J Clin Pathol 1984;37:931-936

Amine and peptide hormone production by lungcarcinoid: a clinicopathological andimmunocytochemical studyFRED T BOSMAN,* A BRUTEL DE LA RIVIERE,t RWM GIARD4 AAJ VERHOFSTAD,§G CRAMER-KNIJNENBURGt

From the *Department ofPathology, University ofLimburg, the tDepartment of Cardiothoracic Surgery,University ofLeiden, the tDepartment ofPathology, University ofLeiden, and the §Department ofAnatomyand Embryology, University ofNijmegen, The Netherlands

SUMMARY A consecutive series of 38 lung carcinoid tumours (36 surgical and two necropsyspecimens) was studied. Histopathological features and amine and peptide hormoneimmunoreactivity were correlated with gross characteristics (size, location) and clinical data.Peripheral carcinoids were detected a decade later than central carcinoids and tended to bebigger. In general, the histological characteristics of peripheral and central carcinoids were simi-lar; atypical features, however, were more common in peripheral carcinoids.Most carcinoids contained many argyrophilic cells (58%). Although argentaffinic cells were

not found, serotonin immunoreactive cells were present in 32% of the tumours. Peptide hormoneimmunoreactivity (adrenocorticotrophic hormone (ACTH), calcitonin, somatostatin, gastrin)was rare. In one case massive ACTH production had caused clinically manifest Cushing's syn-drome. In two other cases few ACTH immunoreactive cells were found and in one case calcitoninimmunoreactive cells were present.The relative rarity of hormone production in lung carcinoids and the predominantly benign

course of the tumour preclude the use of peptide hormone production as a prognostic indicator.

Carcinoid tumours are relatively rare in the lung,comprising less than 1% of all pulmonary neo-plasms.' Most of these tumours arise in the mainbronchi, but carcinoids may also occur in theperiphery of the lung. Central (bronchial) carcinoidsusually show architectural and cytonuclear featurestypical of these neoplasms.2 Peripheral carcinoids,however, regularly show more anaplastic featuressuch as increased nuclear pleomorphism and hyper-chromasia, increased mitotic activity, and architec-tural disorganisation.3 Furthermore, spindle cellvariants occur more often. These atypical carcinoidshave been reported to metastatise more fre-quently.4 5 Based on these findings atypical car-cinoids have been tentatively designated as inter-mediate between typical carcinoids and theneuroendocrine type of small cell carcinoma.6-8Although carcinoid tumours are regarded as

neuroendocrine tumours,9 their association withclinically evident endocrine syndromes is relatively

Accepted for publication 25 April 1984

rare.'0 Carcinoid syndrome due to serotonin produc-tion in a bronchial carcinoid is extremely rare.Immunocytochemical evidence of peptide hormoneproduction has been reported both with'" and with-out clinically evident endocrine syndromes."I- ' Sys-tematic studies of production of amines and peptidehormones by bronchial carcinoids are relativelyscarce.'4 We have therefore investigated the occur-rence of serotonin and a variety of neurohormonalpeptides in a series of bronchial carcinoids usingindirect immunoperoxidase and immunofluores-cence techniques. The histochemical results werecorrelated with gross (location, size) and micros-copic (architecture, cytonuclear features) charac-teristics. The results indicate that peptide hormoneproduction is relatively uncommon, but serotoninproduction (without clinically evident carcinoid syn-drome) occurs much more often.

Material and methods

From the files of the department of pathology 36931

on June 15, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.8.931 on 1 A

ugust 1984. Dow

nloaded from

http://jcp.bmj.com/

Bosman, Riviere, Giard, Verhofstad, Cramer-Knijnenburg

specimens of carcinoid lung tumour obtained duringsurgery (Department of Thoracic Surgery, Univer-sity Medical Centre, Leiden) and two specimensfound incidentally at necropsy were collected (cov-ering the period 1956 to 1981). All tissue specimenshad been fixed in 4% formaldehyde (neutral andbuffered after 1971) and embedded in paraffin. Ofthe available material (usually only one or twoblocks were available) one representative block wasselected for histochemical studies.From each block serial 4 ,tm sections were cut.

Sections were stained for argentaffin cells accordingto Solcia et al" and for argyrophil cells accord-ing to Grimelius.'6 The presence of peptidehormone immunoreactivity was investigated byimmunocytochemistry using an indirect peroxidaselabelled antibody method.'" Rabbit antisera toadrenocorticotrophic hormone (ACTH), calcitonin,somatostatin, and gastrin were used. Specificity ofthese antisera was tested in a solid phase enzymeimmunoassay,'7 8 and all antisera were found to bemonospecific. Controls for the immunocytochemicalstaining included replacement of the primary anti-body by non-immune rabbit serum or by immuneserum preabsorbed with the appropriate peptidehormone. In these control sections staining wasnever seen.

Serotonin immunoreactivity was tested with indi-rect immunofluorescence. Specificity of the rabbitantiserotonin antibody and details of the immuno-staining procedure and control procedures have allbeen described previously.'9 The presence ofimmunoreactivity for any one of the above men-tioned substances was correlated with size, location,and histological appearances. Histological evalua-tion was performed by two of us and included semi-quantitative scoring for the presence of infiltrativegrowth, growth pattern (nodular, trabecular, acinar,or non-distinct), nuclear pleomorphism, and theoccurrence of mitoses. In addition the presence ofcalcification or ossification, or both, and of amyloidwas noted.

Statistical analysis of the data was performedusing Wilcoxon's test or the X2 test.

Results

Two centrally located bronchial carcinoids wereincidental necropsy findings. In one of these patientsthe cause of death was not related to the carcinoid.The other patient presented with Cushinges syn-drome and died of postoperative complications aftertrans-sphenoidal resection of a presumed ACTHproducing pituitary adenoma. The Cushing's syn-drome, however, appeared to be caused by ectopicproduction of ACTH by the bronchial carcinoid,

Table 1 Type ofsurgical resection

Enucleation 1Segmentectomy 5Lobectomy 1 5Bilobectomy 3Pneumectomy 4Lobectomy with sleeve resection 2Sleeve resection only 4

Table 2 Location, age and sex distribution, and size ofsurgically resected lung carcinoid tumours

Peripheral Central

Total 19 15Men 10(53%) 3 (20%)*Women 9 (47%) 12 (80%)*

Mean age (yr) (range) 50-1 (20-75) 41-6 (17-65)tMedian age (yr) 56 44Size (cm) 1-9 (0-3-4) 1-6 (0-7-25)t

*Difference not significant (X2 test).tp = 005 (Wilcoxon's test).tRange.which, owing to its small size, had not been detected.

Pathological specimens were available from 38patients: 23 women (61%) and 15 men (39%) witha mean age of 43 years (women 40 years, men 47years). For 34 patients clinical data were available:24 (70%) presented with respiratory symptomsrelated to the tumour. Carcinoid syndrome had notoccurred in any of the patients, although twopatients had raised urinary concentrations of5-hydroxyindoleacetic acid. In one patient a carotidartery chemodectoma had been resected previously.

Table 1 lists the surgical procedures used. Lobec-tomy was the most frequently chosen operative pro-cedure (45% of patients). Clinical and macroscopi-cal characteristics of the tumours are given in Table2. Peripheral tumours were found slightly moreoften than central tumours (19 v 15). Peripheraltumours occurred equally often in men and women( 10 v 9). Central tumours, however, were four timesmore common in women (12 v 3). Peripheraltumours tended to be slightly bigger than centraltumours 1.9 v 1*6 cm; this difference, however,was not significant. Peripheral tumours occurred atlater age than central tumours.

Microscopically, most tumours were fairly cir-cumscript, and sometimes surrounded by a pseudo-capsule. Invasion into surrounding structures wasnot found more often in peripheral tumours than incentral tumours. Most carcinoid tumours showed amixture of two or more of the characteristic growthpatterns (nodular, trabecular, acinar, or non-distinct). Nuclear pleomorphism was usually slight;in six tumours, all peripheral, moderate nuclearpleomorphism was present. A noticeable increase inmitotic activity was found in only two tumours, bothperipheral. Tumours composed predominantly of

932

on June 15, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.8.931 on 1 A

ugust 1984. Dow

nloaded from

http://jcp.bmj.com/

Functional activity of bronchial carcinoids

spindle cells were not encountered in this series.Calcification was found in five tumours, in four incombination with osseous metaplasia (Fig. 1). Allbut one of these tumours were peripheral. Amyloidwas identified (congo red positive with green bi-refringence) in six tumours (three central, threeperipheral).

In 28 cases sufficient material was available foradditional histochemical and immunohistochemicalstaining; the findings are listed in Table 3. None ofthe tumours contained argentaffin positive cells. Inmost tumours (22) argyrophilic cells were found.Occasionally, most of the tumour cells wereargyrophilic. More frequently, however,argyrophilia was only focal (Fig. 2), and in sometumours a few argyrophilic cells were found scat-tered between negative cells. Argyrophilic tumourstended to occur more often in central than inperipheral tumours (p < 0.06). No correlation wasfound between growth pattern and argyrophilia.Scattered serotonin immunoreactive cells werefound in 12 tumours (Fig. 3). Serotonin tended to

933

Table 3 Histochemical characteristics of28 lung carcinoidtumours

Peripheral Central Total

n % n %

Argentaffin +ve 0 0 0Argyrophil +ve 10 67 12 92 22*Serotonin +ve 5 33 7 54 12tACTH +ve 2 13 1 8 3Calcitonin +ve 1 7 0 1Somatostatin +ve 0 0 0Gastrin +ve 0 0 0Total 15 13 28

*p = 0-06 (XI test).tNot significant (XI test).+ve = positive.

occur more often in central (54%) than inperipheral (33%) carcinoids, although this differ-ence was not significant. No correlation was foundbetween argyrophilia and serotonin immunoreactiv-ity. Immunoreactive peptide hormones were rarelyencountered: ACTH immunoreactivity occurred intwo peripheral and one central carcinoids and cal-

aW,*. W,Ab~~~~.,*2s*w#i 4 41 *,t

~~~*~At'%-,1X

~~ 4 V~~.

,tiw;Bf,.) , *~ .s . f - W A

'.-s...Z9t\X-%.wj2¢*wffl}.ei*_*-

Fig.1~ 'o feosin. Original magnifin x. (

x250.(c) A

41,41 * 4 .1t

~~~4~~~~4'~~~~~~ ~~~~ 41 4~~~~~~'1.4~ ~ 4l

'5' ., 4.5~~~~~~~~~~~~~1FigHisologcalfeatresf lug crcinid tmour. () Noularand rabcula groth ptten. Hematxyli aneoiriiamgiicto x0 b)Sm tmu, youlerdtalHemtxli n esn Oiialmgifcto250()Foclosseus meaplasi. Haeatoxyln andeosin.Origial maniiato x0

on June 15, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.8.931 on 1 A

ugust 1984. Dow

nloaded from

http://jcp.bmj.com/

Bosman, Riviere, Giard, Verhofstad, Cramer-Knijnenburg

44

Fig. 2 Argyrophil positive cells in a lung carcinoidtumour. Note intercellular heterogeneity in silver reactivity.Grimelius. Original magnification X400.

citonin immunoreactivity in one peripheral car-cinoid. In the ACTH producing carcinoid with clini-cally evident Cushinges syndrome the majority oftumour cells showed immunoreactivity (Fig. 4). Inthe other ACTH reactive carcinoid and in the cal-citonin reactive carcinoid a few immunoreactivecells were scattered among negative tumour cells(Fig. 5). The limited number of peptide hormoneproducing tumours precluded an analysis of a poss-ible correlation between growth pattern and peptidehormnone production.

Discussion

Neuroendocrine cells are found in the mucosa of thehuman respiratory tract.""2 In these cells produc-tion of serotonin, calcitonin, bombesin, and leu-enkephalin has been documented.22 It is generallybelieved that lung carcinoids, small cell tumours,neuroendocrine carcinomas, and presumably alsoperipheral pulmonary tumourlets derive from thesecells.'13Based on clinical and histopathological charac-

teristics some investigators626 have distinguishedbetween typical carcinoids, which comprise the neo-plasms with the morphological characteristics classi-cally ascribed to carcinoids, and atypical carcinoids.

Fig. 3 Serotonin immunoreactivity in lung carcinoid. Onlyscattered cells are stained. Indirect immunofluorescence,antiserotonin. Original magnification x250.s; t *. W, 4

V ~ ~ ~ w

*e;t_ - = * @~~Ak

4~* .

*. * - ,E * *W

4 ~~~~~~bVftFig. 4 ACTH immunoreactivity in a lung carcinoid withclinical evidence of Cushing's syndrome. Note theoccurrence ofa few intensely stained cells. Indirectimmunoperoxidase staining, anti-ACTH. Originalmagnification x250.

934

on June 15, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.8.931 on 1 A

ugust 1984. Dow

nloaded from

http://jcp.bmj.com/

Functional activity of bronchial carcinoids

Fig. 5 Scattered calcitonin immuno

carcinoid. Indirect immunoperoxida.

Original magnification x250.

In this latter category the neopldominantly solid (nodular or tratern, foci of necrosis, markedlhyperchromatic (occasionally sjlei, and numerous mitoses. The ]carcinoids is significantly worsecarcinoids.6 Based on these findthat atypical carcinoids might cof neoplasms intermediate becinoids and small cell carcinom

Since atypical carcinoids apmore often in the periphery ofspecial attention to possible (central and peripheral carcinoi(found a slight overall preponiwhich was caused solely by the icinoids occurred four times m(compared with men. With regarperipheral carcinoids were detelater than central carcinoids (myears). Central carcinoids, ipolypoid endobronchial growth.ory symptoms relatively early abetter chance of early detectioncinoids. The finding that centrabe smaller than peripheral carcthis assumption. Overall archit

clear features of peripheral and central carcinoidswere remarkably similar, but we did find several dif-ferences. Firstly, atypical nuclear features occurredonly in peripheral carcinoids (nuclear pleomorphismin six and increased mitotic activity in two of 19tumours). Secondly, calcification occurred moreoften in peripheral than in central carcinoids and,finally, argyrophilia or serotonin immunoreactivity,or both, tended to occur more often in peripheralthan in central carcinoids.The studies of Ranchod and Levine- on spindle

cell carcinoids and of Mills et a6 on atypical car-cinoids suggest that a majority of peripheral pulmo-nary carcinoid tumours show these features. Ourfindings indicate that, although atypical featuressuch as nuclear pleomorphism and increased mitoticactivity do occur more frequently in peripheral thanin central pulmonary carcinoids, the majority ofperipheral carcinoid tumours seem to be morpho-logically indistinguishable from central carcinoidtumours. Whether or not the location of pulmonarycarcinoids is relevant for prognosis remains uncer-tain and needs to be determined in a large series

in a lung with long follow up.oreactivity in a lung In the present series serotonin immunoreactivityse, anticalcitonin.occurred relatively frequently, as was reportedpreviously'4 despite the fact that none of thetumours contained argentaffin positive cells.Immunocytochemistry appears to be more sensitive

lasms showed a pre- for the detection of serotonin than argentaffin stain-Lbecular) growth pat- ing.27 The number of serotonin immunoreactive cellsly pleomorphic and was usually rather low, which may explain why thepindle shaped) nuc- carcinoid syndrome does not occur in pulmonaryprognosis of atypical carcinoids. Peptide hormone immunoreactivity wasthan that of typical rarely encountered, although testing for a wider

lings it was proposed range of neurohormonal peptides might haveonstitute a category revealed more hormonally active tumours. Indeed,-tween typical car- recent reports'3 14 indicate that neurohormonal pep-a.6 20 tides such as bombesin and pancreatic polypeptideopear to be located may be detected more frequently. Our findings withthe lung,56 we paid regard to calcitonin immunoreactivity are in dis-differences between agreement with those of Cooney et al, " who detectedds in our series. We calcitonin (using our antibody) in six of 22 pulmo-derance in women, nary carcinoids, four of which also showed foci offact that central car- ossification. Since in the present study ossificationore often in women did not occur together with calcitonin production wed to age distribution, believe that a causal relation as postulated bycted about a decade Cooney et al" is unlikely. ACTH production, whichlean age 50-1 v 41-6 may rarely cause Cushings syndrome,'" appears towhich often show be equally rare in pulmonary carcinoids. The paucity, may cause respirat- of peptide hormone immunoreactivity in these neo-nd therefore stand a plasms precludes conclusions regarding possiblethan peripheral car- relations between hormonal activity and long termil carcinoids tend to prognosis. Considering the predominantly benigninoids is in line with course of these neoplasms,' 1' however, it is unlikelytectural and cytonu- that neurohormonal production will appear to be an

935

on June 15, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.8.931 on 1 A

ugust 1984. Dow

nloaded from

http://jcp.bmj.com/

Bosman, Riviere, Giard, Verhofstad, Cramer-Knijnenburgimportant prognostic factor.

In conclusion, we have shown that althoughperipheral and central pulmonary carcinoids areremarkably similar, atypical histological features tendto occur predominantly in peripheral tumours. Pul-monary carcinoids often produce serotonin, butpeptide hormone immunoreactivity is relativelyuncommon.

References

'Godwin RA. Carcinoid tumors. Cancer 1975;36:56-9.2 Jones RA, Dawson IMP. Morphology and staining patterns of

endocrine cell tumours in the gut, pancreas and bronchus andtheir possible significance. Histopathology 1977; 1:137-50.

Bonikos DS, Bensch KG, Janiplis RW. Peripheral pulmonarycarcinoid tumors. Cancer 1976;37: 1469-77.

4 Churg A. Large spindle cell variant of peripheral bronchial car-cinoid tumor. Arch Pathol Lab Med 1977;101:216-8.

5 Ranchod M, Levine GD. Spindle cell carcinoid tumors of thelung. A clinicopathological study of 35 cases. Am J Surg Pathol1980;4:315-3 1.

6 Mills SE, Walker AN, Copper PH, Kron IL. Atypical carcinoidtumor of the lung. A clinicopathological study of 17 cases. AmJ Surg Pathol 1982;6:643-54.

7Fisher ER, Palekar A, Paulson JD. Comparative histopathologi-cal, histochemical, electron microscopic and tissue culturestudies of bronchial carcinoids and oat cell carcinomas of lung.J Clin Pathol 1978;69:165-72.

8 McDowell EM, Barrett LA, Trump BF. Observations on smallgranule cells in adult human bronchial epithelium and in car-cinoid and oat cell tumors. Lab Invest 1976;34:202-6.

9 Corrin B. Lung endocrine tumours. Invest Cell Pathol1980;3: 195-206.

Cohen RB, Toll GD, Castleman B. Bronchial adenomas in Cush-ing's syndrome, their relation to thymomas and oat cell car-cinomas associated with hyperadrenocorticison. Cancer1960; 13:812-7.

"Cooney T, Sweeney EC, Luke D. Pulmonary tumours: a com-parative regional study. J Clin Pathol 1979; 32:1100-9.

12Curz E, Chan W, Kay JM, Camberlain DW. Immunoperoxidasestaining for serotonin, bombesin, calcitonin and leu-enkephalin in pulmonary tumorlets, bronchial carcinoids andoat cell carcinomas. Lab Invest 1982;46: 16A.

3 Gould VE, Linnoila RI, Memoli V, Warren WH. Neuroendo-crine cells and neuroendocrine neoplasms of the lung. Pathol

Ann 1983;18:287-330.Yang K, Wick T, Taylor I, Cheng L, Lewin KJ. Pulmonary car-

cinoid. Immunohistochemical demonstration of brain-gut pep-tides. Cancer 1983;52:819-23.

Solcia E, Capella C, Vassallo G. Lead hematoxylin as a stain forendocrine cells. Significance of staining and comparison withother methods. Histochemie 1969;20: 116-26.

16 Grimelius L. A silver nitrate stain for a2-cells in human pancrea-tic islets. Acta Societatis Medicorum Upsalienses 1968; 73:243-70.

'' Bosman FT, Louwerens JWK. APUD cells in teratomas. Am JPathol 1981; 104:174-80.

18 Bosman Fr, Cramer-Knijnenburg G, Van Bergen-Henegouw J.Efficiency and sensitivity of indirect immunoperoxidasemethods. Histochemistry 1983; 77:185-94.

9 Verhofstad AAJ, Steinbusch HWM, Joosten HWJ, Penke B,Varga J, Goldstein M. Immunocytochemical localization ofnoradrenaline, adrenaline and serotonin. In: Polak JM, vanNoorden S, eds. Immunocytochemistry: Practical applicationsin pathology and biology. Bristol: Wright and Sons,1983:143-68.

20 Lauwerijns JM, Peuskens JC. Neuroepithelial bodies(neuroreceptor of secretory organs) in human infant bronchialand bronchiolar epithelium. Anat Rec 1972; 172:471-82.

21 Lauwerijns JM, Goddeerin P. Neuroepithelial bodies in thehuman child and adult lung. Am Rev Resp Dis 1975;111:469-76.

22 Lauwerijns JM, De Bock V, Verhofstad AAJ, Steinbusch HWM.Immunohistochemical localization of serotonin in intrapulmo-nary neuroepithelial bodies. Cell Tissue Res 1982; 226:215-23.

23 Becker KL, Monaghan KG, Silva 0. Immunocytochemical local-ization of calcitonin in Kulchitsky cells of the lung. Arch PatholLab Med 1980; 104:196-8.

24 Cutz E, Chan W, Track VS. Bombesin, calcitonin and leu-encephalin immunoreactivity in the human lung. Experienta198 1;37:765-7.

25 Wharton J, Polak JM, Bloom SR. Bombesin-like immunoreactiv-ity in the lung. Nature 1978;273:769-70.

26 Antigoni MG, Woolner LB, Beratz RE. Atypical carcinoidtumors of the lung. J Thorac Cardiovasc Surg 1972; 64:413-21.

27 Cuello AC, Wells C, Chaplin AJ, Milstein C. Serotoninimmunoreactivity in carcinoid tumours demonstrated by amonoclonal antibody. Lancet 1982;i:771-3.

Requests for reprints to: Dr FT Bosman, Department ofPathology, Biomedical Center, University of Limburg, POBox 616, 6200 MD Maastricht, The Netherlands.

936

on June 15, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.8.931 on 1 A

ugust 1984. Dow

nloaded from

http://jcp.bmj.com/