fatigue (49%), and diarrhea (46%) (Table 2). Compared with other previously reported patients, the toxicities in the thyroid group in this study were similar to all study patients (19). The most common grade 3 event was rash in the original phase I study and likewise for the thyroid population. Five patients with DTC or MTC came off study early due to drug-related toxicity; two experienced grade 3 lipase elevation without symptoms but had continued lipase elevation after dose reduction and were taken off study, and three patients had continued grade 2–3 rash despite dose reduction and were also taken off study. Fourteen of the 35 thyroid patients (40%) had a dose reduction at some point during the study due to toxicities, with six of the 14 dose reductions secondary to grade 3 toxicities, whereas the remaining dose reductions were secondary to grade 2-related toxicities as required by protocol. A total of 362 cycles of therapy were administered to all patients, and the median dose of both drugs was at the maximum tolerated dose of tipifarnib 100 mg bid and sorafenib 400 mg qam and 200 mg qpm, taking into consideration the dose reductions.

Table 2.

Toxicities

Category	Grade	n (%)
		Sorafenib 400 mg/qam 200 mg qpm		Total (n = 35)
		Tipifarnib 100 bid (n = 31)	Tipifarnib 200 bid (n = 4)
Cardiac
Hypertension	1 or 2	4 (13)	1 (25)	5 (14)
	3	1 (3)	0 (0)	1 (3)
Hypotension	1 or 2	1 (3)	0 (0)	1 (3)
Left ventricular failure	1 or 2	1 (3)	0 (0)	1 (3)
Dermatology/skin
Hand foot skin reaction	1 or 2	11 (35)	2 (50)	13 (37)
Rash	1 or 2	13 (42)	1 (25)	14 (40)
	3	3 (10)	1 (25)	4 (11)
Pruritus	1 or 2	2 (6)	1 (25)	3 (9)
Alopecia	1 or 2	8 (26)	0 (0)	8 (23)
Cellulitis	1 or 2	1 (3)	0 (0)	1 (3)
Squamous cell carcinoma of the skin	1 or 2	0 (0)	0 (0)	0 (0)
	3	1 (3)	0 (0)	1 (3)
Metabolic/laboratory
Bilirubin	1 or 2	3 (10)	0 (0)	3 (9)
AST	1 or 2	5 (16)	1 (25)	6 (17)
ALT	1 or 2	2 (6)	0 (0)	2 (6)
Lipase	1 or 2	1 (3)	2 (50)	3 (9)
	3	6 (19)	0 (0)	6 (17)
	4	1 (3)	0 (0)	1 (3)
Hypoalbuminemia	1 or 2	0 (0)	1(25)	1 (3)
Alkaline phosphatase	1 or 2	4 (13)	0 (0)	4 (11)
Hypophosphatemia	1 or 2	1 (3)	0 (0)	1 (3)
	3	1 (3)	1 (25)	2 (6)
Hyponatremia	1 or 2	1 (3)	0 (0)	1 (3)
Hypokalemia	1 or 2	1 (3)	0 (0)	1 (3)
Hyperglycemia	1 or 2	2 (6)	1 (25)	3 (9)
Neurology
Neuropathy	1 or 2	12 (39)	1 (25)	13 (37)
Constitutional symptoms
Fatigue (asthenia, lethargy, malaise)	1 or 2	14 (45)	3 (75)	17 (49)
	3	1 (3)	1 (25)	2 (6)
Headache	1 or 2	1 (3)	1 (25)	2 (6)
Weight loss	1 or 2	3 (10)	0 (0)	3 (9)
Oral pain	1 or 2	1 (3)	0 (0)	1 (3)
Dehydration	1 or 2	0 (0)	1 (25)	1 (3)
Dry mouth	1 or 2	1 (3)	0 (0)	1 (3)
Myalgias	1 or 2	4 (13)	1 (25)	5 (16)
Dizziness	1 or 2	1 (3)	0 (0)	1 (3)
Hot flashes	1 or 2	1 (3)	0 (0)	1 (3)
Chills/rigors	1 or 2	1 (3)	0 (0)	1 (3)
Gastrointestinal
Anorexia	1 or 2	5 (16)	1 (25)	6 (17)
	3	1 (3)	0 (0)	1 (3)
Nausea	1 or 2	2 (6)	2 (50)	4 (11)
Vomiting	1 or 2	1 (3)	0 (0)	1 (3)
Diarrhea	1 or 2	12 (39)	3 (75)	15 (43)
	3	1 (3)	0 (0)	1 (3)
Mucositis	1 or 2	4 (13)	1 (25)	5 (16)
Abdominal pain/discomfort	1 or 2	2 (6)	0 (0)	2 (6)
Gastritis	1 or 2	0 (0)	0 (0)	0 (0)
	3	0 (0)	1 (25)	1 (3)
Gastric Hemorrhage	1 or 2	0 (0)	0 (0)	0 (0)
	3	0 (0)	1 (25)	1 (3)
Perforation: GI duodenum	1 or 2	0 (0)	0 (0)	0 (0)
	3	1 (3)	0 (0)	1 (3)
Blood/bone marrow
Leukopenia	1 or 2	2 (6)	0 (0)	2 (6)
Lymphopenia	1 or 2	1 (3)	1 (25)	2 (6)
	3	2 (6)	0 (0)	2 (6)
Platelets decreased	1 or 2	1 (3)	1 (25)	2 (0)
Eye
Blurry vision	1 or 2	1 (3)	0 (0)	1 (3)

ALT, alanine aminotransferase; AST, aspartate aminotransferase; GI, gastrointestinal.

Tumor responses

Overall, 25 of 35 MTC and DTC patients reached their first restaging; 10 of 13 enrolled patients (77%) with MTC reached a first restaging. Figure 1 depicts best response and mutation status in all patients with MTC and DTC. Most of the DTC patients (17 of 22) had clear progression by RECIST, with a median follow-up time of 2.6 months before study entry (data not shown). Likewise, most of our MTC patients (11 of 13) had PD by RECIST before study entry, with a median time of follow up of 3 months before study entry, indicating a relatively aggressive course (data not shown). Patients who did not reach first restaging were considered progressors and graphed as having a 21% tumor increase. The PR rate in the MTC patients was 38% (five of 13), and the rate of stable disease (SD) of at least 6 months was 31% (four of 13) for a total 69%. Duration of the PR was 9+, 12, 13, 16+, and 34+ months. One of the five patients with a PR had a family history of multiple endocrine neoplasia type 2 or familial MTC, whereas the others had sporadic disease. All MTC patients who had a PR had a decline in their calcitonin levels (Fig. 2), but the degree of change did not correlate 100% with depth of response. As in the example in Supplemental Fig. 1 (published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org), we observed that most MTC patients had an initial drastic decline in calcitonin level and then a leveling out of the calcitonin level with fluctuation at a stabilized level.

Fig. 1.

Waterfall plot of best response in 35 patients with MTC (black) and DTC-papillary (gray), DTC-follicular (white), and DTC-poorly differentiated (striped). Patients taken off treatment before their first restaging due to toxicity, withdrawal of consent, or clinical progression were considered treatment failures and were graphed as having a 21% increase in tumor size. Mutation analysis is reported above each patient's response for DTC-papillary (B-Raf mutations), DTC-follicular (KRAS mutations), and MTC (RET mutations). Mutation analysis was performed either in blood (B) or paraffin tissue (T) when available, or listed as not available (NA) or not detected (ND). *, Sporadic mutation, variant of uncertain significance; **, patient with syndrome of multiple endocrine neoplasia type 2B (MEN2B) (typical facial features, mucosal neuromas, bilateral pheochromocytomas, and megacolon).

Fig. 2.

Tumor markers (CEA and calcitonin) in 10 MTC patients and best response to therapy. Percent change in CEA (red bars) and calcitonin (green bars) from baseline, corresponding to the patient's best response (blue bars) were plotted for each patient. Tumor marker thyroglobulin in 15 DTC patients and best response to therapy is also presented. Percent change in thyroglobulin (orange bars) corresponding to the patient's best response (purple bars) was plotted for 13 patients. CEA, Carcinoembryonic antigen.

Fifteen of 22 (68%) patients with DTC reached a first restaging. The PR rate in DTC was 4.5% (one of 22); the rate of SD of at least 6 months was 36% (eight of 22). Figure 2 also depicts the thyroglobulin response and corresponding best response to therapy. With the exception of patients 5 and 6, all patients with measurable thyroglobulin had decreased thyroglobulin levels. Thyroglobulin levels in patients 1 and 4 were not plotted due to thyroglobulin antibodies and undetectable levels, respectively. Patients 1 and 5 had very short-lived SD responses lasting approximately 1 and 4 months, respectively. Generally, the thyroglobulin level dropped in patients with disease regression, but the degree of change did not correlate with the degree of regression. Examples of tumor responses in patients with MTC and papillary thyroid cancer are shown in Supplemental Fig. 2. Of interest is the cavitation of lesions seen in the metastatic papillary thyroid patient (Fig. 2B).

TTF, PFS, and OS

TTF, PFS, and OS for all 13 MTC patients and 22 DTC patients were calculated using a Kaplan-Meier analysis. The median TTF in patients with MTC was 14.6 months [95% confidence interval (CI) = 12.8 months, not reached (NR)], and median PFS was 15 months (95% CI = 14.6 months to N) (Fig. 3). The median OS has not been reached, but at 24 months, OS was 88% (not shown). The median TTF in the 22 DTC patients was 9 months (95% CI = 5.0–19.8 months) (not shown), and median PFS was 20 months (95% CI = 9.0 to NR) (Fig. 3). Similarly to the MTC patients, OS has yet to be reached, but at 24 months, OS in the DTC patients was 79% (not shown). For both disease groups combined, the median PFS was 18 months, and OS at 24 months was 80%.

Fig. 3.

Progression-free survival (PFS) in patients with MTC (dashed line), DTC (gray line), and both diseases combined (black line).

RET, B-Raf, and KRAS mutational analysis

In MTC patients with available blood or paraffin-embedded tissue, eight of 13 patients (61.5%) had germline or somatic RET mutations detected in blood, paraffin-embedded tissue, or both (Fig. 1). Two patients with MTC who had blood available had germline RET mutations in blood, with one of them having a concordant mutation in tissue. One other patient had a variant of uncertain clinical significance detected in the blood. Five of seven patients who had available paraffin-embedded tissue had a somatic exon 11 RET mutation. 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 629–630 and has been previously reported by us (33). All the mutations were located in the extracellular cysteine-rich domain, which can cause ligand-independent homodimerization and RET kinase constitutive activation. There was no clear correlation with type or presence of RET mutation with RECIST response.

B-Raf mutation analysis of the V600E mutation was done in available tissue samples in 12 of 16 DTC-papillary thyroid patients (Fig. 1). The V600E mutation was present in eight of the 12 patients (66%), but there was no clear correlation between response and presence of the V600E mutation. KRAS mutational analysis was also obtained in two of five patients with DTC-follicular thyroid, but neither had the mutation present (Fig. 1).

Discussion

Our study demonstrated activity for the combination of tipifarnib and sorafenib in patients with DTC, with a 5% (one of 22) PR and 36% (eight of 22) rate of SD of at least 6 months, and if we include only those who reach restaging, the PR and SD rates were 15% (one of 5) and 53% (eight of 15), respectively. Previously reported results of two phase II trials with single-agent sorafenib demonstrate similar tumor responses. These trials included all thyroid cancer subtypes; however, when analyzed for DTC patients (excluding MTC and anaplastic thyroid cancer), the study of Gupta-Abramson et al. (13) shows 68% SD and 32% PR. The analysis by Kloos et al. (16) of DTC patients shows 12% PD, 61% SD, and 15% PR in patients with PTC and 9% PD, 82% SD, and no PR in patients with FTC. When the data are combined for PTC and FTC patients, the PD rate is 13%, SD 74%, and PR 13%. However, six patients who were deemed not assessable were excluded from this analysis. These studies have reported PFS in DTC patients of 21 and 15 months, whereas the PFS in our DTC cohort was 20 months. We observed significant responses in patients with MTC, with 10 of 13 patients achieving prolonged SD of at least 6 months (31%) or a PR (38%) (duration = 9–34+ months). Although the 13 MTC patients are a small cohort, our response rate is higher than a phase II study of sorafenib alone, with a PR rate of 6% and SD of 88% in sporadic MTC patients, but the small number of patients in each study precludes statistical comparison (12). Our PR rate is higher if only the patients who were restaged are considered, 50% of patients (five of 10). Our PR rate is equivalent to that with XL184 (44% PR rate), and higher than the recently reported motesanib study (2% PR rate) (15, 34). The median PFS in the phase II sorafenib trial by Lam et al. (12) was 17.9 months, whereas ours was 15 months. However, again we included all patients in our analysis, whereas the study by Lam et al. (12) excluded one patient with clinical progression who did not have restaging films. Of the eight patients with sporadic MTC who had RET testing in the tumor, only two (25%) had a 918 mutation, which is a more aggressive tumor type. Other studies have reported a higher incidence of this mutation, which may indicate that those studies recruited patients with more aggressive disease (12, 14, 15). On the other hand, the clinical behavior of our patients before enrollment suggested aggressive disease. An interesting observation was the initial rapid decline, gradual leveling out, and fluctuation of calcitonin levels in most of the responders with MTC as in the example shown in Supplemental Fig. 2. This fluctuation did not reflect any alteration by RECIST, suggesting that a change in therapeutic regimen may not be warranted if there is a rise in calcitonin level with no concomitant change in tumor size.

MTC can be hereditary (25%) or sporadic (75%), with the molecular hallmark of the hereditary form being germline point mutations in the RET protooncogene. A subset of patients with sporadic MTC, especially those with more aggressive disease, demonstrate mutations in RET kinase in their tumors (35). In our study, of the 10 MTC patients restaged by RECIST, three of five with a PR and available tissue had either an activating somatic or germline RET mutation. Thus, it is unclear whether or not having an activating RET mutation is necessary or sufficient for a significant response. Likewise, it is too early to conclude if the activity of tipifarnib and sorafenib in MTC was due entirely to sorafenib's inhibition of RET because the RET kinase pathway is complex. Ligand activation of RET kinase activates a cascade of signaling pathways, e.g. Janus kinase, MAPK, c-Jun, NH₂-terminal kinase, Ras/Raf/MAPK, nuclear factor-κB, the phosphatidylinositol 3-kinase/AKT pathways, and phosphorylated signal transducer and activator of transcription 3 (p-STAT 3) (35).

It is intriguing to note that although no statistical comparison can be made between our study results and those of the MTC study of sorafenib alone, our PR rate of five of 13 (38%) was higher than two of 21 (9.5%) in the sorafenib alone study (12). Several hypotheses come to mind. FTase inhibitors have not shown clear activity in KRAS or N-Ras-mediated tumors, and this is thought to be due to alternative prenylation of Ras by geranylgeranylation (36). This, however, does not preclude the idea that other molecules that are activated by RET, such as AKT and p-STAT3, might be inhibited by tipifarnib and could have played a role in increased inhibition of the RET pathway (37). Indeed, activity of tipifarnib in breast and prostate cancer may have been due to AKT and p-STAT3 inhibition (38, 39). Because FTase inhibitors can also inhibit AKT and p-STAT3 activation, it is possible that tipifarnib inhibited those pathways, thereby increasing sorafenib's activity against RET. Additional analysis of peripheral blood mononuclear cells and tumor biopsies by reverse protein phase array and signaling profile analysis in our patients might yield further insights into the pathways involved.

The increase in dosing of the combination, and perhaps efficacy, was limited by the most common dose-limiting toxicity, rash. Rash is a common toxicity of sorafenib, but rarely seen with tipifarnib (27, 40). The rate of severe rash seen in the combination relative to the rate of rash seen with single agent administration suggests possible drug synergy. Patients who developed severe grade 3 rash reported a maculopapular rash, at times with severe pruritis and fever, as reported for our initial phase I study (19). Most patients on the study with rash had both drugs held and received steroids and antipruritic medications until the rash resolved to grade 1 or lower, and a dose reduction, per protocol dose level requirements, was instituted. It was interesting to note that 14 of the 35 patients (40%) had a dose reduction on study. At first glance, this dose reduction rate seems high; however, the rate was, rather, a reflection of the study parameters, which required dose reduction after any grade 2 or greater drug-related event. Only six of the 14 patients who were dose reduced developed grade 3 toxicities. One could argue that a higher response rate may have been achieved if some of the grade 2 event patients had not been dose reduced.

Our study did have several limitations. For instance, it is difficult to make a direct comparison with the results of other studies, given the small number of patients in our study.

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. The number of MTC patients was small; however, their high rate of response suggests that this combination warrants further investigation in MTC patients. Studies with larger number of MTC patients are needed to better assess response rates and determine whether response is based on the activity of one drug, i.e. sorafenib, or whether the suppression of multiple pathways by the two-drug combination contributes to the salutary effects seen in our study.