A phase II, open-label, multicenter study to evaluate the antitumor efficacy of CO-1.01 as second-line therapy for gemcitabine-refractory patients with stage IV pancreatic adenocarcinoma and negative tumor hENT1 expression

Abstract

Background

Nucleotide transporters such as human equilibrative nucleoside transporter-1 (hENT1) play a major role in transporting gemcitabine into cells. CO-1.01 (gemcitabine-5′-elaidate) is a novel cytotoxic agent consisting of a fatty acid derivative of gemcitabine, which is transported intracellularly independent of hENT1. CO-1.01 was postulated to have efficacy as a second-line treatment in gemcitabine-refractory pancreatic adenocarcinoma in patients with negative tumor hENT1 expression.

Methods

Eligibility criteria included patients with either a newly procured or archival biopsy tumor confirming the absence of hENT1 and either gemcitabine-refractory metastatic pancreas adenocarcinoma or with progression of disease following resection during or within 3 months of adjuvant gemcitabine therapy. Patients were treated with intravenous infusion of CO-1.01 dosed at 1250 mg/m² on Days 1, 8, and 15 of a 4-week cycle. The primary end point was disease control rate (DCR).

Results

Nineteen patients were enrolled of which 18 patients were evaluable for efficacy assessment. Thirteen patients (68%) had liver metastases, 6 (32%) had lymph node metastases, and 10 (53%) had lung metastases. Two of 18 patients (11%) achieved disease control. The median survival time was 4.3 (95% CI 2.1–8.1) months. All patients experienced at least one treatment-related adverse event with the majority of events being mild or moderate.

Conclusion

This study did not meet its primary endpoint and no efficacy signal was identified for CO-1.01 in treating progressive metastatic pancreas adenocarcinoma.

Introduction

Pancreas adenocarcinoma is a challenging malignancy with a rising incidence and mortality rate. An estimated 45,220 individuals in the United States will be diagnosed with pancreatic cancer in 2013 and approximately 38,460 will die from the disease [1]. At the time of diagnosis, most patients will have advanced inoperable disease. Gemcitabine-based therapies, including most recently gemcitabine and nab-paclitaxel, and FOLFIRINOX are current standard front-line therapies for patients with unresectable pancreatic adenocarcinoma [2–5]. However, for most, progression of disease occurs sooner rather than later and the value of second and subsequent lines of therapy remains to be established.

A review by Custodio et al. [6] of prior clinical trials investigating second-line therapies in gemcitabine-pretreated patients with advanced pancreatic cancer showed that the average progression-free survival (PFS) ranged from 1.3 to 5.25 months. A median overall survival (OS) of 6 months was reported in a comprehensive review of previously published clinical trials by Rahma et al. [7] for patients receiving salvage chemotherapy after front-line gemcitabine based regimens. The current study evaluated the hypothesis that primary gemcitabine-refractory disease may be a function of inefficient drug delivery.

Gemcitabine is a highly hydrophilic nucleoside analogue, which requires specific membrane transporter proteins to efficiently enter cells. Previous studies have shown that a high flux transporter named human equilibrative nucleoside transporter-1 (hENT1) plays a major role in transporting gemcitabine into cells [8]. Previous in vitro studies have further demonstrated that ENT1-deficient cells can be resistant to cytotoxic nucleosides such as gemcitabine [9]. hENT1, is also expressed variably on pancreatic tumor cells, and has been identified as being a predictive marker of outcome for gemcitabine therapy based on retrospective analyses [10–13]. Spratlin et al. [13] evaluated tumor blocks from patients with gemcitabine-treated pancreatic cancer and found that patients lacking hENT1 tumor immunostaining had a median survival of 4.0 months compared to a median survival of 13.0 months in patients with uniformly detectable hENT1 staining. Giovannetti et al. [12] further showed that patients with various stages of pancreatic cancer treated with gemcitabine-based chemotherapy when grouped based on median gene expression level of hENT1, that patients with the lowest expression level of hENT1 had the poorest survival with a median of 12.4 months versus 22.3 months for the high-expression-level groups. Farrell et al. [10] also demonstrated a correlation between hENT1 expression level and disease-free and overall survival (hazard ratio 0.36 [p = 0.003] and 0.47 [p = 0.04], respectively) in patients treated with gemcitabine after surgical resection in the RTOG 9704 adjuvant study [11]. A recent study by Neoptolemos et al. [14] used samples collected from the adjuvant ESPAC 1 and 3 randomized trials [15,16] and demonstrated that patients with low hENT1 expression had a significantly lower median survival (17.1 versus 26.2 months) compared to patients with high hENT1 expression for those who received adjuvant gemcitabine after undergoing surgical resection for pancreatic cancer.

CO-1.01 (gemcitabine-5′ -elaidate) is a novel experimental cytotoxic agent consisting of a fatty acid derivative of gemcitabine. CO-1.01 has been previously tested for antitumor activity and was shown to have similar or improved oncologic activity over gemcitabine in a number of xenograft models. CO-1.01 exerts its anti-cancer activity intracellularly and is independent of nucleoside transporters such as hENT1 to enter tumor cells. Therefore, it is hypothesized that pancreatic cancer patients with low or no meaningful expression of tumor hENT1 may have improved outcomes if treated with CO-1.01 compared to gemcitabine. Thus, the current study was designed to specifically evaluate the efficacy of second-line CO-1.01 treatment in gemcitabine-refractory patients with Stage IV pancreatic adenocarcinoma and negative tumor hENT1 expression.

Methodology

Study design

This was an open-label, single arm, non-randomized, multicenter, phase II study to evaluate CO-1.01 as second-line therapy in patients with measurable stage IV pancreatic ductal adenocarcinoma and negative tumor hENT1 expression whose best response to front-line therapy, measured radiographically after initial treatment was progressive disease (PD). The primary endpoint of the study was disease control rate (DCR; best response of complete response [CR], partial response [PR], or stable disease [SD]). The DCR endpoint was used because significant responses to second-line treatment in gemcitabine refractory patients whose disease progresses on first-line gemcitabine treatment are limited and stable disease in this second-line setting may infer a signal of relevance. Secondary endpoints of the study included progression-free survival, overall survival, objective response rate, and duration of overall response. The study used a Simon two-stage design [17], in which up to 35 patients (18 in stage 1 and 17 in stage 2) were to be treated pending activity. Stage 2 was to be initiated if three or more patients exhibited disease control (CR, PR or SD) in stage 1. Characteristics of this Simon 2-stage design include: 5% probability of accepting a poor drug, 90% probability of accepting a good drug, DCR of 10% for a poor drug, and a DCR of 30% for a good drug. In addition, a promising drug will have 7 or more patients with disease control (CR, PR or SD) out of 35 patients at the end of stage 2. All eligible patients were to receive and continue CO-1.01 until disease progression or unacceptable toxicity. After treatment discontinuation, patients were monitored for safety, cancer and survival status.

Eligibility criteria

Patients ≥18 years of age, Eastern Co-operative Oncology Group (ECOG) performance status ≤1, and with gemcitabine-refractory metastatic ductal adenocarcinoma of the pancreas, defined as having at least one measurable lesion according to RECIST 1.1 criteria based on a computerized tomography scan ≤28 days prior to CO-1.01, whose first-line treatment included at least 3 doses of gemcitabine (as monotherapy or combination therapy) with the last dose administered at least two weeks prior to CO-1.01, and with radiological best response of disease progression after the start of 1st-line treatment (no radiological stable disease or better response permitted) were eligible. Patients who experienced progression of disease during neo-adjuvant gemcitabine-based therapy were also eligible. Patients who had completed previous adjuvant therapy without progression, then subsequently had a radiologic best response of disease progression on first-line gemcitabine for metastatic disease were also eligible. Eligibility also required sufficient tumor tissue, formalin-fixed, paraffin embedded (FFPE) tissue from either a core biopsy (preferred) or fine needle aspiration cytology to demonstrate the absence of hENT1 expression (>50% of tumor cells with no hENT1 staining) using a research-based immunohistochemical (IHC) assay. This IHC assay was composed of a rabbit monoclonal antibody (SP120; Ventana Medical Systems, Tucson, AZ) developed against hENT1 and was previously validated in accordance with the US Food and Drug Administration guidelines. FFPE tissue was stained with hematoxylin and eosin or antibody. Staining with hematoxylin and eosin determined tumor presence in each sample. Staining with Ventana SP120 anti-hENT1 and a negative rabbit control antibody were used to determine hENT1 expression status. Patients who had stable disease, partial response or complete response to any first-line gemcitabine based therapy, or received any first-line therapy that did not contain gemcitabine or any prior second-line treatment were ineligible. The Institutional Review Board at each participating site reviewed the study and patients gave written informed consent prior to any study procedures.

Study treatment

CO-1.01 (15 mg/mL) was administered as a 30-min intravenous infusion at a dose of 1250 mg/m² on Days 1, 8, and 15 of a 4-week cycle. Patients who had stable disease or better following two cycles with acceptable tolerance could continue CO-1.01 at the same or an increased dose (1400 mg/m²) thereafter.

Assessments

During treatment, patients underwent serial imaging assessments for antitumor efficacy and drug safety evaluation. DCR, duration of response, and survival data were calculated from date of study enrollment with imaging every 8 weeks interpreted with Response Evaluation in Solid Tumor Criteria (RECIST 1.1) and/or with date of death. Adverse events (AE) and laboratory abnormalities were assessed and graded from the time informed consent was obtained through 28 days after the last dose of CO-1.01 using the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) grading system version 4.0. Following treatment discontinuation, patients were followed at two-month intervals to assess survival time. Overall survival (OS) was determined from date of study enrollment to date of death.

Statistical methods

The tumor-evaluable population was defined as all patients who received at least one dose of CO-1.01, had measurable disease at baseline, and had at least one post-baseline tumor assessment. The safety population consisted of all patients who received at least one dose of CO-1.01 at any time. All safety data (adverse events, vital signs, clinical laboratory evaluations, ECOG performance status, dose modifications and concomitant medications/procedures) were collected. All statistical analyses were conducted with the SAS^® System, version 9.1.

Results

Study population

Nineteen patients were enrolled at 17 centers in the United States. Eighteen patients met the stringent eligibility criteria around prior gemcitabine therapy and are included in the efficacy assessment. One additional patient achieved disease control at 8 weeks but was retrospectively adjudicated to not have met the eligibility criteria, as the patient had not progressed during or within 3 months of adjuvant gemcitabine therapy. This patient was therefore evaluable for safety only. All 19 patients ultimately discontinued treatment with the majority, 13 (68%), discontinuing treatment due to disease progression.

Patient demographics and baseline clinical characteristics are summarized in Table 1. The median age was 62 years (range 46–82). The majority of patients were female (63%), white (84%), and had an ECOG Performance Status of 1 (74%). Twelve (63%) patients had previously had surgery for pancreatic cancer; six (32%) had received adjuvant therapy and/or immunotherapy. Thirteen patients (68%) had liver metastases, 6 (32%) had lymph node metastases, and 10 (53%) had lung metastases.

Table 1.

Patient demographic and baseline characteristics.

All patients (N = 19)
Age	Mean (Range)	64 (46–82)
Gender	Female	12	63%
	Male	7	37%
Race	White	16	84%
	Black or African-American	2	10%
	Asian	1	5%
ECOG Performance Status	0	5	26%
	1	14	74%
Location of Primary Tumor	Head of Pancreas	11	58%
	Body of Pancreas	6	32%
	Other	2	10%
Number of Metastatic Sites	1	5	26%
	2	9	47%
	3	5	26%
Metastatic Sites	Liver	13	68%
	Lymph Node	6	32%
	Peritoneum	4	21%
	Lung	10	53%
	Bone	1	5%
CA 19–9	Median (Range)	3027 (6–338,790)
Prior Pancreatic Cancer Surgery		12	63%
Prior Radiation Therapy		3	16%
Prior adjuvant Chemotherapy/ Immunotherapy		6	32%

Efficacy

Nineteen patients were treated in stage 1 of the study of which 18 patients were eligible for the primary study endpoint. Two of 18 patients (11%) achieved disease control (one patient had a partial response and one patient achieved stable disease), and therefore the pre-specified criterion for proceeding to stage 2 of the study was not met and the study closed after stage I. Fifteen (79%) patients had died at the time of data analyses and data were censored for 4 (21%) patients. The median progression-free survival was 1.7 months and the median overall survival time was 4.3 (95% CI 2.1–8.1) months.

Safety and adverse event profile

All 19 patients experienced at least one adverse event (AE), of which 16 (84%) experienced an AE that was grade 3 or higher (Table 2). Ten (53%) patients had an interruption, dose delay or dose reduction of C0-1.01 with thrombocytopenia reported as the most frequent cause (6 patients, 32%). Five patients (26%) discontinued therapy, of which 3 stopped related to treatment toxicity (grade 3 thrombocytopenia, hemolytic uremic syndrome, and hyperbilirubinemia). No therapy-related mortality was observed. Four patients died within 30 days after the last dose of CO-1.01, all attributed to disease progression.

Table 2.

Grade 3 or higher toxicity attributed to C0-1.01.

Organ system	N = 19
Blood and Lymphatic System
Neutropenia, leukopenia, or lymphopenia	5	26%
Anemia	4	21%
Thrombocytopenia	3	16%
Hemolytic Uremic Syndrome	1	5%
Cardiac and Vascular
Tachycardia	1	5%
Hypotension	1	5%
Deep Venous Thrombosis/Pulmonary Embolism	4	21%
Pulmonary
Dyspnea	1	5%
Gastrointestinal
Abdominal Pain	1	5%
Ascites	1	5%
Vomiting	1	5%
Infectious Events
Pneumonia/Sepsis	1	5%
Cholangitis	1	5%
Musculoskeletal
Back Pain/Spinal fracture	1	5%
Psychiatry
Confusion	1	5%
Constitutional
Asthenia, fatigue, or fever	5	26%
Laboratory Abnormalities
Hyponatremia	4	21%
Hypoalbuminemia	1	5%
Hypocalcemia	1	5%
Hyperglycemia	2	10%
Increased aspartate aminotransferase	2	10%
Increased alanine aminotransferase	2	10%
Increased alkaline phosphatase	2	10%
Increased bilirubin	2	10%

Discussion

Limited treatment options are available for patients with progressive metastatic pancreatic cancer who have primary refractoriness to gemcitabine. Several non-randomized phase II trials [18,19] and one randomized phase III trial [20] have evaluated second line treatment options such as gemcitabine with oxaliplatin, capecitabine with oxaliplatin or 5-flurouracil with oxaliplatin in gemcitabine pre-treated progressive pancreas adenocarcinoma patients. While these treatments demonstrated a degree of efficacy, the number of patients enrolled in each of these studies was small. The current trial was designed to evaluate the efficacy of CO-1.01 as a second-line therapy for patients with metastatic pancreatic cancer in a very poor prognostic situation with primary gemcitabine refractory disease whose tumor did not express hENT1. This study utilized an innovative trial design with biomarker selection, and a Simon 2 stage design to stop the trial in stage 1 if less than three patients failed to achieve disease control (CR + PR + SD). While CO-1.01 was mostly tolerable, the study failed to meet its goal with only two of 18 patients achieving disease control and the study was terminated after the first stage. These results could be accounted for by several factors, including lack of study drug efficacy, patient selection, tumor heterogeneity and hENT1 assay not being predictive of response to gemcitabine and CO-1.01 in metastatic pancreatic cancer.

Although unlikely, it is possible that the hENT assay or the scoring system was not reliable. The antibody is highly sensitive and specific and was validated according to current FDA guidelines. In addition, the samples were sent to a central laboratory and read by experienced reference pathologists. However, the fact that the assay was not identical as that used in studies by Farrell and colleagues [10], and also different to that used in the Neoptolemos studies [14] poses the possibility that the biomarker could have failed to correctly identify hENT1 low patients. Furthermore, tumor heterogeneity may have been a factor as the hENT1 analysis was evaluated on a single biopsy specimen. We know from previous molecular analysis in patients with metastatic renal cell cancer that gene expression signatures of good and poor prognoses can be detected in different regions of the same tumor [21]. The trial required core biopsies when possible and fine needle aspiration cytology specimens when not feasible/available. This might have led to some specimens labeled as hENT1 negative, though other regions of the tumor might have been positive for the biomarker. This is challenging in any biomarker driven study as obtaining multiple tumor biopsies is not feasible.

The inclusion criterion of the study were very strict relative to other second line studies and only included patients with best response of disease progression after first line treatment with gemcitabine with no radiological stable disease or better allowed at entry. These criteria selected patients with disease that carried a very poor prognosis and likely multiple upregulated resistance pathways to therapy.

Drug efficacy could have also been a contributing factor to the negative results seen in this study. CO-1.01 was hypothesized to bypass low hENT1 expression on pancreas cancer cells and achieve higher and more effective intracellular concentrations. However, pancreas cancer cells are often also surrounded by extensive hypovascular stroma. This type of tumor microenvironment may have posed significant limitations to the ability to deliver CO-1.01 to the cancer cells in the first place, thereby diminishing drug efficacy.

The role of hENT1 as a predictive biomarker of gemcitabine activity was demonstrated in a retrospective analysis of patients receiving adjuvant gemcitabine as part of the RTOG 9704 [10] and ESPAC 3 [14] studies. However, recently The Low hENT1 Adenocarcinoma of the Pancreas (LEAP) trial was the first study to evaluate in a prospective fashion, the role of hENT1 as a biomarker in patients with metastatic pancreatic cancer. The overall survival of patients with low hENT1 expression was similar in patients treated with CO-1.01 or gemcitabine (5.7 versus 6.1 months). Furthermore, in the gemcitabine arm, there was no difference in overall survival between the hENT1 high and low subgroups [22]. The earlier retrospective studies are prone to inherent biases and it is also possible that hENT1 may have greater prognostic value in the adjuvant setting [14] but a limited role in the metastatic setting.

In conclusion, this was a translational study using prospective tissue procurement, a promising biomarker that was predictive of response to gemcitabine in retrospective studies, and had a novel endpoint of disease control. While the study did not meet its pre-specified criterion of disease control rate, the characteristics of the current trial are important in future studies of pancreas adenocarcinoma and provide an efficient platform for biomarker and drug evaluation.