Eur Radiol (2003) 13:729–733 DOI 10.1007/s00330-002-1443-x
G. Herlin B. Persson M. Bergström B. Långström P. Aspelin
Received: 26 October 2001 Revised: 21 February 2002 Accepted: 6 March 2002 Published online: 24 May 2002 © Springer-Verlag 2002
G. Herlin (✉) · B. Persson · B. Långström P. Aspelin Department of Diagnostic Radiology, Huddinge University Hospital, 141-86 Stockholm, Sweden e-mail:
[email protected] Tel.: +46-8-585-8000 Fax: +46-8-7114840 M. Bergström Uppsala University PET Centre, Uppsala, Sweden
NUCLEAR MEDICINE
11C-harmine
as a potential PET tracer for ductal pancreas cancer: in vitro studies
Abstract Our objective was to find a tracer in diagnosing human pancreatic cancer using positron emission tomography (PET). For this purpose in vitro test of pancreatic tissues with autoradiography was used. Autoradiography was performed with 11C-harmine (a MAO-A-inhibitor) with and without competitive inhibition. Tissue preparations were obtained from normal human pancreas and pancreatic cancer. The uptake was compared with rat brain or pig brain, tissues with high expression of MAO-A. Nine autoradiography studies on 16 samples from five different human pancreatic cancers gave a significant level of specific binding of 11C-harmine in 13, and 3 samples did not give a significant level of
Introduction The diagnosis of pancreatic carcinoma in humans is difficult despite many imaging methods including US, CT, endoscopic retrograde cholangiopancreatography (ERCP) and magnetic resonance tomography (MRT). Furthermore, these methods mostly visualise the tumour as a low-signal area in the normal pancreatic tissue, where the latter generally is enhanced by different contrast media. Finding a method that enhances the pancreatic cancer tissue per se as compared with the surrounding normal pancreas tissue and the often associated chronic pancreatitis would therefore be of great value. Ductal pancreas cancer is frequently associated with distant metastases as well as with infiltrative spread in the abdomen. For an optimal planning of treatment also these features should ideally be identifiable; hence,
specific binding of 11C-harmine. All 16 samples were analysed with autoradiography. Compared with rat brain, the uptake in the human cancers varied between 9 and 43% except for one tissue preparation which had a too low value for measurement. This study shows expression of MAO-A in human pancreatic cancer. This is readily characterised in vitro. The potential use of 11C-harmine in the diagnosis of pancreatic cancer using PET might be limited, but further PET studies are necessary. Keywords Tomography · Emission-computed · Autoradiography · Harmine · Pancreatic neoplasms
methods are searched which could identify both primary tumours and metastases. Numerous reports have recently shown high accuracy with positron emission tomography (PET) using 18F-fluorodeoxyglucose for the detection of primary and metastatic pancreas cancer [1, 2, 3, 4]. In a series of in vitro and in vivo studies, utilising PET methodology on different tumours and using a variety of PET tracers, we noticed that one of theses tracers, 11C-harmine (a substance used for the characterisation of the enzyme monoamine oxidase A (MAO-A), [5, 6, 7], had a significant uptake both in metastases and in primary tumours. This finding motivated us to perform a series of in vitro tests where the binding of a PET tracer, 11C-harmine, to human pancreatic cancers and normal pancreas tissue from humans, pigs and rats was investigated by means of frozen section autoradiography.
730
The aim of the study was to evaluate whether this tracer had sufficient specific binding to pancreatic cancer tissue in vitro, suggesting that it could be used in the diagnosis of pancreatic cancer and metastases with PET. In this study there was specific uptake of 11C-harmine in human pancreatic cancer in vitro. 11C-harmine
Materials and methods Frozen-section autoradiography Tumour samples were obtained at surgery, frozen at –70°C and transported to Uppsala University PET Centre Preclinical Laboratory where they were stored at –70°C. The tissue preparations used included normal human pancreas from one patient, human pancreas cancer from five patients and normal pig and rat pancreas. For reference measurements brains from pigs and rats were used. The brain from pigs and rats are well known to have high levels of MAO-A [8, 9]. In the beginning of the study our comprehension was that rat brain and pig brain had approximately the same uptake of 11Charmine, which is the reason why some of the preparations were examined with pig brain as a reference instead of rat brain. This showed later not to be true; we therefore compared the uptake of pig brain to rat brain, and the uptake in pig brain compared with rat brain was 48%. From this comparison it is possible to calculate the relative uptake of 11C-harmine to rat brain in those tissue samples that were only compared with pig brain. The tissue and the tumour specimens were cut in 20-µm sections and adhered to object slide glasses. This procedure was performed at a temperature of –20°C. The glasses with frozen sections were stored at –20°C until the experiments were performed. The radioactive labelled substance was synthesised at the PET centre. Frozen-section autoradiography was performed as previously described [6, 7, 10]. Briefly, the glass slides with sections were first pre-incubated in TRIS-HCl buffer at room temperature for 15 min. The PET tracer 11C-harmine was added at concentrations 2 and 10 nM. For competitive inhibition, unlabelled harmine was used at 10-µM concentration. This latter experiment should indicate if specific binding was at hand as opposed to non-specific binding. The incubation time with the radioactive labelled substance was 20, 30 or 45 min, followed by a washing procedure of 3×2, 3×5 or 3×7 min. The sections were then dried for 7 min. The slide glasses with the sections were placed on a phosphor imager plate covered with a thin plastic film and exposed for 20, 30, 40 or 60 min inside the cassette [6, 7, 10, 11]. The phosphor imager plate was scanned by the laser system and read into the computer. The autoradiograms were presented on the screen and the evaluations performed with the aid of the Image Quant programme [6, 7]. Nine autoradiography studies were performed with 11C-harmine. Calculations Regions of interest were manually outlined on the images to represent tumour tissue as well as a background region outside the tumour section. The average counts were measured separately for each region on a relative scale. On each slide there were 2–12 sections (mean 4 sections) from the tissue preparations used and an average from these were calculated. From this value the background value was subtracted to give a value representing the total uptake [6, 7].
In order to find the specificity of the uptake, all experiments included parallel sections incubated with tracer plus non-labelled substance at sufficiently high concentration to give complete blocking of the target receptor or enzyme. These latter sections indicate non-specific binding, which was subtracted from the total binding, giving a value of specific binding (=PC). The degree of competitive inhibition in percent was obtained by dividing the specific value with the total value, i.e. 100×PC/(P-Bg), where PC is the value of specific binding, P is the value of total binding and Bg is the background value [6, 7]. Identically treated frozen sections of rat brain or pig brain were used as a reference to standardise the uptake between the different experiments and for calculating the binding of tumour or normal pancreas in percent of the binding in reference tissue. In the aforementioned specimens, rat brain had the highest enzyme expression and consequently the highest binding of 11C-harmine and was here set to 100% and the uptake from the other specimens were compared with rat brain in percent. Four of the five human pancreatic cancers were compared with rat brain and one of the five human pancreatic cancers were compared with pig brain. Hence, 100×PC/RBC for rat brain and 100×PC/PBC for pig brain were calculated, where PC is the specific value for the tissue preparation examined, RBC is the specific value for rat brain during the same autoradiography performance and PBC the corresponding specific value for pig brain [7]. The statistical method used was the Student’s t-test.
Results Autoradiography Nine autoradiography studies were performed with 11C-harmine on 16 samples from five human pancreatic cancers. A significant difference (p
