- Email: [email protected]
Immunohistochemical demonstration of ribonuclease and amylase in normal and neoplastic parotid glands
Original Contributions Immunohistochemical Demonstration of Ribonuclease and Amylase in Normal and Neoplastic Parotid Glands DEBRA J. MORLEY, BS, JONATHAN E. HODES, MD, JEANNE CALLAND, BA, AND M. E. HODES, MD, PhD Antibodies specific for bovine ribonuclease A (antiRNase A) were raised in rabbits, and immunologic cross-reactivity between bov i n e RNase A and h u m a n salivary gland RNase was demonstrated. The antiRNase A served as the primary antibody in the peroxidase-antiperoxidase immunohistochemleal technique. Paraffin blocks of five normal human parotids and 20 parotid tumors were examined. In normal parotid and in cases of cystadenoma lymp.homatosum, immunoreactive RNase was localized in the ductal epithelium, evidence of the ductal cell origin of these benign tumors. RNase immunoreactivity was noted in the adenomatous structures and in cells isolated in the myxoid matrix of pleomorphlc adenomas, which supports recent evidence of an epithelia! origin o f these tumors. Malignant acinar cells of acinlc cell carcinoma were strongly positive for immunoreactive RNase, while acinar cells of normal parotid were uniformly negative. This expression of the gene for RNase A probably represents a loss of differentiation (i.e., control) of the neoplastic acinar cells. Further evidence for this hypothesis was obtained by treating these tumors with an antihuman salivary amylase antibody, which is localized in normal acinar cells. No immunoreactive amylase was observed. The results support the idea that immunoreactivity need not accompany enzyme activity, as the presence of immunoreactive RNase was noted in all neoplastic tissues examined. Immunohistochemical localization of two antigens in the same tissue demonstrates the varied biochemical changes associated with parotid neoplasia. HUM PATIIOL 14:969--973, 1983.
nant tissues have shown a loss, some benign tumors have manifested normal activity. 5 A further relationship between RNase and neoplasia was indicated with a report of increased levels of serum RNase in h u m a n pancreatic cancer, 9 although this has not been an invariable finding. 1~ EIeT vated serum RNase has also been noted with many other human cancers, including ovarian carcinoma,ll leukemia, 12 and salivary gland tumors, is It should be stressed that serum RNase levels are reliable only in persons with normal kidney function, since various kidney disorders a p p e a r to interfere with normal RNase excretion. 9 Although there is convincing evidence of a correlation between RNase activity and cancer, the source of serum RNase in both the normal and neoplastic states is unknown. Previous immunohistochemical studies in o u r laboratoryl4 have concerned amylase in parotid neoplasia. Whereas amylase is localized in the acinar cells of normal human parotid, neoplastic parotid invariably shows no amylase immunoreactivity. This study utilized antibodies to two antigens, one of which is localized in the ductal epithelium (RNase), the other in acinar cell cytoplasm (amylase). This affords a unique opportunity to examine the cellular origin of various parotid tumors.
Ribonucleases play a vital role in the nucleic acid metabolism of mammalian cells. This role includes cleavage o f primary transcription products, maturation or degradation of various types of intracellular RNA, and digestion. Many authors have suggested that a perturbation in normal RNA metabolism may be related to the neoplastic process. The numerous demonstrations o f loss o f nuclease activity in a variety of malignant tissues have been reviewed. 1 The techniques include digestion of polynucleotide substrate films, 2 immunohistochemical localization, a,4 and biochemical assays. 5,6 An increase in RNase inhibitor has been found in malignant cells, concomitant with the loss of RNase activity. TM Interestingly, while malig-
MATERIALSAND METHODS
Three antibody preparations were used: 1) Commercial bovine ribonuclease A (RNase A), with which rabbits were immunized. Specificity of the antiserum was shown by a single precipitin arc after double diffusion in agar and after immunoelectrophoresis, as well as by absorption of the antibody with RNase A. 2) Amylase antibody, which was prepared and characterized as previously described.14 3) Rabbit a~tihuman pancreatic RNase antiserum. Each preparation served as the primary antibody in the peroxidase-antiperoxidase (PAP) method of Sternberger.15 All tissues were pathologic specimens obtained at Wishard Memorial Hospital, Indianapolis. Normal human parotid salivary glands were obtained from routine autopsies, from two women (aged 27 and 54 years) and three men (aged 24, 51, and 62 years). Benign and malignant tumors of the parotid were obtained as su/~gical specimens (table 1). The tissues
Received July 12, 1982, from the Departments of Medical Genetics and Pathology, Indiana University School of Medicine, Indianapolis, Indiana. Revision accepted for publication September 20, 1982. Supported by Graduate Fellowship NSF SPI 80-19138 from the National Science Foundation. Address correspondence and reprint requests to Dr. M. E. Hodes: Department of Medical Genetics, Riley Research 241, Indiana University School of Medicine, 702 Barnhill Dr, Indianapolis, IN 46223.
969
HUMAN PATHOLOGY
IP , ' : " 1 ~ ,
', - . . - .
" .,--~\
~'.'5:,,-
~-
' ~
-'.:i:,
~.
.:;.
Volume 14, No. 11 (November 1983)
.; .
FIGURE '1 (left). Normal human parotid salivary gland. Immunohistochemical localization of RNase to the ductal epithelium. (PAP technique: antiRNase dilution 1:20; hematoxylin counterstain, x 250.) FIGURE 2 (right). Normal human parotid salivary gland, lmmunohistochemical localization of amylase to ti~e acinar cells. (PAPtechnique; antiomylase dilution 1:100; hematoxylin counterstain, x 250.)
were fixed in 10 per cent neutral buffered formalin and processed with standard histologic dehydration and paraffin embedding techniques. Sections were cut at 4 p.m, placed on albuminized slides, and rehydrated through a series o f graded alcohols into p h o s p h a t e - b u f f e r e d saline, p H 7.2 to 7.4. Sections were incubated in a dilute Pb 2+ solution (10-6M) after the method of Zan-Kowalczewska et al. 16 to inactivate the RNase inhibitor. This was followed by a 30-minute incubation in 0.03 per cent peroxide in methanol to block e n d o g e n o u s peroxidase activity. Nonspecific antibody binding was prevented by 30-minute incubation in 10 per cent normal goat serum. T h e primary antibody was used at concentrations between 1:10 and 1:50 for maximal staining. Incubation was for one to four hours at room temperature. Goat antirabbit IgG (1:20) was followed by rabbit PAP conjugate (Miles Laboratories) (1:50). A solution of 3,3'diaminobenzidine tetrahydrochloride in 0.02 per cent peroxide was placed on the slides for 8 to 10 minutes. Slides were washed, counterstained with hematoxylin, dehydrated, and mounted for viewing under a light microscope. Negative controls were utilized with every immunoperoxidase procedure and consisted of consecutive sections treated with 1) no primary antibody, 2) no goat antirabbit IgG, 3) no primary antibody and no goat antirabbit IgG, 4) preimmune rabbit serum as primary antibody, and 5) absorbed antiserum as primary antibody. Bovine pancreas sections and h u m a n pancreas and parotid gland served as positive controls.
the ductal epithelium, whereas acinar cells were uniformly negative (fig. 1). The localization patterns using antiRNase A serum or antihuman pancreatic RNase serum as the primary antibody were identical, and henceforth the term antiRNase serum will refer to both. Amylase, by contrast, was localized to acinar cells only (fig. 2). TABLE 1. Immunoperoxidase Localization of RNase and Amylase in 20 Parotid Tumors
Age (years)/ Sex Pleomorphic adenoma (n = 10)
RESULTS
Ten sections of each of the five normal human parotid glands manifested immunoreactive RNase in
RNase
Amylase
+ + + + + + + +
-
40/M
+
-
51IM
35lF 64/M 59/M 61/M 42lF 57/M
+ + + + + + + + + + + + + + +
-
Cystadenoma lymphomatosum (Warthin's tumor) (n = 7)
58/F 6 I/M 8 I/M 24lF 31IF 58/M 46/M
+ + + + + +++ + + + + + + + +
-
A c i n i c cell carcinoma (n = 3)
61/M 52lF 49/M
+ + + + + + + + ++++
-
19/M 36/F 56/F
Immunoperoxidase Staining Intensity
-
-
--z
-
-
KEY: - - , n e g a t i v e ; + , f a i n t l y p o s i t i v e ; + + , m o d e r a t e i n t e n sit)'; + + + , h i g h i n t e n s i t y ; + + + + , e x t r e m e l y h i g h i n t e n s i t y .
970
RIBONUCLEASEAND AMYLASEIN PAROT1DGLAND TISSUE[Morley et al.)
At least seven replicate sections o f each tumor were stained for RNase, and three for amylase. The results pertain to all sections of each case studied. Two types of benign parotid salivary gland tumors, pleomorphic adenoma (n = 10) and cystadenoma lymphomatosum (n = 7) were subsequently examined (table 1). Immunoreactive RNase was observed in the neoplastic ductal epithelium in cystadenoma l y m p h o m a t o s u m (fig. 3), whereas immunoreactive amylase was not present. Acinar cells adjoining the tumor were positive for immunoreactive amylase. RNase was localized to the apical cytoplasm in both adenomatous cells (fig. 4) and isolated cells of the mucinous matrix of pleomorphic adenomas (fig. 5). These tumors were devoid of immunoreactive amylase. T h r e e specimens of malignant parotid acinic cell carcinoma were also examined (table 1). While acinar cells of normal parotid are negative for RNase, these neoplastic acinar cells were strongly positive (figs. 6 and 7). DISCUSSION
FIGURE 3. Cystadenoma lymphomatosum [Warthin's tumor) of human parotid salivary gland, showing adjacent normal and tumor tissue. Duct cells of normal parotid and of tumor tissue are positive for immunoreactive RNase. [PAPtechnique; antiRNase dilution 1:20; hematoxylin counterstain, x 100.)
T h e role of RNase in the neoplastic process is ductal cells, the acquisition of RNase immunoreacnot well established. A diminution or loss of RNase tivity by neoplastic acinar cells suggests dedifferentiaactivity wit h malignant transformation has, however, tion toward ductal epithelium. This hypothesis is furbeen amply demonstrated in many tissues by various ther corroborated by the demonstrated lack of imbiochemical and histochemical techniques. 1,3,5We are munoreactive amylase by the same neoplastic cells. not aware of any immunohistochemical localization Amylase is always localized in normal acinar cells 14 of RNase in parotid neoplasia. and is a differentiated function of this cell populaIn this study, immunoreactive RNase was localtion. ized in the duct-like cells of cystadenoma lymphoIt is noteworthy t h a t Daoust and Amano 21 rematosum, a benign tumor thought to originate from ported a loss of RNase activity associated with parotid the ductal epithelium./7 Since normal parotid ductal neoplasia. If they are correct, our results suggest that cells manifest immunoreactive RNase, the neoplastic immunoreactivity does not imply enzyme activity, since process evidently does not affect RNase immunoimmunoreactive RNase was localized in every type of reactivity in epithelial cells. parotid neoplasia examined. Interestingly, Chavez and T h e cell of origin of pleomorphic adenomas has Sheraga 22 have demonstrated separate antigenic and been controversial. 18,19 Some authors have held that catalytic sites in bovine RNase A, a molecule that is myoepithelial cells of mesenchymal origin constitute closely related to and cross-reacts with antihuman the major component of the myxoid stroma of these RNase. 2~ This possible discrepancy between enzyme tumors, whereas the adenomatous structures are epactivity, and immunoreactivity raises an interesting ithelial in origin, hence the term "mixed tumor. ''2~ point worth further study. Unfortunately, RNase is Dardick et al) 9 suggest that evidence is mounting for no longer active in the fixed pathologic specimens an epithelial-derived cell of origin for pleomorphic used in tiffs investigation. adenomas, a cell with the potential for muhidirecGordon and Myers 4 found no immunoreactive tional differentiation. Their ultrastructural observarat liver RNase in liver tumors. RNase A is distinctly tions of cells in the myxoid region of these tumors different from liver-spleen RNase, 23 and we have indicated characteristics of duct epithelium as well as not yet examined the latter species in our specimens. evidence of squamous metaplasia. Our immunohisHansen et al. 3 examined rat pancreas and rat pantochemical results with the "ductal-cell marker" ancreatic acinar cell carcinoma by immunofluorescence tigen RNase are consistent with this hypothesis. Both techniques. They reported the presence of RNase A the adenomatous and isolated cells of the myxoid main both normal and malignant acini, although dimintrix in pleomorphic adenomas are indeed positive for ished immunofluorescence in the malignant acini was RNase. This supports a ductal epithelial origin for observed. We have not examined human pancreatic these tumors. acinar tumors with our RNase A antibody, partly beNormal parotid acinar cells do not contain imcause these tumors are extremely rare. munoreactive RNase, whereas those of acinic cell carCytoplasmic RNase is normally accompanied by cinoma are strongly positive. Since we have shown an endogenous inhibitor that effectively blocks its acRNase to be an antigen localized to normal parotid 971
HUMAN PATHOLOGY ....
,~-~
-_~"7 ~
~-,
~....u~
" ,~t~.~.'s
4, .
.
"~.
"lfi~
-~
~'O~.
"
~
"*-
'-'~,~.-'~
~
--
:.;L
~"-~,')"
, -
~,L/,1~
A~i~'-;'.~'~O~
Volume 14, No. '11 [November 1983] ,
"-
. . . .
"~ ~
"~
,~,7.
~
r..~Z~.,
-
....
TM
" ~
,~. ~
.
~-
' ~
.~
"~-.'.-"o
"
'k
,
~''~u
'~
",~'"
.
~,
.
.
.
.
.
*
..,.,z~,.'.:".~.
~5
: "'-..~rz
__
",
"-,,,./,
.,:_."~J~!,,l~k.\Ul~,~*~..-~
~.~.~ i / ~ ' l
,::.<, ,~
.~
~.,~,
, :~
-'~ A ~ ~,
FIGURE 4 (top ;eft]. Pleomorphic adenoma [mixed tumor] of human parotid salivary gland, lmmunoreactive RNase Is localized to the apical region of adenomatous cells. (PAPtechnique; antiRNase dilution 1:20; hematoxylin counterstain, x 250.] FIGURE 5 (top right]. Pleomorphlc adenoma [mixed tumor) of human parotid salivary gland. Immunohistochemical localization of RNase to pleomorphlc cells of the matrix as well as to adenomatous cells. [PAP technique; antiRNase dilution 1:20; hematoxylin counterstain, x 250.] FIGURE 6 [bottom left]. Acinic cell carcinoma of human parotid salivary gland. Malignant acinar ceils demonstrate immunoreactive RNase. [PAP technique; antiRNase dilution 1:20; hematoxylin counterstain, x 250.] FIGURE 7 [bottom right]. Acinic cell carcinoma of human parotid salivary gland. Malignant border lies diagonally, showing transition of normal parotid acinar cells [upper right] into neoplastic acinar cells [lower left]. A duct stained with RNase appears in the upper right. [PAP technique; anti-RNase dilution 1:20; hematoxyiin counterstain, x 250.]
tivity. Treatments with para-chloromercuribenzoate or a dilute solution of Pb z+ have been shown to serve equally well in inactivating the inhibitor. ]6 Preincubation of fixed tissue sections in a Pb 2+ solution produced no effect on RNase localization patterns. Increased levels of RNase inhibitors have been reported in neoplastic cells. 5,7 Our results indicate that, even if present, inhibitor binding does not affect the antigenic site(s) of the enzyme. It appears possible that a distinction must be made between enzyme activity and immunoreactivity, as in-
dicated above with RNase in neoplastic tissues. Assays of tumor tissue for enzyme activity should be accompanied by an immunohistochemical search for a possibly nonfunctional but cross-reactive protein. The use of RNase A as an antigenic marker for ductal epithelium provides further evidence for the ductal cell origin o f some parotid tumors. Amylase proved a useful adjunct marker for acinar cell differentiation. Combined immunohistochemical results with the two antigens illustrate the well-known association of biochemical alterations and neoplasia. 972
RIBONUCLEASEAND AMYLASEIN PAROTIDGLAND TISSUE[Morley et al.)
Acknowledgment. Rabbit a n t i h u m a n pancreatic RNase antiserum was kindly provided by Dr.Joachim Weichmann, D e p a r t m e n t of Biological Chemistry, University o f California School of Medicine, Los Angeles. REFERENCES 1. Daoust R, deLamirande G: Ribonucleases and neoplasia. Subcell Biochem 4:185, 1975 2. Daoust R: Histochemical comparison of focal losses of RNase and ATPase activity in preneoplastic rat livers. J Histochem Cytochem 27:653, 1979 3. Hansen L, Mangkornkanok/Mark M, Reddy J: Immunohistochemical localization of pancreatic exocrine enzymes in normal and neoplastic pancreatic acinar epithelium of rat. J Histochem Cytochem 29:309, 1981 4. Gordon J, Myers J: Ribonucleases of the rat III: localization by immunofluorescence. Biochim Biophys Acta 113:187, 1966 5. Taper H, Brucher J, Fort L: Activity of alkaline and acid nucleases in tumors of the human central nervous system: histochemical study. Cancer 28:482, 1971 6. Brody S, Balis M: Ribonuclease and deoxyribonuclease activities during normal and neoplastic'growth. Nature 182:940, 1958 7. Chakravorty A, Busch H: Alkaline ribonuclease and ribonuclease inhibitor in nuclear and nucleolar preparations from normal and neoplastic tissues. Cancer Res 27:789, 1967 8. Nagao M, Ichikawa Y: Systemic effect of mouse leukemia on ribonuclease and its inhibitor of host liver and spleen. Gann 60:279, 1969 9. Reddi K, Holland J: Elevated serum ribonuclease in patients with pancreatic cancer. Proc Natl Acad Sci USA 73:2308, 1976 : 10. Peterson L:Serum RNase in the diagnosis of pancreatic carcinoma. Proc Natl Acad Sci USA 76:2630, 1979
973
11. Sheid B, Lu T, Pedrinan L, et al: Plasma ribonuclease: a marker for the detection of ovarian cancer. Cancer 39:2204, 1977 12. Drake W, Schmukler M, Pendergrast W Jr, et ah Abnormal profile of human nucleolytic activity as a test for cancer. J Natl Cancer Inst 55:1055, 1975 13. Chretien P, Matthews W, Twomey P: Serum ribonucleases in cancer: relation to rumor histology. Cancer 31:175, 1973 14. Hodes J, Hull M, Karn R, et ah Immunohistochemical demonstration of the loss of immunoreactive amylase from neoplastic human salivary gland. Experientia 37:184, 1981 15. Sternberger L: Immunocytochemistry. Englewoood Cliffs, NJ, Prentice Hall, 1979 16. Zan-Kowalczewska M, Sierakowska H, Bardon A, et ah Specificities of rat alkaline RNase and cytochemical localization of pancreatic-like RNases. Biochim Biophys Acta 341:138, 1974 17. Thompson A, Bryant H: Histogenesis of papillary cystadenoma 1)'mphomatosmn (Warthin tumor) of the parotid salivary gland. Am J Pathol 26:807, 1950 18. Grishman E: Histochemical analysis of mucopolysaccharides occurring in mucus-producing tumors. Cancer 5:700, 1952 19. Dardick I, Van Nostrand A, Phillips A: Histogenesis of salivary gland pleomorphic adenoma (mixed tumor) with an evaluation of the role of the myoepithelial cell. HuM PATtlOL 13:62, 1982 20. Welsh R, Meyer A: Mixed tumors of human salivary gland: histogenesis. Arch Pathol 85:433, 1968 21. Daoust R, Amano H: Ribonuclease and deoxyribonuclease activities in experimental and human tumors by the histochemical substrate film method. Cancer Res 23:131, 1963 22. Chavez L, Sheraga H: Location of the antigenic determinants of bovine pancreatic ribonuclease. Biochemistry 18:4386, 1979 23. Neuwelt E, Schmukler M, Niziak M, et al: The immunological characterization of several human ribonucleases by using primary binding tests. Biochem J 163:419, 1977
nant tissues have shown a loss, some benign tumors have manifested normal activity. 5 A further relationship between RNase and neoplasia was indicated with a report of increased levels of serum RNase in h u m a n pancreatic cancer, 9 although this has not been an invariable finding. 1~ EIeT vated serum RNase has also been noted with many other human cancers, including ovarian carcinoma,ll leukemia, 12 and salivary gland tumors, is It should be stressed that serum RNase levels are reliable only in persons with normal kidney function, since various kidney disorders a p p e a r to interfere with normal RNase excretion. 9 Although there is convincing evidence of a correlation between RNase activity and cancer, the source of serum RNase in both the normal and neoplastic states is unknown. Previous immunohistochemical studies in o u r laboratoryl4 have concerned amylase in parotid neoplasia. Whereas amylase is localized in the acinar cells of normal human parotid, neoplastic parotid invariably shows no amylase immunoreactivity. This study utilized antibodies to two antigens, one of which is localized in the ductal epithelium (RNase), the other in acinar cell cytoplasm (amylase). This affords a unique opportunity to examine the cellular origin of various parotid tumors.
Ribonucleases play a vital role in the nucleic acid metabolism of mammalian cells. This role includes cleavage o f primary transcription products, maturation or degradation of various types of intracellular RNA, and digestion. Many authors have suggested that a perturbation in normal RNA metabolism may be related to the neoplastic process. The numerous demonstrations o f loss o f nuclease activity in a variety of malignant tissues have been reviewed. 1 The techniques include digestion of polynucleotide substrate films, 2 immunohistochemical localization, a,4 and biochemical assays. 5,6 An increase in RNase inhibitor has been found in malignant cells, concomitant with the loss of RNase activity. TM Interestingly, while malig-
MATERIALSAND METHODS
Three antibody preparations were used: 1) Commercial bovine ribonuclease A (RNase A), with which rabbits were immunized. Specificity of the antiserum was shown by a single precipitin arc after double diffusion in agar and after immunoelectrophoresis, as well as by absorption of the antibody with RNase A. 2) Amylase antibody, which was prepared and characterized as previously described.14 3) Rabbit a~tihuman pancreatic RNase antiserum. Each preparation served as the primary antibody in the peroxidase-antiperoxidase (PAP) method of Sternberger.15 All tissues were pathologic specimens obtained at Wishard Memorial Hospital, Indianapolis. Normal human parotid salivary glands were obtained from routine autopsies, from two women (aged 27 and 54 years) and three men (aged 24, 51, and 62 years). Benign and malignant tumors of the parotid were obtained as su/~gical specimens (table 1). The tissues
Received July 12, 1982, from the Departments of Medical Genetics and Pathology, Indiana University School of Medicine, Indianapolis, Indiana. Revision accepted for publication September 20, 1982. Supported by Graduate Fellowship NSF SPI 80-19138 from the National Science Foundation. Address correspondence and reprint requests to Dr. M. E. Hodes: Department of Medical Genetics, Riley Research 241, Indiana University School of Medicine, 702 Barnhill Dr, Indianapolis, IN 46223.
969
HUMAN PATHOLOGY
IP , ' : " 1 ~ ,
', - . . - .
" .,--~\
~'.'5:,,-
~-
' ~
-'.:i:,
~.
.:;.
Volume 14, No. 11 (November 1983)
.; .
FIGURE '1 (left). Normal human parotid salivary gland. Immunohistochemical localization of RNase to the ductal epithelium. (PAP technique: antiRNase dilution 1:20; hematoxylin counterstain, x 250.) FIGURE 2 (right). Normal human parotid salivary gland, lmmunohistochemical localization of amylase to ti~e acinar cells. (PAPtechnique; antiomylase dilution 1:100; hematoxylin counterstain, x 250.)
were fixed in 10 per cent neutral buffered formalin and processed with standard histologic dehydration and paraffin embedding techniques. Sections were cut at 4 p.m, placed on albuminized slides, and rehydrated through a series o f graded alcohols into p h o s p h a t e - b u f f e r e d saline, p H 7.2 to 7.4. Sections were incubated in a dilute Pb 2+ solution (10-6M) after the method of Zan-Kowalczewska et al. 16 to inactivate the RNase inhibitor. This was followed by a 30-minute incubation in 0.03 per cent peroxide in methanol to block e n d o g e n o u s peroxidase activity. Nonspecific antibody binding was prevented by 30-minute incubation in 10 per cent normal goat serum. T h e primary antibody was used at concentrations between 1:10 and 1:50 for maximal staining. Incubation was for one to four hours at room temperature. Goat antirabbit IgG (1:20) was followed by rabbit PAP conjugate (Miles Laboratories) (1:50). A solution of 3,3'diaminobenzidine tetrahydrochloride in 0.02 per cent peroxide was placed on the slides for 8 to 10 minutes. Slides were washed, counterstained with hematoxylin, dehydrated, and mounted for viewing under a light microscope. Negative controls were utilized with every immunoperoxidase procedure and consisted of consecutive sections treated with 1) no primary antibody, 2) no goat antirabbit IgG, 3) no primary antibody and no goat antirabbit IgG, 4) preimmune rabbit serum as primary antibody, and 5) absorbed antiserum as primary antibody. Bovine pancreas sections and h u m a n pancreas and parotid gland served as positive controls.
the ductal epithelium, whereas acinar cells were uniformly negative (fig. 1). The localization patterns using antiRNase A serum or antihuman pancreatic RNase serum as the primary antibody were identical, and henceforth the term antiRNase serum will refer to both. Amylase, by contrast, was localized to acinar cells only (fig. 2). TABLE 1. Immunoperoxidase Localization of RNase and Amylase in 20 Parotid Tumors
Age (years)/ Sex Pleomorphic adenoma (n = 10)
RESULTS
Ten sections of each of the five normal human parotid glands manifested immunoreactive RNase in
RNase
Amylase
+ + + + + + + +
-
40/M
+
-
51IM
35lF 64/M 59/M 61/M 42lF 57/M
+ + + + + + + + + + + + + + +
-
Cystadenoma lymphomatosum (Warthin's tumor) (n = 7)
58/F 6 I/M 8 I/M 24lF 31IF 58/M 46/M
+ + + + + +++ + + + + + + + +
-
A c i n i c cell carcinoma (n = 3)
61/M 52lF 49/M
+ + + + + + + + ++++
-
19/M 36/F 56/F
Immunoperoxidase Staining Intensity
-
-
--z
-
-
KEY: - - , n e g a t i v e ; + , f a i n t l y p o s i t i v e ; + + , m o d e r a t e i n t e n sit)'; + + + , h i g h i n t e n s i t y ; + + + + , e x t r e m e l y h i g h i n t e n s i t y .
970
RIBONUCLEASEAND AMYLASEIN PAROT1DGLAND TISSUE[Morley et al.)
At least seven replicate sections o f each tumor were stained for RNase, and three for amylase. The results pertain to all sections of each case studied. Two types of benign parotid salivary gland tumors, pleomorphic adenoma (n = 10) and cystadenoma lymphomatosum (n = 7) were subsequently examined (table 1). Immunoreactive RNase was observed in the neoplastic ductal epithelium in cystadenoma l y m p h o m a t o s u m (fig. 3), whereas immunoreactive amylase was not present. Acinar cells adjoining the tumor were positive for immunoreactive amylase. RNase was localized to the apical cytoplasm in both adenomatous cells (fig. 4) and isolated cells of the mucinous matrix of pleomorphic adenomas (fig. 5). These tumors were devoid of immunoreactive amylase. T h r e e specimens of malignant parotid acinic cell carcinoma were also examined (table 1). While acinar cells of normal parotid are negative for RNase, these neoplastic acinar cells were strongly positive (figs. 6 and 7). DISCUSSION
FIGURE 3. Cystadenoma lymphomatosum [Warthin's tumor) of human parotid salivary gland, showing adjacent normal and tumor tissue. Duct cells of normal parotid and of tumor tissue are positive for immunoreactive RNase. [PAPtechnique; antiRNase dilution 1:20; hematoxylin counterstain, x 100.)
T h e role of RNase in the neoplastic process is ductal cells, the acquisition of RNase immunoreacnot well established. A diminution or loss of RNase tivity by neoplastic acinar cells suggests dedifferentiaactivity wit h malignant transformation has, however, tion toward ductal epithelium. This hypothesis is furbeen amply demonstrated in many tissues by various ther corroborated by the demonstrated lack of imbiochemical and histochemical techniques. 1,3,5We are munoreactive amylase by the same neoplastic cells. not aware of any immunohistochemical localization Amylase is always localized in normal acinar cells 14 of RNase in parotid neoplasia. and is a differentiated function of this cell populaIn this study, immunoreactive RNase was localtion. ized in the duct-like cells of cystadenoma lymphoIt is noteworthy t h a t Daoust and Amano 21 rematosum, a benign tumor thought to originate from ported a loss of RNase activity associated with parotid the ductal epithelium./7 Since normal parotid ductal neoplasia. If they are correct, our results suggest that cells manifest immunoreactive RNase, the neoplastic immunoreactivity does not imply enzyme activity, since process evidently does not affect RNase immunoimmunoreactive RNase was localized in every type of reactivity in epithelial cells. parotid neoplasia examined. Interestingly, Chavez and T h e cell of origin of pleomorphic adenomas has Sheraga 22 have demonstrated separate antigenic and been controversial. 18,19 Some authors have held that catalytic sites in bovine RNase A, a molecule that is myoepithelial cells of mesenchymal origin constitute closely related to and cross-reacts with antihuman the major component of the myxoid stroma of these RNase. 2~ This possible discrepancy between enzyme tumors, whereas the adenomatous structures are epactivity, and immunoreactivity raises an interesting ithelial in origin, hence the term "mixed tumor. ''2~ point worth further study. Unfortunately, RNase is Dardick et al) 9 suggest that evidence is mounting for no longer active in the fixed pathologic specimens an epithelial-derived cell of origin for pleomorphic used in tiffs investigation. adenomas, a cell with the potential for muhidirecGordon and Myers 4 found no immunoreactive tional differentiation. Their ultrastructural observarat liver RNase in liver tumors. RNase A is distinctly tions of cells in the myxoid region of these tumors different from liver-spleen RNase, 23 and we have indicated characteristics of duct epithelium as well as not yet examined the latter species in our specimens. evidence of squamous metaplasia. Our immunohisHansen et al. 3 examined rat pancreas and rat pantochemical results with the "ductal-cell marker" ancreatic acinar cell carcinoma by immunofluorescence tigen RNase are consistent with this hypothesis. Both techniques. They reported the presence of RNase A the adenomatous and isolated cells of the myxoid main both normal and malignant acini, although dimintrix in pleomorphic adenomas are indeed positive for ished immunofluorescence in the malignant acini was RNase. This supports a ductal epithelial origin for observed. We have not examined human pancreatic these tumors. acinar tumors with our RNase A antibody, partly beNormal parotid acinar cells do not contain imcause these tumors are extremely rare. munoreactive RNase, whereas those of acinic cell carCytoplasmic RNase is normally accompanied by cinoma are strongly positive. Since we have shown an endogenous inhibitor that effectively blocks its acRNase to be an antigen localized to normal parotid 971
HUMAN PATHOLOGY ....
,~-~
-_~"7 ~
~-,
~....u~
" ,~t~.~.'s
4, .
.
"~.
"lfi~
-~
~'O~.
"
~
"*-
'-'~,~.-'~
~
--
:.;L
~"-~,')"
, -
~,L/,1~
A~i~'-;'.~'~O~
Volume 14, No. '11 [November 1983] ,
"-
. . . .
"~ ~
"~
,~,7.
~
r..~Z~.,
-
....
TM
" ~
,~. ~
.
~-
' ~
.~
"~-.'.-"o
"
'k
,
~''~u
'~
",~'"
.
~,
.
.
.
.
.
*
..,.,z~,.'.:".~.
~5
: "'-..~rz
__
",
"-,,,./,
.,:_."~J~!,,l~k.\Ul~,~*~..-~
~.~.~ i / ~ ' l
,::.<, ,~
.~
~.,~,
, :~
-'~ A ~ ~,
FIGURE 4 (top ;eft]. Pleomorphic adenoma [mixed tumor] of human parotid salivary gland, lmmunoreactive RNase Is localized to the apical region of adenomatous cells. (PAPtechnique; antiRNase dilution 1:20; hematoxylin counterstain, x 250.] FIGURE 5 (top right]. Pleomorphlc adenoma [mixed tumor) of human parotid salivary gland. Immunohistochemical localization of RNase to pleomorphlc cells of the matrix as well as to adenomatous cells. [PAP technique; antiRNase dilution 1:20; hematoxylin counterstain, x 250.] FIGURE 6 [bottom left]. Acinic cell carcinoma of human parotid salivary gland. Malignant acinar ceils demonstrate immunoreactive RNase. [PAP technique; antiRNase dilution 1:20; hematoxylin counterstain, x 250.] FIGURE 7 [bottom right]. Acinic cell carcinoma of human parotid salivary gland. Malignant border lies diagonally, showing transition of normal parotid acinar cells [upper right] into neoplastic acinar cells [lower left]. A duct stained with RNase appears in the upper right. [PAP technique; anti-RNase dilution 1:20; hematoxyiin counterstain, x 250.]
tivity. Treatments with para-chloromercuribenzoate or a dilute solution of Pb z+ have been shown to serve equally well in inactivating the inhibitor. ]6 Preincubation of fixed tissue sections in a Pb 2+ solution produced no effect on RNase localization patterns. Increased levels of RNase inhibitors have been reported in neoplastic cells. 5,7 Our results indicate that, even if present, inhibitor binding does not affect the antigenic site(s) of the enzyme. It appears possible that a distinction must be made between enzyme activity and immunoreactivity, as in-
dicated above with RNase in neoplastic tissues. Assays of tumor tissue for enzyme activity should be accompanied by an immunohistochemical search for a possibly nonfunctional but cross-reactive protein. The use of RNase A as an antigenic marker for ductal epithelium provides further evidence for the ductal cell origin o f some parotid tumors. Amylase proved a useful adjunct marker for acinar cell differentiation. Combined immunohistochemical results with the two antigens illustrate the well-known association of biochemical alterations and neoplasia. 972
RIBONUCLEASEAND AMYLASEIN PAROTIDGLAND TISSUE[Morley et al.)
Acknowledgment. Rabbit a n t i h u m a n pancreatic RNase antiserum was kindly provided by Dr.Joachim Weichmann, D e p a r t m e n t of Biological Chemistry, University o f California School of Medicine, Los Angeles. REFERENCES 1. Daoust R, deLamirande G: Ribonucleases and neoplasia. Subcell Biochem 4:185, 1975 2. Daoust R: Histochemical comparison of focal losses of RNase and ATPase activity in preneoplastic rat livers. J Histochem Cytochem 27:653, 1979 3. Hansen L, Mangkornkanok/Mark M, Reddy J: Immunohistochemical localization of pancreatic exocrine enzymes in normal and neoplastic pancreatic acinar epithelium of rat. J Histochem Cytochem 29:309, 1981 4. Gordon J, Myers J: Ribonucleases of the rat III: localization by immunofluorescence. Biochim Biophys Acta 113:187, 1966 5. Taper H, Brucher J, Fort L: Activity of alkaline and acid nucleases in tumors of the human central nervous system: histochemical study. Cancer 28:482, 1971 6. Brody S, Balis M: Ribonuclease and deoxyribonuclease activities during normal and neoplastic'growth. Nature 182:940, 1958 7. Chakravorty A, Busch H: Alkaline ribonuclease and ribonuclease inhibitor in nuclear and nucleolar preparations from normal and neoplastic tissues. Cancer Res 27:789, 1967 8. Nagao M, Ichikawa Y: Systemic effect of mouse leukemia on ribonuclease and its inhibitor of host liver and spleen. Gann 60:279, 1969 9. Reddi K, Holland J: Elevated serum ribonuclease in patients with pancreatic cancer. Proc Natl Acad Sci USA 73:2308, 1976 : 10. Peterson L:Serum RNase in the diagnosis of pancreatic carcinoma. Proc Natl Acad Sci USA 76:2630, 1979
973
11. Sheid B, Lu T, Pedrinan L, et al: Plasma ribonuclease: a marker for the detection of ovarian cancer. Cancer 39:2204, 1977 12. Drake W, Schmukler M, Pendergrast W Jr, et ah Abnormal profile of human nucleolytic activity as a test for cancer. J Natl Cancer Inst 55:1055, 1975 13. Chretien P, Matthews W, Twomey P: Serum ribonucleases in cancer: relation to rumor histology. Cancer 31:175, 1973 14. Hodes J, Hull M, Karn R, et ah Immunohistochemical demonstration of the loss of immunoreactive amylase from neoplastic human salivary gland. Experientia 37:184, 1981 15. Sternberger L: Immunocytochemistry. Englewoood Cliffs, NJ, Prentice Hall, 1979 16. Zan-Kowalczewska M, Sierakowska H, Bardon A, et ah Specificities of rat alkaline RNase and cytochemical localization of pancreatic-like RNases. Biochim Biophys Acta 341:138, 1974 17. Thompson A, Bryant H: Histogenesis of papillary cystadenoma 1)'mphomatosmn (Warthin tumor) of the parotid salivary gland. Am J Pathol 26:807, 1950 18. Grishman E: Histochemical analysis of mucopolysaccharides occurring in mucus-producing tumors. Cancer 5:700, 1952 19. Dardick I, Van Nostrand A, Phillips A: Histogenesis of salivary gland pleomorphic adenoma (mixed tumor) with an evaluation of the role of the myoepithelial cell. HuM PATtlOL 13:62, 1982 20. Welsh R, Meyer A: Mixed tumors of human salivary gland: histogenesis. Arch Pathol 85:433, 1968 21. Daoust R, Amano H: Ribonuclease and deoxyribonuclease activities in experimental and human tumors by the histochemical substrate film method. Cancer Res 23:131, 1963 22. Chavez L, Sheraga H: Location of the antigenic determinants of bovine pancreatic ribonuclease. Biochemistry 18:4386, 1979 23. Neuwelt E, Schmukler M, Niziak M, et al: The immunological characterization of several human ribonucleases by using primary binding tests. Biochem J 163:419, 1977