JOURKAL OF PATHOLOGY, VOL.
161: 145-1 51 (1990)
EXPRESSION OF GASTRIN-RELEASING PEPTIDE (HUMAN BOMBESIN) GENE IN LARGE CELL UNDIFFERENTIATED CARCINOMA OF THE LUNG Q. A. HAMID, B. CORRIN, A. DEWAR, H. HOEFLER* AND M. N. SHEPPARD
Departments of Pathology?National Heart and Lung Institute. Brompton Hospital, London, U.K.; *Graz Universiiy, Austria Received 8 December 1989 Accepted 2 February 1990
SUMMARY Production of the growth factor gastrin-releasingpeptide (GRP) or human bombesin has been shown to be a feature of neuroendocrinetumours of the lung, particularly small cell carcinoma, and is possibly responsible for the characteristically rapid growth of this tumour. Large cell undifferentiated carcinoma of the lung (LCC) is also characterized by rapid growth and there is increasing evidence that some LCCs exhibit neuroendocrine differentiation. We therefore investigated GRP/bombesin immunoreactivity and the expression of GRP gene in ten LCCs. Histologically, all were composed of large cells with abundant cytoplasm, open nuclei, and prominent nucleoli, and there was no evidence of squamous, glandular, or neuroendocrine differentiation. At the ultrastructural level, most showed squamous or glandular differentiation but none contained neuroendocrine granules. None of the tumours showed immunoreactivity for GRP/bombesin but seven of the ten showed a focal hybridization signal when treated with 32P-labelledGRP cRNA probes, indicatingthe presence of GRP mRNA. This wasconfirmed by northern blot analysis. This study shows for the first time that GRP gene is expressed in LCC. The production of GRP may contribute to the aggressive behaviour of LCC. KEY WORDS-Bombesin,
gastrin-releasing peptide, large cell carcinoma, lung, gene and in situ hybridization.
INTRODUCTION Gastrin-releasing peptide (GRP) o r human bombesin is an important growth regulatory factor. It is a 27-amino acid peptide that was first isolated from the stomach and later found to be widely distributed in the central nervous system and the gastrointestinal and respiratory tracts.I4 Its presence in the respiratory system has attracted attention for two reasons. The first is its high expression in early fetal lung. This decreases gradually throughout the neonatal period and thereafter so that G R P is hardly detectable in the adult lung, suggesting that it has a possible role in the growth and development of the The second is the association of GRP Addressee for correspondence: Professor B. Corrin, Department of Pathology, Naiional Heart and Lung Institute, London SW3 6LY.U.K.
0022-341 7/90/060145-07$05.00 0 1990 by John Wiley & Sons, Ltd
with neuroendocrine tumours of the lung, particularly its secretion by small cell carcinoma. Here G R P is possibly acting as an autocrine growth fact ~ r . ~Experimental -'~ studies support this by showing that G R P stimulates the rowth of normal and neoplastic respiratory cells, I F 1 3 whilst monoclonal antibodies t o G R P block its binding to cellular receptors and inhibit the growth of small cell carcinoma in v i m and the growth of small cell carcinoma xenografts in vivo.13 The cloning of G R P cDNA7-I6and the development of molecular biological techniques have opened up new ways of investigating GRP production in lung tumours. Recently we have shown that GRP mRNA is highly expressed in small cell carcinomas of the lung, although a large proportion lack GRP/bornbesin immunoreactivity owing to poor storage of the mature protein."." However, the production of G R P in non-small cell carcinoma of the
I46
Q. A. HAMID ET A L .
lung has not been investigated so extensively. yet there is increasing evidence that a proportion of LCCs exhibit neuroendocrine features. LCC forms about 10-20 per cent of lung turn our^.'^ It is an aggressive neoplasm and the 5-year survival is only 5 per cent overall’0 and 12 per cent if operable.” Growth factors such as GRPibombesin have been identified by immunocytochemistry in a minority ( 10 per cent) of LCCs.”.” The aim of this study was to investigate the expression of GRPjbombesin immunoreactivity and GRP gene in LCC using immunocytochemical and molecular biological techniques.
described.“ The cDNA of G R P was subcloned with reverse orientation into the polylinker region of pSP64 vector and the resulting plasmid linearized with Xba I prior to transcription. A radio-labelled probe complementary to the coding sequence (cRNA robe) of G R P was transcribed in the presence of [P‘PIUTP and SP6 polymerase. The resulting 700-base cRNA was digested to an average size of 175 bases each by alkaline hydr~lysis.’~ A sense probe was constructed by subcloning the G R P cDNA into SP65 vector. to give a sequence identical to G R P mRNA following transcription. In situ h>+hridization
MATERIALS AND METHODS Tissue preparation Surgically resected LCCs of the lung ( n = 10) were collected fresh from the operating theatre of the Brompton Hospital and four pieces from each tumour were taken and fixed or frozen immediately. One was fixed in 10 per cent formalin for conventional histology and immunocytochemistry; the second was fixed in sodium cacodylate buffered 2.5 per cent glutaraldehyde for electron microscopy; the third was fixed in 4 per cent paraformaldehyde for in situ hybridization; and the fourth was frozen in liquid nitrogen for northern blot analysis. Conventional light and electron microscop). Haematoxylin and eosin stained paraffin sections of each tumour were examined. For electron microscopy the 4 1 8 h glutaraldehyde-fixed tissues were rinsed in buffer. post-fixed in osmium tetroxide, embedded in araldite. and ultrathin sections stained with uranium and lead after selecting suitable areas from 1 um thick toluidine blue-stained sections. II?in~unoc?.tochemistr?. Tumour tissue from each case was fixed for 12 h in 10 per cent formalin. embedded in paraffin. and 5 pm thick sections were stained with antibody to GRP,’bombesin ( 1 i2000 dilution, Cambridge Research Biochemicals. Cambridge. U.K.) using the peroxidase-antiperoxidase method of Sternberger PI Sections from human fetal lung were used as a positive control.
Probe construction Recombinations containing sequences corresponding to G R P were isolated as previously
For in situ hybridization the tissue was fixed in freshly prepared 4 per cent paraformaldehyde for 4 h and then washed in phosphate-buffered saline (PBS) containing 15 per cent sucrose. and cryostat blocks were prepared. Sections were cut at a thickness of 15 pm, mounted on poly-L-lysine-coated slides, and allowed to dry overnight at 37°C. For hybridization the preparations were permeabilized by soaking in the following solutions: 0.1 mol/l glycine/PBS ( 5 min), 0.3 per cent Triton X-100 in PBS ( 1 5 min), 1 ,ug/ml proteinase K, 0.1 mol/l Tris, and 50 mmol/l ethylenediamine tetraacetic acid (EDTA)solution (20min at 37°C). Thesections were then treated with 4 per cent paraformaldehyde for 5 rnin to stop the reaction and immersed in 0.25 per cent acetic anhydride in 0.1 molil triethanolamine solution to minimize non-specific background binding. Prehybridization was carried out with 50 per cent formamide in 2 x standard saline citrate (2 x SSC) for 30 rnin at 37°C. Hybridization was carried out with 6 ng probeisection (approximately lo6 cpm/section) diluted in a buffer containing 50 per cent formamide, 2 x SSC, 10 per cent dextran sulphate, 0.25 per cent bovine serum albumin, 0.25 per cent Ficol 400, 0.25 per cent polyvinyl pyrrolidone 300, 250mmol/l Tris, 0.5 per cent sodium pyrophosphate. and 0.5 per cent sodium dodecyl sulphate and 250 pg/ml denatured salmon sperm DNA. Hybridization was carried out overnight in a moist chamber at 42°C. After hybridization the sections were rinsed in decreasing concentrations of SSC (4xSSC-0.5xSSC). 30 rnin each at 40°C. Non-specifically bound, single-stranded probe was removed by treatment with a solution containing 20 pg/ml RNase A. 0.5 molil sodium chloride, 10 mmol/l Tris, and 1 mmol/l EDTA for 30 min at 40°C. After dehydrating the sections in graded alcohols, they were dipped in Ilford K5 emulsion,
147
BOMBESIN m R N A IN L A R G E C E L L C A N C E R
Table I-Electron microscopy. immunocytochemistry. in situ hybridization. and northern blot analysis of ten large cell carcinomas of the lung G R P mRNA localization Case No. 1
2 3 4 5
6 7 8 9 10 t
Ultrastructural differentiation
G RP Iborn besin immunoreactivity
Adenosquamous Adenosquamous Glandular Glandular None Squamous Clara cell Glandular Mixed Clara and type I1 cell Squamous
by in situ hybridization
by northern blotting
+ +
NP
+ +
+
+ + -
+-
NP
+-
NP
+
NP NP
=positive: - =negative: N P = n o t performed.
dried, and exposed for 3 days at 4°C before being developed. GRP mRNA hybridization was controlled positively by hybridizing the 3’P-labelled GRP cRNA probe with sections of a small cell carcinoma of the lung known for its high expression of G R P mRNA by northern blotting. Negative controls were provided by separate sets of sections of LCC being hybridized with probe identical to the coding strand of mRNA of GRP (sense probe). Northern blot ana/ysis Total RNA was purified from five cases of LCC (Table I) by the guanidinium isothiocyanate method.” The RNA was subjected to electrophoresis on 1 per cent agarose gel containing 2.2 M formaldehyde in 50 per cent formamide, transferred nitrocefiulose~ and hybridized with 106cpmOf 7’P-labelled GRP cRNA p r ~ b e . ’ ~
RESULTS Histologically, the LCCs were all composed of large cells with abundant cytoplasm, open nuclei, and prominent nucleoli, and there were no features suggesting squamous, glandular, or neuroendocrine differentiation (Fig. I). Table I shows the ultrastructural features: there was evidence of glandular, squamous, or adenosquamous differentiation in most, but no neurosecretory granules were identified (Figs 2 and 3).
Fig. I-Large cell undifferentiated carcinoma of the lung characterized by tumour cells having abundant cytoplasm and nuclei with an open chromatin pattern and prominent nucleoli
None of the LCCs showed any degree of immunoreactivity to GRPibombesin antibody in contrast to thestrongimmunoreactivity observed in the positive control neuroendocrine cells of fetal lung. Hybridization between radio-labelled cRNA probes and mRNA encoding GRP was demonstrated by specific deposits of silver grains in the emulsion overlying tumour cells. Using 3’P-labelled GRP cRNA probe, an intense selective labelling was observed over the cytoplasm of tumour cells in seven of the ten LCCs investigated (Table I). The
148
Q.A. HAMID E T A L .
Fig. 2-Large cell undifferentiated carcinoma showing prominent tonofilaments, ultrastructural evidence of squamous differentiation. Elsewhere the same tumour also showed ultrastructural evidence of glandular differentiation (see Fig. 3). EM
Fig. 3-The Same large cell undifferentiated carcinoma as Fig. 2. here showing m,crovill, into a small lumen. ultrastructural evidence ofglandular differentiation,EM 17 000
x 170OO
tumour cells which showed positive hybridization were either arranged in groups or scattered singly among other tumour cells (Fig. 4). Strong positive hybridization was detected in the positive control small cell carcinoma. Only background levels of silver grains were detected in all preparations treated with GRP mRNA (sense probe) (Fig. 5). Northern blot analysis confirmed the specificity of the signal and the probe. Single bands corresponding to G R P mRNA were detected among the total RNA isolated in three or five LCCs (Fig. 6 ) . DISCUSSION In this study we have used immunocytochemical and molecular biological techniques to demonstrate GRPand itsmRNA in LCC. The results provideevidence of the localizationof GRP gene transcriptsin seven often such tumours,asdemonstrated by in siru hybridization.
Fig. &Autoradiograph of a cryostat section of large cell undifferentiated carcinoma. In siru hybridization with a '?P-labelled CRNA probe coding for GRP mRNA. Note the cytoplasmic localization of GRP mRNA
149
BOMBESIS mRNA IN LARGE CELL CANCER
Fig. 5-Autoradiograph of a cryostat section of large cell crrcinoma showing in situ hybridization with a radio-labelled GRP sense (control) probe. Note the absence of specific signal
The term large cell carcinoma of the lung is applied to anaplastic tumours which by light microscopy show no attempt at maturation.” A although by electron microscopy many show squamous or glandular differentiation. or both.” ‘O Recently i t has been shown by electron microscopy3’ and immunocytochemistry” that a proportion of LCCs exhibit neuroendocrine features. It has been suggested that LCCs with neuroendocrine features are associated with a particularly poor prognosis.’2 The presence of neurosecretory granules in turnour cells represents an advanced stage of Fig. &Northern (RNA) blot analysis oftotal RNA froma large undifferentiated carcinoma (B). Hybridization was carried neuroendocrine differentiation and their demon- cell out with a ’2P-labelledcRNA probe coding for GRP mRNA. A stration at ultrastructural level is not always easy. similar but stronger hybridization signal was obtained with a On the other hand. the value of immunocyto- small cell carcinoma used as a positive control (A). Arrows chemistry in demonstrating neuroendocrine differ- indicate the site of GRP mRNA (approximately 900 base pairs) entiation of lung tumours is limited by the lack of a specific general neuroendocrine marker. The most widely used are neuron specific enolase and protein G R P has been linked to the growth of normal and gene peptide 9.5. but both have frequently been neoplastic pulmonary neuroendocrine cells and reported in non-neuroendocrine tumours.2‘.7i.32 shown to be an autocrine growth factor for small ~ only a marker Conversely. negative immunoreactivity for a par- cell carcinoma of the l ~ n g . ~It-is’ not ticular peptide or protein does not necessarily mean of neuroendocrine differentiation, but can also prothat it is not being synthesized. In siru hybridization vide valuable information on the biological activity is a more effective way of studying such biological of a tumour. i7.33-35 GRP is generally regarded as activity and differentiation of tumour cells. This is being specific for neuroendocrine tumours of the well illustrated in our study. as we were able to lung.3’ Although it has been reported in a minority demonstrate GRP mRNA but not GRP,’bombesin of LCCs,”.” we were unable to demonstrate GRP/ bombesin immunoreactivity in any of our ten cases. immunoreactivity in seven of ten LCCs. GRP is a peptide which has been reported in the However. we demonstrated G R P mRNA in seven lung almost exclusively in neuroendocrine cells and out of the ten, indicating the presence of G R P gene neuroendocrine tumours. with particularly high transcription and possible production of the levels being secreted by small cell c a r c i n ~ m a . ~ , corresponding ~.~ peptide. The discrepancy between
B
150
Q. A. HAMID ETAAL.
the results of immunocytochemistry and in situ differentiation by electron microscopy failed to give hybridization could be explained by any of the fol- a hybridization signal. This observation is suplowing. (1) GRP may be produced but rapidly ported by the demonstration of GRP mRNA in released, so that although the gene is active its prod- cases of very poorly differentiated adenocarcinoma uct cannot be detected within the cells. (2) The GRP (unpublished observations), indicating that adenogene transcript might produce other fragments of carcinoma of the lung might also have a tendency the preprobombesin molecule that cannot be toward neuroendocrine differentiation and GRP detected by the antibody used. This has been shown production. In conclusion, we have shown for the to be the case in small cell carcinoma of the lung, first time that the gene for GRP production is where the C-flankingpeptide of GRP is expressed at actively expressed in LCC.This may be relevant to a higher level than GRP it~e1f.I~ Antibodies to dif- the aggressive behaviour and poor prognosis of this ferent fragments of the preprohormone molecule tumour. might be better for investigatingpeptide gene products in malignant turn our^.^^ (3) There may be tranREFERENCES scription of the gene but blockage or abnormal translation of the mRNA. 1 McDonald TH, Jornvall H, Ghatei M, Bloom SR,Mutt V. CharacterOnly seven of our ten LCCs were positive for ization of an avian gastric (proventricular) peptide having sequence GRP mRNA. The absence of an in situ hybridizhomology with the porcine gastrin-releasing peptide and the amphibian peptides bornbesin and alytain. FEBS Lett 1979; 122 ation signal in the remaining three cases might indicate that these tumours were at different stages of 1 4548. McGregor G, Woodhams P. OShaughnessy D, Ghatei M, Polak JM. differentiation or growth and there was no stimulus Bloom SR. Developmental changes in bombesin. substance P. somatostatin and vasoactive intestinal polypeptide in the rat brain. to initiate transcription of the GRP gene. However, Left 1982;28: 21-27 limited sensitivity of in siru hybridization cannot be 3 Neurosci Dockray GJ, Vaillant C, Walsh, JH. The neuronal origin of bombesinexcluded, especially in the presence of low copies of like immunoreactivity in the rat gastrointestinal tract. Neuroendocrinot o g ~19794 1561-1568. the message. Wharton J, Polak JM, Bloom SR. e f ul. Bombesin-like immunoreacTo ensure that the in situ hybridization signal was 4 tivity in the lung. Nulure 1978;273: 769-770. specific we used radio-labelled cRNA probes, which 5 CUUE, Chan W, Track NS. Bombcsin, calcitonin and leuenkephalin immunoreactivity in endocrine cells of the human lung. Experienfiu are the most specific and We also used 1981;37: 765-767. a number of control procedures including the use of 6 Price J. Nieuwenhuijzen Kruxman AC, Doniach I. Howlett TA, a sense probe (a probe which has an identical Besscr GM, Rees LH. Bornbesin-like peptides in human endocrine tumors: quantitation, biochemical characterization. and secretion. J sequenceto GRP mRNA) and RNase pretreatment. C/in EndocrinoiMefub 1985:60:1097-1 103. The absence of signals following these procedures is 7 Spindcl ER. Sunday ME. Hoefler H, Wolfe HJ, Habner JF, Chin evidence of the specificity of the reaction. To ensure WW. Transient elevation of mRNAs encoding gastrin-releasing pep tide (GRP), a putative pulmonary growth factor, in human fetal lung. probe specificity,the probe was applied to a smallcell JCCIinInvesr 1987;80: 1172-1179. carcinoma known for its GRP mRNA expression. 8 Erisman MD, Linnoila R1, Hernandez 0, Diaugustine RP, Lazarus LH. Human lung small cell carcinoma contains bombesin. Proc Nut/ The strong positive signal obtained here provided USA1982;7% 2379-2383. further support for the specificity of the hybridiz- 9 AcOdSci Moody TW, Pert CB, Gazdar AF. Carney DN, Minna JD. High levels ation. The specificity of the probe and the hybridizof intracellular bombesin characterize human small cell carcinoma. Science 1981;214124&1248. ation signal were further confirmed by northern blot VE. Linnoila R-I, Memoli V, Warren WH. Neuroendocrine analysis, which showed mRNA that corresponded 10 Could components of the bronchopulmonary tract: hyperplasia, dysplasias, to GRP in three of the five cases investigated by this and neoplasms. Lab Invest 1983:4% 519-537. technique. The negative results obtained in the other 11 Could VE. Linnoila R-I, Memoli VA. Warren WH. Neuroendocrine cells and neuroendocrine neoplasms of the lung.Puthol Annu 1983;18: two cases positive by in situ hybridization might be 287-291. due to dilution of GRP mRNA in the tumour’s total 12 Warren WH, Memoli VA, Could VE. Immunohistochemical and ultrastructural analysis of bronchopulmonary neuroendocrine RNA. Using northern blotting, the level of GRP UIlrusfruct Pufhol 1984;7:185-199. mRNA in LCCs appears to be less than in small cell 13 neoplasms. Cuttitta F, Carney DN, Mulshine I, et ul. Bombesin-like peptide can carcinomas. However, it was noteworthy that the function as autocrine growth factor in human small cell cancer. Nufure 1985; 316 823-826. pattern of mRNA distribution at the insitu level was 14 Weber S.Zuckerman JE,Bostwick DG, Bensch KG, Sikic BI. Raffin similar, suggesting that GRP played a similar role in TA. Gastrin-releasinn Deptide is a selective mitoeen for small cell the two tumours. It may be relevant that most of carcinoma in v i m . J 5th h w s f 1985;75 306309.the cases that showed positive hybridization also 15. Willey JC, Lechner JR, Hams CC. Bombesin and the C-terminal tetradecapeptide of gastrin-releasing peptide are growth factors for showed glandular differentiation, as seen in adenonormal human bronchial cpitheiial cells. Exp Cell Res 1984; 153 carcinomas. The two cases that showed squamous 245-248.
BOMBESIN mRNA IN LARGE CELL CANCER 16. Spindel ER,Zilberberg MD, Habner JF, Chin WW .Two prohornones
forgastrin-reledsingpeptideareencodedby twomRNAdiffering by 19 nucleotides. f r o c N o d AcodSci U S A 1986: 83: 19-23. 17. Hamid Q . Addis BJ. Springall DR. er al. Expression of the C-terminal peptide of human pro-bomksin in 361 lung endocrine turnours. a reliable marker and possible prognostic indicator for small cell carcinoma. Virchows Arch A 1987;411: 185-192. 18. Hamid Q, Bishop A. Springall D, er al. Detection of human probornbesin mRNA in neuroendocrine (small cell) carcinoma of the lung. Cancer 1989;63 266-271. 19. Edinburgh Lung Cancer Group. Patients presenting with lung cancer in South East Scotland. Thorox 1987: 4 2 853-857. 20. Huhti E.Sutinen S, Saloheimo M. Survival among patients with lung cancer. An epidemiologic study. Am Rev Respir Dis 1981; 124: 13-16. 2 I . Said JW. Vimadakdl S, Nash G, er a / . lmmunoreactive neuron-specific enolase. bornbesin, and chromogranin as markers for neuroendocrine lung turnours. Hum fafhol1985:1 6 2 3 6 2 4 0 . 22. Piehl MR, Gould VE, Warren WH, er al. lmmunohistochemical identification ofexocrine and neuroendocrine subsets of largecell lung carcinomas. farhol Res f r a c f 1988; 183: 675-5682. 23. Sternberger LA, Hardy PH Jr, Cuchlin JJ, Meyer HG. The unlabelled antibody enzyme method of immunocytochemistry. Preparation and properties of soluble antigen-antibody complex (horseradish peroxidastantiperoxidase) and its use in the identification of spirochetes. J Hisrochem Cyrochem 1970; 1 8 31 5-324. 24. Hamid Q. Wharton J , Terenghi G , er al. Localisation of arterial natriuretic peptide mRNA and immunoreactivity in rat heart and humanatrialappendage. f r o c NadAcadSciUSA 1987;84:6760-6764. 25. Chirgwin JM, Prrybyla AE. MacDonald RJ. Rutter WJ. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemisrry 1979; 18 5294-5299. 26. World Health Organization. International Histological Classification of Tumours. No. 1. Histological Typing of Lung Tumours. 2nd ed. Geneva: WHO, 1981. 27. Churg A. The fine structure of large cell undifferentiated carcinoma of the lung. Evidence for its relation to squamous cell carcinomas and adenocarcinomas. Hum farhol1978; 9:143-156.
151
28. Hone A, Ohta M. Ultrastructural features of large cell carcinoma of the lung with reference to the prognosis ofpatients. Hum Parhol1981; 1 2 423-432. 29. Albain KS, True LD, Golomb HM. Hoffman PC, Little AG. Large cell carcinoma of the lung. Ultrastructural differentiation and clinicopathologiccorrelations. Cancer 1985; 56: 1618-1623. 30. Kodama T , Shimosato Y,Koide T, Watanabe S, Teshima S. Large cell carcinoma of the lung-ultrastructural and immunohistochemical studies. Jpn JClin Oncoll985; 15: 431-441. 31. Rode J, Dhillon AP, Doran JF. Jackson P, Thompson RJ. PGP9.5-a new marker for human neuroendocrine tumours. Hisloparhology 1985: 9: 147-158. 32. Addis BJ, Hamid Q,lbrahim NBN, Fahey M, Bloom SR, Polak JM. lmmunohistochemical markers of small cell carcinoma and related neuroendocrine tumours of the lung. J fafhol1987; 153 137-150. 33. Roth KA, Evans CJ, Weber E. Barchas JD, Bostwick DG, Bensch KG. Gastrin-releasingpeptide-related peptides in a human malignant lungcarcinoid tumor. Cancer Res 1983; 43: 541 1-5415. 34. Carney DN, Gazdar AF, Bepler G, ef a / . Establishment and identification of small cell lung cancer cell lines having classic and variant features. Cancer Re5 1985; 4 5 2913-2923. 35. Polak JM, Bloom SR. Theneuroendocrine system and its turnours. In: Frank LM, Teich NMC, eds. Introduction to the Cellular and Molecular Biology of Cancer. Oxford: Oxford University Press, 1986; 327-331. 36. Hamid Q, Bishop A, Sikri K, Varndell I, Bloom SR, Polak JM. Immunocytochemical characterization of 10 pancreatic turnours, associated with the glucagonoma syndrome, using antibodies to separate regions of the proglucagon molecule and other neuroendocrine markers. Hisropafhology 1 9 8 6 10: 119-133. 37. CoxKH,DeLeonDV,AngererLM,AngererRC. DetectionofmRNA in sea urchin embryos by in sifuhybridisation using asymmetric RNA probe. Dev E d 1984; 101:485-502. 38. Hoefler H,Childers H, Montminy MR, k h a n RM, Goodman RH. Wolfe HJ. 'In sifu' hybridisation methods for the detection of somatostatin mRNA in tissue sections using antisense RNA probes. Hisrochem J 1986; I 8 597-604.
161: 145-1 51 (1990)
EXPRESSION OF GASTRIN-RELEASING PEPTIDE (HUMAN BOMBESIN) GENE IN LARGE CELL UNDIFFERENTIATED CARCINOMA OF THE LUNG Q. A. HAMID, B. CORRIN, A. DEWAR, H. HOEFLER* AND M. N. SHEPPARD
Departments of Pathology?National Heart and Lung Institute. Brompton Hospital, London, U.K.; *Graz Universiiy, Austria Received 8 December 1989 Accepted 2 February 1990
SUMMARY Production of the growth factor gastrin-releasingpeptide (GRP) or human bombesin has been shown to be a feature of neuroendocrinetumours of the lung, particularly small cell carcinoma, and is possibly responsible for the characteristically rapid growth of this tumour. Large cell undifferentiated carcinoma of the lung (LCC) is also characterized by rapid growth and there is increasing evidence that some LCCs exhibit neuroendocrine differentiation. We therefore investigated GRP/bombesin immunoreactivity and the expression of GRP gene in ten LCCs. Histologically, all were composed of large cells with abundant cytoplasm, open nuclei, and prominent nucleoli, and there was no evidence of squamous, glandular, or neuroendocrine differentiation. At the ultrastructural level, most showed squamous or glandular differentiation but none contained neuroendocrine granules. None of the tumours showed immunoreactivity for GRP/bombesin but seven of the ten showed a focal hybridization signal when treated with 32P-labelledGRP cRNA probes, indicatingthe presence of GRP mRNA. This wasconfirmed by northern blot analysis. This study shows for the first time that GRP gene is expressed in LCC. The production of GRP may contribute to the aggressive behaviour of LCC. KEY WORDS-Bombesin,
gastrin-releasing peptide, large cell carcinoma, lung, gene and in situ hybridization.
INTRODUCTION Gastrin-releasing peptide (GRP) o r human bombesin is an important growth regulatory factor. It is a 27-amino acid peptide that was first isolated from the stomach and later found to be widely distributed in the central nervous system and the gastrointestinal and respiratory tracts.I4 Its presence in the respiratory system has attracted attention for two reasons. The first is its high expression in early fetal lung. This decreases gradually throughout the neonatal period and thereafter so that G R P is hardly detectable in the adult lung, suggesting that it has a possible role in the growth and development of the The second is the association of GRP Addressee for correspondence: Professor B. Corrin, Department of Pathology, Naiional Heart and Lung Institute, London SW3 6LY.U.K.
0022-341 7/90/060145-07$05.00 0 1990 by John Wiley & Sons, Ltd
with neuroendocrine tumours of the lung, particularly its secretion by small cell carcinoma. Here G R P is possibly acting as an autocrine growth fact ~ r . ~Experimental -'~ studies support this by showing that G R P stimulates the rowth of normal and neoplastic respiratory cells, I F 1 3 whilst monoclonal antibodies t o G R P block its binding to cellular receptors and inhibit the growth of small cell carcinoma in v i m and the growth of small cell carcinoma xenografts in vivo.13 The cloning of G R P cDNA7-I6and the development of molecular biological techniques have opened up new ways of investigating GRP production in lung tumours. Recently we have shown that GRP mRNA is highly expressed in small cell carcinomas of the lung, although a large proportion lack GRP/bornbesin immunoreactivity owing to poor storage of the mature protein."." However, the production of G R P in non-small cell carcinoma of the
I46
Q. A. HAMID ET A L .
lung has not been investigated so extensively. yet there is increasing evidence that a proportion of LCCs exhibit neuroendocrine features. LCC forms about 10-20 per cent of lung turn our^.'^ It is an aggressive neoplasm and the 5-year survival is only 5 per cent overall’0 and 12 per cent if operable.” Growth factors such as GRPibombesin have been identified by immunocytochemistry in a minority ( 10 per cent) of LCCs.”.” The aim of this study was to investigate the expression of GRPjbombesin immunoreactivity and GRP gene in LCC using immunocytochemical and molecular biological techniques.
described.“ The cDNA of G R P was subcloned with reverse orientation into the polylinker region of pSP64 vector and the resulting plasmid linearized with Xba I prior to transcription. A radio-labelled probe complementary to the coding sequence (cRNA robe) of G R P was transcribed in the presence of [P‘PIUTP and SP6 polymerase. The resulting 700-base cRNA was digested to an average size of 175 bases each by alkaline hydr~lysis.’~ A sense probe was constructed by subcloning the G R P cDNA into SP65 vector. to give a sequence identical to G R P mRNA following transcription. In situ h>+hridization
MATERIALS AND METHODS Tissue preparation Surgically resected LCCs of the lung ( n = 10) were collected fresh from the operating theatre of the Brompton Hospital and four pieces from each tumour were taken and fixed or frozen immediately. One was fixed in 10 per cent formalin for conventional histology and immunocytochemistry; the second was fixed in sodium cacodylate buffered 2.5 per cent glutaraldehyde for electron microscopy; the third was fixed in 4 per cent paraformaldehyde for in situ hybridization; and the fourth was frozen in liquid nitrogen for northern blot analysis. Conventional light and electron microscop). Haematoxylin and eosin stained paraffin sections of each tumour were examined. For electron microscopy the 4 1 8 h glutaraldehyde-fixed tissues were rinsed in buffer. post-fixed in osmium tetroxide, embedded in araldite. and ultrathin sections stained with uranium and lead after selecting suitable areas from 1 um thick toluidine blue-stained sections. II?in~unoc?.tochemistr?. Tumour tissue from each case was fixed for 12 h in 10 per cent formalin. embedded in paraffin. and 5 pm thick sections were stained with antibody to GRP,’bombesin ( 1 i2000 dilution, Cambridge Research Biochemicals. Cambridge. U.K.) using the peroxidase-antiperoxidase method of Sternberger PI Sections from human fetal lung were used as a positive control.
Probe construction Recombinations containing sequences corresponding to G R P were isolated as previously
For in situ hybridization the tissue was fixed in freshly prepared 4 per cent paraformaldehyde for 4 h and then washed in phosphate-buffered saline (PBS) containing 15 per cent sucrose. and cryostat blocks were prepared. Sections were cut at a thickness of 15 pm, mounted on poly-L-lysine-coated slides, and allowed to dry overnight at 37°C. For hybridization the preparations were permeabilized by soaking in the following solutions: 0.1 mol/l glycine/PBS ( 5 min), 0.3 per cent Triton X-100 in PBS ( 1 5 min), 1 ,ug/ml proteinase K, 0.1 mol/l Tris, and 50 mmol/l ethylenediamine tetraacetic acid (EDTA)solution (20min at 37°C). Thesections were then treated with 4 per cent paraformaldehyde for 5 rnin to stop the reaction and immersed in 0.25 per cent acetic anhydride in 0.1 molil triethanolamine solution to minimize non-specific background binding. Prehybridization was carried out with 50 per cent formamide in 2 x standard saline citrate (2 x SSC) for 30 rnin at 37°C. Hybridization was carried out with 6 ng probeisection (approximately lo6 cpm/section) diluted in a buffer containing 50 per cent formamide, 2 x SSC, 10 per cent dextran sulphate, 0.25 per cent bovine serum albumin, 0.25 per cent Ficol 400, 0.25 per cent polyvinyl pyrrolidone 300, 250mmol/l Tris, 0.5 per cent sodium pyrophosphate. and 0.5 per cent sodium dodecyl sulphate and 250 pg/ml denatured salmon sperm DNA. Hybridization was carried out overnight in a moist chamber at 42°C. After hybridization the sections were rinsed in decreasing concentrations of SSC (4xSSC-0.5xSSC). 30 rnin each at 40°C. Non-specifically bound, single-stranded probe was removed by treatment with a solution containing 20 pg/ml RNase A. 0.5 molil sodium chloride, 10 mmol/l Tris, and 1 mmol/l EDTA for 30 min at 40°C. After dehydrating the sections in graded alcohols, they were dipped in Ilford K5 emulsion,
147
BOMBESIN m R N A IN L A R G E C E L L C A N C E R
Table I-Electron microscopy. immunocytochemistry. in situ hybridization. and northern blot analysis of ten large cell carcinomas of the lung G R P mRNA localization Case No. 1
2 3 4 5
6 7 8 9 10 t
Ultrastructural differentiation
G RP Iborn besin immunoreactivity
Adenosquamous Adenosquamous Glandular Glandular None Squamous Clara cell Glandular Mixed Clara and type I1 cell Squamous
by in situ hybridization
by northern blotting
+ +
NP
+ +
+
+ + -
+-
NP
+-
NP
+
NP NP
=positive: - =negative: N P = n o t performed.
dried, and exposed for 3 days at 4°C before being developed. GRP mRNA hybridization was controlled positively by hybridizing the 3’P-labelled GRP cRNA probe with sections of a small cell carcinoma of the lung known for its high expression of G R P mRNA by northern blotting. Negative controls were provided by separate sets of sections of LCC being hybridized with probe identical to the coding strand of mRNA of GRP (sense probe). Northern blot ana/ysis Total RNA was purified from five cases of LCC (Table I) by the guanidinium isothiocyanate method.” The RNA was subjected to electrophoresis on 1 per cent agarose gel containing 2.2 M formaldehyde in 50 per cent formamide, transferred nitrocefiulose~ and hybridized with 106cpmOf 7’P-labelled GRP cRNA p r ~ b e . ’ ~
RESULTS Histologically, the LCCs were all composed of large cells with abundant cytoplasm, open nuclei, and prominent nucleoli, and there were no features suggesting squamous, glandular, or neuroendocrine differentiation (Fig. I). Table I shows the ultrastructural features: there was evidence of glandular, squamous, or adenosquamous differentiation in most, but no neurosecretory granules were identified (Figs 2 and 3).
Fig. I-Large cell undifferentiated carcinoma of the lung characterized by tumour cells having abundant cytoplasm and nuclei with an open chromatin pattern and prominent nucleoli
None of the LCCs showed any degree of immunoreactivity to GRPibombesin antibody in contrast to thestrongimmunoreactivity observed in the positive control neuroendocrine cells of fetal lung. Hybridization between radio-labelled cRNA probes and mRNA encoding GRP was demonstrated by specific deposits of silver grains in the emulsion overlying tumour cells. Using 3’P-labelled GRP cRNA probe, an intense selective labelling was observed over the cytoplasm of tumour cells in seven of the ten LCCs investigated (Table I). The
148
Q.A. HAMID E T A L .
Fig. 2-Large cell undifferentiated carcinoma showing prominent tonofilaments, ultrastructural evidence of squamous differentiation. Elsewhere the same tumour also showed ultrastructural evidence of glandular differentiation (see Fig. 3). EM
Fig. 3-The Same large cell undifferentiated carcinoma as Fig. 2. here showing m,crovill, into a small lumen. ultrastructural evidence ofglandular differentiation,EM 17 000
x 170OO
tumour cells which showed positive hybridization were either arranged in groups or scattered singly among other tumour cells (Fig. 4). Strong positive hybridization was detected in the positive control small cell carcinoma. Only background levels of silver grains were detected in all preparations treated with GRP mRNA (sense probe) (Fig. 5). Northern blot analysis confirmed the specificity of the signal and the probe. Single bands corresponding to G R P mRNA were detected among the total RNA isolated in three or five LCCs (Fig. 6 ) . DISCUSSION In this study we have used immunocytochemical and molecular biological techniques to demonstrate GRPand itsmRNA in LCC. The results provideevidence of the localizationof GRP gene transcriptsin seven often such tumours,asdemonstrated by in siru hybridization.
Fig. &Autoradiograph of a cryostat section of large cell undifferentiated carcinoma. In siru hybridization with a '?P-labelled CRNA probe coding for GRP mRNA. Note the cytoplasmic localization of GRP mRNA
149
BOMBESIS mRNA IN LARGE CELL CANCER
Fig. 5-Autoradiograph of a cryostat section of large cell crrcinoma showing in situ hybridization with a radio-labelled GRP sense (control) probe. Note the absence of specific signal
The term large cell carcinoma of the lung is applied to anaplastic tumours which by light microscopy show no attempt at maturation.” A although by electron microscopy many show squamous or glandular differentiation. or both.” ‘O Recently i t has been shown by electron microscopy3’ and immunocytochemistry” that a proportion of LCCs exhibit neuroendocrine features. It has been suggested that LCCs with neuroendocrine features are associated with a particularly poor prognosis.’2 The presence of neurosecretory granules in turnour cells represents an advanced stage of Fig. &Northern (RNA) blot analysis oftotal RNA froma large undifferentiated carcinoma (B). Hybridization was carried neuroendocrine differentiation and their demon- cell out with a ’2P-labelledcRNA probe coding for GRP mRNA. A stration at ultrastructural level is not always easy. similar but stronger hybridization signal was obtained with a On the other hand. the value of immunocyto- small cell carcinoma used as a positive control (A). Arrows chemistry in demonstrating neuroendocrine differ- indicate the site of GRP mRNA (approximately 900 base pairs) entiation of lung tumours is limited by the lack of a specific general neuroendocrine marker. The most widely used are neuron specific enolase and protein G R P has been linked to the growth of normal and gene peptide 9.5. but both have frequently been neoplastic pulmonary neuroendocrine cells and reported in non-neuroendocrine tumours.2‘.7i.32 shown to be an autocrine growth factor for small ~ only a marker Conversely. negative immunoreactivity for a par- cell carcinoma of the l ~ n g . ~It-is’ not ticular peptide or protein does not necessarily mean of neuroendocrine differentiation, but can also prothat it is not being synthesized. In siru hybridization vide valuable information on the biological activity is a more effective way of studying such biological of a tumour. i7.33-35 GRP is generally regarded as activity and differentiation of tumour cells. This is being specific for neuroendocrine tumours of the well illustrated in our study. as we were able to lung.3’ Although it has been reported in a minority demonstrate GRP mRNA but not GRP,’bombesin of LCCs,”.” we were unable to demonstrate GRP/ bombesin immunoreactivity in any of our ten cases. immunoreactivity in seven of ten LCCs. GRP is a peptide which has been reported in the However. we demonstrated G R P mRNA in seven lung almost exclusively in neuroendocrine cells and out of the ten, indicating the presence of G R P gene neuroendocrine tumours. with particularly high transcription and possible production of the levels being secreted by small cell c a r c i n ~ m a . ~ , corresponding ~.~ peptide. The discrepancy between
B
150
Q. A. HAMID ETAAL.
the results of immunocytochemistry and in situ differentiation by electron microscopy failed to give hybridization could be explained by any of the fol- a hybridization signal. This observation is suplowing. (1) GRP may be produced but rapidly ported by the demonstration of GRP mRNA in released, so that although the gene is active its prod- cases of very poorly differentiated adenocarcinoma uct cannot be detected within the cells. (2) The GRP (unpublished observations), indicating that adenogene transcript might produce other fragments of carcinoma of the lung might also have a tendency the preprobombesin molecule that cannot be toward neuroendocrine differentiation and GRP detected by the antibody used. This has been shown production. In conclusion, we have shown for the to be the case in small cell carcinoma of the lung, first time that the gene for GRP production is where the C-flankingpeptide of GRP is expressed at actively expressed in LCC.This may be relevant to a higher level than GRP it~e1f.I~ Antibodies to dif- the aggressive behaviour and poor prognosis of this ferent fragments of the preprohormone molecule tumour. might be better for investigatingpeptide gene products in malignant turn our^.^^ (3) There may be tranREFERENCES scription of the gene but blockage or abnormal translation of the mRNA. 1 McDonald TH, Jornvall H, Ghatei M, Bloom SR,Mutt V. CharacterOnly seven of our ten LCCs were positive for ization of an avian gastric (proventricular) peptide having sequence GRP mRNA. The absence of an in situ hybridizhomology with the porcine gastrin-releasing peptide and the amphibian peptides bornbesin and alytain. FEBS Lett 1979; 122 ation signal in the remaining three cases might indicate that these tumours were at different stages of 1 4548. McGregor G, Woodhams P. OShaughnessy D, Ghatei M, Polak JM. differentiation or growth and there was no stimulus Bloom SR. Developmental changes in bombesin. substance P. somatostatin and vasoactive intestinal polypeptide in the rat brain. to initiate transcription of the GRP gene. However, Left 1982;28: 21-27 limited sensitivity of in siru hybridization cannot be 3 Neurosci Dockray GJ, Vaillant C, Walsh, JH. The neuronal origin of bombesinexcluded, especially in the presence of low copies of like immunoreactivity in the rat gastrointestinal tract. Neuroendocrinot o g ~19794 1561-1568. the message. Wharton J, Polak JM, Bloom SR. e f ul. Bombesin-like immunoreacTo ensure that the in situ hybridization signal was 4 tivity in the lung. Nulure 1978;273: 769-770. specific we used radio-labelled cRNA probes, which 5 CUUE, Chan W, Track NS. Bombcsin, calcitonin and leuenkephalin immunoreactivity in endocrine cells of the human lung. Experienfiu are the most specific and We also used 1981;37: 765-767. a number of control procedures including the use of 6 Price J. Nieuwenhuijzen Kruxman AC, Doniach I. Howlett TA, a sense probe (a probe which has an identical Besscr GM, Rees LH. Bornbesin-like peptides in human endocrine tumors: quantitation, biochemical characterization. and secretion. J sequenceto GRP mRNA) and RNase pretreatment. C/in EndocrinoiMefub 1985:60:1097-1 103. The absence of signals following these procedures is 7 Spindcl ER. Sunday ME. Hoefler H, Wolfe HJ, Habner JF, Chin evidence of the specificity of the reaction. To ensure WW. Transient elevation of mRNAs encoding gastrin-releasing pep tide (GRP), a putative pulmonary growth factor, in human fetal lung. probe specificity,the probe was applied to a smallcell JCCIinInvesr 1987;80: 1172-1179. carcinoma known for its GRP mRNA expression. 8 Erisman MD, Linnoila R1, Hernandez 0, Diaugustine RP, Lazarus LH. Human lung small cell carcinoma contains bombesin. Proc Nut/ The strong positive signal obtained here provided USA1982;7% 2379-2383. further support for the specificity of the hybridiz- 9 AcOdSci Moody TW, Pert CB, Gazdar AF. Carney DN, Minna JD. High levels ation. The specificity of the probe and the hybridizof intracellular bombesin characterize human small cell carcinoma. Science 1981;214124&1248. ation signal were further confirmed by northern blot VE. Linnoila R-I, Memoli V, Warren WH. Neuroendocrine analysis, which showed mRNA that corresponded 10 Could components of the bronchopulmonary tract: hyperplasia, dysplasias, to GRP in three of the five cases investigated by this and neoplasms. Lab Invest 1983:4% 519-537. technique. The negative results obtained in the other 11 Could VE. Linnoila R-I, Memoli VA. Warren WH. Neuroendocrine cells and neuroendocrine neoplasms of the lung.Puthol Annu 1983;18: two cases positive by in situ hybridization might be 287-291. due to dilution of GRP mRNA in the tumour’s total 12 Warren WH, Memoli VA, Could VE. Immunohistochemical and ultrastructural analysis of bronchopulmonary neuroendocrine RNA. Using northern blotting, the level of GRP UIlrusfruct Pufhol 1984;7:185-199. mRNA in LCCs appears to be less than in small cell 13 neoplasms. Cuttitta F, Carney DN, Mulshine I, et ul. Bombesin-like peptide can carcinomas. However, it was noteworthy that the function as autocrine growth factor in human small cell cancer. Nufure 1985; 316 823-826. pattern of mRNA distribution at the insitu level was 14 Weber S.Zuckerman JE,Bostwick DG, Bensch KG, Sikic BI. Raffin similar, suggesting that GRP played a similar role in TA. Gastrin-releasinn Deptide is a selective mitoeen for small cell the two tumours. It may be relevant that most of carcinoma in v i m . J 5th h w s f 1985;75 306309.the cases that showed positive hybridization also 15. Willey JC, Lechner JR, Hams CC. Bombesin and the C-terminal tetradecapeptide of gastrin-releasing peptide are growth factors for showed glandular differentiation, as seen in adenonormal human bronchial cpitheiial cells. Exp Cell Res 1984; 153 carcinomas. The two cases that showed squamous 245-248.
BOMBESIN mRNA IN LARGE CELL CANCER 16. Spindel ER,Zilberberg MD, Habner JF, Chin WW .Two prohornones
forgastrin-reledsingpeptideareencodedby twomRNAdiffering by 19 nucleotides. f r o c N o d AcodSci U S A 1986: 83: 19-23. 17. Hamid Q . Addis BJ. Springall DR. er al. Expression of the C-terminal peptide of human pro-bomksin in 361 lung endocrine turnours. a reliable marker and possible prognostic indicator for small cell carcinoma. Virchows Arch A 1987;411: 185-192. 18. Hamid Q, Bishop A. Springall D, er al. Detection of human probornbesin mRNA in neuroendocrine (small cell) carcinoma of the lung. Cancer 1989;63 266-271. 19. Edinburgh Lung Cancer Group. Patients presenting with lung cancer in South East Scotland. Thorox 1987: 4 2 853-857. 20. Huhti E.Sutinen S, Saloheimo M. Survival among patients with lung cancer. An epidemiologic study. Am Rev Respir Dis 1981; 124: 13-16. 2 I . Said JW. Vimadakdl S, Nash G, er a / . lmmunoreactive neuron-specific enolase. bornbesin, and chromogranin as markers for neuroendocrine lung turnours. Hum fafhol1985:1 6 2 3 6 2 4 0 . 22. Piehl MR, Gould VE, Warren WH, er al. lmmunohistochemical identification ofexocrine and neuroendocrine subsets of largecell lung carcinomas. farhol Res f r a c f 1988; 183: 675-5682. 23. Sternberger LA, Hardy PH Jr, Cuchlin JJ, Meyer HG. The unlabelled antibody enzyme method of immunocytochemistry. Preparation and properties of soluble antigen-antibody complex (horseradish peroxidastantiperoxidase) and its use in the identification of spirochetes. J Hisrochem Cyrochem 1970; 1 8 31 5-324. 24. Hamid Q. Wharton J , Terenghi G , er al. Localisation of arterial natriuretic peptide mRNA and immunoreactivity in rat heart and humanatrialappendage. f r o c NadAcadSciUSA 1987;84:6760-6764. 25. Chirgwin JM, Prrybyla AE. MacDonald RJ. Rutter WJ. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemisrry 1979; 18 5294-5299. 26. World Health Organization. International Histological Classification of Tumours. No. 1. Histological Typing of Lung Tumours. 2nd ed. Geneva: WHO, 1981. 27. Churg A. The fine structure of large cell undifferentiated carcinoma of the lung. Evidence for its relation to squamous cell carcinomas and adenocarcinomas. Hum farhol1978; 9:143-156.
151
28. Hone A, Ohta M. Ultrastructural features of large cell carcinoma of the lung with reference to the prognosis ofpatients. Hum Parhol1981; 1 2 423-432. 29. Albain KS, True LD, Golomb HM. Hoffman PC, Little AG. Large cell carcinoma of the lung. Ultrastructural differentiation and clinicopathologiccorrelations. Cancer 1985; 56: 1618-1623. 30. Kodama T , Shimosato Y,Koide T, Watanabe S, Teshima S. Large cell carcinoma of the lung-ultrastructural and immunohistochemical studies. Jpn JClin Oncoll985; 15: 431-441. 31. Rode J, Dhillon AP, Doran JF. Jackson P, Thompson RJ. PGP9.5-a new marker for human neuroendocrine tumours. Hisloparhology 1985: 9: 147-158. 32. Addis BJ, Hamid Q,lbrahim NBN, Fahey M, Bloom SR, Polak JM. lmmunohistochemical markers of small cell carcinoma and related neuroendocrine tumours of the lung. J fafhol1987; 153 137-150. 33. Roth KA, Evans CJ, Weber E. Barchas JD, Bostwick DG, Bensch KG. Gastrin-releasingpeptide-related peptides in a human malignant lungcarcinoid tumor. Cancer Res 1983; 43: 541 1-5415. 34. Carney DN, Gazdar AF, Bepler G, ef a / . Establishment and identification of small cell lung cancer cell lines having classic and variant features. Cancer Re5 1985; 4 5 2913-2923. 35. Polak JM, Bloom SR. Theneuroendocrine system and its turnours. In: Frank LM, Teich NMC, eds. Introduction to the Cellular and Molecular Biology of Cancer. Oxford: Oxford University Press, 1986; 327-331. 36. Hamid Q, Bishop A, Sikri K, Varndell I, Bloom SR, Polak JM. Immunocytochemical characterization of 10 pancreatic turnours, associated with the glucagonoma syndrome, using antibodies to separate regions of the proglucagon molecule and other neuroendocrine markers. Hisropafhology 1 9 8 6 10: 119-133. 37. CoxKH,DeLeonDV,AngererLM,AngererRC. DetectionofmRNA in sea urchin embryos by in sifuhybridisation using asymmetric RNA probe. Dev E d 1984; 101:485-502. 38. Hoefler H,Childers H, Montminy MR, k h a n RM, Goodman RH. Wolfe HJ. 'In sifu' hybridisation methods for the detection of somatostatin mRNA in tissue sections using antisense RNA probes. Hisrochem J 1986; I 8 597-604.
