New Series N o 179- ISSN 0346-6612
From the Departm ents o f O torhinolaryngology - Head and Neck Surgery, H istology and Cell Biology, and Anatomy,
University o f U m eå, Um eå, Sweden
SALIVARY GLAND NEOPLASMS
Studies on the cytoskeleton, the secretory apparatus and the nuclear D NA content
by
Hans Gustafsson
Umeå 1986
SALIVARY GLAND NEOPLASMS
Studies on the c y toskele ton, the secretory apparatus and the nuclear DNA content
Akademisk avhandling som med v e d e rb ö rlig t t i l l s t å n d av Rektorsäm
b e te t vid Umeå uni v e r s it é t för avläggande av medicine doktorsexamen kommer a t t o ffe n tlig e n försvaras i Rosa salen, Tandläkarhögskolan, 9 t r , fredagen den 12 december 1986, kl 09.00.
av
HANS GUSTAFSSON
UMEÅ 1986
New S e r i e s No 179 - ISSN 0346-6612
Salivary gland neoplasms. Studies on the cytoskeleton, the secretory apparatus and the nuclear DNA content.
Hans Gustafsson, Departments of Otorhinolaryngology - Head and Neck Surgery, Histology and Cell Biology, and Anatomy, University of Umea, S- 901 85 Umeå, Sweden
ABSTRACT
The heterogeneity of s a liv a ry gland neoplasms have made c l a s s i f i c a t i o n and prognostication of these tumours sometimes d i f f i c u l t , and the i n troduction of techniques, such as enzyme and carbohydrate histochemis
t r y and elec tro n microscopy have only to a c e r ta in exten t increased our knowledge in these re s p e c ts . In the present study immunohistochemical methods have been used to i d e n tif y interm ediate filam ent p ro tein s (IFP) in normal f e ta l and a d u lt p a r o tid glands, as well as in sa liv a ry neo
plasms. The interm ediate filam ents (IF) make up the cytoskeleton in eucaryotic c e l l s . E p ith e lia l t i s s u e contains IF composed of d i f f e r e n t c y to k eratin s (CK 1-19) w h ils t mesenchymal t i s s u e generally contains IF composed of vimentin, and the IFP p attern is very s ta b le even during cell transform ation. I t would thus be p ossible to f u r th e r c l a r i f y the h isto g en esis of s a liv a ry neoplasms by id e n tif y in g IFP, in addition the IFP pattern would probably be useful in tumour typing. Furthermore, ul t r a s t r u c t u r a l cytochemical s tu d ie s , mi ero spec to r photometry on nuclear DNA as well as enzyme sec reto ry stu d ies of c e r ta in tumour types were c a r r ie d out, in order to f u r th e r c h a ra c te riz e the biology of sa liv a ry neoplasms.
The immunohistochemical i n v e s tig a tio n s showed th a t in normal p a ro tid t i s s u e , the d i f f e r e n t cell types d if f e r e d in IFP expression: ac in a r c e l l s express mainly CK 18 and myoepithelial c e l l s mainly CK 17 and 19, w h ilst duct c e l l s contained a broad range of CK. Vimentin could in ad
d itio n to CK be detected in myoepithelial c e l l s and basal c e l l s of ex
c r e to ry ducts. Fetal p a ro tid c e l l s showed a sim ila r CK pattern as mature duct c e l l s . In a d d itio n , vimentin could be found in some basal c e l l s of the terminal tubules of the f e ta l glands. S alivary neoplasms could be divided into three types with regard to t h e i r IFP p a tte rn : 1. Acinic cell carcinomas showed a CK-pattern sim ila r to normal acinar c e l l s but a co-expression of CK and vimentin was pre sent in some c e l l s . 2. Adenoid c y s tic carcinomas, mixed tumours and basal ce ll adenomas showed a CK-pattern of normal duct or myoepithelial c e l l s . The p e r i pheral c e l l s were a lso vimentin p o s itiv e . 3. Mucoepidermoid carcinomas and adenocarcinomas had a s im ila r CK-pattern as duct c e l l s , and no t u mour c e l l s contained vimentin. This in d ic a te s t h a t typing of IFP may be useful fo r subgrouping of s a liv a ry neoplasms.
By ste re o lo g ic a l measurements, the c e l l s of ac in ic ce ll carcinomas were found to be very s im ila r to normal p a r o tid acinar c e l l s . Furthermore, they contained amylase and a f t e r stim ulation by norepiphrine a s e c re tory response was induced, with a r i s e in i n t r a c e l l u l a r cAMP as well as a re le a se of amylase. By sin g le cell measurements of nuclear DNA con
t e n t , no diffe renc e was found between a c in ic ce ll carcinomas with de
f i n i t e m etastasis and those without recurrence, both in p a r a ffin sec
tio n s and cy tological smears.
Key words: Salivary gland neoplasms, interm ediate fila m e n ts, immuno- histochem istry, u l t r a s t r u c t u r e , DNA, development, h is to g e n e s is , progno
sis
UMEÂ UNIVERSITY MEDICAL DISSERTATIONS New S e r i e s No 179 - ISSN 0346-6612
From the Departments of Otorhinolaryngology - Head and Neck Surgery, Histology and Cell Biology, and Anatomy,
University of Umeå, Umeå, Sweden
SALIVARY GLAND NEOPLASMS
Studies on the cytoskeleton, the sec reto ry apparatus and the nuclear DNA content
by
HANS GUSTAFSSON
UMEÂ 1986
CONTENTS
ABBREVIATIONS ... 4
ABSTRACT ... 5
ORIGINAL PAPERS ... 6
INTRODUCTION ... 7
C la s s i f i c a t i o n of s a liv a ry neoplasms ... 8
S tru c tu re of the sa liv a ry glands ... 10
Tumour h isto g en esis ... 14
New approaches in c e ll b i o l o g y ... 15
OBJECTIVES OF THE STUDY ... 18
MATERIAL AND METHODS ... 19
Morphology (I-VI) ... 19
Imraunohistochemistry (I-V) ... 19
Immunoblots (IV) ... 20
Immunostaining (I-V) ... 20
Stereological methods (VI) ... 20
Studies on s e c re tio n and c y c lic AMP accumulation (VI) ... 21
Amylase and c y c lic AMP assay (VI) ... 21
DNA measurement (VII) ... 21
RESULTS 23 Cytoskeletal p ro te in s in normal s a liv a ry glands ( I-IV ) ... 23
Cytoskeleton of developing p a r o tid glands (IV) ... 23
Cytoskeleton of benign s a liv a r y neoplasms ( I , I I I , V) ... 23
Cytoskeleton of malignant s a liv a ry neoplasms ( I I , I I I , V) ... 24
Morphometry and sec reto ry behaviour of ACC (VI) ... 27
DNA le v e ls in a c in ic c e ll carcinomas (VII) ... 27
3
DISCUSSION ... 28
Intermediate filam ent p ro te in s in normal s a liv a ry glands ... 28
Intermediate filam ents in f e ta l p a r o tid glands ... 31
Intermediate filam ents in sa liv a ry gland neoplasms ... 32
Aspects of h isto g en esis as a basis fo r tumour t y p i n g ... 36
The stim u lu s-se cre to ry coupling in a c in ic ce ll cancer ... 37
Measurement of DNA content and determination of long-term prognosis ... 38
GENERAL SUMMARY AND CONCLUSIONS ... 40
ACKNOWLEDGEMENTS ... 43
REFERENCES ... 44
PAPERS I-VII
ABBREVIATIONS
PA - pleomorphic adenoma (mixed tumour) MA - monomorphic adenoma
ACC - ac in ic c e ll carcinoma MEC - mu coepidermoid carcinoma
CPA - carcinoma in pleomorphic adenoma (carcinoma in mixed tumours) AdCC - adenoid c y s tic carcinoma
SDC - s a liv a r y duct carcinoma Ab(s) - antibody (a ntibodies)
mAb(s) - monoclonal antibody (a n tib o d ies) pAb(s) - polyclonal antibody (a ntibodies) IF(s) - interm ediate filam ent (filam ents) IFP(s) - interm ediate filam ent pro tein (p ro te in s) CK(s) - c y t o k e r a t i n ( s ) , p r e k e r a tin ( s )
kD - kiloDalton
5
ABSTRACT
The heterogeneity of s a liv a r y gland neoplasms have made c l a s s i f i c a t i o n and prognostication of these tumours sometimes d i f f i c u l t , and the i n troduction of techniques, such as enzyme and carbohydrate histochemis
t r y and e lec tro n microscopy have only to a c e r ta in ex te n t increased our knowledge in these re s p e c ts . In the present study immunohistochemical methods have been used to id e n tif y interm ediate filam ent p ro te in s (IFP) in normal f e ta l and a d u lt p a r o tid glands, as well as in s a liv a r y neo
plasms. The interm ediate filam ents (IF) make up the cytoskeleton in eucaryotic c e l l s . E p ith e lia l t i s s u e contains IF composed of d i f f e r e n t c y to k eratin s (CK 1-19) w h ils t mesenchymal t i s s u e generally contains IF composed of vimentin, and the IFP p a tte rn is very s ta b le even during ce ll transform ation. I t would thus be p o ssib le to f u r th e r c l a r i f y the histo g en esis of s a liv a r y neoplasms by id e n tif y in g IFP; in addition the IFP p a tte rn would probably be useful in tumour typing. Furthermore, u l t r a s t r u c t u r a l cytochemical s t u d i e s , microspectorphotometry on nuclear DNA as well as enzyme se c re to ry studies of c e r t a i n tumour types were c a r r ie d o u t, in order to f u r th e r c h a r a c te r iz e the biology of s a liv a r y neoplasms.
The immunohistochemical in v e s tig a tio n s showed t h a t in normal p a r o tid t i s s u e , the d i f f e r e n t ce ll types d if f e r e d in IFP expression: ac in a r c e l l s express mainly CK 18 and myoepithelial c e l l s mainly CK 17 and 19, w h ils t duct c e l l s contained a broad range of CK. Vimentin could in ad
d itio n to CK be detected in myoepithelial c e l l s and basal c e l l s of ex
cr e to ry ducts. Fetal p a r o tid c e l l s showed a sim ila r CK p a tte rn as mature duct c e l l s . In a d d itio n , vimentin could be found in some basal c e l l s of the terminal tubules of the f e ta l gland. S alivary neoplasms could be divided in to th ree types with regard to t h e i r IFP p a tte r n : 1. Acinic ce ll carcinomas showed a CK-pattern sim ila r to normal a c in a r c e l l s but a co-expression of CK and vimentin was present in some c e l l s . 2. Adenoid c y s tic carcinomas, mixed tumours and basal c e ll adenomas showed a CK-pattern of normal duct or myoepithelial c e l l s . The p e r i pheral c e l l s were a ls o vimentin p o s itiv e . 3. Mucoepidermoid carcinomas and adenocarcinomas had a s im ila r CK-pattern as duct c e l l s , and no t u mour c e l l s contained vimentin. This in d ic a te s t h a t typing of IFP may be useful fo r subgrouping of s a liv a r y neoplasms.
By ste re o lo g ic a l measurements, the c e l l s of a c in ic c e ll carcinomas were found to be very s im ila r to normal p a r o tid ac in a r c e l l s . Furthermore, they contained amylase and a f t e r stim ulation by norepiphrine a s e c re tory response was induced, with a r i s e in i n t r a c e l l u l a r cAMP as well as a re le a s e of amylase. By sin g le c e ll measurements of nuclear DNA con
t e n t , no d iffe re n c e was found between a c in ic ce ll carcinomas with de
f i n i t e m etastasis and those without re currence, both in p a r a ffin sec
tio n s and cytological smears.
Key words: S alivary gland neoplasms, Intermediate fila m e n ts, immuno- histochem istry, u l t r a s t r u c t u r e , DNA, development, h is to g e n e s is , progno
s i s
ORIGINAL PAPERS
This th e s i s i s based upon the following p u b lic a tio n s , reference to which w ill be made by c i t a t i o n of the ap p ropriate Roman numerals.
I Gustafsson H, Kjörell U, Carlsöö B.
Cytoskeletal p ro te in s in oncocytic tumours of the p a r o tid gland.
Arch Otolaryngol 1985;111:99-105.
II Gustafsson H, Carlsöö B, K jörell U, Thornell L-E.
Immunohistochemical and u l t r a s t r u c t u r a l observations on adenoid c y s t i c carcinoma of sa liv a ry glands with special reference to in termediate filam ents and proteoglycan p a r t i c l e s .
Acta Otolaryngol 1986;102:152-60.
II I Gustafsson H, Carlsöö B, Kjörell U, Thornell L-E.
U l t r a s tr u c tu r a l and immunohistochemical aspects of carcinoma in mixed tumours.
Am J Otolaryngol 1986;7:218-30.
IV Gustafsson H, Kjörell U, Eriksson A, Virtanen I , Thornell L-E.
D is trib u tio n of interm ediate filam ent p ro te in s in f e ta l and a d u lt p a r o tid glands in man.
Submitted
Y Gustafsson H, Virtanen I , Thornell L-E.
The interm ediate filam ent p ro te in p a tte r n of sa liv a ry gland c a r c i nomas as revealed with monoclonal a n tib o d ie s.
Submitted
VI Gustafsson H, Carlsöö B, Henriksson R.
U l t r a s tr u c tu r a l morphology and sec re to ry behaviour of a c in ic cell carcinoma.
Cancer 1985;55:1706-10.
YII Gustafsson H, Lindholm C, Carlsöö B.
DNA cytophotometry of a c in ic c e ll carcinoma and i t s r e la tio n to prognosis.
Acta Otolaryngol (in p r e s s ) .
7
INTRODUCTION
Salivary gland neoplasms, a r i s i n g in major or minor s a liv a r y and mucous glands, c o n s t i t u te a very heterogeneous group of tumours. For a long time, d iffe re n c e s in opinion with regard to the h isto g e n e sis of these le sio n s as well as the d i f f i c u l t i e s encountered in d i f f e r e n t i a t i n g malignant tumours from benign ones have prevented a u n iv e rsa lly a p p lic able c l a s s i f i c a t i o n .
Other d i f f i c u l t i e s in e s t a b l i s h i n g such a c l a s s i f i c a t i o n are:
1. These tumours are r a r e ; the annual incidence fo r malignant neoplasms being only 1.0/100,000 in h a b ita n ts (Cancer Incidence in Sweden, 1958- 1981).
2. They show a wide h is to lo g ic a l v a r i a b i l i t y , not only between d i f f e r ent tumours but a ls o within one and the same tumour. A sin g le neoplasm can simultaneously show d i f f e r e n t i a t i o n of not only squamous and gland
u la r c e l l s , but a lso of a v a r ie ty of mesenchymal t i s s u e types.
3. The morphology and c l i n i c a l behaviour do not always c o r r e l a t e . Neo
plasms, such as a c in ic c e ll tumours - with c e l l s very akin to normal acinar c e l l s - or well d i f f e r e n t i a t e d mucoepidermoid tumours, may occa
s io n a lly m etastasize . Moreover, even in cases of "benign" mixed tu mours, secondary m etastasis has been reported.
Foote & Frazell (1953) were the f i r s t to c l a s s i f y a la rg e s e r ie s of t u mours on the basis of t h e i r b iologic al behaviour, and t h e i r work s t i l l forms the basis fo r c l a s s i f i c a t i o n of s a liv a r y tumours. The works of for example, Eneroth (1964), Evans & Cruickshank (1970) and S e i f e r t &
Donath (1976), have f u r th e r increased our knowledge. In 1972 the c u r
re n tly most used c l a s s i f i c a t i o n was e s ta b lis h e d (WHO, Thackray A Sobin) (Table I ) .
Table I . C la s s i f i c a t i o n of s a liv a ry neoplasms (WHO 1972) I , E p ith e lia l tumours
A. Adenomas
1. Pleoaorptilc adenowa (mixed tumour). A circumscribed tumour char
a c te r iz e d m icroscopically by i t s pleomorphic or "mixed" appearance;
c l e a r l y recognizable e p i t h e l i a l t i s s u e being intermingled with t i s s u e of mucoid, rrçyxoid, or chondroid appearance. (74%)*
2. Nonoaorphlc adénouas: The monomorphic adenomas are benign growths in which the epithelium forms a very r e g u la r, usually glandu
l a r , p a tte rn and in which th ere is no evidence of the mesenchyme-like t i s s u e t h a t i s so c h a r a c t e r i s t i c a component of the pleomorphic adeno
ma.
(<a) Adenolymphoma: A tumour composed of glandular and often c y s tic s t r u c t u r e s , sometimes with p a p illa r y c y s tic arrangement, lin e d by c h a r a c t e r i s t i c e o s in o p h ilic epithelium . The stroma contains a v a r i able amount of lymphoid t i s s u e with f o l l i c l e s . (4.1%)*
(b) Oxyphilic adenoma: A benign tumour c o n s is tin g of la rg e c e l l s with granular e o sin o p h ilic cytoplasm (oncocytes). There is very l i t t l e stroma, and although occasionally lymphocytes may be p re sen t, lymphoid f o l l i c l e s are ra re ly seen, in c o n tr a s t to adenolymphoma.
(0.9%)*
(£) Other types of monomorphic adenomas. There are other types of ade
noma composed, in varying p ro p o rtio n s, of e p i t h e l i a l c e l l s of duct- lin in g type and of c e l l s resembling those of a basal ce ll skin t u mour (myoepithelial c e l l s ) , arranged in a tu b u la r, a lv e o la r or t r a b e c u la r p a t t e r n , with or without cysts or p a p illa r y formations.
In one type the c e l l s and t h e i r arrangement so resemble t h a t of a b a s a l- c e ll carcinoma of the skin t h a t i t has been named a b a s a l- c e ll adenoma. There i s a ls o a very ra re type of tumour containing d u c t - l i k e s t r u c t u r e s with an e p i t h e l i a l lin in g surrounded by c le a r c e l l s in the p o sitio n normally occupied by the myoepithelial c e l l s .
9
B. Mucoepidermoid tumour
A tumour c h a ra c te riz e d by the presence of squamous c e l l s , mucus-secret
ing c e l l s , and c e l l s of interm ediate type. (4.9%)*
C. Acinic ce ll tumour
A tumour c o n s is tin g of c e l l s sim ila r to the serous c e l l s of s a liv a ry glands. (2.7%)*
D. Carcinoma
1. Adenoid c y s t ic carcinoma: An i n f i l t r a t i v e malignant tumour hav
ing a very c h a r a c t e r i s t i c c r ib r if o rm appearance. The tumour c e l l s are of two types, d u c t- lin in g c e l l s and c e l l s of myoepithelial type, and are arranged as small d u c t- lik e s tr u c tu r e s or la r g e r masses of myoepi
t h e l i a l c e l l s disposed around c y s t i c spaces to give a c r ib rifo rm or l a c e - l i k e p a tte r n . (4.7%)*
2. Adenocarcinoma: A malignant e p i t h e l i a l tumour showing some tubule or p a p illa r y glandular formation. No remnants of pleomorphic adenoma are p re se n t. (2.3%)*
3. Epidermoid carcinoma: A malignant e p i t h e l i a l tumour with c e l l s forming k e ra tin or having i n t e r c e l l u l a r bridges; mucous sec retion is not p re se n t. (0 . 8%)*
4. U ndifferentiated carcinoma: A malignant tumour of e p i t h e l i a l s tr u c tu r e t h a t is too poorly d i f f e r e n t i a t e d to be placed in any of the other groups of carcinoma. (3.9%)*
5. Carcinoma in pleomorphic adenoma (malignant mixed tumour): Tu
mours showing d e f in i t e evidence of malignancy, such as invasive growth and c ytological changes appropriate to carcinoma, and in which areas c h a r a c t e r i s t i c of pleomorphic adenoma can s t i l l be found. (1.5%)*
I I . Non e p i t h e l i a l tumours I I I . U n c la ssifie d tumours IV. A llied conditions
*Frequency of a ll s a liv a ry neoplasms (from Biörklund & Eneroth, 1980).
A drawback with the WHO c l a s s i f i c a t i o n i s the lack of discrim ination between benign and malignant le s io n s , due to the use of an interm ediate term, tumour. The two types of neoplasms c l a s s i f i e d under t h i s heading u sually behave benignly. However, high grade mucoepidermoid tumours have a very bad prognosis and fo r a c in ic c e ll tumours the 20-year de
term inate survival r a te is only 56% (Eneroth e t al 1966). For both these tumour types, most authors nowadays use the designation "cancer"
(Blanch 1974, Batsakis 1979, Biörklund & Eneroth 1980).
New e n t i t i e s , such as s a liv a r y duct carcinoma (K leinsasser e t al 1968), c l e a r c e ll carcinoma (e .g . Echewarria 1967), e p ith e lia l-m y o e p ith e lia l carcinoma of in t e r c a l a t e d ducts (Donath e t al 1972) and terminal duct carcinoma (Batsakis e t al 1983a) have been introduced, making modern c l a s s i f i c a t i o n even more voluminous (Table I I ) .
S tru ctu re of the s a liv a ry glands
All the major and minor s a liv a ry glands have the same p rincipal s t r u c t u r e . A large excretory duct is repeatedly dichotomously divided into smaller ducts term inating in to a c in i . The stroma divides the parenchyme into lobes and lo b u le s. Vessels and nerves mainly follow the ducts.
Current t e x t books (Geneser 1985, Fawcett 1986) divide s a liv a r y glands in to serous, mucous or seromucous. Serous glands are the p a r o tid , and the p o s te rio r lingual glands of von Ebner. Mixed mucous and serous are the submandibular, su b lin g u al, l a b i a l , buccal, a n t e r i o r lingual and the p o s te r io r lingual (except von E bner's ) . The gl osso p alatin e and p a la tin e and the glands of the root of the tongue are pure mucous glands. Some
"serous" c e l l s , however, a c tu a lly contain material which has histochem- ical c h a r a c t e r i s t i c s of mucous.
The s a liv a ry ducts are divided in to i n t e r c a l a t e d ducts close to the a c i n i , s t r i a t e d ducts which are purely i n t r a l o b u l a r and excretory ducts located e x tr a lo b u la r ly .
11
Table I I . C l a s s i f i c a t i o n of e p i t h e l i a l sa liv a ry gland tumours (from Batsakis 1979)
Type of lesion V ariations
Benign Mixed tumour (pleomorphic adenoma)
P a p illa r y cystadenoma lymphomatosum (Warthin's tumour)
Oncocytoma (oncocytosis) Monomorphic tumours
Basal c e ll adenoma Glycogen rich adenoma (?) Clear c e ll adenoma Membranous adenoma Myoepithelioma Sebaceous tumours
Adenoma Lymphadenoma
P a p illa ry ductal adenoma (papilloma) Benign lymphoepithelial lesion U nc la ssifie d
Malignant Carcinoma ex pleomorphic adenoma (carcinoma a r is i n g in a mixed tumour)
Malignant mixed tumour (biphasic malignancy) Mucoepidermoid carcinoma
Low-grade
Intermedi ate-grade High-grade
Adenoid c y s t i c carcinoma Acinous ce ll (a c in ic ) carcinoma Adenocarcinoma
Mucus-producing adenopapillary and non-papillary carcinoma
S alivary duct carcinoma (ductal carcinoma) Other adenocarcinomas
Oncocytic carcinoma (malignant oncocytoma)
Clear ce ll carcinoma (non-mucinous and glycogen-con
ta in in g or non-glycogen-containing) Primary squamous ce ll carcinoma
Hybrid basal c e ll adenoma/adenoid c y s t i c carcinoma U n d iffe re n tia te d carcinoma
E p ith e lia l-m y o e p ith e lia l carcinoma of in te r c a l a t e d ducts
Miscellaneous (includes sebaceous, S tensen1s duct, melanoma and carcinoma ex lymphoepithelial le sio n ) M etastatic
U nclassified
A c in i. The epithelium of the acini c o n s is ts of mucous and serous c e l l s (Fig. l a ) ; sometimes intermingled within the same a c in i . P erip h e ra lly located myoepithelial c e l l s a ls o c o n s t i t u te an important p art of the s a liv a r y acinus. The serous c e l l s are pyramidal in shape and contain e x tensive rough endoplasmic reticulum a t the base and an apex loaded with zymogen granules (PAS p o s itiv e but mucicarmine neg a tiv e ). Mucous c e l l s appear generally more "swollen" but have p r in c ip a lly the same a r c h i t e c t u r e . A confluent network of mucinous substances pushes the nucleus b a sa lly . In routine pre p ara tio n s the c e l l s appear pale, but the granules are s ta in e d by both PAS and mucicarmine.
Myoepithelial c e l l s are found between the acinar or duct c e l l s and the basal lamina (Figs l a , b). They are spindle-shaped, with numerous long th in p r o je c tio n s , and contain la rg e numbers of cytoplasmic filam en ts.
The myoepithelial c e l l s are f a r more numerous in the submandibular gland than in the p aro tid . They are found spread from the acini to the d i s t a l p a r t of the s t r i a t e d duct (C utler e t al 1977).
Ducts. I n te r c a la te d ducts are composed of small cuboidal c e l l s s u r
rounding a duct lumen. They are poor in organelles and are often en
closed by myoepithelial c e ll p ro je c tio n s (Fig. l b ) . The s t r i a t e d ducts are composed of c y lin d ric a l c e l l s with a larg e nucleus. In the basal p a r t of these c e l l s , p a r a l l e l rows of numerous mitochondria (Fig. l c ) , separated by invaginations of the ce ll membrane, are c le a r ly apparent.
Small zymogen granules can be seen a p ic a lly (Cutler e t al 1977). The excretory ducts are locate d e n t i r e ly within the connective t i s s u e se p ta , and are lin e d by p s e u d o s tr a tif ie d epithelium th a t is gradually transformed in to s t r a t i f i e d squamous epithelium.
Embryology. As numerous th e o rie s on the o rig in of s a liv a ry neoplasms have concerned u n d if f e r e n tia te d c e l l s , a r e c a p itu la tio n of the develop
ment is necessary. The p a r o tid anlage a r is e s from the epithelium of the stomodeal ectoderm. The anlagen of the other major glands a r is e near the zone where the buccopharyngeal membrane once marked the ju nction of ectoderm and entoderm (Batsakis 1979). The development of the a c in a r- duct u n it has been most thoroughly studied in the submandibular gland of r a t and hamster. The p a tte r n of d i f f e r e n t i a t i o n , however, appears
Fig. 1 a) Serous (S) and mucous (M) c e l l s of a sublingual acinus. (ME = myoepithelial c e l l , b) Submandibular i n t e r c a l a t e d duct, c) Typical s t r i a t e d duct c e l l s with numerous mitochondria.
s im ila r fo r most s a liv a ry glands, i r r e s p e c t i v e of sp ec ie s, although d iffe re n c e s in timing may occur. In b r i e f , the hamster submandibular gland a t g e sta tio n a l day 13 i s composed of p lu r i p o t e n t i a l ( u n d if f e r e n t
ia te d ) stem c e l l s . At day 14, cytom orphodifferentiation in to "terminal tu b u le s " , composed of sec reto ry c e l l s , myoepithelial ce ll precursors and u n d if f e r e n tia te d stem c e l l s take place. From the u n d if f e r e n tia te d c e l l s of the terminal tubules the "terminal buds" a r is e which are f o r e runners of the a c i n i . Both the terminal tubules and the buds are sup
ported by sh o rt s ta lk s which l a t e r are transformed in to in t e r c a l a t e d d ucts. The acini are f u l l y developed 2-3 weeks post partum (Chaudhry e t al 1985).
Myoepithelial c e ll precursors in the terminal tubules are ric h in rib o somes, polyribosomes, and mitochondria. A fter b ir th a s i g n i f i c a n t i n crease in myoepithelial c e ll pre c u rso rs, d i s t r i b u te d between the basal lamina and other c e l l s of the terminal tubules occurs. Deposits of de
l i c a t e cytofilam ents can be seen i n t r a c e l l u l a r l y sh o rtly a f t e r b i r t h . After two weeks the cytoplasms are v i r t u a l l y f i l l e d with myofilaments.
The development of myoepithelial c e l l s thus occurs in several d is c r e te steps p rio r to the synthesis and deposition of myofilaments and once t h i s r e l a t i v e l y advanced stage of maturation is reached the c e l l s lose t h e i r a b i l i t y to divide (Chaudhry e t al 1983a).
In man a larg e number of c i l i a t e d c e l l s occur ea rly in the p rim itiv e ducts but by week 20 the only remnants are within the a c in i . Triangular myoepithelial c e l l s are found as ea rly as week 12, but before week 25 they contain only minute amounts of myofilaments (Donath e t al 1978).
Tumour h isto g en esis
The polymorphism of s a liv a r y neoplasms has given r i s e to a number of th e o rie s concerning t h e i r o r ig in . Tumour c l a s s i f i c a t i o n in general is based on the presumed o rig in of the n eoplasia, e .g . malignant tumours from mesenchymal t i s s u e are c a lle d sarcomas, whereas e p i t h e l i a l malig
nancies are c a lle d carcinomas or cancers.
The f i r s t h is to lo g ic a l d e sc rip tio n s of s a liv a ry neoplasms were pu b lish ed by B i llr o th (1859) and Virchow (1863). The neoplasms were presumed to a r i s e from connective t i s s u e elements and were named enchondromas.
15
The theory of a mesenchymal o rig in p e r s is te d fo r decades. The term mixed tumour was coined by Minssen (1882) to in d ic a te a dual e p i t h e l i a l and mesenchymal o r ig in . Later on an e p i t h e l i a l o rig in theory became more and more prominent and the name was changed to pleomorphic adenoma (W illis 1953). Of the e p i t h e l i a l c e l l s , the rnyoepithelial c e ll has p a r
t i c u l a r l y a t t r a c t e d i n t e r e s t as a p rogenitor (fo r review, see e .g . Mylius 1960, Hamperl 1970, Hübner e t al 1971, Batsakis e t al 1983b).
The issu e of tumour o rig in i s clo se ly a sso c ia te d with s a liv a ry gland cell regeneration and renewal. In man, the reg en erativ e and p r o l i f e r a t i v e capacity i s proposed to re sid e within the i n t e r c a l a t e d duct (Batsakis 19,79). A fter p a r ti a l ac in a r c e ll degeneration, induced by e th io n in e , the recovery phase manifests both ac ina r c e ll p r o l i f e r a t i o n and the appearance of ac in a r c e l l s as a r e s u l t of the d i f f e r e n t i a t i o n of the most terminal portion of the s a liv a ry duct system. However, a f t e r complete ac in a r d e stru c tio n following io nizing ra d ia tio n th ere is no evidence of ac in a r regeneration even a f t e r 75 weeks (Leeb 1975).
Present th e o rie s mostly favour u n d if f e r e n tia te d duct reserve c e l l s as progenitors (Eversole 1971, Batsakis 1980, Chaudhry e t al 1983a, b, 1984, 1985, 1986). On the other hand, Dardic s tr e s s e d the renewing po
t e n t i a l of ac ina r as well as duct c e l l s and proposed t h a t the o rig in for mixed tumours was from a mature duct or ac in a r c e ll or a complete a c in a r-d u c t u n it ra th e r than an u n d if f e r e n tia te d duct c e ll (Dardic e t al 1982, 1983a, b).
New approaches in ce ll biology
Today's pathology is based prim arily on morphological stu d ie s of c e l l s and t i s s u e s . The intro d u ctio n of molecular markers may change the na
tu re of d iagnostic pathology from a su b je c tiv e a p p r a is a l, whose accura
cy i s dependent upon the s k i l l and experience of the individual patho
l o g i s t , to a multi méthodologie and a n a ly tic science based on o b je c tiv e ly demonstrable data.
Immunohistochemistry. Immunological techniques have opened up g re a t p o s s i b i l i t i e s for the i d e n t i f i c a t i o n of d i f f e r e n t p ro te in s and o ther tumour products in both serum and tumour t i s s u e . Two major types of
probe have been used to id e n tif y such ce ll ty p e - s p e c if ic components:
a n tibodies and, to a l e s s e r e x te n t, nucleic acid sequences. Up to now most stu d ies have been concerned with the presence of e .g . onco-fetal a n tig e n s , m e ta llo p ro te in s, hormones, enzymes and a number of other nor
mal membrane and i n t r a c e l l u l a r c o n s titu e n ts (fo r stu d ies concerning s a liv a r y gland tumours, see e .g . S e i f e r t & C a se litz 1983, Murase e t al 1985, Warner e t al 1985).
The s e l e c t i v i t y of c e l l s in which the antigens appear is a p r e r e q u is ite for a wide a p p l i c a b i l i t y of whatever immunological marker is employed.
Furthermore, the expression of the antigen in v e s tig a te d must be r e l a t i v e l y constant during c e ll transform ation, otherwise f a ls e -n e g a tiv e r e s u l t s may be too frequent. Moreover, the antigen expression must d i vide c e l l s in a meaningful manner, e .g . d iscrim inate between d i f f e r e n t types of t i s s u e or d i f f e r e n t i a t e benign le sio n s from malignant ones.
Studies on interm ediate filam ent p ro te in s (IFP) have in these respects been promising.
D iff e re n t types of p ro tein s c o n s t i t u te these filam ents and by immuno
lo g ical methods f iv e c la s s e s have been d istin g u ish e d . U1 t r a s t r u c t u r a l l y the filam ents cannot be separated as they all have a diameter of 7-11 nm. The t i s s u e s p e c i f i c i t y of IF i s remarkable. Cytokeratins (CKs) are found in e p i t h e l i a l c e l l s of ectodermal or endoderma! o r ig in . Most mesenchymal c e l l s contain vimentin but, in s t r i a t e d and some vascular smooth muscle c e l l s , desmin is a lso found. Neurofilaments and g l ia l f i b r i l l a r y a c id ic protein (GFA) are found in neurons and g lia l t i s s u e , re s p e c tiv e ly . Vimentin, desmin and GFA are s in g le p ro te in s . On the other hand, th re e d i f f e r e n t neurofilament p roteins e x i s t and as f a r as CKs are concerned, a larg e group of a t l e a s t 19 d i f f e r e n t pro tein s are p re se n tly known in man. During the present i n v e s tig a tio n s , the p a tte rn s of CKs in d i f f e r e n t e p i t h e l i a were c h a ra c te riz e d (Moll e t al 1982, 1983) and were found to be co n stan t. The pattern of IFPs, both as a group and the p a tte rn of individual CKs, have been found to be highly s ta b le during n eo p la stic transform ation (Moll e t al 1982, Quinlan e t al 1985). These p ro p e rtie s would make IFPs valuable to o ls fo r studies of h isto g e n e sis and tumour typing (Gabbiani e t al 1981, Ramaekers e t al
17
1982, 1983, 1985b, Osborn & Weber 1982, 1983, Kahn e t al 1983, M i e t t i nen e t al 1984b, Cooper e t al 1985, Virtanen e t al 1985).
With regard to s a liv a ry gland and s a liv a ry neoplasm IFP p a t t e r n s , a number of stu d ies have been performed using d i f f e r e n t immunohistochemi- cal techniques. The r e s u lt s on the normal p a ro tid have, however, been divergent, making evaluation of the tumour s ta in in g d i f f i c u l t .
Measurements of nuclear DNA c o n te n t. The in troduction of q u a n tita tiv e microspectrophotometric methods by Caspersson (1936), made determina
tio n of DNA content of individual c e l l s p o ssib le . At present a number of techniques fo r DNA determination e x i s t : in ce ll suspensions, flow cytometry can be used and in sec tio n s or smear, sin g le ce ll measure
ments are used.
For a number of neoplasms a c o r r e la tio n between c e l l u l a r DNA content and prognosis has been documented. For s a liv a ry neoplasms, Eneroth &
Z etterberg (1973, 1974a, b, 1975, 1976, Eneroth e t al 1974) found a c o r r e la tio n between malignancy and DNA content. However, only a few stu d ies have been performed on the prognostic value of DNA le v e ls in s a liv a r y gland tumours.
Receptor c h a ra c te riz a tio n and stim u lu s-se c re tio n coupling. A vast num
ber of sec reto ry products have been described in tumour t i s s u e . Simi
l a r l y , receptors in a few neoplasms have been c h a ra c te riz e d and in mam
mary neoplasms, for in stan ce , estrogen receptor content has been c o r
r e la t e d to prognosis and d i f f e r e n t i a t i o n (H a rtv eit e t al 1980, Mohammed e t al 1986). Only a few d e sc rip tio n s of adrenoceptors, prim arily in p i t u i t a r y adenomas, have been published (H eisler e t al 1983, Reisine e t al 1983, Ramsdell e t al 1985). Likewise, i n t r a c e l l u l a r events during the sec reto ry process, such as changes in c y c lic AMP l e v e l s , have been studied (de Carni I l i e t al 1980, Reisine e t al 1983, H e isler e t al 1983). Comparisons of such re a c tio n s with t h a t of the t i s s u e o rig in may be a way of determining the degree and d ire c tio n of d i f f e r e n t i a t i o n .
Furthermore, the above discussed stu d ies on tumour t i s s u e may be a way of eva lu a tin g the influence of in t e r a c t io n between ce ll types in normal t i s s u e when the t i s s u e of o rig in i s complex with regard to ce ll type and in n erv a tio n .
OBJECTIVES OF THE PRESENT STUDY
The main o b jectiv es of the p re sent study were to determine:
- the interm ediate filam ent p a tte r n of developing and mature normal s a liv a r y gland
- whether the IFP p attern of s a liv a ry neoplasms i s r e la te d to any par
t i c u l a r cell (a d u lt or f e t a l ) with regard to histo g en esis
- whether the IFP p a tte rn can be used fo r typing of s a liv a ry gland tumours
- whether c e r t a i n tumour c e l l s which clo sely resemble normal acinar c e l l s respond to sec reto ry stim uli
- whether the nuclear DNA content of n eo p lastic c e l l s in a c in ic ce ll carcinoma has any p r e d ic tiv e value in s e le c tin g p a tie n ts with good pro
gnosis as r e f l e c t e d by long-term s u rv iv a l.
19
MATERIAL AND METHODS
Morphology (I-VI)
Minute specimens were taken from s a liv a ry t i s s u e and tumour t i s s u e a t the time of the operation and f u r th e r processed. Fetal p a ro tid t i s s u e was obtained from aborted fe tu se s (IY). For immunological s t u d ie s , specimens were immediately frozen by immersion in isopentane precooled with liq u id n itrogen. For e le c tro n microscopy, specimens were immedi
a te ly fixed in 2.5% or 4% glutaraldehyde in phosphate ( I - I I I , VI) or cacodylate ( I I ) b u ff e r, rin sed in b u ffe r, p o s t-fix e d in 1% O S O 4 , de
hydrated in graded ethanol s o lu tio n s , propylene oxide and embedded in Yestopal or Epon 812 r e s in .
For i d e n t i f i c a t i o n of mucous substances, t i s s u e from adenoid c y s tic carcinomas was fixed in 2.5% glutaraldehyde in cacodylate buffer with 2 000 p.p.m. ruthenium red added overnight, rinsed in b u ffer containing 1 000 p.p.m. ruthenium red, p o st-fix e d in 1% O SO 4 with 500 p.p.m.
ruthenium red, dehydrated and embedded in Epon ( I I ) . Specimens were a lso processed to demonstrate p erioda te re a c tiv e substances by the p eriodic acid - chromic acid - s i l v e r methenamine method and the phos- photungstic acid method ( I I ) . U ltra th in sec tio n s were cut using an LKB Ultrotome and were mounted on copper or gold grids and examined e i t h e r unstained or a f t e r s ta in in g with uranyl a c e ta te and lead c i t r a t e in a P h ilip s EM 300 or a Jeol CX 100 e lec tro n microscope.
Immunohistochemistry (I-V)
Antibodies;. Three mouse monoclonal antibodies d ire c te d ag a in st cyto- k e r a tin s were used (PKK1, PKK2 and PKK3). These antibodies have been c h a ra c te riz e d elsewhere (Holthöfer e t al 1983, 1984).
Two mouse monoclonal antib o d ies to vimentin were used: one was r a is e d a g a in st bovine lens vimentin (vim 9) and was purchased from Sanbio (N istelrhode, the Netherlands) and the other monoclonal antibody was a g a in st vimentin from human f i b r o b l a s t s from clone FW24 BA6 (vim 24) (Labsystems Oy, H e lsinki, F inla nd).
Guinea pig and r a b b it polyclonal a n tibodies a g a in st human sole k e r a tin s were prepared according to Kjörell & Östberg (1984) and DNase 1 a n t i bodies were prepared according to Wang & Goldberg (1978), Preimmuniza
tio n serum or sera from non-immunized animals were a ls o processed as control samples ( I - I I I ) .
Immunoblots (IV)
Pieces of normal p a r o tid t i s s u e were f i r s t e x tra c te d in 0.5% Triton X-100 (50 mM Tris-HCl, pH 7 .4 , supplemented with p ro te o ly s is i n h i b i t o r s ) . T h e re a fte r, the samples were dissolved in e le c tro p h o re s is sample bu ffer (Laemmli 1970) and were subjected to polyacrylamide gel e l e c t r o phoresis in the presence of sodium dodecyl sulphate. Then, the polypep
t i d e s were t r a n s f e r r e d onto n i t r o c e ll u l o s e sheets according to Towbin e t al (1979). Amidoblack was used fo r protein s ta in in g . Culture super
n atan t of the monoclonal an tibodies was used for immunoreaction. The binding of the antibodies was detected by peroxidase-coupled ra b b it and mouse antiserum (1:100, Dakopatts, Glostrup, Denmark). The r e s u lt s with p a r o tid t i s s u e were i n te r p r e te d by using cy to sk e le ta l preparation of standard cell lin e s as a c o n tro l: UCF7 and A431 c e l l s with a known p a t
tern of cy to k eratin expression (Quinlan e t al 1985).
Immunostaining (I-V)
S e ria l frozen sec tio n s were used fo r immunostaining a f t e r acetone f i x a tio n a t -20°C fo r 5 min. The primary an tibodies were used in a concen
t r a t i o n of 20-50 mg/1 a t room temperature. The DNase 1 anti-DNase 1 i n cubation has been described by Wang & Goldberg (1978). A fter thorough washing, the sec tio n s were exposed to flu o re scein isothiocyanate (FITC coupled) or tetramethylrhodamine isothiocyanate (TRITC conjugated) l a b e lle d antibodies or processed by the peroxidase-anti peroxidase (PAP)-method (Sternberger e t al 1970). The sec tio n s were mounted in Moviol 4-88 (Hoechst, F rankfurt am Main, West Germany), pH 8.5 and viewed in a L eitz Orthoplan microscope equipped with e p iflu o re s c e n t o p tic s or an Olympus VAN0X microscope.
Ste reo lo g ica l methods (VI)
The ste re o lo g ic a l analyses were c a r r ie d out on randomly taken e lec tro n micrographs of p a r o tid a c in a r c e l l s and tumour c e l l s using Wei b e l ' s
21
(1979) point counting method employing a multipurpose g rid . The d ia meters of the sec reto ry granules in the ele c tro n micrographs were mea
sured in a Zeiss TGZ-3 p a r t i c l e s i z e analyser. The ste re o lo g ic a l ca lc u
l a t i o n s are described in d e ta il in VI. I t should be pointed out t h a t values obtained by ste re o lo g ic a l methods are su b je c t to various in h e r
ent e r r o r s , due mainly to t i s s u e p re p a ra tiv e procedures including sec
tio n in g . In comparative s t u d i e s , however, these e r r o r s diminish in im
portance.
Studies on se c re tio n and c y c lic AMP accumulation (VI)
Gland specimens were t r a n s f e r r e d to a Krebs-Henseleit bicarbonate buf
f e r (pH 7 . 4 ) . The b u ffer was supplemented with pyruvate, glutamate and fumarate as well as albumin (1 mg/ml) and glucose (0.6 mg/ml). Before incubation a piece of t i s s u e was taken fo r analyses of t o t a l anylase and c y c lic AMP c o n ten ts. The se c re to ry stu d ies in v i t r o were performed with a batch-incubati on technique described in d e ta il elsewhere (VI;
Carlsöö e t al 1981).
Amylase and c y c lic AMP assay (VI)
Amylase was assayed by a micromodification of the 3 , 5 - d i n i t r o s a l i c y l a t e method (Danielsson 1974). One u n i t of amylase i s defined as the a c t i v i t y l i b e r a t i n g reducing groups which correspond to 1 umol of maltose monohydrate per minute a t 25°C. Amylase re le a se was expressed as the percentage of amylase in medium in r e la t io n to the t o t a l anjylase con
t e n t in medium plus homogenate. Cyclic AMP was e x tra c te d from the t i s sue and assayed radio-immunologically, t h i s being expressed as pmol/mg t i s s u e dry weight. For d e t a i l s , see VI.
DNA measurement (VII)
Representative areas of the o rig in a l paraffin-embedded tumour specimens were se le c te d and two consecutive se c tio n s (4 urn) were c u t. One sec tion was s ta in e d with haematoxylin-eosin, the other was Feulgen-stained by the method of Duijndam & van Duijn (1973) a f t e r acid hydrolysis in 5 N HC1 fo r 1 hour. Computerized stage scanning cytophotometry was perform
ed using a L eitz MPV 3 cytophotometer and the HISTOSCAN program (B je l- kenkrantz e t al 1981). Within each specimen a re p r e s e n ta tiv e area of the tumour was s e le c te d and w ithin t h i s area n u c le i, not overlapping
each o th e r, were syste m a tic a lly measured. About 100 tumour c e l l s and about 50 control c e l l s were measured. The control c e l l s , i . e . lympho
cytes and/or ac in a r c e l l s , were used to define the normal, d ip lo id (2C) DNA value.
For f i v e a c in ic ce ll tumours, DNA an a ly sis was performed on a s p ira tio n biopsy smears. A fter destai n i ng in methanol and hydrolysis in 5 N HC1, the smears were stain ed in an 0.02% ac ri f i a v ine-S02 solu tio n (Tanke &
van Ingen 1980). Measurements were made using a L eitz MPY 3 cytophoto- meter and the FLU0RA program (Bjelkenkrantz e t al 1983). In these spec
imens, granulocytes were used as control c e l l s .
To d is tin g u is h d ip lo id /n e a r d ip lo id from aneuploid ce ll populations, the percentage of tumour c e ll nuclei with DNA values exceeding 2.5C was determined fo r each specimen. Furthermore, mean and modal values were c a lc u la te d from the histograms.
23
RESULTS
Cytoskeletal p ro te in s in normal s a liv a r y glands (I-IV)
In sec tio n s of normal s a liv a r y glands la b e lle d with CK-pAbs, intense s ta in in g was found in a ll types of duct and myoepithelial c e l l s , w h ils t acinar c e l l s were sta in e d mainly a t the apex and l a t e r a l side ( I - I I I ) . A sim ila r re action was obtained fo r CK-mAb PKK1. CK-mAb PKK2 re acted strongly with myoepithelial and duct c e l l s , w h ils t PKK3 re acted moder
a te ly with duct and ac ina r c e l l s (IV).
Vimentin pAbs did not r e a c t with any e p i t h e l i a l or myoepithelial c e l l , nor did mAb vim 9 ( I I - I Y ) . On the co n tra ry , vim 24 c l e a r l y s ta in e d nrçyo- e p i t h e l i a l and basal c e l l s of excretory ducts (IY). No d iffe ren c e was detected between p a r o tid , submandibular, or p a la ta l glands, or between mucous or serous ac in a r c e l l s . Actin was detected 1n myoepithelial c e l l s and a t luminar surfaces of s t r i a t e d duct and ac in a r c e l l s by the DNase 1/antiDNase 1 method. Along the i n t e r c e l l u l a r c a n a l i c u l i , s t a i n ing was a ls o evident between ac in a r c e l l s ( I ) . At those lo c a tio n s where IFP or a c tin could be found immunologically, fin e filam ents could a ls o be i d e n t i f i e d by e le c tro n microscopy.
Cytoskeleton of developing p a r o tid glands (IY)
CK and vimentin were determined by mAbs in developing p a r o tid gland.
In general, a moderate to strong s ta in in g was found in a ll f e ta l gland c e l l s fo r PKK1 and PKK2, a t g e sta tio n a l week 18. For PKK3, the s ta in in g i n t e n s i ty was somewhat weaker. The s ta in in g re a c tio n s were most Intense in the la r g e r ducts fo r PKK1 and PKK2. At g e sta tio n a l week 22, the d i f ferences in s t a i n a b i l i t y between la rg e and small ducts was more pro
nounced, although a ll e p i t h e l i a l c e l l s s t i l l sta in e d by a l l CK-mAbs. In terminal tubules and buds, s c a tt e r e d t r i a n g u l a r basal c e l l s were found s ta in a b le with both vim 9 and vim 24. In la r g e r ducts no vimentin con
ta in in g c e l l s were found.
Cytoskeleton of benign s a liv a r y neoplasms ( I , I I I , Y)
In mixed tumours CK-pAbs re acted with v i r t u a l l y a ll tumour c e l l s ( I I I ) , and mAbs PKK1 and PKK2 re acted strongly with v i r t u a l l y a ll tumour c e l l s , while PKK3 gave a f a i n t e r reactio n and the re actio n was most pronounced a t the duct lumina.
Y i menti n was d etected by pAbs ( I I I ) and mAbs (vim 9, vim 24) in c e l l s of both mesenchymal and e p i t h e l i a l c h a ra c te r. In the e p i t h e l i a l a r e a s , i t was mainly the peripheral c e l l s of cords t h a t were re a c tiv e .
Basal ce ll adenomas gave with mAbs almost the same re action as did the e p i t h e l i a l areas of mixed tumours (V, unpublished d a ta ) .
The CK a n tib o d ie s , PKK1 and PKK2, were strongly r e a c tiv e , while a f a i n t to moderate PKK3 re action was found a t duct lumina. Vimentin (vim 9, vim 24) was detected in peripheral c e l l s .
In oncocytomas two ce ll types were d istin g u ish ed : one more numerous and la rg e ce ll type and one small elongated type. Both types reacted strongly with PKK1 and weakly with PKK3, while the l a t t e r ce ll type was much more re a c tiv e to both CK-pAbs and PKK2 than was the former. Vimen
t i n was not d etected in tumour c e l l s (pAbs, vim 9, vim 24), and a c tin (DNase 1/ antiDNase 1) was only revealed a t luminal surfaces ( I , unpub
lis h e d d a ta ) .
In the e p i t h e l i a l p a r ts of adenolymphomas the small basal c e ll and the la rg e apical c e l l s re acted d i f f e r e n t l y with CK-pAbs: the basal c e l l s were in te n se ly sta in e d and the apical c e l l s moderately so ( I ) . With CK-mAbs t h i s d iffe re n c e was l e s s marked, with PKK1 and PKK2 the d i f f e r ence being barely v i s i b le while with PKK3 the basal c e l l s were le s s r e a c tiv e (unpublished d a ta ) . Vimentin was not d etected in tumour c e l l s (pAbs, vim 9, vim 24). Actin (DNase 1/antiDNase 1) was found only a t the apical surface of the columnar c e l l s ( I ) .
Cytoskeleton of malignant s a liv a ry neoplasms ( I I , I I I , V)
By using pAbs, CK was detected in adenoidcystic carcinomas and c a rcin o mas in mixed tumours ( I I , I I I ) .
CK and vimentin were determined by mcAbs in 16 consecutive s a liv a ry tumours which were i n i t i a l l y judged as malignant. The neoplasms were r e c l a s s i f i e d by one p a th o lo g is t and the r e l i a b i l i t y of the diagnoses evaluated. Ten diagnoses were judged as d e f in i t e (2 AdCC, 3 MEC, 2 ACC, 2 SDC and 1 poorly d i f f e r e n t i a t e d c l e a r - c e l l adenocarcinoma) while fo r
25
two other tumours i n i t i a l l y judged as AdCC, the d i f f e r e n t i a l diagnosis of basal c e ll adenoma and mixed tumour, re s p e c tiv e ly , could not be com
p le te ly ruled out. One tumour could have been MEC or a mucous-producing adenocarcinoma. For one probable ACC, the d i f f e r e n t i a l diagnosis was a m e ta s ta s is . One neoplasm was probably a MEC and f i n a l l y one AdCC was r e c l a s s i f i e d as a carcinoma in mixed tumour, where the malignant p a r t was th a t of an adenoid c y s tic carcinoma. The immunoreactivity of mAbs a g a in st CKs and vimentin i s given in Table I I I .
With re sp e c t to PKK2, PKK3 and vim 9 r e a c t i v i t y , the tumours could be divided in to three groups.
Group A co nsisted of neoplasms with stronger s t a i n a b i l i t y fo r PKK3 than for PKK2. They a ls o contained vim 9 p o s itiv e tumour c e l l s .
This group co n sisted of two ACC and one probable ACC.
In Group B neoplasms, a markedly stro n g er re action fo r PKK2 was noted than fo r PKK3. For vim 9 the c e ll population was non-homogeneous. In the periphery, towards the basal membrane, c e l l s were strongly vim 9 p o s itiv e but in the ce n tral po rtio n s and towards sec reto ry lumina, the c e l l s were vim 9 negative. This group of neoplasm contained two AdCC, th ree probable AdCC, one AdCC ex mixed tumour and a lso one probable MEC.
In Group C, PKK2 stain ed more in te n se ly than PKK3. Vim 9 was negative.
The tumours in t h i s group co n siste d of two sa liv a ry duct carcinomas, one c l e a r - c e l l adenocarcinoma and th ree MEC and one MEC or adenocarci
noma.
Polygonal p a r t i c l e s of proteoglycans in pseudocysts of AdCC could be i d e n t i f i e d by electron microscopy a f t e r conventional glutaraldehyde fix a tio n and osmium p o s t - f i x a t i o n . Addition of ruthenium red to the f i x a t i v e r e s u lte d in a more in te n se s ta in in g of the p a r t i c l e s and they could be i d e n t i f i e d without lead and uranium s t a in in g . In specimens fixed in OSO4 without p r io r aldehyde f i x a t i o n , the polygonal p a r t i c le s could not be demonstrated. In p eriodic acid - chromic ac id - s i l v e r methenamine sta in e d , Vestopal-embedded s e c tio n s , the deposits
were found mainly along i n t r a c y s t i c collagen bundles and basal lamina.
A f a i n t d iffu s e s ta in in g was r e g is te r e d over the amorphous or f in e ly f i b r i l l a r ground substance of the c y s t s , but was d i f f i c u l t to evaluate because of the s iz e of the s i l v e r p a r t i c l e s . Phosphotungstic acid s ta in e d the most l a t e r a l p a r t of the basal lamina network and the amorphous m aterial of the c y s ts .
Of the four carcinomas in mixed tumours in v e stig a te d u l t r a s t r u c t u r a l l y and compared with IFP p a tte r n , squamous d i f f e r e n t i a t i o n with t o n o f i l a - ments and glandular d i f f e r e n t i a t i o n with sec retory granules were simul
taneously found in two carcinomas.
In one neoplasm only glandular c e l l s were found, while in the fourth the c e l l s were devoid of e p i t h e l i a l s t r u c t u r e s , such as basal lamina and desmosomes. In the l a t t e r neoplasm, c e l l s were often elongated and resembled f i b r o b l a s t s ( I I I ) .
The three carcinomas t e s t e d immunohistochemically displayed d i f f e r e n t IFP p a t t e r n s , two with a d i s t i n c t coexpression of vimentin and CK, though both ex h ibited d i f f e r i n g h is to lo g ic a l s tr u c tu r e s : one showed a predominant mesenchymal d i f f e r e n t i a t i o n and the o ther was of a purely e p i t h e l i a l n atu re . In the t h i r d tumour, the c e l l s expressed CKs only ( I I I ) .
Table I I I . Immunoreactivity of monoclonal an tibodies ag a in st CKs and vimentin.
PKK1 PKK2 PKK3 Yim 24 Yim 9
Mixed tumour 2-3 2-3 0-2 0-2* 0-2*
Adenolymphoma 2-3 2-3 1 0 0
Oncocytoma 2 1-3** 1 0 0
Basal ce ll adenoma 3 3 0-1 0-2* 0-2*
Acinic c e ll carcinoma 3 2 3 0-2 0-2
Mucoepidermoid carcinoma 1-3 1-3 0-2 0 0
Adenoid c y s t i c carcinoma 3 3 0-1 0-2* 0-2*
Adenocarcinoma 3 3 2 0-1 0
♦Peripheral c e l l s are v im e n tin -p o sitiv e , ce n tral negative
**Two c e ll types, one strongly s ta in a b le , the o ther weakly
0=no re a c tio n , l=weak re a c tio n , 2=moderate re a c tio n , 3=strong re actio n
27
Morphometry and sec reto ry behaviour of ACC (VI)
For the tumour c e l l s from th ree ACC, compared s t e r e o lo g ic a lly with nor
mal p a r o tid gland ac ina r c e l l s , the morphometric study revealed t h a t the granule density was increased. The c e l l s contained more granules but the diameter was considerably le ss compared to normal c o n tr o ls . No d iffe ren c e in ce ll volume or nuclear volume density could be found.
In the batch incubation system, amylase and cAMP were detected in t i s sue from the three ACC, as well as in normal p a ro tid gland t i s s u e . Even though a high basal (non-stim ulatory) amylase output was noted from t u mour specimens, noradrenalin induced a marked increase in the amylase re le a s e . Noradrenalin stim ulation a lso caused an enhanced cAMP accumu
la tio n in the tumour t i s s u e . Sim ilar q u a n t i t a t i v e a l t e r a t i o n s were ob
served in the normal gland t i s s u e .
DNA le v e ls in a c in ic ce ll carcinomas (VII)
Cytology smears and p a r a f f in blocks were processed fo r DNA measurements from 13 p a tie n ts : 9 p a tie n ts without signs of recurrence in the 6-year follow-up period, 2 with local re cu rre n ces, and 2 p a tie n ts with méta
sta s é s and/or had died of d isease . Measurements on sec tio n s (8, 2, 2 p a tie n ts in the groups, re s p e c tiv e ly ) as well as on smears (3, 0, 2) revealed t h a t a ll tumours were d ip lo id or n ea r-d ip lo id within the range (2C +_ 10%). Two re c u rre n t and two m etastasizing tumours had, however, modal values between 2.2C and 2.3C. The percentages of tumour c e l l s ex
ceeding the 2.5C le v e l , as a measure of n o n -d ip lo id ity , were 6.4% f o r n o n-recurrent, 2.5% for re c u rre n t and for neoplasms with métastasés 16%. Measurements c a rrie d out on smears gave higher corresponding values than those on s e c tio n s , 25% fo r non-recurrent tumours and 24%
for m etastasizing neoplasms.
DISCUSSION
Intermediate filam ent p ro te in s in normal s a liv a ry glands
In order to evaluate tumour behaviour, tumour histo g en esis and tumour d i f f e r e n t i a t i o n , a thorough knowledge of the d i s t r i b u ti o n of immuno
logical markers present in the various normal c e l l s , and how these s tr u c tu r e s develop and mature is required.
C y to k eratin s. The composition of c y to k e ra tin s v arie s between d i f f e r e n t kinds of e p i t h e l i a l c e l l s . By peptide mapping and two-dimensional e l e c tr o p h o r e s is , catalogues of the CK-content of d i f f e r e n t e p i t h e l i a have been e s ta b lis h e d (Moll e t al 1982). Cytokeratins can, according to t h e i r charges, immunoreacti vi t i e s , mRNA h y b rid iz a tio n , peptide mapping p a t t e r n , aminoacid sequences and r e la tio n s h ip to wool k e r a tin s be d i v i ded in to an a c id ic type I and a ( n e u tr a l - t o - ) basic type II (Cooper e t al 1985, Quinlan e t al 1985). Tissue d i s t r i b u ti o n data have e s ta b lis h e d t h a t a t l e a s t one member from each i s present in a ll e p i t h e l i a , and i t has been proposed t h a t one a c id ic and one basic cy to k eratin is a p re re q u i s i t e for cyto k eratin filam ent formation (Cooper e t al 1985). There are in d ic a tio n s t h a t s p e c if ic k e r a tin p a ir s e x i s t , where the basic mem
ber is always l a r g e r than the a c id ic one. Each p a i r i s s p e c if ic fo r a c e r ta in e p i t h e l i a l t i s s u e type (Cooper e t al 1985).
Our immunoblots have shown, t h a t vimentin as well as CKs are present in s a liv a ry gland t i s s u e . The immunohistochemical s tu d ie s f u r th e r revealed t h a t the CK composition of various s a liv a r y c e l l s d i f f e r s . Duct c e l l s contain a complex of CKs strongly s ta in a b le with a ll CK-Abs (PKK1, PKK2, PKK3 and a number of pAbs) w h ils t the acinar c e l l s are most re ac
t i v e fo r antibodies a g a in st CK18 (mAbs PKK1 and PKK3). The myoepithe
l i a l c e l l s a ls o contain CKs, but these do not r e a c t with the antibody s p e c if ic a g a in st CK18 (mAb PKK3). Q u a lita tiv e ly , the IFP composition i s b a s ic a lly s im ila r in the p a r o tid , the submandibular and the p a la ta l glands.
Squamous e p i t h e l i a mainly expresses la r g e r c y to k eratin s as 50/58 kD CKs supplemented by another p a ir according to d i f f e r e n t i a t i o n ; k e r a t i n ized epidermis contain a ls o 56.5/65-67 kD CKs, cornea contains addi-
29
t i o n a l l y 55/64 kD CKs and esophageal or tongue type of e p î t h e l l a makes 51/59 kD CKs. Simple epithelium expresses 2 to 4 o f the sm allest CKs (45/52 kD and 40 and 54 kD CKs). Complex e p i t h e l i a l s tr u c tu r e s such as glands usually express a mixture of CK p a tte r n s of simple and s t r a t i fie d e p i t h e l i a (Cooper e t al 1985). The exact CK composition in n\yoepi- t h e l i a l c e l l s has not y e t been determined.
With regard to the polymorphism of c y to k e r a tin s , i t i s understandable th a t minor d iffe ren c es in antigen preparation and immunization may give d i f f e r e n t r e s u lt s regarding CK l o c a l i z a ti o n . The immunological findings of d i f f e r e n t stu d ies on normal p a r o tid gland are summarized in Table IV. Polyclonal antibodies have been used by C a s e litz e t al (1982), Krepier e t al (1982), Nathrath e t al (1982), Saku e t al (1984), Kahn e t al (1985), Takai e t al (1985) and Hosaka e t al (1985) as well as in p a r ts of the pre sent study ( I , II and I I I ) . The only stu d ie s in which monoclonal antibodies were used are the p re sent study (IV), Erlandsson e t al (1984), Palmer e t al (1985) and C a s e litz e t al (1986) (Table I I I ) . Erlandsson e t al (1984), however, using form alin-fixed pepsin- digested p a r a f f in s e c tio n s , were unable to demonstrate CKs in a c in a r and myoepithelial c e l l s . Palmer e t al (1985) used a mAb (16a) s p e c if ic fo r 45-46 kD CK and found r e a c t i v i t y only in basal excretory duct c e l l s . The s ta in in g p a tte r n of a mAb s p e c if ic fo r CK 18 (mAb CK 5) was most re cen tly reported on by C a s e litz e t al (1986a), and the s ta in in g p a tte rn in s a liv a r y glands was v i r t u a l l y id e n tic a l to our findings with PKK3 (IV). In the present study (IV), several mAbs were employed and CKs could be demonstrated in a ll e p i t h e l i a l and myoepithelial c e l l s .
Vimentin, in c o n tr a s t to c y to k e r a tin s , i s a sin g le gene product and one would consequently expect a more uniform s ta in in g p a t t e r n . Most s tu d ie s f a i l to d e te c t vimentin in e p i t h e l i a l and myoepithelial c e l l s . However, C a s e litz e t al (1982) found a p o s itiv e s ta in in g in basal t r i a n g u l a r c e l l s of ducts, and Erlandsson e t al (1984) and Kahn e t al (1985) found r e a c t i v i t y within a few myoepithelial c e l l s . Hosaka e t al (1986) claim the existence of vimentin in a ll ductal c e l l s of s a liv a r y glands. How
ever, they used non-digested formaldehyde-fixed p a r a f f in blocks and pAbs. They had s im ila r lo c a tio n s fo r a ll other antigens t e s te d and judging by published p i c t u r e s , t h e i r r e s u l t s may well be a r t i f a c t s due
to o v e rsta in in g . In a ll these s tu d ie s , pAbs were employed. Using a mAb (vim 24) we were able to demonstrate vimentin in both these ce ll types.
The other mAb (vim 9) sta in e d n e ith e r e p i t h e l i a l nor myoepithelial c e l l s (IV). The most probable explanation fo r t h i s d i s p a r i t y in vimen- t i n r e a c t i v i t y between the two monoclonal antibodies may be some kind of antigen masking, fo r example, by phosphorylation. By immunoblotting i t was f u r th e r shown (V) t h a t with vim 24, no c r o s s- re a c tio n occurred with any other polypeptide. The findings in a ll stu d ie s on vimentin in
s a liv a r y gland c e l l s are summarized in Table V.
Table IV. Cell types in which CKs have been detected
Type of Ab
AC MEC ICDC SDC EDC EDBC
Krepier e t al 1982 pAbs + + + + ? ?
C a s e litz e t al
1981, 1982 pAbs - + + + + +
1986a, b Ln 5 + + + + + +
Kl 1 + + + + + +
CK 5 + - + + + +
Nathrath e t al AK1 _ + _ . -
1982 AK + + + + + ?
Kahn e t al 1985 pAbs + + + + + +
Saku e t al 1984 pAbs + - + + + +
Erlandsson e t al AE1 . _ + + + +
1984 AE2 - - - - - -
AE3 - - + + + +
Palmer e t al pAbs - - + + + +
1985 16a - - - +
Takai e t al 1985 pAbs - - + + ? ?
Hosaka e t al 1986 pAbs - - + + + +
Gustafsson e t al pAbs + + + + + +
1985, 1986 PKK1 + + + + + +
PKK2 - + + + + +
PKK3 + — + + + +
AC * ac in a r c e l l s , MEC = rnyoepithel ia l c e l l s , ICDC = in t e r c a l a t e d duct c e l l s , SDC = s a liv a r y duct c e l l s , EDC = excretory duct c e l l s , EDBC = excretory duct basal c e l l s .
