Gut 1996; 38: 33-39 33

Positive receptor scintigraphy correlates with the presence of somatostatin

receptor subtype 2 Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from

M John, W Meyerhof, D Richter, B Waser, J-C Schaer, H Scherubl, J Boese-Landgraf, P Neuhaus, C Ziske, K Molling, E-O Riecken, J C Reubi, B Wiedenmann

Abstract binding studies using SST-14, SST-28, and scintigraphy (SRS) as ligands, suggested that different is positive in approximately 75% of all somatostatin receptor (sstr) subtypes exist in patients with neuroendocrine gastro- neuroendocrine tumour cells (for review see 4). enteropancreatic tumours. This study Five distinct human sstr subtypes have aimed to identify specific somatostatin recently been identified and characterised by receptor (sstr) subtypes, which are molecular cloning and functional expression responsible for the in vivo binding of the studies.12-20 For sstr subtype 2, splice variants widely used somatostatin analogue, sstr2A and sstr2B have been identified. 14 Thus octreotide in human neuroendocrine six sstr subtypes exist up to now. All have been gastroenteropancreatic tumours. Twelve shown to be coupled to G- and to patients underwent SRS with radio- possess seven putative membrane spanning labelled octreotide. After surgical resec- domains.21 22 sstr subtypes differ in their affini- tion, tumour tissues were analysed in vitro ties for specific ligands such as SST-14, SST- for somatostatin and octreotide binding 28, and the stable somatostatin analogues sites by autoradiography. In addition, for octreotide and (23, for review see 24). Departments of the first time, sstr subtype mRNA expres- Expression studies of single sstr subtypes in Gastroenterology and sion was examined by semiquantitative non-neuroendocrine cells found that Surgery, Benjamin reverse transcription polymerase chain octreotide bound with high affinity to sstr sub- Franklin Medical 24 Centre, Freie reaction (RT-PCR). Tumour tissues from types 2 and with lower affinity to 3 and 5.17 Universitat Berlin, all SRS positive patients were positive by Thus, expression of selected sstr subtypes in Germany autoradiography. Semiquantitative RT- different tumour types may account for the M John sstr2 or results obtained by SRS. H Scheruibl PCR revealed most prominently positive negative http://gut.bmj.com/ J Boese-Landgraf expression in scintigraphically positive Recently, Kubota et a125 studied seven neuro- C Ziske tumours. Two SRS negative tumours con- endocrine gastroenteropancreatic tumours E-O Riecken tained in vitro octreotide binding sites as (pancreas, n=6; intestine, n= 1) using the B Wiedenmann well as high levels of sstrl and sstr2 reverse transcription polymerase chain reaction Institute for Celiular mRNAs. Positive SRS is mainly due to (RT-PCR) and suggested that only sstr2 was Biochemistry and sstr2. sstrl, 3, 4, and probably 5 are less functionally relevant. Interestingly, no expres- Clinical Neurobiology, in tumour Universitats- important for in vivo octreotide binding. sion of sstr5 was observed the tissues on October 2, 2021 by guest. Protected copyright. Krankenhaus False negative scintigraphic results seem studied. Furthermore, no quantitation of sstr- Eppendorf, Hamburg, to be influenced by factors independent of mRNAs was done and nor was the expression of Germany W Meyerhof the expression of specific sstr. sstrs studied at the protein level. In addition, the D Richter (Gut 1996; 38: 33-39) data obtained were not correlated to in vivo binding conditions - that is, SRS. Institute ofPathology, Keywords: somatostatin receptor scintigraphy, We have studied a group ofpatients with neu- Universitfit Berne, somatostatin receptor subtype 2, octreotide, tumours, Switzerland neuroendocrine tumour, gastroenteropancreatic roendocrine gastroenteropancreatic B Waser system, RT-PCR who were all subjected to SRS. To evaluate the J-C Schaer number of false positive and false negative J C Reubi results obtained using SRS, tumour specimens Institute for Medical obtained by surgery from all patients were Virology, Zurich, Neuroendocrine tumour cells of the gastroen- analysed by means of SST-28 and octreotide Switzerland K Molling teropancreatic system contain high affinity autoradiography. In addition, and for the first binding sites for somatostatin (SST) and its time, we used a quantitative RT-PCR method Department of analogues (1-3, for review see 4). In 1989, to evaluate the expression levels of sstr mRNAs Surgery, Universitlltsklinilum Krenning et a15 6 applied radiolabelled octreo- in neuroendocrine tumour tissues. Rudolf Virchow, tide intravenously to patients with these Humboldt Universitlit tumours and showed that they could be zu Berlin, Germany P Neuhaus localised by somatostatin receptor scintigraphy Patients and methods (SRS). Although most of these tumours were Correspondence to: were not TUMOUR TISSUES Dr B Wiedenmann, Freie positive by SRS, approximately 25% Universitat Berlin, Benjamin detected by this technique using octreotide as Tumour tissue was obtained from 12 patients Franklin Medical Center, Department of ligand (7 8, for review see9). Interestingly, some with neuroendocrine gastroenteropancreatic Gastroenterology, of the scintigraphically negative tumours tumours by surgery at the Universitatsklinikum Hindenburgdamm 30, exhibited binding sites for SST but not for Benjamin Franklin (UKBF) and Rudolf D-12200 Berlin, Germany. 11 Additional Virchow Berlin, Informed Accepted for publication octreotide in vitro.10 findings, (UKRV), Germany. 31 May 1995 based on autoradiographic and biochemical consent was obtained from all patients and the 34 3John, Meyerhof, Richter, Waser, Schaer, Scheruibl, et al

study was performed in accordance with the encode human sstr subtypes 1-5 has been standards set by the ethical committee of the achieved by PCR amplification of their corres- UKBF. Tissue samples were quick frozen in ponding cDNAs, using ,B-actin mRNA as a liquid nitrogen and stored at -80°C. Tumour control as recently described.29 Briefly, 1 ,ug of histology was verified by both conventional and total cellular RNA that had been prepared immunohistological methods before either from the different tumour tissues and exhaus- Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from RNA preparation or SST receptor autoradio- tively digested with DNase I was used for graphy. One atypical carcinoid was studied and oligo(dT)-primed reverse transcription. showed a positive immunostaining for neuron cDNAs were amplified with 2.5 U Taq DNA specific enolase, synaptophysin, but not for polymerase (Promega, Madison, USA) in 100 chromogranin A, calcitonin, carcinoembryonic gl reaction assays using the following oligonu- antigen (CEA), or serotonin. Focal positive cleotide primers: immunoreactivity was also obtained for cyto- * sstrl - GCTGAGCAGGACGCCACG and keratin and vimentin. Ki 67 expression varied CAACCTGAGAACCTGGAGTCC, corres- from 3-4% of all tumour cells. By conventional ponding to nucleotide positions 903-923 and histology, tumour cells were polygonally 1119-1140 of the published cDNA arranged. By electron microscopy, numerous sequence12; organelles as well as a few neuroendocrine * sstr2 - CCCCAGCCCTTAAAGGCA- granules (150 nm in diameter) were found. TGT and GGACCCTCCTCAATGGAGA- CC, 874-895 and 1087-1 10712; * sstr3 - GTCAACGTGGTGTGCCCAC- PATIENTS TG and GGGCCCCCGGAGAAGACTG- Patient characteristics and SRS results are AG, 932-952 and 1124'-1 14413; summarised in the Table. SRS was performed * sstr4 - GCCTTGATGCCACCGTCAA- as previously described.7826 To avoid jeop- CC and GGTGCATGTGCCCCCCACT- ardy to patients with carcinoid syndrome, SST AA, 956-976 and 1169-118919; treatment was continued during SRS or * sstr5 - ATCTGTCAACCTGGCCGTGG- surgery, or both.27 CG and CGTCCAGACAGGATCCGGC- AG, 855-875 and 1056-107617; * 3-actin - GGGCATGGGTCAGAAGGA- RNA PREPARATION TT and ATGAGGTAGTCAGTCAGGTC, Total RNA from tumour tissue samples was 173-192 and 591-610.30 purified using a modified protocol based on All reaction mixtures were amplified by 40 the method of Chomczynski and Sacchi.28 cycles at 94°C for two minutes, at 63°C (sstrl Briefly, approximately 10 to 20 mg of tumour and 4), at 61°C (sstr2 and 3), at 65°C (sstr5), tissue were sonication in homogenised by or at 60°C (,-actin) for 80 second and at 72°C http://gut.bmj.com/ guanidine thiocyanate buffer. The solution for two minutes. Aliquots (15 gl) of each was acidified with sodium acetate and amplification reaction were removed after 15, extracted once with phenol/chloroform/ 20, 25, 30, 35,m and 40 cycles, and run on isoamyl alcohol, followed by two extractions 1.5% agarose gels. The sizes ofthe cDNA frag- with chloroform. An equal volume of iso- ments were found to be 230 bp (sstrl, 2, 4), propanol was added to the aqueous phase and 210 bp (sstr3), and 220 bp (sstr5) and cor-

the sample was incubated overnight at responded to the predicted fragment lengths on October 2, 2021 by guest. Protected copyright. -20°C. Pelleted RNA was washed twice with of 237, 233, 212, 233, 221 nucleotides for 75% ethanol, stored in 75% ethanol at sstrl-5, respectively. After blotting on Hybond -80C, and redissolved in RNase-free water N membranes (Amersham, Braunschweig, before use in reverse transcription. Germany), filters were hybridised in 5XSSC, 10XDenhardt's solution, 1 mM EDTA, 100 mg/ml herring sperm DNA, 0.5% (w/v) SEMIQUANTITATIVE PCR sodium dodecylsulphate at 50°C with the Estimates of the relative levels of mRNAs that appropriate 32P-labelled sstr or ,B-actin probe: * sstrl - GCTGGATGGACAACGCCGC- Characteristics ofpatients with gastroenteropancreatic neuroendocrine tumours GG, 1036-105612; * sstr2 - GGCACAGATGATGGGGAGC- Patient Age (y) Sex Primary tumour/resected tissue SRS Treatment GG, 1014'-103412; sstr3 - Atypical carcinoid* 33 F Unknown/liver mt -ve s * TCCCGCCGTGTGCGCAGCC- NE ileal tumour 1 60 F Small intestine/ovary mt -ve s, SMS, IFN AG, log 1-1 11 13; NE ileal tumour 2 a 53 F Small intestine/ovary mt -ve s, SMS - NE ileal tumour 2 b Small intestine/liver mt +ve * sstr4 CCTGCGCTGCTGCCTCCTG- Insulinoma 1 66 F Pancreas/p -ve s GA, 1077-109719; Gastrinoma 1 39 M Pancreas/p +ve s, OMP * sstr5 - Gastrinoma 2 42 F Liver/p +ve s, OMP CGCAAGGGCTCTGGTGCCA- Gastrinoma 3 12 F Duodenum/liver mt +ve s, SMS, OMP, IFN AG, 1014'-1034l7; NE ileal tumour 3 68 F Small intestine/p +ve a, SMS - NE ileal tumour 4 67 M Small intestine/p +ve s, SMS * P-actin CCACACCTTCTACAATGA- NE ileal tumour 5 59 F Small intestine/ivermt +ve a, SMS GC, 302'-32130; PNET 1* 45 M Pancreas/p +ve s, SMS, IFN, ch PNET 2** 54 M Pancreas/liver mt +ve SMS 1 pmollml with specific activities between s, 5.2X 105 and 3-8X 106 cpm/pmol. Filters were All patients were subjected to somatostatin receptor scintigraphy (SRS). finally washed in 6XSSC at either 65°C (sstrl, +ve=SRS positive; -ve=SRS negative. Tumours were functional unless otherwise indicated: 2, 4, and 5), 69°C (sstr3), or 55°C (,-actin). *Non-functional; +serotonin producing. Quantification of the hybridisation signals NE=neuroendocrine; PNET=pancreatic neuroendocrine tumour; F=female; M=male; p=primary tumour; mt=metastasis; s=surgery; SMS=somatostatin; IFN=interferon; were carried out using a Bio-Imaging-Analyzer OMP=omeprazole; ch=chemoembolization. (Fujix BAS 2000, Fuji Photo Film Co, Japan). A Scintigraphy and somatostatin receptor subtypes 35

semilogarithmic plot of pixel values versus cycle itracin (40 pug/ml), and MgCl2 (5 mM) to number showed that amplification was exponen- inhibit endogenous proteases. Non-specific tial between cycles 15 and 30 and then reached a binding was determined by adding 1 p,M plateau. The curves obtained with the P-actin solution of unlabelled [Tyr3]-octreotide or mRNA and the individual sstr mRNAs dis- SST-28. Incubated sections were washed twice played similar shapes but were shifted along the for five minutes in cold incubation buffer con- Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from x axis. This indicates a similar efficiency in the taining 0-25% BSA, then in buffer alone, and amplification reaction and that the initial dried quickly. Finally, the sections were amount of specific cDNA present is crucial to apposed to 3H-Hyperfilms (Amersham, UK) the extent of exponential amplification (see and exposed for one week in x ray cassettes. also 31 for a discussion of quantification). Pixel intensities obtained with sstr specific probes during exponential amplification were expressed Results relative to those obtained with the 3-actin probe. No cross hybridisation to the other sstr cDNAs CLINICAL COURSE AND REDUCTION OF was observed with any of the probes (data not SECRETION IN PATIENTS TREATED WITH SST shown). All amplifications were carried out at ANALOGUES least three times with similar results. Eight of 12 patients with neuroendocrine tumours had been treated with octreotide (3X 100-500 ,ug/d subcutaneously, n=7) or SST AUTORADIOGRAPHY lanreotide (30 mg intramuscularly every 14 Cryostat tissue sections (20 Kum) were pro- days, n=4 or 3X5 mg/d, n=3). Two of the cessed for SST receptor autoradiography as eight patients had been treated with a combi- previously described in detail.'-3 The radio- nation of octreotide (3x250 ,ug/d) subcuta- ligands used were the SST analogue 1251- neously and interferon-alpha (3 X 5 mio. [Tyr3]-octreotide and 1251-[Leu8, D-Trp22, IU/wk). All gastrinoma patients were taking Tyr25]-SST-28. Both ligands were iodinated, omeprazole (20'-40 mgfd). Treatment with purified by high pressure liquid chromato- SST analogues led to a reduction in flushing graphy column and characterised in standard and diarrhoea as well as in serum chromo- binding assays as described previously.1 3 For granin A concentrations in all patients with autoradiography, tissue sections were mounted carcinoid syndrome (n=6). Abnormal serum on precleaned microscope slides and stored at concentrations of serotonin (six of six) or -20°C for at least three days to improve gastrin (one; two not tested) were reduced adhesion of tissue to the slide. Sections were after SST treatment. No correlation between then incubated for two hours at ambient SRS positivity and the control of tumour

temperature in the presence of the iodinated growth was observed. http://gut.bmj.com/ ligand (0.15-0-30X106 dpm/ml, approxi- mately 80-160 pM). The incubation solution was 170 mM Tris-HCl buffer (pH 7.4) con- SRS taining 1% bovine serum albumin (BSA), bac- SRS showed neuroendocrine tumour lesions in eight of 12 patients (Table). The tumours in these eight patients consisted of gastrinomas

(n=3), neuroendocrine tumours of the small on October 2, 2021 by guest. Protected copyright. intestine (n=3), and neuroendocrine pan- creatic tumours (n=2). In four of 12 patients negative SRS results were obtained (Table). 10.0 These four patients suffered from insulinoma (n= 1), atypical carcinoid (liver metastasis and 5 unknown primary tumour, n=1), and func- C.) cc < tional neuroendocrine ileal tumours (n= 2), A one with metastases of the liver and ovary and C') U) one with only an ovarian metastasis. Although the liver metastasis was positive by SRS, the metastases to the ovaries, which were 3 or 4 cm 1.0 in diameter, were negative in the latter two patients. Tumour tissues obtained from all 0 these patients during surgery were studied for o PCR:sstr 12345 12345 12345 12345 12345 the expression of the various sstr subtypes by SRS - - - + SST autoradiography and semiquantitative SST-28 - nt + + ++ RT-PCR. Since the diagnostic reliability of Octreotide - +/++ + ++ Atypical NE ileal Ovary Liver Insulinoma 1 SRS is not compromised by continuous treat- carcinoid tumour 1 NE ileal tumour 2 ment with somatostatin analogues,27 somato- Figure 1: Detection ofspecific somatostatin receptor (sstr) subtype mRNAs in somat ostatin statin treatment was not stopped before SRS. receptor scintigraphy (SRS) negative neuroendocrine gastroenteropancreatic tumour tissues. sstr, relative amounts ofsomatostatin receptor subtype 1 to 5 mRNAs. sstr subtype nnRNAs were by RT-PCR using subtype specific primerpairs, in (A) quantitation 5was analysed WITH SST-28 AND performed in relation to P-actin mRNA determined as internal control. In panel (B) AUTORADIOGRAPHY [-actin cDNA fragments could not be amplified due to low levels ofcorresponding mzRNA, OCTREOTIDE AS SST RECEPTOR LIGANDS sstr expression was evaluated at a qualitative level only. SRS=somatostatin receptorrtide as Binding studies of single sstr subtypes autoradiography on cryostat sections of tumour tissue samples using SST-28 or octre + and in non-neuroen- ligands; somatostatin receptors: -, absence of receptors; +, low to moderate density; ++ transfected expressed high density; nt, not tested A U=arbitrary units. docrine cells have shown that octreotide binds 36 6John, Meyerhof, Richter, Waser, Schaer, Scherubl, et al

larisation, occupation of the ligand binding site, endocytotic rate, etc add to the positive or negative signal in SRS. In addition to SST-28 binding, high levels of octreotide binding were found in all gastrinoma tissues (Fig 2) and in tissues of neuroendocrine ileal and pancreatic Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from c tumours (Figs 3, and 4 A and B). This indi- cates that the observed positive SRS findings cm 3.0 are indeed true positive. C,, cnCO, 2.0 SSTR SUBTYPE mRNA 1.0 The abundance of sstr subtype mRNAs was quantified in relation to 3-actin mRNA expres- sion. In three tissue samples, determination of 0 PCR: sstr 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 ~ 3-actin mRNA was not possible. In these SRS + + + experiments the 3-actin mRNA levels were SST-28 ++++ +++ apparently too low to be detected after 40 Octreotide ++++ Gastrinoma 1 Gastrinoma 2! Gastrinoma 3 cycles of amplification. However, sstr-mRNA species have been amplified (insulinoma, n= 1; Figure 2: Detection ofspecific somatostatin receptor (sstr) subtype mRNAsbastypeomat i receptor scintigraphy (SRS) positive gastrinoma tissue samples. sstr su inmAstatre neuroendocrine ileal tumour n= 12, liver analysed as described in the legend to Figure 1. For abbreviations see iFigure 1. metastasis, n= 1; serotonin producing pan- Somatostatin receptors: + +, high density; + + +, very high density. creatic tumour, PNET 2, n= 1; Figs 1 and 3). Figure 1 shows the RT-PCR data obtained from tumour tissue samples of all patients only to sstr2 with high affinityr24 and to a lesser negative by SRS. Two patients were negative extent to sstr5.17 Using octr(cotide as well as both by SRS and octreotide autoradiography. the universal ('panligand') S'ST-28 as radioli- In one of these patients (atypical carcinoid), gand, specific binding sites were found in most mRNA levels for sstrs were comparably low in tumour tissue samples (Figs 1-4). However, comparison to SRS positive tumour tissues the atypical carcinoid and the insulinoma such as gastrinomas (see below). In the other showed no specific octreotide binding as deter- patient (insulinoma 1), moderate amounts of mined by autoradiography (Fiigs 1 and 5 C-F). sstrl and sstr2 mRNAs but high levels of sstr4 This agrees with the negative SRS results in mRNA were detected at a qualitative level (see these two patients. In contrasit, in two patients above). This suggests that SRS negative metastases of the ovary that ] had been missed tumours do not synthesise sufficient receptor http://gut.bmj.com/ by SRS using octreotide as ligCand were positive protein for adequate visualisation. Only sstr by octreotide autoradiography (Fig 1). Inter- subtype 1 and 2 mRNAs were consistently estingly, the liver metastasis of one of these detected at high levels in all SRS negative cases patients exhibited octreotidle binding sites that showed positive octreotide autoradio- both by SRS and by autora diography. This graphy. Interestingly, in one patient with a suggests that depending on the metastatic neuroendocrine ileal tumour metastatic to the

organ additional factors such as tumour vascu- liver and the ovary, the sstrl and 2 transcripts on October 2, 2021 by guest. Protected copyright. were observed in both metastases (Fig 1). As SRS was A B already pointed out, only positive in 10.0 the liver in this patient. The positive autoradi- ography with octreotide in both metastases (liver and ovary) probably accounts for the sstr2 8.0 I _ subtype.32 In vivo binding of octreotide to sstr subtype 5.0 2 is especially supported by one patient with an < serotonin producing pancreatic tumour _ (PNET 2), in whom octreotide autoradiogra- _ phy was positive and only sstr subtype 2 mRNA was found by RT-PCR (Fig 3).

2.0- Gastrinomas, known to be almost always positive by in vitro autoradiography and by 1.0 1.0 SRS, showed consistently high expression of sstr subtypes 2 and 5 mRNAs (Fig 2). This 0 PCR:sstr 12345 12345 12345 123455 1 2 3 4 5 0 observation, taken together with binding SRS + + + + + studies of octreotide on cultured cells trans- SST-28 ++ ++ ++ ++ ++ fected with various sstr subtypes,24 25 suggests Octreotide ++ ++ ++ ++ ++ that sstr subtypes 2 and perhaps 5 were respon- NE ileal NE ileal NE ileal PNET 1 PNET 2 tumour 3 tumour 4 tumour 5 sible for positive signals in both autoradio- Figure 3: Detection ofspecific somatostatin receptor (sstr) subtype mRJNAs in somatostatin graphy and SRS. receptor scintigraphy (SRS) positive neuroendocrine gastroenteropancrreatic tumour tissue In summary, analysis of the expression of the samples. sstr subtype mRNAs were analysed as described in the legend'to Figure 1. In panel mRNAs of all sstr subtypes in tumour tissues (A) quantitation was performed in relation to 1-actin mRNA determi;ned asznternal (n= 13) from all patients (n= 12) investigated control. In panel (B), [-actin cDNA fragments could not be amplified due to low leveso corresponding mRNA, sstr expression was only evaluated at a qualitative level. For showed a heterogeneous pattern. sstr2 was most abbreviations see Figure 1. frequently found (13 of 13), followed by sstrl Scintigraphy and somatostatin receptor subtypes 37 E Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from

a F http://gut.bmj.com/

Figure 4: In vitro detection ofsomatostatin receptors by autoradiography. Left panels - neuroendocrine (NE) ileal tumour 4. (A) Autoradiogram showing total binding of1251.[ Tyr31-octreotide. (B) Autoradiogram showing non-specific binding. Bar=l mm. Right panels - Insulinoma 1. (C) Autoradiogram showing total binding of T25-rTyr]-octreotide. (D) Autoradiogram showing non-specific binding. (E) Autoradiogram showing total binding of'25I-[Leu8, D-Trp 3 Tyr25]-SST-28. (F) Autoradiogram showing non-specific binding. Bar=1 mm. High density ofSST receptors with a high affinityfor octreotide was detected in the neuroendocrine (NE) ileal tumour 4. In the insulinoma, however, a high density on October 2, 2021 by guest. Protected copyright. ofSST receptors with a high affinityfor SST-28 but notfor octreotide was identified.

(1 1 of 13), sstr5 (seven of 13), and sstr4 (seven 28 and octreotide as SST receptor ligands, and of 13). sstr subtype 3 transcripts were detected in vivo using SRS. This allows us to draw con- only at low levels or not at all throughout all clusions about false positive and false negative tissue samples studied (six of 13) (Figs 1-3). findings obtained by SRS as well as about the sstr subtypes involved in octreotide binding in vivo. SRS was performed in all 12 patients and Discussion failed to detect tumour lesions in four patients Binding sites for SST-14, SST-28, and the clin- (Table, Fig 1). Interestingly liver metastases ically relevant SST analogue octreotide have were positive by SRS whereas an ovarian metas- been found in a large number of neuroen- tasis was negative in the same patient (Table, docrine tumours by in vitro radioligand and Fig 1). It is noteworthy that eight out of 12 autoradiographic techniques.l1 Using SRS, in patients, and particularly all three patients with vivo binding of octreotide to neuroendocrine gastrinomas, were positive by SRS (Table, Figs tumours was initially observed by Krenning 2 and 3). To determine whether the intra- et a15 6 and was later confirned by others.7 8 26 33 venously applied, radiolabelled SST analogue To date, the expression of the known five main octreotide bound to specific receptors in vivo, sstr subtypes (for nomenclature see 24) has been an in vitro analysis of SST binding sites and sstr examined only in a small number of patients subtypes present in the respective tumour with neuroendocrine tumours of the pancreas tissues was performed. As demonstrated by in and intestine by a non-quantitative RT-PCR vitro autoradiography, two of the four patients approach.25 In this study, we have assessed who were diagnosed to be negative by SRS were the composition of sstr subtypes in human true negative for octreotide binding (Fig 1). neuroendocrine tumour tissue by semiquantita- However, in two other patients ovarian metas- tive RT-PCR, by autoradiography using SST- tases that had been missed by SRS despite their 38 John, Meyerhof, Richter, Waser, Schaer, Scheriubl, et al

sizes of 3 or 4 cm were positive for octreotide frequently detected (six of 13) and expression and SST binding by in vitro autoradiography levels were comparably low. (Fig 1). This indicates that at least some neuro- Analysis of tumour tissues homogenates by endocrine tumour lesions are false negative by RT-PCR also included non-tumour cells such SRS. Since one of the two patients with SRS as those derived from connective tissues and negative ovarian metastases had liver metastases blood vessels. Thus, we cannot exclude that Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from that were positive by SRS, we suggest that SRS normal tissue and cells contributed to the posi- positivity may depend not only on the expres- tive signals obtained by RT-PCR. In vitro sion pattern and expression levels ofvarious sstr autoradiography, however, detected only bind- subtypes (see below) but also on the blood ing sites on tumour cells, suggesting that non- supply, locally high concentrations of endoge- tumour cells express SST protein at low levels nous SST, as well as down regulation of SST only. Thus, despite the extremely high sensitiv- binding sites by various factors (for example, ity of RT-PCR, the detected sstr mRNAs may corticosteroids). derive mainly from tumour cells and not from Based on our autoradiographic and SRS others. findings, we expected to find at least one ofthe In summary, our data suggest that specific five subtypes in 10 out of 12 patients studied binding of radiolabelled octreotide as detected by RT-PCR. Semiquantitative RT-PCR by in vivo SRS correlates mainly with binding analysis of sstr expression in the various to sstr2, since the corresponding transcript can neuroendocrine tumour tissues showed that be detected at relatively high levels in neuroen- all contained high levels of sstr2 mRNA, docrine gastroenteropancreatic tumours. except the two tissues mentioned above that Future studies, using subtype specific ligands, were true negative by SRS and autoradio- may help to determine whether sstrl, 3, 4 are graphy (Figs 1-3). To date, it is not known also relevant in the diagnosis of neuroen- which sstr subtype(s) binds octreotide in vivo. docrine tumours, and may lead to improved Transfection studies in cultured non-neuro- symptomatic and antiproliferative SST treat- endocrine cells expressing an individual sstr ment of this disease. subtype suggest that human sstr2 may fulfil The authors thank I Eichhorn (Berlin) and H H Honk this role.2425 Moreover, in vitro binding of (Hamburg) for expert technical assistance. This study was labelled supported in part by grants from the Deutsche Krebshilfe/Dr octreotide to human neuroendocrine M Scheel-Stiftung (W31/91/Wi 1), Verum-Stiftung, and tumours correlated with the presence of sstr2 the Deutsche Forschungsgemeinschaft (SFB 366/A5) to BW; Deutsche Forschungsgemeinschaft (Sche 326/3-1) to HS; subtype mRNA.32 Since only sstr2 mRNA was and Deutsche Forschungsgemeinschaft (SFB 232/B4) to WM detected in tumour tissue of one of our and DR. patients Part of this work was presented at the 10th International (Fig 3) and since the tumour had Symposium on Gastrointestinal Hormones, Santa Barbara, been positive by SRS and autoradiography, we California, August 27-31, 1994 and was published in abstract form (Scherubl H John M, Meyerhof W, Raulf F, Bruns

consider this Ch` http://gut.bmj.com/ the first direct evidence for high Reubi JC, et al. Somatostatin receptor subtypes in neuroen- affinity binding sites of octreotide to sstr2 docrine cell lines and tumor tissues of the gastroenteropancre- in vivo. atic system. Dig Dis Sci 1994; 39 (8): 1787). Gastrinomas generally exhibit high uptake of radiolabelled octreotide in vivo and show 1 Reubi JC, Haecki WH, Lamberts SWJ. Hormone-produc- high incidence of sstr by 9 ing gastrointestinal tumors contain a high density of autoradiography.4 somatostatin receptors. Jf Clin Endocrinol Metab 1987; 65: All of our gastrinoma patients were strongly 1127-34. 2 Reubi JC, Maurer R, von Werder K, Torhorst J, Klijn JGM, positive by SRS, and tumour tissues of these Lamberts SWJ. Somatostatin receptors in human on October 2, 2021 by guest. Protected copyright. patients contained both sstr2 and sstr5 tran- endocrine tumors. Cancer Res 1987; 47: 551-8. scripts (Fig It is 3 Reubi JC, Kvols LK, Waser B, Nagorney DM, Heitz PU, 2). noteworthy that only sstr2 Charboneau JW, Reading CC, Moertel C. Detection of and sstr5 mRNAs were detected in tumour somatostatin receptors in surgical and percutaneous nee- tissue dle biopsy samples of carcinoids and islet cell carcinomas. of one of the three gastrinoma patients Cancer Res 1990; 50: 5969-77. (Fig 2). Since the intensity of receptor scinti- 4 Reubi JC, Laissue J, Waser B, Horisberger U, Schaer JC. gram in this patient was similar Expression of somatostatin receptors in normal, inflamed to that of the and neoplastic human gastrointestinal tissues. Ann NY two other gastrinoma patients expressing addi- Acad Sci 1994; 733: 122-37. tional sstr 5 Krenning EP, Breeman WAP, Kooij PPM, Lameris JS, subtype mRNAs, we suggest that in Bakker WH, Koper JW, Ausema I, Reubi JC, Lamberts addition to sstr2 sstr5 may also bind octreotide SWJ. Localisation of endocrine-related tumours with radioiodinated analogue of somatostatin. Lancet 1989; i: in vivo, since both subtypes possess reported 242-4. high affinity binding sites for octreotide at least 6 Lamberts SWJ, Bakker WH, Reubi JC, Krenning EP. in Somatostatin-receptor imaging in the localization of vitro.24 Other sstr subtypes seem to be less endocrine tumors. NEnglJMed 1990; 323: 1246-9. important for the positivity of SRS. 7 Scheriibl H, Bader M, Fett U, Hamm B, Schmidt-Gayk H, in one with Koppenhagen K, Dop F-J, Riecken E-O, Wiedenmann B. Interestingly, patient a primary, Somatostatin-receptor imaging of neuroendocrine gas- ileal neuroendocrine tumour metastatic to the troenteropancreatic tumors. Gastroenterology 1993; 105: liver and 1705-9. ovary, only sstrl and 2 were found in 8 Zimmer T, Ziegler K, Bader M, Fett U, Hamm B, Riecken all tissues studied (Fig 1), which supports the E-O, Wiedenmann B. Localisation of neuroendocrine 25 tumors of the upper gastrointestinal tract. Gut 1994; 35: suggestion3 that the expression patterns of 471-5. sstr subtypes may be conserved during the 9 Krenning EP, Kwekkeboom DJ, Oei HY, de Jong RJB, Dop FJ, Reubi JC, Lamberts SWJ. Somatostatin receptors in metastatic process in a tumour specific fashion. gastroenteropancreatic tumors: on overview of European Analysis of the expression of the mRNAs of all results.AnnNYAcadSci 1994; 733: 416-24. 10 Lamberts SWJ, Hofland U, van Koetzveld PM, Reubi JC, sstr subtypes in tumour tissues (n= 13) from all Bruining HA, Bakker WH, Krenning EP. Parallel in vivo patients (n= 12) investigated showed a hetero- and in vitro detection of functional somatostatin receptors in human endocrine pancreatic tumors: consequences geneous pattern, sstr2 was most frequently wvith regard to diagnosis, localization, and therapy. J Clin found (13 of 13), followed by sstrl (11 of 13), EndocninolMetab 1990; 71: 566-74. 11 Kvols LK, Browvn ML, O'Connor MK, Hung JC, Hayostek sstr5 (seven of 13) and sstr4 (seven of 13). RJ, Reubi JC, Lamberts SWJ. Evaluation of a radiolabeled Expression of sstr3 transcripts was the least somatostatin analog (I-123 octreotide) in the detection Scintigraphy and somatostatin receptor subtypes 39

and localisation of carcinoids and islet cell tumors. 22 Bell GI, Reisine T`. Molecular biology of somatostatin Radiology 1993; 187: 129-33. receptors. TINS 1993; 16: 34-8. 12 Yamada Y, Post SR, Wang K, Tager HS, Bell GI, Seino S. 23 Rens-Domiano S, Reisine T. Biochemical and functional Cloning and functional characterization of a family of properties of somatostatin receptors. J Neurochemy 1992; human and mouse somatostatin receptors expressed in 58: 1987-96. brain, gastrointestinal tract, and kidney. Proc Natl Acad 24 Bruns C, Weckbecker G, Raulf F, Kaupmann K, Schoeffter Sci USA 1992; 89: 251-5. P, Hoyer D, Lubbert H. Molecular pharmacology of 13 Yamada Y, Reisine T, Law SF, Ihara Y, Kubota A, somatostatin receptor subtypes. Ann NY Acad Sci 1994; Kagimoto S, et al. Somatostatin receptors, an expanding 733: 138-46. Gut: first published as 10.1136/gut.38.1.33 on 1 January 1996. Downloaded from family: cloning and functional characterization of 25 Kubota A, Yamada Y, Kagimoto S, Shimatsu A, Imamura human sstr 3, a protein coupled to adenylyl cyclase. M, Tsuda K, Imura H, Seino S, Seino Y. Identification of Mol Endocrinol 1992; 6: 2136-42. somatostatin receptor subtypes and an implication for the 14 Patel YC, Greenwood M, Kent G, Panetta R, Srikant CB. efficacy of somatostatin analogue SMS 201-995 in treat- Multiple gene transcripts of the somatostatin receptor ment of human endocrine tumors. J Clin Invest 1994; 93: sstr2: tissue selective distribution and cAMP regulation. 1321-5. Biochem Biophys Res Commun 1993; 192: 288-94. 26 Wiedenmann B, Bader HM, Scherubl H, Fett U, Zimmer 15 Corness JD, Demchyshyn LL, Seeman P, van Tol HHM, T, Hamm B, Koppenhagen K, Riecken E-O. Srikant CB, Kent G, Patel YC, Niznik HB. A human Gastroenteropancreatic tumor imaging with somato- somatostatin-14 like peptides. FEBS Let 1993; 321: statin-receptor scintigraphy. Semin Oncol; 1994; 21 279-84. (suppl): 29-32. 16 Xu Y, Song J, Bruno JF, Berelowitz M. Molecular cloning 27 D6rr V, Wurm K, Horing K, Gozman G, Rath U, Bihl H. and sequencing of a human somatostatin receptor, hsstr4. Diagnostic reliability of somatostatin receptor scinti- Biochem Biophys Res Commun 1993; 193: 648-52. graphy during continuous treatment with different 17 Yamada Y, Kagimoto S, Kubota A, Koichiro Y, Masuda K, somatostatin analogs. Horm Met Res 1993; 27 (suppl): Yoshimichi S, Ihara Y, Li Q, Imura H, Seino S, Seino Y. 36-43. Cloning, functional expression and pharmacological 28 Chomczynski P, Sacchi N. Single-step method of RNA characterization of a fourth (hsstr4) and fifth (hsstr5) isolation by acid guanidiniumthyocyanate-phenol-chloro- human somatostatin receptor subtype. Biochem Biophys form extraction. Analytical Biochemistry 1987; 162: Res Commun 1993; 195: 844-52. 156. 18 Demchyshyn LL, Srikant CB, Sunahara RK, Kent G, 29 Wulfsen I, Meyerhof W, Fehr S, Richter D. Expression Seeman P, van Tol HHM, Panetta R, Patel YC, Niznik patterns of rat somatostatin receptor in pre- and HB. Cloning and expression of a human somatostatin-14- postnatal brain and pituitary. J Neurochem 1993; 61: selective receptor variant (somatostatin receptor 4) 1549-52. located on 20. Mol Pharmacol 1993; 43: 30 Ponte P, Ng SJ, Engel J, Gunning P, Kedes L. Evolutionary 894-901. conservation in the untranslated regions of actin mRNAs: 19 Rohrer L, Raulf F, Bruns C, Hofstadter F, Buttner R, DNA sequence of a human beta-actin cDNA. Nucleic Schiule R. Cloning and characterization ofa fourth human Acids Res 1984; 12: 1687-6. somatostatin receptor. Proc Natl Acad Sci USA 1993; 90: 31 Chelly J, Kaplan JC, Maire P, Gautron S, Kahn A. 4196-200. Transcription of the dystrophin gene in human muscle 20 Panetta R, Greenwood MT, Warszynska A, Demchyshyn and non-muscle tissues. Nature 1988; 333: 858-60. LL, Day R, Niznik HB, Srikant CB, Patel YC. Molecular 32 Reubi JC, Schaer JC, Waser B, Mengod. Expression cloning, functional characterization, and chromosomal and localization of somatostatin receptor sstrl, sstr2, localization of a human somatostatin receptor (somato- and sstr3 mRNAs in primary human tumors using statin receptor type 5) with preferential affinity for in situ hybridization. Cancer Res 1994; 54: 3455-9. somatostatin-28. Mol Pharmacol 1994; 45: 417-27. 33 Nocaudie-Calzada M, Huglo D, Deveaux M, Carnaille B, 21 Patel YC, Greenwood MT, Warszynska A, Panetta R, Proye C, Marchandise X. Iodine-123-Tyr-3-octreotide Srikant CB. All five cloned human somatostatin receptors uptake in pancreatic endocrine tumors and in carcinoids (hsstrl-5) are functionally coupled to adenylyl cyclase. in relation to hormonal inhibition by octreotide. J Nucl Biochem Biophys Res Commun 1994; 198: 605-12. Med 1994; 35: 57-62. http://gut.bmj.com/ on October 2, 2021 by guest. Protected copyright. 302 Letters, Book review, Correction

drug induced acute pancreatitis is probably rare and that the disease usually takes a LETTERS TO benign course. Such a retrospective evalua- BOOK tion in a substantial number of patients has THE EDITOR not been done before. REVIEW P G LANKISCH Department ofInternal Medicne, Municipal Hospital ofLiuneburg, D-21339 Luneburg, Germany

Drug induced pancreatitis Inflammatory Bowel Disease. 4th ed. Edited by J B Kirsner, R G Shorter. (Pp EDITOR,-I have read with interest the article 1033; illustrated; £1 16). Baltimore: Williams on drug induced pancreatitis by Lankisch et al Biliary stenting in the management of and Wilkins, 1995. ISBN 0-683-04627-6. (Gut 1995; 37: 565-7). It is unclear to me bile duct stones how the opinion that in 1-4% of patients The fourth edition of 'Kirsner and Shorter' disease was drug induced could have been EDITOR,-We read with interest the leading appears exactly 20 years after the first edition substantiated. article by Dalton and Chapman (Gut 1995; and is a tribute to the extraordinary energy of It is difficult to believe that only 22 of 1613 36: 485-7). Their suggestion that a sphinc- its editors. Compared with the third edition patients were exposed to drugs. What about terotomy may not always be necessary in published in 1985, the current volume has drugs taken by the remaining 1591; is opinion these patients is absolutely valid. We would expanded by over 200 pages and many new enough to exclude possible causation, plainly go a step further in stating that in such authors have been introduced. It remains an not. patients, ifthe size ofthe stone is > 15-20 mm all American book but its perspective of the To estimate the true impact of drug at ultrasound examination or ERCP, then literature is global and generally well bal- induced disease the authors would have had stenting should be performed with a 7 French anced. Inevitably in a book of 41 chapters to conduct a controlled study. Nevertheless, stent without sphincterotomy. This will devoted to two diseases, there is some repeti- the authors have concluded that drug induced prevent the complications associated with tion but this is no bad thing if the volume is acute pancreatitis occurs rarely in clinical endoscopic sphincterotomy, which occur in used for reference or for browsing. The practice. That opinion has not been substan- 8-10% of patients undergoing the largest expansion compared with previous tiated by this study. procedure.' Furthermore, this would prevent editions concems pathogenesis - 11 chapters M J S LANGMAN migration of straight stents. As already men- compared with six in the third edition. This Department ofMedicine, tioned in the leading article, there is no evi- rightly reflects the remarkable explosion of Queen Elizabeth Medical Centre, dence to show so far, that 10 French stents interest in the immunological and inflamma- Queen Elizabeth Hospital, Edgbaston, Birmingham B12 2TH are superior to stents of smaller diameter. tory mechanisms in pathogenesis that has Although 7 French stents tend to clog earlier occurred during the past 10-15 years. than the 10 French ones,2 they may easily This volume provides us with the most be exchanged when blockage occurs. comprehensive account of ulcerative colitis Endoscopic sphincterotomy may however be and Crohn's disease currently available. It is required if multiple stents need to be placed. obsessively referenced and therefore provides Reply Apart from maintaining the flow ofbile and an excellent entry into the original literature. preventing stone impaction and cholangitis, It provides elegant accounts of the experi- EDITOR,-We are grateful for the interest stenting has other benefits too. Placement of mental, immunological, and pathophysio- Professor M J S Langman took in our paper biliary endoprostheses has been shown to logical mechanisms that may be involved in and regret that it has obviously given rise to decrease the size of the stones on follow pathogenesis but also provides detailed some misunderstanding. up.3 4 Moreover, in patients with stricture of accounts of medical and surgical treatment. The preselection of patients diagnosed for the common bile duct, where lithotripsy may The psychosocial problems ofthe diseases are drug induced acute pancreatitis was made by be difficult or impossible, biliary endopros- amplified by a final chapter written from the the centres where they had been treated. All theses may resolve such strictures-7 in addi- perspectives of an affected subject and a charts of patients considered to fall into this tion to decreasing the size of the stone. multiply affected family. group were reviewed by us. We looked only S P MISRA Clinical gastroenterologists will find this for drugs, however, currently held responsible M DWIVEDI book invaluable, those in training will find a for inducing acute pancreatitis. It is possible Department of Gastroenterology, mine of information, and for the IBD MLN Medical College, that in the 135 patients with acute pancreati- Allahabad - 211 001, India specialist it will continue to be a much used tis of unknown aetiology drugs had been book ofreference. It well lives up to its aims as acute in given still unknown to induce pancreati- 1 Cotton PB. Endoscopic management of bile duct described the preface although I am not as tis and thus the incidence of this aetiology stones; (apples and oranges). Gut 1984; 25: sanguine as the editors that the aetiology of might be somewhat higher. 587-97. either disease will be understood within the Prospective studies help to answer open 2 Status evaluation: biliary stents. Gastrointest next five years. Endosc 1992; 38: 750-2. questions but logistic realities pose problems. 3 Chan ACW, Ng EKW, Lai CW, et al. Common D P JEWELL The two questions at issue are: how fre- bile duct stones become smaller after endo- quently does the application of a certain drug scopic biliary stenting. [Abstract]. Gastrointest lead to acute how is Endosc 1995; 41: 393. pancreatitis and, frequent 4 Vallera RA, McGee SG, Shearin M, et al. Biliary drug induced acute pancreatitis among all stents decrease the size ofretained common bile patients with acute pancreatitis? duct stones. [Abstract]. Gastrointest Endosc The first question is impossible to answer. 1995; 41:419. 5 Bourke MJ, Elfant AB, Alhalel R, Kotan P, CORRECTION In view ofthe great number ofpatients receiv- Haber GB. Biliary and pancreatic strictures ing drugs such as frusemide and oestrogen, it complicating endoscopic biliary sphincter- would be impossible to follow up all of them otomy. [Abstract]. Features and endoscopic for and acute management. Gastrointest Endosc 1995; 41: 390. signs symptoms of pancreatitis. 6 Hmeidan A, Jacob J, Sherman S, Lehman GA. Even the second question is difficult to Benign biliary strictures: outcome of endo- An error occurred in the paper by Dr John answer. To make quite sure that the sus- scopic therapy. [Abstract]. Gastrointest Endosc and others (Gut 1996; 38: 33-39). The title of pected drug has really induced the disease, a 1995; 41:399. the paper should read 'Positive somatostatin 7 Hmeidan A, Jacob J, Sherman 5, Lehman GA. re-exposure to the drug in question is neces- Benign biliary strictures: frequency and man- receptor scintigraphy correlates with the sary, something ethically difficult to justify. agement at ERCP. [Abstract]. Gastrointest presence of somatostatin receptor subtype 2 The message of our paper was simply that Endosc 1995; 41: 399. and 5'.