Cancer Therapy Vol 1, 121-131, 2003.
Safety, feasibility and clinical
benefit of localized chemotherapy using microencapsulated cells for inoperable
pancreatic carcinoma in a phase I/II trial
Research
Article
Matthias Lšhr1,2,
Jens-Christian Kršger4, Anne Hoffmeyer2,6, Mathias Freund3,
Johannes Hain6, Albrecht Holle2, Wolfram T. Knšfel8,
Stefan Liebe2, Horst Nizze5, Matthias Renner6,9,
Robert Saller6, Petra MŸller2,6, Thomas Wagner10,
Karlheinz Hauenstein4, Brian Salmons6,9 and Walter H.
GŸnzburg7*
1Department
of Medicine II, Medical Faculty Mannheim, University of Heidelberg, Germany; 2Division
of Gastroenterology, 3Division of Hematology & Oncology,
Department of Medicine, 4Department of Diagnostic and Interventional
Radiology and 5Department of Pathology, University of Rostock,
Rostock, Germany; 6Bavarian Nordic GmbH, Martinsried, Germany, 7Institute
of Virology, University of Veterinary Sciences, Vienna, Austria, 8Department
of Surgery, University of Hamburg, Hamburg, Germany; 9Austrianova,
VeterinŠrplatz 1, Vienna, Austria; 10Division of Hematology &
Oncology, Medical University LŸbeck, LŸbeck, Germany
________________________________________________________________________
*Correspondence: Prof. Walter H. GŸnzburg,
Institute of Virology, University of Veterinary Medicine, Veterinaerplatz 1, A-1210 Vienna,
Austria; Tel.: +43-1-25077-2301; Fax:
+43-1-25077-2390; e-mail: walter.guenzburg@vu-wien.ac.at
Key Words: Pancreatic carcinoma, cell
therapy, gene therapy, ifosfamide, microcapsules, angiography
Received: 10
June 2003; Accepted: 25 June 2003; electronically published: June 2003
Summary
Previous preclinical studies suggested that
implantation of encapsulated, genetically modified cells converting a
chemotherapeutic agent in the vicinity of tumors may represent an effective
treatment for pancreatic cancer. A phase I/II clinical trial was performed to
determine safety, feasability and efficacy of such a targeted, low dose,
chemotherapy in 14 advanced-stage pancreatic cancer patients. Genetically
modified allogeneic cells expressing the enzyme cytochrome P450 2B1
encapsulated in cellulose sulfate polymers were delivered angiographically via
catheter into blood vessels leading to the tumor. These cells locally activate
systemically administered ifosfamide to its active metabolites, whilst
remaining immuno-isolated. Although adverse events were experienced by all
patients, none of these were related to the treatment, with the possible
exception of increased serum lipase 15 days after
CapCell instillation in one patient. According
to the NCI tumor response classification, at the final observation within the
study, 2 of the 14 patients treated had partial remissions (14.3%), 11 patients
had stable disease (78.6%) and one patient died after 8 days. Median
survival was doubled compared to a historic control group (p=.008) and 50% more than usually achieved with
gemcitabine. One year survival, at 36%, was three fold that of the control
group (p=.047) and twice that reported for gemcitabine. Of 13 evaluable
patients, 4 patients reported improvements in pain assessment, with 6 remaining
unchanged (4 of these experienced no pain) and 3 patients experiencing slightly
more pain. Using a worst case scenario, 50% of patients experienced a clinical
benefit whereas in a best case scenario benefit was experienced by 71% of
patients.
I. Introduction
Pancreatic carcinoma ranks as the eighth most frequent
solid cancer in industrialized countries but is the fifth leading cause of
cancer-related deaths (Greenlee et al, 2001). Radical surgery can only be
applied in about 10% of diagnosed cases (Huguier and Mason 1999; Neoptolemos et
al, 2001) and, to date, all efforts to control tumor growth by radiation and/or
chemotherapy have not significantly prolonged survival or reduced tumor load
(Heinemann 2002; Rosenberg 2000). Even the newly introduced chemotherapeutic
agent gemcitabine only marginally prolongs the survival of patients (Burris et
al, 1997). Nevertheless, this agent has rapidly become a standard treatment
because of the additional palliative effect and its ability to improve the
clinical benefit response and quality of life of patients with pancreatic
cancer. Thus, there is a need for new treatment regimes to treat pancreatic
cancer (Hawes et al, 2000). Along with more classical types of treatment,
attempts also have been made to employ gene therapy approaches such as using
suicide genes encoding enzymes that are able to convert a prodrug to itÕs
active, tumor toxic form (Aspinall and Lemoine 1999; GŸnzburg et al, 2002;
Rosenberg 2000).
The
chemotherapeutic agent, ifosfamide, has been shown to have potentially
therapeutic effects for pancreatic cancer (Loehrer et al, 1985). In a phase II
trial in which 1.6g/m2/day ifosfamide was administered for 5 days to
21 evaluable patients, 7 Stable Diseases with mostly none severe, grade 1-2
toxicity was reported (Wils et al, 1993). In another study (Keizer et al,
1995), where up to 1.5g/m2/day ifosfamide was given as a 10 day
continuous i.v. infusion to patients with various tumor types, of six patients
with pancreatic cancer, one showed a partial response and a second evidenced a
tumor reduction of 45%. Major side effects observed were leukopenia with
granulocytopenia, whilst subjective side-effects included nausea/vomiting and
fatigue (probably related to neurotoxicity). More encouraging clinical effects
have been observed in other trials where medium doses of ifosfamide (1-2g/ml)
have been investigated, but this is accompanied by medium grade toxicity
profiles. In an initial study by Gad-El-Mawla and colleagues, where 2g/m2
were given for 5 days, all but two patients developed haemorrhagic cystitis.
However there were 6 partial responses in 10 patients (Gad-El-Mawla and Ziegler
1981). A further study revealed that of 25 patients receiving daily doses of
1.8g for five days, 1 patient showed a complete remission and 14 patients
showed partial remision (Gad-El-Mawla 1986). However, these patients suffered
from grade 3 alopecia (100%), grade 1 anaemia (100%) and leukopenia (30%).
Thus, it is to be expected that higher doses (2-3g/ml) of ifosfamide may show
even greater efficacy, but that this will be associated with possibly
unacceptable levels of toxicity.
Ifosfamide is a prodrug that requires activation by
liver specific cytochrome enzymes, such as the 2B1 isoform (CYP2B1) to generate
tumor toxic metabolites (Dirven et al, 1996). Unfortunately, the short
half-life of these metabolites in plasma (Cerny et al, 1991b; Kurowski and Wagner
1993), coupled with the distance that they have to travel, require high
systemic levels of ifosfamide to achieve therapeutic levels in the tumor.
Indeed, these levels are so high as to lead to unacceptable side effects
(Loehrer et al, 1985). Local activation of ifosfamide at the site of the tumor
should, in contrast, result in good local cytotoxic activity, and at the same
time low systemic ifosfamide concentrations, thus resulting in only minimal
systemic side effects (Chen and Waxman 2002). Local activation may be achieved
by introducing encapsulated human 293 cells genetically modified to overexpress
CYP2B1 at the site of the tumor. Encapsulation in cellulose sulfate allows
allogeneic cells to survive in vivo, by protecting them from host immune attack
as well as by physically constraining them to the site where they are required
(Dautzenberg et al, 1999). Previous experiments had revealed that injection of
such encapsulated CYP2B1 expressing cells into pre-established tumors in a nude
mouse model of human pancreatic carcinoma (Lšhr et al, 1994), resulted in
complete tumor regression in about 20% of mice and a significant anti-tumor
effect in the remaining mice (Lšhr et al, 1998). Nevertheless, this route of
application may not be suitable for patients and so a further study was
performed to demonstrate the feasibility of intra-arterial placement of
micro-encapsulated cells into blood vessels leading to the pig pancreas (Kršger
et al, 1999; Lšhr et al, 2003). Based upon these encouraging preclinical data,
a phase I/II clinical trial was initiated involving patients with inoperable
pancreatic carcinoma to assess the feasibility, safety, and tolerability of
this new treatment modality (Lšhr et al, 1999). We have recently described the
results obtained concerning safety and efficacy in a brief report (Lšhr et al,
2001). Here, data is presented concerning the outcome and clinical benefit of
this treatment.
II.
Patients and methods
A. Patients, trial design and approval
The study
was planned as an open, prospective, single-arm, single center phase
I/II-study, following the German gene therapy working group (DAG-GT)
recommendations. The protocol was approved by the state ethics committee, the
gene therapy board of the German Medical Association and published (Lšhr et al,
1999), in line with the recommendation of the German working party on gene
therapy indicating the approval of all regulatory bodies. The study was opened
on the 28th July 1998 and closed on the 20th September 1999. The trial was
conducted in full accordance with good clinical practice guidelines (ICH-GCP).
B. Patient
enrollment
A total of 17 patients were
enrolled in the trial between July 1998 and April 1999 (Table 1) from the 51 patients screened during the study period. Reasons for
non-enrolment were previous chemotherapy (n = 8), pancreatic surgery (n = 13),
poor general condition (n = 18), unwillingness to participate (n = 5), or death
(n = 7). Criteria for entering the study included an inoperable pancreatic
adenocarcinoma stage III-IV (UICC) (Hermanek et al, 1997), as determined by
histology and measured by CAT scan and only patients who had not received prior
chemotherapy were enrolled (Lšhr et al, 1999). During the preparation period,
clinical data were collected and a baseline CAT scan of the abdomen was
performed. The patients were scheduled for the initial celiac angiography with
capsule placement (day 0). On day 1, the patients were monitored for evidence
of any clinically relevant adverse reactions, e.g. allergic, and/or pancreatitis.
The levels of serum amylase, lipase, lactate, lactate dehydrogenase, and liver
enzymes, as well as complete blood cell count were determined. Systemic
chemotherapy commenced on day 2 with 1g/m2 body surface of
ifosfamide (Holoxan¨) in 250 ml 0.9% normal saline being given as a 1-hour
intravenous infusion on three consecutive days. This was accompanied by a 60%
dose equivalent of the uroprotector MESNA (Uromitexan¨) given as three i.v.
injections. This regimen was repeated at days 23-25 for all patients except 5
and 17 who only received one round of ifosfamide. Toxicity was measured based
on the WHO/NCI guidelines on common toxicity criteria. Control CAT scans were
scheduled for weeks 10 and 20, respectively. During the final visit, a control
angiography was performed. On the initial CAT scan, the scan demonstrating the
largest diameter of the primary tumor was identified and the area measured.
Using appropriate landmarks, an identical scan was used for comparison. CAT
scans were evaluated by two unrelated radiologists, one of whom was not
involved in the study. Standard NCI criteria for evaluating tumor growth were
used to assess stable disease (SD), partial remission (PR), and minor response
(MR). After formally finishing the study, patients were followed up on an
ambulatory basis with three-monthly visits. Besides measuring tumor size by CAT
scan, the need for pain medication and the quality of life was monitored using
questionnaires established for pancreatic diseases (Bloechle et al, 1995). A
clinical benefit score based upon variables including Karnofsky score, body
weight, pain and analgesic consumption was also calculated from this data. Pain
intensity was measured on a visual analogue scale ranging from 0 (no pain) to
100 (the most imaginable intensive pain), in steps of 10. Analgesics
consumption was assessed using another scale in which 0 indicated no regular
administration of analgesic, whereas scores of 25, 50 and 75 indicated
administration of non-steroidal anti-inflammatory drugs (NSAID) or opiates
several times per year (25), per month (50) or per week (100) (Bloechle et al,
1995).
C.
Historical patient collective
Survival data of a
retrospective (historic) control group and the treatment group of this study
were compared. A historic control group was established from an evaluation of
all patients (n=35) with pancreatic carcinoma admitted to the Division of
Gastroenterology, Rostock during the years 1996 to 1998 who were not treated by
tumor resection. Of these patients, 1 had UICC stage I, 2 stage III, and 33
stage IV pancreatic carcinoma, respectively. Seven underwent palliative
surgery, 10 received palliative chemotherapy, 24 needed biliary drainage (ERCP
or PTCD), and 19 received best supportive care (in addition to biliary drainage
or surgery). One stage IV patient was excluded since no date of death was
available for this patient. Though the selection criteria for treated patients
could not be applied completely to the historic
control group, the historic controls and the treated patients were comparable
in clinical diagnoses and initial symptoms of the disease (jaundice, abdominal
pain were most frequent) and also with respect to median age (63 years in both
cohorts) and gender (male patients: 74.3% in historic controls and 64.3% in
treated patients).
D. Production of clinical grade CapCell¨
The cytochrome P450 2B1 (CYP2B1)
expression construct (Lšhr et al, 1998), as well as the good laboratory
practice (GLP) production and characterization of the CYP2B1 expressing 293
cell clone (22P1G) (Gunzburg et al, 1999) have been described previously. Cells
were amplified under good manufacturing practice (GMP) conditions (Q-One,
Glasgow, Scotland, UK) and encapsulated in polymers of cellulose sulphate using
an apparatus from Inotech (Dottikon, Switzerland) (Dautzenberg et al, 1999).
The encapsulated cells (CapCell¨) were washed twice with plain RPMI cell
culture medium (Gibco/BRL) and stored at 4¡C. Cell viability was determined
using the Life&Dead viability kit (MobiTec, Braunschweig, Germany). Necessary
quality control tests required for release included sterility and a
demonstration that the CapCell¨ were both mycoplasma and endotoxin free. The
mechanical stability of the capsules was determined and the potency of the
encapsulated cells was determined in a cell toxicity bioassay (Lšhr et al,
2002).
E.
Angiography
Visualization of the vasculature
leading to the pancreatic tumor was performed by angiography in a standard
manner with the transfemoral approach (Seldinger technique). Digital
subtraction angiography of the celiac trunk, superior mesenteric artery,
splenic artery, common hepatic artery and, if necessary for identification of
tumor leading vessels, of the gastroduodenal artery, was performed with a 4
French introducer system (Terumo), 4 French visceral catheters with a inner
diameter of 0,038" (Cordis) and a monomer nonionic contrast medium (Imeron
300, BYK, Gulden). The most appropriate tumor access was determined by relating
tumor localization in CAT scans to the vessel anatomy. Supraselective
catheterization of an artery leading into the tumor was performed with a
coaxial 2.3 French microcatheter system (Cordis) (Kršger et al, 1999; Lšhr et
al, 2003). The optimal approach to the tumor vasculature was gained through the
inferior pancreatoduodenal artery, the dorsal pancreatic artery and/or the
superior pancreatic head branches of the gastroduodenal artery. After
documentation of the correct microcatheter placement in a non-occluding
position, 300 CapCells were instilled slowly one by one with the blood flow in
13 patients. An additional patient received 250 capsules due to limited space
in the tumor artery. The patency of the cannulated vessel was controlled
periodically by fluoroscopy, followed by a control angiography of the target
vessel region. The catheter and introducer systems were then removed, the
puncture site compressed for 15 minutes, and a compression tape put in place
for 6 hours. Diagnostic angiography visualising the peritumoral vessels was
repeated in the same manner during the final visit (week 20).
F. Quality of Life
A quality of
life core questionnaire for cancer patients, QLQ-C30, has been validated in
several languages (Aaronson et al, 1993; Fayers et al, 1999; Hjermstad et al,
1998; Klee et al, 1997; Sprangers et al, 1993), but the module for pancreatic
carcinoma is still under development with respect to reliability, sensibility
against changes, and multicultural validation (Fayers et al, 1999). Therefore,
in this study an unauthorised version of the core questionnaire and a German
quality of life scale for pancreas patients was used which had been published
(Bloechle et al, 1995). The quality of life-data were documented independently
from the safety and efficacy data by filling-out an independent questionnaire
by the patient. Thus, the assessment of the quality of life data did not
interfere with the routine documentation of the adverse events that were
reported by the patient. The quality of life core questionnaire was analyzed in
analogy to the prescriptions of the EORTC (Fayers et al, 1999). Quality of Life
data were available from the baseline evaluation for all 14 patients and for
analysis of change from 8 patients. The analysis was strictly performed
according to the EORTC recommendations (Fayers et al, 1999).
III. Results
Each
patient enrolled in the trial received 300 cellulose sulfate capsules
(CapCellÒ) except patient 12 who received 250 CapCell¨ by angiographic
placement (day 0) into a suitable artery feeding a primary, inoperable tumor
(stage III-IV). Each capsule had an average diameter of 0.8 mm and contained
around 104 cells (Lšhr et al, 1999). An appropriate artery leading
into the tumor could be supraselectively cannulated (Figure 1) in 14 of the 17 patients entering the
study (Table 1). Two
patients developed severe infections before the start of the trial and had to
be treated by other means, whilst angiography was not successful in one
patient.
