A Phase I Study of 5-Fluorouracil/Leucovorin and Arsenic Trioxide for Patients with Refractory/Relapsed Colorectal Carcinoma

Home , Arsenic trioxide, Fluorouracil, Oxaliplatin

Published OnlineFirst May 25, 2010; DOI: 10.1158/1078-0432.CCR-09-2590

Cancer Therapy: Clinical Clinical Cancer Research A Phase I Study of 5-Fluorouracil/Leucovorin and Arsenic Trioxide for Patients with Refractory/Relapsed Colorectal Carcinoma

Bach Ardalan1, Pochi R. Subbarayan1, Yipsel Ramos2, Michael Gonzalez2, Anthony Fernandez2, Dmitry Mezentsev1, Isildinha Reis3, Robert Duncan3, Lisa Podolsky1, Kelvin Lee5, Mayra Lima1, and Parvin Ganjei-Azar4

Abstract Purpose: This Phase I study was designed to determine a safe combination dose of 5-fluorouracil (5-FU) and arsenic trioxide (ATO) to treat 5-FU–resistant relapsed/refractory colorectal cancer patients. We studied the effect of ATO in the downregulation of thymidylate synthase (TS) in peripheral blood mononuclear cells and in tumor biopsies. Experimental Design: ATO was administered for 5 consecutive days during the first week and twice during weeks 2 to 3 and once on week 4. 5-FU/leucovorin (LV) was administered on days 8, 15, and 22. A modified accelerated titration design was used. 5-FU was dose escalated first followed by a planned dose increase for ATO. Results: No dose-limiting toxicities were seen in seven patients who received 0.15 mg/kg ATO; grade 3 toxicities were as follows: neutropenia 1, diarrhea 1, and bowel obstruction 1. In patients receiving 0.20 mg/kg ATO, grade 3 toxicities were QTc prolongation 1, fatigue 4, alkaline phosphatase elevation 2, diarrhea 2, and peripheral edema 1. TS gene expression in peripheral blood mononuclear cell decreased in all patients. Eight tumors were biopsied, four showed TS downregulation, three showed upregulations, and one did not change. Estimated median progression-free survival and overall survival were 3.1 and 13.9 months, respectively. In patients who showed TS increase or no change versus TS reduction, estimated median progression-free survival was 2.6 versus 7.9 months (P = 0.188) and overall survival was 8.6 versus 11.7 months (P = 0.44), respectively. Conclusions: Thus, we determined 0.20 mg/kg ATO, 2,600 mg/m2 5-FU, and 500 mg/m2 leucovorin (LV) to be the recommended phase II dose. Clin Cancer Res; 16(11); 3019–27. ©2010 AACR.

For the last 40 years, 5-fluorouracil (5-FU) has been the as a single agent has shown a 10% response rate in pa- most active chemotherapeutic agent for colorectal cancer tients with CRC (1, 3). However, the combination of Iri- (CRC). The reported response rate to single-agent 5-FU notecan, Fluorouracil, and leucovorin increased the ranges from 10% to 30% (1). Yet, the median survival response rate to ∼40% and the median survival to 14.8 of patients with advanced CRC has remained ∼12 months months(4).FOLFIRIhasshownanincreaseinmedian (2). In the past decade, new agents, oxaliplatin, irinotecan survival to ∼20 months. Similarly, oxaliplatin as a FOL- (CPT-11), bevacizumab, panitumumab, and cetuximab, FOX regimen resulted in an increase in median survival have been introduced into the clinical practice. Irinotecan to 19.5 months (5). Thus, it is only when these agents are combined with 5-FU that significant improvements Authors' Affiliations: 1Division of Hematology and Oncology, in response rates and median survival are seen. In the past Department of Medicine, Departments of 2Biology and 3Epidemiology and Public Health and Sylvester Biostatistics Core Resource, and 2 years, no new agents have been approved in the treat- 4Department of Pathology, Miller School of Medicine, University of ment of CRC. Therefore, it seems likely that the develop- Miami, Miami, Florida; and 5Department of Immunology, Roswell Park ment of 5-FU resistance in CRC cells will have substantial Cancer Institute, Buffalo, New York and adverse effects on these promising combination regi- Note: http://www.clinicaltrials.gov/ct2/show/NCT00449137?term=Arsenic mens. A major mechanism of 5-FU resistance in CRC is +trioxide+AND+Miami&rank=2 overexpression of the dTMP-synthesizing enzyme thymi- Clinical Trial Number: NCT00449137. dylate synthase (TS). Therefore, strategies to avoid cellular Corresponding Author: Bach Ardalan, Division of Hematology and On- resistance to fluoropyrimidines may create new treatment cology, Department of Medicine, Miller School of Medicine, University of Miami, Miami, FL 33136. Phone: 305-243-6608; Fax: 305-243-4975; options for 12,000 to 18,000 patients annually. We have E-mail: [email protected]. found in vitro that low concentrations of arsenic trioxide doi: 10.1158/1078-0432.CCR-09-2590 (As2O3; ATO) effectively inhibited TS protein and gene ©2010 American Association for Cancer Research. expression, and reversed 5-FU resistance in CRC cell lines

www.aacrjournals.org 3019

Ardalan et al.

venous catheter (port-a-cath) for the administration of Translational Relevance chemotherapy and patient consent to undergo two sepa- rate biopsies for the determination of TS tumor signatures. This is a phase I study designed to determine a safe Patients required an Eastern Cooperative Oncology Group combination dose of 5-fluorouracil (5-FU) and arsenic performance status of 0 to 2, WBC of ≥3,000 cells/mm3, – trioxide (ATO) to treat 5-FU resistant relapsed/refractory and platelet count of ≥100,000 cells/mm3. Patients must colorectal cancer patients (CRC). Overexpression of the have shown adequate renal function as documented by TS enzyme thymidylate synthase ( ) is correlated to the serum creatinine ≤1.5 × ULN, bilirubin ≤1.5 mg/dl, and – failure of 5-FU based therapy. We hypothesize that the aspartate aminotransferase and alkaline phosphatase level TS ATO downregulation of expression will lead to the re- ≤5 times the normal limit. sensitization of refractory CRC to 5-FU. We determined 2 2 Patients receiving an investigational drug within 4 0.20 mg/kg ATO, 2,600 mg/m 5-FU, and 500 mg/m weeks before the start of study were ineligible. Metastatic leucovorin to be the recommended phase II dose. Man- disease to the central nervous system, history of prior neu- ageable toxic side effects were seen in patients with meta- rologic disorders (grade 3 or higher by the National Can- static CRC receiving the above regimen. There is a cer Institute (NCI) Common Toxicity Criteria, in particular TS significant decrease in gene expression in correlative seizure disorders), and/or peripheral neuropathy of ≥grade studies of peripheral blood mononuclear cells of all pa- 2 also excluded patients from study participation. Patients tients. Furthermore, four of eight tumor biopsies showed who were of childbearing age and refused pregnancy test- TS downregulation of gene expression. Successful outcome ing, or were pregnant, were not able to participate in this – of this study will lead to a new treatment option for 5-FU study. All patients gave informed consent before under- resistant relapsed/refractory colorectal cancer patients. going protocol-specific evaluations and treatment.

Treatment plan and study design One treatment cycle spanned 5 weeks, which included 4 (6). Furthermore, we have found that patients with meta- weeks of treatment followed by 1 week of rest. A fixed static CRC receiving ATO as a single agent showed a signif- dose of ATO was administered on days 1 to 5, 8, 11, icant decrease in TS gene expression in correlative studies 15, 18, and 22 through i.v. infusion over 1 to 4 hours. of their peripheral blood mononuclear cells (PBMC; A loading dose was administered on days 1 to 5 to down- ref. 7). We hypothesized that a combination of ATO and regulate TS expression levels. After the first week, the ad- 5-FU may benefit the patients with refractory CRC. Thus, ministration of ATO twice weekly was sufficient to we initiated this phase I study. The primary objective of maintain TS downregulation. The assigned dose of 5-FU this study was to determine a combination dose of 5- and 500 mg/m2 of leucovorin were administered over FU/leucovorin and ATO that could be safely administered 24 hours by i.v. infusion on days 8, 15, and 22 following to patients. The secondary objective was to determine if TS ATO administration. At the discretion of the principal in- downregulation occurred in tumors as well as in the vestigator, patients were given subsequent cycles of treat- PBMC of patients receiving the proposed treatment. In this ment provided they had recovered from toxicity and article, we summarized the outcome of this clinical trial. established evidence of treatment benefit.

Patients and Methods Dose escalation Dose escalation was conducted for the two main drugs: Eligibility criteria 5-FU and ATO. Dose escalation was based on the occurrence Patients 18 years and older with histologically confirmed of dose-limiting toxicity (DLT) during the first-treatment stage IV (TanyNanyM1), advanced, refractory, metastatic cycle (5 wk, including the 1 wk of rest). DLT was defined CRC were eligible. CRC patients with k-ras wild-type who as grade 3 or higher nonhematologic toxicity, or grade 4 initially responded and then failed at least three different hematologic toxicity as per NCI Common Terminology standard chemotherapy regimens were considered refractory. Criteria for Adverse Events v3. Initially, ATO was admin- In patients with mutated k-ras, failure to respond to at least istered at a fixed dose of 0.15 mg/kg, whereas 5-FU was two different standard chemotherapy regimens was consid- escalated. The doses of 5-FU tested were 1,600, 2,000, ered refractory. In our institution, traditionally, patients 2,300, and 2,600 mg/m2. Dose escalation began with an diagnosed with advanced CRC received Avastin and FOLFIRI. accelerated phase; each cohort consisted of one patient Avastin and FOLFOX combination was used as second line per dose level. In the absence of DLT in the first treatment therapy. When both the regimens proved ineffective, patients cycle, the starting dose of 5-FU for each subsequent cohort were treated with Erbitux and CPT-11. ATO and 5-FU com- was one dose level higher than that used for the previ- bination therapy was offered only to patients who failed at ously treated patient. Therefore, enrollment of patients least all three or two regimens, respectively. Other eligibility was spaced at least 5 weeks apart, allowing for the evaluation criteria included disease progression, bidimensionally mea- of first-cycle dose-limiting toxicities. The accelerated phase surable disease, biopsiable tumor, life expectancy of >2 of 5-FU was concluded at 2,600 mg/m2, which was earlier months, and accessibility for the placement of a central reported by us to be the recommended phase II dose for

3020 Clin Cancer Res; 16(11) June 1, 2010 Clinical Cancer Research

Arsenic Trioxide and 5-FU/Leucovorin for CRC

5-FU (8). No patient experienced any DLT at 0.15 mg/kg days 1, 5, 8, 11, 15, 18, and 22, in royal blue top tubes. ATO in combination with an escalated dose of 5-FU. We Pharmacokinetic analysis was done using PK Solutions then proceeded with the escalation of ATO to 0.20 mg/kg v2.0 software. Arsenic parameters were correlated with in combination with 5-FU at 2,600 mg/m2. TS levels, disease response, and recurring toxicities.

Patient evaluation and follow-up Statistical analysis All patients underwent prestudy evaluations that included Patient characteristics were summarized using descrip- medical history, a serum-human chorionic gonadotropin tive statistics: counts, percentages for categorical variables, pregnancy test for women of childbearing age, a physical median, and range for continuous variables. Response examination with Eastern Cooperative Oncology Group rates were reported with corresponding 95% confidence performance status score, hematology and chemistry pro- intervals (95% CI) computed by the exact binomial method. files, a urinalysis and a carcinoembryonic antigen mea- Progression-free survival (PFS) was measured from date of surement. A cardiac profile (electrocardiogram) and enrollment in the study to the earliest date of relapse, pro- initial imaging/diagnostic studies with corresponding tumor gression, or death from any cause. Overall survival (OS) assessment were done within 4 weeks before initiating was measured from the time of enrollment to the study till therapy. death. Event-free patients were censored at the last docu- Evaluations were done at different intervals throughout mented date of event-free status. PFS and OS were estimated the study. Electrocardiogram, history, physical, and hema- by the Kaplan-Meier method. The log-rank test was used tology and biochemistry parameters were obtained before to evaluate the potential prognostic effect of TS changes. therapy on days 1, 8, 15, and 22, as well as when clinically Serial elemental arsenic levels were obtained before and indicated. Plasma arsenic levels were drawn on preinfusion post-ATO injection twice weekly for the first 3 weeks of the days 1, 5, 8, 11, 15, 18, and 22. Carcinoembryonic antigen treatment cycles with an additional evaluation on week 4. measurements were obtained on week 4 of every cycle and The corresponding data were fit to a single-compartmental as clinically warranted. Imaging studies and tumor response model with exponential washout. Statistical analysis was were assessed after the completion of every two cycles. done in SAS version 9.2. Response and progression were evaluated using the new international criteria proposed by the Response Evalua- Results tion Criteria in Solid Tumors Committee (9). Those patients who achieved complete, partial, or stable disease had con- Patient characteristics firmatory tumor assessment 4 weeks postresponse. Twelve patients participated in this phase I study. All pa- Posttreatment evaluations namely history, physical exam, tients had metastatic CRC that had initially responded and hematology and biochemistry profiles, carcinoembryonic subsequently progressed through the following regimens: antigen measurement, imaging and diagnostic studies, FOLFOX or FOLFIRI, and in combination with bevacizu- and tumor response assessments were carried out on all mab, panitumumab, cetuximab, and irinotecan. Patients' patients no longer receiving therapy. These were carried ages ranged from 32 to 81 years, with a median of 56 out at the end of months 3, 6, and 12 and then annually. years. The majority of patients were white (67%), male (83%), and non-Hispanic (67%). Patient characteristics Pharmacokinetics and pharmacodynamics are detailed in Table 1. The first seven patients received The goal of the correlative studies was to determine 0.15 mg/kg ATO and patients 8 to 12 received 0.20 mg/kg the effect of ATO on TS expression in tumor and in ATO (Table 1). Accordingly, patient no. 1 received 1,600 PBMCs, on serum plasma arsenic levels, and the relation- mg/m2 5-FU,patientno.2startedwith2,000mg/m2 ship between these biological markers. TS and β-actin 5-FU, patient no. 3 started with 2,300 mg/m2 5-FU, and (internal control) expressions in PBMC were obtained be- patient no. 4 and subsequent patients were administered fore treatment start (baseline) and on days 1, 5, 8, 11, 15, 2,600 mg/m2 5-FU. The average number of treatment cycle 18, and 22. Fine needle aspirates of liver or lung meta- received per patient was 6 with a range of 2 to 10 cycles. stases were obtained within 1 week before the start of treatment and on day 23 of cycle one. All fine needle Toxicities aspirates were confirmed for malignant cells before fur- All patients were monitored for signs of cardiac toxic- ther processing. ity (persistent hypertension, arrhythmias, and weight Peripheral blood samples (PAXgene Blood RNA tube, gain), neuropathy (lethargy and peripheral neuropathy), Qiagen) were obtained preinfusion on day 1 and 1 hour hepatotoxicity, nephrotoxicity, and skin toxicity. During post-ATO infusion on days 1, 5, 8, 11, 15,18 and 22. RNA cycle one, there were one episode each of grade 2 abdom- isolation, processing, and gene expression analysis were as inal pain and peripheral edema. No DLTs (nonhematolo- detailed under Subbarayan et al. (7). gic grade 3 or all grade 4 toxicities) were observed in any It is possible that concurrent 5-FU administration may of the seven patients who were treated with 0.15 mg/kg of alter ATO pharmacokinetics. The plasma elemental arsenic ATO (Table 2). levels were monitored (Nichols Laboratory) 1 hour prein- Five patients were treated with 0.20 mg/kg ATO and fusion on day 1 and 1 hour post-ATO administration on 2,600 mg/m2 5-FU.Onlyoneofthesefivepatients

www.aacrjournals.org Clin Cancer Res; 16(11) June 1, 2010 3021

Ardalan et al.

tration of plasma arsenic reached ∼100 μg/L and total area Table 1. Patient characteristics and dose under the curve equaled 831 (r2 = 0.84; Fig. 1A and B). cohorts With 0.20 mg/kg ATO, the peak concentration reached ∼160 μg/L and total area under the curve equaled 1,429 n % (r2 = 0.96; Fig. 1C and D). As seen in Fig. 1, serum arsenic levels remained higher throughout the course of treatment Age (y) in patients who received 0.20 mg/kg ATO compared with <40 1 8.3 those who received 0.15 mg/kg. 40-49 5 41.7 50-59 0 0.0 Response 60-69 4 33.3 TS mRNA expression was measured by real-time PCR in ≥70 2 16.7 PBMCs. All 12 patients, regardless of receiving the lower Median, 56; range, 32-81 0.15 mg/kg or the higher 0.20 mg/kg ATO, showed a sig- Sex nificant decrease in TS gene expression in PBMC (proof of Male 10 83.3 our hypothesis). These levels remained low throughout Female 2 16.7 the entire treatment (Fig. 2A). Within a week of cessation Ethnicity of treatment, there was partial recovery of TS mRNA in the Hispanic or Latino 4 33.3 PBMC (data not shown). Non-Hispanic 8 66.7 Tumor biopsies were obtained from eight patients with Race metastatic liver or lung disease, and measured for TS White 8 66.7 mRNA levels. Results showed that by the end of cycle Black or African-American 2 16.7 one of treatment, ATO suppressed the tumor TS mRNA Asian 2 16.7 level by an average of 66% in four patients (Fig. 2B). Dose cohort The TS mRNA levels increased in three patients by an av- 0.15 ATO + 1,600 5-FU 1 8.3 erage of 71.2% and showed no change in one patient. 0.15 ATO + 2,000 5-FU 1 8.3 Based on radiological findings, two patients showed sta- 0.15 ATO + 2,300 5-FU 1 8.3 bilization of disease without clear evidence of tumor 0.15 ATO + 2,600 5-FU 4 33.3 shrinkage. The two patients with stabilization of disease 0.20 ATO + 2,600 5-FU 5 41.7 showed TS downregulation in tumor tissues (−62.2, −40.1; Fig. 2B). Because patients failed numerous previ- experienceda DLT in the first treatment cycle. This was a ous chemotherapy regimens, we have considered the sta- grade 3 episode of QTc prolongation. During cycle one, bilization of disease as a response, yielding to a response there were also three episodes of grade 2 fatigue, one ep- rate of 16.7% (Table 3). isode each of grade 2 peripheral edema, and abdominal Based on all 12 patients, the estimated median PFS was pain. Treatment-related toxicities became more common 3.1 months (95% CI, 2.1-11.9 mo). The median OS was with subsequent cycles and seemed to accumulate as treat- 13.9 months (95% CI, 3-19.3 mo; Fig. 3A and B). Addi- ment continued at higher doses of 5-FU. No grade 4 toxi- tionally, we compared PFS and OS between four patients cities were observed. ATO was fixed for each patient; who showed TS reduction and four patients with TS in- however, the dose of 5-FU could increase at each two crease or no change. In patients who showed a decrease cycles in the absence of toxicity. Overall, grade 2 fatigue in TS mRNA levels, the estimated median PFS was 7.9 and nausea were the most common toxicities. They in- months (95% CI, 2.1-11.9 mo), whereas median PFS creased with the escalation of ATO dose but did not cause was 2.6 months (95% CI, 2.1-4.5 mo) in patients with interruption or changes in treatment. Peripheral edema TS increase or no change (Fig. 3C). However, given the and weight gain were frequently observed at 0.20 mg/kg small number of patients, the difference in PFS was not ATO. Anorexia and constipation was observed with in- statistically significant (log-rank P = 0.188). We also exam- creasing doses of ATO. Three patients at 0.15 mg/kg ined the OS between patients in whom TS mRNA was re- ATO developed herpes zoster and were successfully treated duced and not reduced. The estimated OS medians were with acyclovir. As reported in previous studies, ATO is 11.7 months (95% CI, 5-16.2 mo) and 8.6 months known to reactivate herpes zoster (10). (95% CI, 3.0-13.9 mo), respectively. In this small cohort of patients, the observed survival advantage was not statis- Pharmacokinetics tically significant (log-rank P = 0.44; Fig. 3D). Serial elemental arsenic levels were obtained before and post-ATO injection twice weekly for the first 3 weeks of the Discussion treatment cycles, with an additional evaluation on week 4. The data were fit to a single-compartmental model with 5-FU has been the most active chemotherapeutic agent exponential washout. Plasma arsenic levels peaked at the for CRC and the mainstay of treatment for the last 40 conclusion of week 1, consisting of 5 consecutive days of years. However, development of 5-FU resistance in CRC loading with ATO. At 0.15 mg/kg ATO, the peak concen- patients is a major challenge. Strategies to reverse cellular

3022 Clin Cancer Res; 16(11) June 1, 2010 Clinical Cancer Research

Arsenic Trioxide and 5-FU/Leucovorin for CRC

resistance to 5-FU by ATO may create a treatment choice oxygen species involving both caspase-dependent and cas- for 12,000 to 18,000 patients annually, who otherwise pase-independent pathways (15, 16). However, it has have no alternative options. We and others have shown been observed that ATO has diverse mechanisms of anti- that the development of 5-FU resistance in metastatic neoplastic effect in a wide range of tumors (16). Recently, CRC patients is highly associated with gene amplification ATO has been examined for potential treatment options in (2) and overexpression of intratumoral TS mRNA (11). several solid tumor cell lines targeting various mechanisms Thus, strategies to reduce TS overexpression are likely to (17, 18). It is important to realize the above in vivo arsenic restore 5-FU sensitivity. concentrations compare favorably to our in vitro concen- ATO was recently rediscovered in the modern era as a trations. Evaluation is based only on the limited number highly active agent for the treatment of chemotherapy/ of patients in our study. It is our understanding that ATO ATRA-resistant acute promyelocytic leukemia (12, 13). is a cytostatic agent and promotes downregulation of the Clinical trials done in the United States resulted in 92% tumor TS gene expression. complete response rates (14). Thereafter, ATO was ap- Although all these studies have shown successful results, proved by the Food and Drug Administration and has be- no trial had yet studied the effects of ATO on TS gene come treatment of choice for acute promyelocytic expression and 5-FU sensitivity. Previously, ATO has only leukemia patients. ATO is now widely studied in the treat- been used as a single agent and, as a result, has proved ment of multiple myeloma. We recently reported in a pilot ineffective in the treatment of CRC. We found that ATO study that ATO (0.25 mg/kg) plus ascorbic acid, a glutathi- reduces PBMC and tumor TS mRNA levels. one-depleting agent, was well tolerated in patients with re- We have previously reported trials on biochemical fractory/relapsed multiple myeloma (15). We found that modulation to treat CRC. When compared with bio- ATO kills myeloma cells through the generation of reactive chemical modulation, TS gene suppression may allow

Table 2. Treatment-related toxicities by grade and ATO dose

Cycle 1 Cycles 2-10 ATO 0.15 mg/kg ATO 0.20 mg/kg ATO 0.15 mg/kg ATO 0.20 mg/kg (7 patients) (5 patients) (7 patients) (5 patients) Grade Grade Grade Grade 1 2 3 Total 1 2 3 Total 1 2 3 Total 1 2 3 Total (no. of episodes) (no. of episodes) (no. of episodes) (no. of episodes)

Blood/bone marrow Anemia 2 2 1 1 Neutropenia 1 1 Cardiovascular Qtc prolongation 1 1 1 1 2 2 2 1 1 2 (arrythmia) Sinus tachycardia 2 2 Constitutional Abdominal pain 1 1 1 1 6 6 symptoms Fatigue 4 4 4 3 7 1 1 3 9 4 16 Weight gain 2 2 1 1 8 8 Dermatology/skin Mucositis 2 2 Herpes zoster 1 2 3 Gastrointestinal Anorexia 3 3 4 2 6 Constipation 2 2 1 1 3 3 6 1 7 Diarrhea 1 1 2 2 1 1 1 3 Nausea/vomiting 3 3 1 1 3 1 4 6 2 8 Small bowel 11 obstruction Lymphatics Peripheral edema 1 1 3 1 4 1 1 3 6 1 10 Metabolic/ Alk. phosphatase 22 laboratory elevation Proteinuria 11 Musculoskeletal Muscle weakness 11 Total 12 2 1 15 17 5 1 23 14 3 3 20 35 28 9 72

NOTE: Patient may have multiple episodes of toxicity. DLT was defined as grade 3 or higher nonhematologic toxicity, or grade 4 hematologic toxicity as per NCI Common Terminology Criteria for Adverse Event v3.

www.aacrjournals.org Clin Cancer Res; 16(11) June 1, 2010 3023

Ardalan et al.

for better efficacy of treatment (19). Biochemical modu- at mRNA level, remaining suppressed for the duration of lation is a method that leads to a “biochemical switch,” therapy. At reduced TS levels, tumors are very sensitive to e.g., in pyrimidine biosynthetic pathway leading to small quantities of FdUMP, inhibiting DNA synthesis for de novo shutdown, therefore augmenting the salvage path- prolonged periods. Lower TS levels create the prime envi- way. This will lead to the further production of FdUMP, ronment for 5-FU to be used, thus effectively maximizing which could bind to “wild-type” TS and inhibit DNA syn- tumor response. Therefore, we decided to perform a study thesis. TS is shown to auto regulate at the translational level. combining ATO with 5-FU to suppress TS gene expression Sequestration of TS by FdUMP eliminates this autoregu- for the treatment of CRC. lation, thereby resulting in increased TS production. Addi- Previously, we completed in vitro studies assessing the tionally, during cell cycle, fresh TS is continuously formed. sensitivity of a panel of CRC cell lines [HT29, SW480 sen- Both the above mechanisms could “override” the FdUMP sitive to 5-FU, and HT29FU (resistant to 5-FU)] to differ- present in the cells and allow DNA synthesis to resume, ent concentrations of ATO. We assessed the effects of ATO preventing biochemical modulation to be successful. concentration on TS expression. Downregulation of TS On the other hand TS gene suppression shuts down gene expression in HT29FU cells is most marked between transcription. This mechanism works regardless of wild- 0.2 and 0.6 μmol/L. This downregulation in TS mRNA is type or mutated TS. We have shown ATO to inhibit TS similarly reflected in TS protein levels.

Fig. 1. Mean serum arsenic levels. A, predicted and observed serum arsenic levels at ATO dose of 0.15 mg/kg; pretreatment and posttreatment. B, correlation between predicted and observed serum arsenic levels at ATO dose of 0.15 mg/kg. C, predicted and observed serum arsenic levels at ATO dose of 0.2 mg/kg; pretreatment and posttreatment. D, correlation between predicted and observed serum arsenic levels at ATO dose of 0.2 mg/kg.

3024 Clin Cancer Res; 16(11) June 1, 2010 Clinical Cancer Research

Arsenic Trioxide and 5-FU/Leucovorin for CRC

Fig. 2. A, the loading and maintenance dose of ATO (As2O3) rendered suppression of TS mRNA in PBMC for the first 22 d of treatment (cycle 1). In the second and subsequent cycles, TS mRNA remained suppressed in PBMC (data not shown). The results are expressed as mean normalized expression, which is the ratio of β-actin (internal control) over TS mRNA. Patients receiving 0.15 or 0.2 mg/kg of ATO showed TS mRNA suppression as shown in this figure. B, in eight patients, tumor biopsies were done before and post–first-cycle treatment. In four patients, there was a reduction in TS mRNA expression of whom two showed disease response (*, stable disease). Of the remaining patients, TS increased in three and remained unchanged in one, none of whom showed clinical response. Pre– and post–first-cycle treatment, tumor biopsies were obtained. Total RNA was isolated using Tri-Reagent (Sigma). cDNA was synthesized using M-MLV (Promega) and dT16 primer. Reverse transcription-PCR was done in a 96-well plate on the GeneAmp Sequence Detection System 5700 (Applied Biosystem) with SYBR Green PCR master mix and appropriate primers. β-Actin gene was used as the internal reference in calculating mean normalized expression.

Cell detachment, viability assays, and molecular marker one-sided paired t test) following treatment in the four studies suggested the efficacy of ATO as an anti-CRC agent, patients whom blood samples were obtained. Only buffy particularly against 5-FU–resistant cells. This formed the coat was analyzed for TS regulation; no tumor biopsies rationale for our previous pilot study of ATO plus ascorbic were taken in that study (7). In our current NCI-funded acid as a single-agent regimen in 5-FU–resistant relapsed/ phase I study, TS gene expression decreased in PBMC refractory CRC patients (7). No significant disease re- samples in all patients who received either 0.15 mg/kg sponses were seen in that study. However, TS mRNA ex- or 0.20 mg/kg ATO. pression in the PBMCs was significantly reduced (P = 0.03, In our previous study, using ATO at 0.25 mg/kg as a single agent, we observed a higher number of grade 3 toxicities (7). In our current phase I study, we have encountered Table 3. Response (best overall response by less toxicities with repeated treatments at 0.20 mg/kg ATO and 5-FU at 2,600 mg/m2. Therefore, we have determined Response Evaluation Criteria in Solid Tumors 0.20 mg/kg ATO, 2,600 mg/m2 5-FU, and 500 mg/m2 leu- criteria) covorin to be the recommended phase II dose. Pharmacokinetic studies were also done on plasma arse- n % 95% CI nic levels. Treatment regimen maintained serum arsenic le- Stable disease 2 16.7 2.1-48.4 vels at the therapeutic range over the course of treatment. Progressive disease 9 75.0 42.8-94.5 When these results were analyzed in relation to patient Not evaluated 1 8.3 0.2-38.5 toxicity, a correlation could be seen between higher serum arsenic levels and increased toxicity.

www.aacrjournals.org Clin Cancer Res; 16(11) June 1, 2010 3025

Ardalan et al.

Fig. 3. PFS and OS. A, median PFS is 3.1 mo (95% CI, 2.1-11.9 mo). B, median OS is 13.9 mo (95% CI, 3-19.3 mo). C, median PFS in patients showing an increase or no change in TS is 2.6 mo (95% CI, 2.1-4.5 mo) versus 7.9 mo (95% CI, 2.1-11.9 mo) in patients showing decrease in TS (P = 0.188). D, median OS in patients showing an increase or no change in TS is 8.6 mo (95% CI, 3 -13.9 mo) versus 11.7 mo (95% CI, 5 -16.2 mo) in patients showing decrease in TS (P = 0.44).

The conduct of phase I trial showing the combination of Disclosure of Potential Conflicts of Interest ATO and 5-FU has acceptable toxicities and has led to the suppression of TS mRNA in PBMC (proof of our hypothe- No potential conflicts of interest were disclosed. sis). Furthermore, we observed the suppression of TS mRNA Acknowledgments in 50% of the patients examined. Two of the four patients TS whose tumor mRNA were reduced showed clinical re- We thank Melissa Oh and Alexandra Thulin for their help in editing this sponse (stable disease). Although this is a small phase I article. study, we do not typically report disease-free survival or OS. However, it is of great interest to observe an increase Grant Support in OS benefit in all the patients enrolled in this study. This TS NCI 1R21CA117116-01A2 grant (B. Ardalan). was more pronounced in patients whose tumor level de- The costs of publication of this article were defrayed in part by the creased (P = 0.44). To confirm the results of this small cohort payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to of patients, we plan to conduct a phase II trial to examine the indicate this fact. combination of 0.20 mg/kg ATO, 2,600 mg/m2 5-FU, and 2 500 mg/m leucovorin in relapsed/refractory CRC patients. Received 09/24/2009; revised 03/09/2010; accepted 04/02/2010; published OnlineFirst 05/25/2010.

References 1. Braun AH, Achterrath W, Wilke H, et al. New systemic frontline treat- 2. Wang TL, Diaz LA, Jr., Romans K, et al. Digital karyotyping identifies ment for metastatic colorectal carcinoma. Cancer 2004;100:1558–77. thymidylate synthase amplification as a mechanism of resistance to

3026 Clin Cancer Res; 16(11) June 1, 2010 Clinical Cancer Research

Arsenic Trioxide and 5-FU/Leucovorin for CRC

5-fluorouracil in metastatic colorectal cancer patients. Proc Natl 11. Bathe OF, Franceschi D, Livingstone AS, et al. Increased thymidylate Acad Sci U S A 2004;101:3089–94. synthase gene expression in liver metastases from colorectal carci- 3. Diaz-Rubio E, Sastre J, Zaniboni A, et al. Oxaliplatin as single agent noma: implications for chemotherapeutic options and survival. Can- in previously untreated colorectal carcinoma patients: a phase II mul- cer J Sci Am 1999;5:34–40. ticentric study. Ann Oncol 1998;9:105–8. 12. Niu C, Yan H, Yu T, et al. Studies on treatment of acute promyelo- 4. Saltz LB, Cox JV, Blanke C, et al, Irinotecan Study Group. Irinotecan cytic leukemia with arsenic trioxide: remission induction, follow-up, plus fluorouracil and leucovorin for metastatic colorectal cancer. and molecular monitoring in 11 newly diagnosed and 47 relapsed N Engl J Med 2000;343:905–14. acute promyelocytic leukemia patients. Blood 1999;94:3315–24. 5. Goldberg RM, Sargent DJ, Morton RF, et al. A randomized controlled 13. Shen L, Chen TX, Wang YP, et al. As2O3 induces apoptosis of the trial of fluorouracil plus leucovorin, irinotecan, and oxaliplatin combi- human B lymphoma cell line MBC-1. J Biol Regul Homeost Agents nations in patients with previously untreated metastatic colorectal 2000;14:116–9. cancer. J Clin Oncol 2004;22:23–30. 14. Soignet SL, Frankel SR, Douer D, et al. United States multicenter 6. Subbarayan PR, Lee K, Ardalan B. Arsenic trioxide suppresses thy- study of arsenic trioxide in relapsed acute promyelocytic leukemia. midylate synthase in 5-FU-resistant colorectal cancer cell line ht29 J Clin Oncol 2001;19:3852–60. in vitro re-sensitizing Cells to 5-FU. Anticancer Res. In press 2010;30. 15. Grad JM, Bahlis NJ, Reis I, et al. Ascorbic acid enhances arsenic 7. Subbarayan PR, Lima M, Ardalan B. Arsenic trioxide/ascorbic acid trioxide-induced cytotoxicity in multiple myeloma cells. Blood therapy in patients with refractory metastatic colorectal carcinoma: 2001;98:805–13. a clinical experience. Acta Oncol 2007;46:557–61. 16. McCafferty-Grad J, Bahlis NJ, Krett N, et al. Arsenic trioxide uses 8. Spicer DV, Ardalan B, Daniels JR, et al. Reevaluation of the maximum caspase-dependent and caspase-independent death pathways in tolerated dose of continuous venous infusion of 5-fluorouracil with myeloma cells. Mol Cancer Ther 2003;2:1155–64. pharmacokinetics. Cancer Res 1988;48:459–61. 17. Frohlich E, Czarnocka B, Brossart P, et al. Antitumor effects of arsenic 9. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to eval- trioxide in transformed human thyroid cells. Thyroid 2008;18:1183–93. uate the response to treatment in solid tumors. European Organiza- 18. Pettersson HM, Pietras A, Munksgaard PM, et al. Arsenic trioxide is tion for Research and Treatment of Cancer, National Cancer Institute highly cytotoxic to small cell lung carcinoma cells. Mol Cancer Ther of the United States, National Cancer Institute of Canada. J Natl 2009;8:160–70. Cancer Inst 2000;92:205–16. 19. Jason TL, Figueredo R, Ferguson PJ, et al. ODN 491, a novel anti- 10. Nouri K, Ricotti CA, Jr., Bouzari N, et al. The incidence of recurrent sense oligodeoxynucleotide that targets thymidylate synthase, ex- herpes simplex and herpes zoster infection during treatment with ar- erts cell-specific effects in human tumor cell lines. DNA Cell Biol senic trioxide. J Drugs Dermatol 2006;5:182–5. 2008;27:229–40.

www.aacrjournals.org Clin Cancer Res; 16(11) June 1, 2010 3027

A Phase I Study of 5-Fluorouracil/Leucovorin and Arsenic Trioxide for Patients with Refractory/Relapsed Colorectal Carcinoma

Bach Ardalan, Pochi R. Subbarayan, Yipsel Ramos, et al.

Clin Cancer Res 2010;16:3019-3027. Published OnlineFirst May 25, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-09-2590

Cited articles This article cites 18 articles, 8 of which you can access for free at: http://clincancerres.aacrjournals.org/content/16/11/3019.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/16/11/3019.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/16/11/3019. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.