- © 2002 by American Society of Clinical Oncology
Phase II Study of Anti–Gastrin-17 Antibodies, Raised to G17DT, in Advanced Pancreatic Cancer
- B. T. Brett,
- S. C. Smith,
- C. V. Bouvier,
- D. Michaeli,
- D. Hochhauser,
- B. R. Davidson,
- T. R. Kurzawinski,
- A. F. Watkinson,
- N. Van Someren,
- R. E. Pounder and
- M. E. Caplin
- From the Department of Medicine, Royal Free Hospital National Health Service Trust; Department of Oncology, Royal Free and University College Medical School; Department of Surgery, The Middlesex Hospital; and Chase Farm Hospital, London, United Kingdom; and Aphton Corporation, Woodland, CA.
- Address reprint requests to Martyn Caplin, MD, Centre for Gastroenterology, Royal Free Hospital National Health Service Trust, Hampstead, London NW3 2QG, United Kingdom; email: m.caplin{at}rfc.ucl.ac.uk
Abstract
PURPOSE: The prognosis for advanced pancreatic cancer remains poor. Gastrin acts as a growth factor for pancreatic cancer. We describe the first study of the antigastrin immunogen G17DT in pancreatic cancer. Our aims were to determine the antibody response, safety, tolerability, and preliminary evidence of efficacy of G17DT in advanced pancreatic cancer.
PATIENTS AND METHODS: Thirty patients with advanced pancreatic cancer were immunized with three doses of either 100 μg or 250 μg of G17DT.
RESULTS: In the whole group, 20 (67%) of 30 patients produced an antibody response. The 250-μg dose resulted in a significantly greater response rate of 82% compared with 46% for the 100-μg group (P = .018). The most significant side effects, seen in three patients, were local abscess and/or fever. The median survival for the whole group from the date of the first immunization was 187 days; median survival was 217 days for the antibody responders and 121 days for the antibody nonresponders. The difference in survival between the antibody responders and nonresponders was significant (P = .0023).
CONCLUSION: Patients with advanced pancreatic cancer are able to mount an adequate antibody response to G17DT. The 250-μg dose is superior to the 100-μg dose, and it appears to be generally well tolerated. Antibody responders demonstrate significantly greater survival than antibody nonresponders. Phase III studies are currently underway in order to determine efficacy.
PATIENTS WITH pancreatic cancer usually present with advanced disease and have a poor prognosis.1 The median survival depends on the stage of the disease at presentation, but is usually on the order of 4 to 6 months for metastatic disease.1-3 Current treatments for inoperable pancreatic cancer make little impact on the median survival.1-3 Gemcitabine is currently recognized as the most effective chemotherapy agent for this condition. The median survival of patients on the first two clinical trials using gemcitabine was between 3.85 months and 5.65 months from the date of the first treatment.2,3 However, it has been shown to offer some improvement in symptoms compared with fluorouracil when a symptom score called the “clinical benefit response” was used.2-5
Before the introduction of gemcitabine, fluorouracil (5-FU) was the most commonly used chemotherapeutic agent for pancreatic cancer. Some uncontrolled studies have suggested an improved survival rate with combination chemotherapy of 5-FU with other agents, usually in a triple-therapy regimen, but controlled studies have failed to confirm enhanced survival.2,3,6,7
New agents for pancreatic cancer are needed. Several novel strategies are under active investigation. These include attempts to modulate growth factor activity, antiangiogenic factors, vascular inhibitors, antistromal factors, and immunotherapy.8-13 Of these novel agents, metalloproteinase agents such as marimastat have been suggested to be as effective as gemcitabine in terms of median survival, but although there was no significant difference, the median survival for the gemcitabine group was longer than each of the three dosage groups of marimastat.13
Immunotherapy involves the stimulation of a patient’s immune system to achieve antitumor activity. Attempts can be made to enhance the immune response to tumor antigens, nonspecifically stimulate the immune system, or induce the formation of antibodies with antitumor activity.8,9
Our study has used one such novel agent called G17DT (Aphton Corporation, Woodland, CA). This agent is an immunoconjugate of the amino-terminal sequence of gastrin-17 linked by means of a spacer peptide to diphtheria toxoid. It is given as an intramuscular vaccination and induces the formation of antibodies that neutralize both amidated gastrin-17 (G-17) and precursor glycine extended G-17 (Gly–G-17).14 Gastrin has been shown to have trophic effects on a variety of gastrointestinal and nongastrointestinal neoplastic cells.14-18 Furthermore, increasing evidence has demonstrated that the precursor forms of gastrin, such as glycine-extended gastrin, may be even more potent trophic agents than fully processed amidated gastrin.18-21
Gastrin is expressed in the fetal pancreas and probably acts as a trophic agent.22 In the normal adult pancreas, gastrin is largely expressed in islet cells.23,24 In pancreatic adenocarcinoma, gastrin isoforms have been demonstrated, representing so-called oncofetal expression.23-25 We and others have demonstrated the presence of cholecystokinin-B/gastrin receptors in 90% of pancreatic cancer resection tissue, precursor gastrin forms in 55% to 90% of resection tissue, and fully processed amidated gastrin in 25% of pancreatic cancer resection tissue.23-25 In vitro studies and in vivo animal models have shown that gastrin stimulates the proliferation of human pancreatic cancer cells.15,16,26 We have demonstrated that antigastrin antibodies raised against G17DT inhibit the proliferation of pancreatic cancer cells.26 Other studies have also demonstrated that antigastrin agents including antibodies and cholecystokinin-B/gastrin receptor antagonists inhibit the growth of pancreatic cancer cells.15,16 Early human studies of G17DT in patients with other forms of cancer have demonstrated that G17DT appears to be safe and well tolerated.27 The aims of this study were to determine the antibody response of patients with proven pancreatic cancer to G17DT, to evaluate the safety and tolerability of G17DT in these patients, and to look for preliminary evidence of efficacy.
PATIENTS AND METHODS
Study Design
The study was an open, multiple-dose, single-center study of G17DT in patients with proven pancreatic cancer. Patients were initially recruited to receive three doses of 100 μg of G17DT on weeks 0, 2, and 6. The Royal Free Hampstead NHS Trust Ethical Practices Subcommittee approved the study. In view of a poor antibody response to 100 μg of G17DT, a request was subsequently submitted to the ethics committee, after 14 patients were recruited, to increase the dose to 250 μg.
End Points
The formal trial period was 16 weeks, and the major end points of the study were assessment of antigastrin antibodies by 8 weeks and any adverse effects attributable to G17DT. The minor end points were survival, pain scores, analgesia consumption, tumor load as determined by imaging, quality-of-life scores, changes in weight, and CA 19-9.
Patients
Patients were referred for possible entry to the study by their clinician after a firm clinical diagnosis of locally advanced or metastatic pancreatic cancer had been made. All patients had been informed of the diagnosis before referral. Patients were given an information sheet and had the opportunity to ask additional questions regarding the study. Eligibility criteria included the following: proven pancreatic cancer by histologic or cytologic criteria; no chemotherapy for 4 weeks before study entry or planned for the duration of the study; no concomitant or anticipated use of radiotherapy for the duration of the study; no immunosuppressant drugs in the past month or concomitant use for the duration of the study; patients older than 18 years of age; life expectancy anticipated to be at least 2 months; no history of other malignant disease except for nonmelanomatous skin cancer or in situ carcinoma of the uterine cervix; no women who are pregnant or lactating; no patients taking part in another study involving an investigational or licensed drug or device in the 3 months preceding enrollment or during the study; and no patients with hemoglobin less than 9.5 g/dL or WBC count less than 3.5 × 109/L or platelet count less than 100 × 109/L. Written informed consent was obtained when each patient was eligible to enter the study.
Study Plan
The study consisted of a screening visit followed by seven additional visits over 16 weeks. At the end of week 16, each patient was offered the opportunity of continued follow-up and further G17DT injections as required. On every visit, the signs and symptoms of pancreatic cancer, vital signs, concurrent illnesses, concomitant medication, weight, Karnofsky performance index, pain (visual analog scale), analgesic use, CA 19-9, and C-reactive protein were assessed along with hematologic and biochemical parameters as part of toxicology screening. Subjective pain scores and analgesic consumption using diary cards were recorded by each patient daily. Anti–G-17 antibodies were assessed on each visit after the screening visit.
Anti–G-17 antibodies.
Serum samples were taken and stored at −20°C during each visit. The serum was analyzed by Huntington Laboratories (Huntington, United Kingdom) using an enzyme-linked immunosorbent assay. An antibody response was defined as an antibody level that had a binding capacity of 25.6 × 103 pmol/L of G-17 or more (this minimum level was defined as 1 unit). The serum levels of G-17 in humans would normally lie between 10 and 50 pmol/L, giving an antibody excess of at least 500-fold with 1 unit of anti–G-17 antibodies.
Local reactions to G17DT.
Each injection site was assessed at each visit for erythema, swelling, and pain. Adverse reactions were scored using criteria adapted from the Cancer Therapy Evaluation Program common toxicity criteria, version 2.0. Injection site reactions within these criteria are not designed primarily for vaccinations, and therefore are not applicable to this context without minor changes. Abscess was included in grade 3, and pain with swelling was included in the definition of grade 2. The reactions therefore were scored as follows: 0 = no reaction; 1 = pain or itching or erythema; 2 = pain with swelling or inflammation; and 3 = abscess, ulceration, or necrosis or other complications that are prolonged, or require surgery or hospital admission.
Adverse events.
All adverse events were recorded in detail. On each occasion, the examining physician recorded a documented assessment regarding the likelihood that any adverse event was attributable to the G17DT. Concomitant illnesses were also recorded.
General assessment.
Weight in kilograms was recorded on each visit, along with height on the first visit. To assess pain, each patient was given a diary card based on the Memorial Pain Assessment Card visual analog scale card described by Fishman et al28 and used by Burris et al.2,3 This card was designed to allow patients to record the maximum pain they experienced on a daily basis. In order to standardize analgesic consumption, patients were converted, if possible, to Oramorph (Boehringer Ingelheim Ltd, Bracknell, Berks, United Kingdom) and MST continuous (Napp Pharmaceuticals, Cambridge, United Kingdom) with additional paracetamol as required. Other analgesic compounds were withdrawn. This allowed analgesic consumption to be quantified using the diary card, in terms of daily morphine consumption. Karnofsky performance index was used to assess performance.
Survival.
The date of death for each patient in the study was recorded and verified. Median survival estimates were calculated using the Kaplan-Meier method. The survival of the two dosage groups was compared using the log-rank test.
Computed tomography scanning.
Computed tomography (CT) scans were obtained for the first six patients at weeks 0 and 16. After the dose of G17DT was increased from 100 μg to 250 μg, with ethical approval, all subsequent patients were assessed at weeks 0, 8, and 16, in order to allow a more accurate ascertainment of time to progression. A consultant radiologist who was unaware of each individual patient’s antibody response assessed the scans. The maximum diameter in two planes was determined and the product of the scores recorded. Progression was defined as an increase in the sum of the products of the lesions of more than 25% or the appearance of a new lesion. Patients who failed to attend for a CT scan because of deterioration in their clinical status were also deemed to have disease progression. Regression was defined as a decrease in the sum of the products of the maximum diameters of each lesion by more than 25% with the development of no new lesions.
CA 19-9.
The CA 19-9 was measured from blood samples taken at each visit. A decrease in CA 19-9 was defined as a 50% reduction and an increase as an increase by 10% or more. Stability was defined as any change less than these two definitions.
European Organization for Research and Treatment of Cancer Quality-of-Life Questionnaire C-30 Version 2.0.
Patients were asked to complete the European Organization for Research and Treatment of Cancer Quality-of-Life Questionnaire C-30 Version 2.0 and PAN26 quality of life scores at visit 2, week 0. The first 10 patients were just asked to repeat the questionnaires at visit 8, week 16. All subsequent patients were asked to perform this task at week 0 and then at 4-week intervals for the duration of the study.
RESULTS
Demography and Baseline Characteristics
The baseline characteristics of the subjects in the two dosage groups were similar (Table 1). The median age of the patients in the whole group was 63 years of age. The sex distribution consisted of six female and 24 male patients. In the whole group, 15 (50%) of 30 had stage IV disease. There were five patients with stage IV disease in the 100-μg group and 10 patients with stage IV disease in the 250-μg group. The median Karnofsky performance index, CA 19-9, weight, height, and body mass index are also listed in Table 1. The median number of days from tissue diagnosis to enrollment on the study was 36.5 days, with a mean of 68.1 days.
Table 1. Demographic Data and Stage at Presentation
Other Interventions
Four patients underwent enteral stenting for duodenal stenosis before week 16. Two were antibody responders and two were antibody nonresponders. One patient had a surgical bypass for gastric outlet obstruction during the study. She had a complicated postoperative course and died 134 days after her first injection; she was an antibody responder. One patient received gemcitabine after the collection of all study data at week 16 but before closure of the study. Her survival data are censored from the date of the first dose of chemotherapy.
Antibody Response
In the whole group, 20 (67%) of 30 patients mounted an antibody response by week 8. In the 100-μg group, six (46%) of 13 responded by week 8 (Tables 2 and 3⇓). In the 250-μg group, 14 (82%) of 17 patients mounted an antibody response by week 8. The difference between the response rate in the two dosage groups was statistically significant (P = .018). No patient with an undetectable titer at week 8 went on to achieve detectable antibody levels.
Table 2. Antibody Response According to Dose and Time
Table 3. Baseline Characteristics of Antibody Responders and Nonresponders
Safety, Tolerability, and Adverse Events
Three patients had a systemic reaction to G17DT (Table 4). All three cases were associated with marked antibody responses and grade 3 local reactions. Patient no. 10 developed a sterile abscess after the second injection. The third injection provoked hospital admission with fever, pain, and recurrent abscess requiring surgical drainage.
Table 4. Local Injection Site Reactions
Patient no. 19 developed fever and pain after the second dose of G17DT requiring hospitalization. Patient no. 30 developed a reaction to the first injection, with induration, swelling, and pain but without fever. The second injection provoked a sterile abscess that resolved spontaneously and did not require hospitalization. It is notable that patient no. 10 was still alive at 435 days; patient no. 19 survived for 288 days and patient no. 30 was still alive at 175 days. Patients no. 10 and 30 gained weight from week 0 to week 16.
Four patients had a grade 2 reaction to the injection and 11 patients suffered from a grade 1 reaction. Among the antibody responders, 16 (80%) of 20 suffered a grade 1 or greater local reaction to the immunization, whereas only two (20%) of 10 antibody nonresponders suffered from an adverse local reaction.
Preliminary Evidence of Efficacy
Survival.
The median survival of the whole group from the date of the first injection with G17DT was 187 days (95% CI, 141.45 to 232.55 days), and median survival from the date of diagnosis or recurrence was 244 days (95% CI, 193.8 to 294.2 days) (Table 5 and Fig 1).
Table 5. Survival as Determined by Kaplan-Meier
Fig 1. Survival: Kaplan-Meier plot of antibody responders and antibody nonresponders from first injection to closure of study.
The median survival from the first injection for the antibody responders was 217 days (95% CI, 177.1 to 256.9 days) and for the antibody nonresponders 121 days (95% CI, 38.9 to 203.1 days); this difference is statistically significant (log-rank test, P = .0023). The median survival from diagnosis was 265 days (95% CI, 234.6 to 295.4 days) for the antibody responders and 159 days (95% CI, 128.6 to 189.4 days) for the antibody nonresponders. In both the antibody responder and the antibody nonresponder groups, one patient died before week 8. The median survival from the first injection was 206 days (95% CI, 133.9 to 278.1 days) in the 100-μg group and 187 days (95% CI, 168.8 to 205.2 days) in the 250-μg group; the difference is not statistically significant (log-rank test, P = .78).
Weight.
Over the 16-week period, one patient had more than 5% gain in weight over baseline and was an antibody responder. Eight patients had stable weight over the 16 weeks (defined as < 5% weight gain or weight loss), of which six were antibody responders. None of these patients had edema or ascites.
Pain scores and analgesic consumption.
In the whole group, pain and analgesia scores improved in one (3.3%) of 30, 19 (63%) of 30 were stable for ≥ 4 weeks, and 14 (47%) of 30 were stable for ≥ 8 weeks. Among the antibody producers, the pain and analgesia scores improved in one (5%) of 20, were stable for ≥ 8 weeks in 11 (55%) of 20, and were stable for ≥ 4 weeks in 15 (75%) of 20. Among the antibody nonresponders, zero (0%) of 10 improved, four (40%) of 10 were stable for 4 weeks, and two (20%) of 10 were stable for 8 weeks.
Karnofsky performance index.
At week 16, 10 (33%) of 30 patients had a Karnofsky performance index (KPI) of ≥ 80, and 18 (60%) of 30 had a KPI of ≥ 70. Nine (90%) of 10 patients with a KPI of ≥ 80 at week 16 were antibody responders. From week 0 to week 16, 16 (53%) of 30 patients remained stable (≤ 10-point change in KPI) and one (3%) of 30 had improved (≥ 10-point increase in KPI).
CT imaging.
From week 0 to week 16, five patients had stable disease; all were antibody responders. These patients represent 17% of the whole group and 25% of antibody responders.
CA 19-9.
Twenty-eight of 30 patients had an elevated CA 19-9 level during the study. Four patients demonstrated stable CA 19-9 levels during the course of the study, with no increase over baseline at week 16 compared to baseline (excluding the effects of cholestasis). The median CA 19-9 level at baseline was 688.5 for the whole group, compared with 3,168.5 for the antibody nonresponders and 239.5 for the antibody responders.
Quality-of-Life Scores
There was no significant difference in any of the individual components or in the group scores (global, physical, role, emotion, cognition, or social) of the European Organization for Research and Treatment of Cancer questionnaire between the antibody responders and nonresponders at baseline. There was also no significant difference between either the antibody responders or the antibody nonresponders between baseline and week 8. Among antibody responders, there was no significant change from baseline (week 0) to last recorded result in four of five of the group scores (global, physical, role, emotion, cognition, or social), with a deterioration in global well-being (Mann-Whitney test, P = .048).
Patient Excluded From Analysis
One patient was excluded from the analysis. She had stage I ampullary carcinoma and not locally advanced pancreatic cancer but was deemed nonresectable by the referring physician because of comorbid disease. She completed the trial according to the protocol, but we have not included her in the data or analysis. In brief, the excluded patient was an antibody responder by week 6, she was still alive at the close of the study (379 days from diagnosis and 311 days from her first injection), and she gained weight (> 5% of baseline body weight) from week 0 to week 16. She had minor injection site pain (grade 1). No other patients who received G17DT are excluded from the data.
DISCUSSION
The therapeutic options for patients with pancreatic cancer are limited.2,8-10,12 Only a small proportion of patients present at an early enough stage to consider surgery with curative intent.1 The majority of patients have inoperable disease, and current treatments have little impact on survival.1-4,6,7 Novel therapeutic strategies are desperately needed.8-13 We describe the first study of one novel strategy, the use of immunotherapy to modify growth factor activity.8-10,18 G17DT is an immunogen that induces antibodies that neutralize both G-17 and precursor glycine-extended G-17.14,18
There are theoretical concerns regarding the immune responsiveness of patients with advanced pancreatic cancer with use of the novel approach of immunotherapy. Patients with advanced colorectal cancer are able to mount an adequate antibody response to G17DT, but this has not previously been assessed in patients with pancreatic cancer.27 We have shown that the majority (82.4%) of patients with inoperable pancreatic cancer are able to mount an antibody response to G17DT at a dose of 250 μg and that this dose is significantly superior to 100 μg (P = .018). In pancreatic cancer, because the prognosis is poor, it is desirable for any treatment to take effect as soon as possible. Our study suggests that there is only a few weeks’ delay from the first treatment to an adequate antibody response. Over 50% of those who mounted an antibody response did so by week 4. By week 6, 13 (72%) of 18 in the 250-μg group had made an antibody response. The majority (87%) of patients who are going to respond to the 250-μg dose have done so by week 6. The peak antibody titer occurs at around week 8 and then tends to wane at a variable rate over the ensuing weeks.
G17DT was generally well tolerated, but three patients did have significant adverse reactions related to the immunogen. Two of these patients did not receive the three doses as scheduled in the protocol because of the severity of the adverse reactions, but they still mounted an antibody response above the average for the study group as a whole. Interestingly, the patients with these adverse reactions appeared to fair well clinically. In one of these patients, there was a fall in the CA 19-9 level, with evidence of tumor regression visible on CT scanning by week 22 (not included in results above because the shrinkage did not occur by week 16). All three of the patients who had these adverse reactions survived for longer than the median survival of the whole group, two were still alive at the time of study closure, and one had survived for over 18 months. It is possible that a severe reaction is more likely to occur in a fitter patient who is more able to mount a vigorous immune response to the immunogen—in other words, a marked adverse reaction may be a marker for better survival rather than a cause of improved survival. It also possible that a severe reaction produces a clinical benefit in a nonspecific way by inducing generalized immune hyperresponsiveness. The rationale behind G17DT is that antigastrin antibodies have activity against patients’ tumors by abrogating the gastrin autocrine growth pathway, resulting in clinical improvement. A higher antibody titer may then lead to a more significant clinical improvement.
Local reactions at injection sites were relatively common, but these were usually mild and hardly ever precipitated the use of additional analgesia. When local reactions were noted, most patients described them as less troublesome than travel vaccinations.
In an uncontrolled study, it is not possible to prove whether a therapy is efficacious or not. The median survival of the whole group (187 days from first injection or 244 days from diagnosis or recurrence) compares favorably with that seen in comparable studies.2,3,6,7,13 Furthermore, the significantly longer survival among the antibody responders compared with the nonresponders provides further support for the theory that the G17DT immunogen may have antitumor activity. There was a greater proportion of patients with stage IV disease in the antibody nonresponders (70%) compared with the antibody responders (40%) and a higher median CA19-9 in the nonresponders, although a higher proportion of nonresponders were in the 100-μg group. Part of the reason for a decreased survival in the antibody nonresponders may be that nonresponse is a marker for more advanced disease. A more advanced disease state may impair humoral immunity.
The improved weight status in antibody responders is interesting. It is possible that the antigastrin effects of the antibodies induced by G17DT may influence satiety.29 It may also reduce gastric acidity and reduce gastric or esophageal inflammation and therefore improve appetite.
In planning future approaches to pancreatic cancer therapy, the combination of G17DT with chemotherapy will be assessed, as there is evidence of synergy in an animal model.30 In order to determine the true efficacy of G17DT in the treatment of pancreatic cancer, phase III studies involving G17DT alone and in combination with chemotherapy have begun, and the results of these studies will determine the usefulness of this novel form of therapy.
Acknowledgments
Supported by grant no. BXB8 from Aphton Corporation, Woodland, CA.
- Received November 30, 2001.
- Accepted June 24, 2002.
