Transitional Cell Carcinoma Characterization © 1990 S. Karger A G . Basel 1015-2008/90/0586-0304$2.75/0
Pathobiology 1990;58:304-311
Discrimination of Cell Types in Primary Transitional Cell Carcinoma by Monoclonal Anti-Cytokeratin Antibodies1,2 Edward A. Reedya 3, Kristin L. Colombo-Burkeb, James H. ResaiC 1 Department of Pathology, School of Medicine, University of Maryland at Baltimore, and bUniversity of Maryland Medical System, Department of Anatomic Pathology, Baltimore, Md., USA
Key Words. Cytokeratin • Bladder • Transitional cell carcinoma • Cytology • Immunocytochemistry • Antibodies • Differentiation Abstract. The morphologic discrimination between low and high grade malignant tumor cells arising in the uri nary bladder is a major diagnostic problem in cytopathology. Using immunochemical peroxidase staining of cytokeratins (CKs) of human bladder exfoliative cytology specimens, we have been able to disciminate between transi tional cell carcinoma cells, atypical cells and normal bladder cells. Commercially available monoclonal antibodies used in this study were: anti-CK 13 (Sigma K8.12), anti-CK 5, 7 and 8 (Sigma K8.13), anti-CK 19 (Sigma K4.62), and anti-CK 18 (Transformation Res. 1091). When normal bladder cells are stained with these antibodies, CK 18 is specific for superficial cells; CK 13 is specific for the basal type cells and CKs 5, 7, 8 and 19 are expressed by all urothelial cell types. Four cases diagnosed by cytopathological criteria as 'atypical' and 7 diagnosed as ‘positive’ (various grades) were used in this study. Cytologic diagnosis was confirmed by histopathology in 7 cases. Tissue was not available for histopathology in 4 cases. Malignant cells with a ‘basal’ or ‘immature’ phenotype preferentially stained with CK 13 and were associated with increased metastatic or malignant potential. Patients with higher grade tumors had more cells positive for CK 13 than did patients with atypical or lower grade malignancies. Patients with well-differentiated, low grade tumors predominantly shed cells that expressed CK 18 and CK 19. High grade, inva sive bladder lesions were characterized by many cells expressing CK 13, and fewer cells expressing CK 19. The combination of diagnosis by morphologic criteria on Papanicolaou-stained specimens with immunochemical char acterization of the same cells for CKs facilitate an accurate diagnosis of bladder lesions.
Exfoliative cytology of bladder washings and voidings has been shown to be an effective means of monitoring patients for bladder carcinoma [10]. Cytopathological techniques have been enhanced by applying mono- and polyclonal antibodies to cytologic preparations from var ious body sites to detect specific lesions or cell types [4, 9, 12, 18]. Segasothy et al. [ 17] have advocated the use of monoclonal antibodies to identify different cell types found in urine to aid in the diagnosis of nephritis. Normal human urothelium is known to be a ‘com plex’ epithelium, characterized by ‘simple’ and ‘squa mous’ epithelium cytokeratin (CK) subtypes [13]. The transitional epithelium of the urinary bladder (urothe
lium) is composed of several specialized cell types. Pre vious studies have shown that different cell layers of the urothelium express various CK subtypes which can be detected and discriminated from each other by immuno chemical and biochemical methods [13, 14, 16]. Bladder cytology from washings, voidings and cathe terizations is an effective means of evaluating bladder tumors. Immunocytochemical techniques can be easily applied to all these types of specimens. Since alcohol fix1 Supported in part by funding from the National Disease Re source Interchange (NDRI). 2 A portion of these results were presented in abstract format at the 1989 American Society of Cytology Annual Meeting. 3 Submitted by E.A.R. to the University of Maryland as partial fulfillment for the degree of doctor of philosophy.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
Introduction
Cytokeratin Expression in Bladder Cytopreparations
Materials and Methods Specimens All cytology and surgical pathology specimens used in this study were obtained from the University of Maryland Hospital, Depart ment of Pathology. The protocols were reviewed and approved by the Human Volunteers Research Committee at the University of Maryland. Ethanol-fixed, Papanicolaou (Pap)-stained cytoprepara tions were examined and characterized by standard criteria [5, 8], Abnormal diagnoses were correlated with surgical specimens from 7 of the 11 cases, and the cytology diagnosis and the surgical pathol ogy diagnosis were compared. Normal human bladder was obtained at immediate autopsy [19] within 1 h of somatic death from a 19-year-old white male with no previous history of urinary tract disease. Urinary bladder mucosa was fixed in 100% (absolute) ethanol, and processed and sectioned for light microscopy. Histologic sections were evaluated microscop ically by HE staining. Next, serial 5-p.m paraffin sections were cut and stained using the immunohistochemical technique described below. Stains Used Pap Stain. The Pap stain is the standard cytological multichromatic stain made up of aqueous hematoxylin and ethanolic counter stains. Specimens included in this study were stained in the follow ing manner: Gill’s hematoxylin, 90 s; tap water rinse until clear; 95% ETOH. 60s; OG-6, 2 min; 95% ETOH. 60s; modified EA, 5 min; 95% ETOH, two changes, 60s each; 100% ETOH. two changes, 60 s each; xylene, two changes, 3 min each. Specimens were initially screened and diagnosed using this stain.
Immunochemical Stains. Monoclonal mouse anti-CK antibodies were used in an indirect immunochemical procedure for both tissue sections and cytopreparations. When tissue sections were stained, the tissue was deparaffinized with three changes of xylene and the xylene was removed with three changes of 100% ethanol. For immunostaining, both tissue sections and cytology specimens were treated with 3% H 2 O 2 for 20 min, taken to water, then rinsed in PBS (pH 7.4) for 5 min. Nonspecific protein binding was blocked on tis sue sections by incubating with normal nonimmune rabbit serum (Dako Corp., Carpinteria. Calif.) for 15 min. The primary anti bodies used were K8.12 (anti-CK 13), K8.I3 (anti-CK 5, 7 and 8). K4.62 (anti-CK 19) (Sigma Chemical Corp., St. Louis, Mo.), and 1091 (anti-CK 18) (Transformation Research, Framingham, Mass.). All primary antibodies were used at a dilution of 1:20 in PBS except 1091, which was 1:100, and all were incubated for 2 h at room tem perature. All primary antibody reactivities were determined by Western blot analysis using normal human urothelial CKs [see 16]. Specimens were washed with 50 mM Tris-HCl buffer (pH 7.6) be tween antibody incubation steps. Rabbit anti-mouse IgG horserad ish peroxidase-conjugated secondary antibody (Dako) was used at a 1:80 dilution in PBS and incubated for 30 min at room temperature. CK filaments were localized by incubating for 10 min with 0.05% (w/v) 3,3'-diaminobenzidine (Sigma) in Tris buffer freshly prepared and filtered with 0.03% (v/v) H;02. Specimen Evaluation The cytopathology diagnosis was performed on Pap-stained specimens with criteria described by Frost [5] and Kem [8]. An ‘atypical’ diagnosis was based on the presence of slightly to moder ately irregular and enlarged nuclei with hyperchromasia and in creased chromatin granularity. A ‘suspicious’ diagnosis was made when nuclei exhibited coarser chromatin granules, marked hyper chromasia, an increased nuclcar/cytoplasmic (N/C) ratio, and when an abundant number of atypical cells or cell groups were present. A specimen was judged ‘positive’ for malignancy if the cells present exhibited severely atypical nuclear morphology, a very high N/C ratio, and many atypical cells or cell groups, or numerous mitotic figures (fig. 4c). Immunostained specimens were evaluated by counting 500 cells from each slide on each case. Urothelial cells were scored as staining positive or negative, and the percent positive cells were determined. The immunostaining results, cytology diagnosis and surgical pathol ogy diagnosis were compared and correlated.
Results Normal human urothelium is composed of three dis tinct cell types: large, often multinucleated superficial (umbrella) cells at the lumenal surface; intermediate cells, and small basal cells with dense nuclei lying on the basal lamina (fig. 1). In ethanol-fixed paraffin sections of normal human bladder, only the superficial cells stain positive with the antibody 1091 (fig. 2a, table 1). Based on Western blot ting results with this antibody [16], superficial cells
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
ation is used predominantly for cytological preparations, retrospective studies are possible and appropriate for immunocytochemical staining. Prior Papanicolaou staining does not interfere with antibody binding [18]. Our laboratory previously has shown in a rat bladder tumor model that, as bladder carcinomas progress, the developing tumors express unique CK patterns; and CK expression was positively associated with the invasive potential of these induced primary bladder tumors [16]. This report describes how this data, applied to human transitional cell carcinomas, aids in the diagnosis and treatment of the patient. In this study, cytopreparations (direct smear or cytospin) of bladder voidings, washings and catheterizations were immunochemically stained with a panel of four commercially available monoclonal anti-CK antibodies and correlated with the cytopathology diagnosis. These results were then compared to the histopathologic diag nosis. Using the panel of antibodies utilized by Reedy et al. [ 16], different urothelial cell types could be identified in immunostained cytopreparations. The cell types iden tified in the immunostained cytopreprations were corre lated with the histopathologic diagnosis.
305
306
Reedy/Colombo-Burke/Resau
Table 1. CK subtypes and antibody reactivities in normal human bladder CK
Antibody
Histological location
5. 7. 8 19 13 18
Sigma K8.13 Sigma K4.62 Sigma K8.12 TR 1091
all layers all layers basal superficial
Fig. 1. Smear preparation of normal human urinary bladder from a 19-year-old white male, 2 h postmortem, with no history of urinary tract disease. Numbered arrowheads indicate the different cell types present: large superficial cells are often multinucleated (arrowhead I), intermediate cells (arrowhead 2), and small basal cells with dense nuclei (arrowhead 3). Pap stain. X 40.
Fig. 2. Immunostained ethanol-fixed normal human bladder mucosa from a 19-year-old white male, 2 h postmortem, with no history of urinary tract disease, a Antibody 1091 (anti-CK 18) reacts specifically and intensely with the superficial cells of the urothelium. b Antibody K8.12 (anti-CK 13) reacts specifically with the
d basal and some intermediate ceils of the urothelium. c Antibody K4.62 (anti-CK 19) reacts with all cell layers of the urothelium. d Antibody K8.13 (anti-CK 5, 7 and 8) reacts intensely with all cell layers of the urothelium. Indirect immunoperoxidase, hematoxylin counterstain. X 40.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
b
307
Cytokeratin Expression in Bladder Cytopreparations
express CK 18, while other cells in the urothelium ex press CK 18 but in amounts below the levels of detection with immunohistochemical techniques [13]. Patients with a papillary carcinoma of the bladder had 49% (pa tient 5) to 85% (patient 2) of cells in the urine positive for CK 18 (table 2). In normal bladder, these cells corre spond to those found on the lumenal surface of the blad der (fig. 2a, 3a, table 1). When antibody K8.12 was used to stain ethanol-fixed paraffin-embedded sections of normal human bladder, only the basal layers of cells were reactive (fig. 2b, table 1). Therefore, based on the reported reactivities of this antibody [7, 16], these cells express CK 13. In immunostained cytopreparations, the percentage of cells posi tive for basal CK (CK 13) ranged from 34% (patient 3) to 100% (patients 4 and 7). Generally, the number of
positive cells increased as the histologic grade (malignant potential) increased (patients 1-3, 5-7). Patients 4 and 7 had 100% of cells positive for basal CKs. By histopathology, patient 4 had a carcinoma in situ in addition to a papillary transitional cell carcinoma. Patient 7 was diag nosed to have a nonpapillary transitional cell carcinoma by histopathology (table 2). Previous reports have established the reactivity of the antibody K4.62 as specific for CK 19 [6. 16], Immuno histochemical results of normal bladder stained with the antibody K4.62 indicate that all cell layers in the urothe lium express CK 19 (fig. 2c, table 1). Of patients whose surgical diagnosis was known, 6 of 7 had 100% of the cells counted as positive for CK 19. The 1 patient (No. 7) that did not have all cells positive was diagnosed to have an invasive transitional cell carcinoma (table 2).
b
Fig. 3. Immunostained cytopreparations of a bladder washing from a 66-year-old male (patient 1). See table 2 for diagnoses. The following immunostaining pattern is typical of low grade, noninvasive, papillary transitional cell carcinoma, a Antibody 1091 (anti-CK 18) intensely stains the large, multinucleated superficial (‘umbrella’) cells, while the squamous cells (arrows) are unreactive. b A group of small cells resembling basal cells are nonreac-
tive with the antibody 1091. c Cells resembling superficial cells are unstained with the antibody K.8.12 (arrows) (anti-CK 13), while similar appearing cells and a group of small basal-like cells stain intensely, d A urothelial cell is positive when stained with the antibody K4.62 (anti-CK 19), but squamous cells (arrows) are negative. Indirect immunoperoxidase, hematoxylin counterstain. X 40.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
d
Reedy/Colombo-Burke/Resau
308
Table 2. Comparison of immunochemical staining results with cytology and surgical diagnosis Age/sex
Method
Cytology screening diagnosis
Surgical pathology diagnosis
Immunochemical staining percent positive cells/500 cells 1091
K.8.12
K8.I3
K.4.62
1
66/M
wash
atypical transitional cells present
grade II papillary transitional cell carcinoma, no invasion
83
56
100
100
2
67/M
wash
atypical transitional cells present
grade II papillary transitional cell carcinoma, no invasion
85
63
100
100
3
63/M
cath
suspicious for malignancy: grade I papillary transitional groups of atypical cell carcinoma, no invasion transitional cells are present
67
34
100
100
4
67/M
wash
positive for malignancy: cells are present which represent a transitional cell carcinoma
flat transitional cell carcinoma in situ and grade I papillary transitional cell carcinoma, no invasion
69
100
100
100
5
73/M
wash
positive for malignancy: cells are present which represent a transitional cell carcinoma
grade II papillary transitional cell carcinoma, no invasion
49
53
100
100
6
71/M
void
positive for malignancy: cells are present which represent a transitional cell carcinoma
grade II/III papillary transitional cell carcinoma, no invasion
51
62
100
100
7
39/M
cath
positive for malignancy: cells are present which represent a transitional cell carcinoma
grade III infiltrating transitional cell carcinoma with squamous features, smooth muscle invasion
50
100
100
43
Based on immunohistopathology, all cell layers of the urothelium stain strongly with antibody K.8.13 in etha nol-fixed paraffin sections (fig. 2d, table 1). Squamous cells found in cytopreparations stain weakly or not at all (fig. 3a, b). Only cells that displayed a urothelial pheno type and stained strongly with K8.13 were included in this data. Cells that possessed a squamous morphology (flat with pyknotic nucleus) or were negative for K.8.13 were excluded from the cells counted because they were npt considered urothelial cells. Therefore, all cells (100%) included in this study’s data stained positive with the antibody K8.13 (table 2).
Discussion Previous studies have attempted to identify CKs as markers for transitional cell carcinoma. Baker et al. [1] correlated the relative amount of CK 18 in transitional
cell carcinoma using a radioimmunoassay (RIA) system. Basta et al. [2] used an ELISA system to detect the pres ence of keratin in the urine of patients with transitional cell carcinoma. Both groups identified CK as a marker of urothelial malignancy but did not specifically identify cancer cells. More definite results regarding the expression of spe cific CKs in bladder carcinomas have been presented by Ramaekers et al. [14, 15], Cintorino et al. [3], and Moll et al. [13], Generally, all the reports confirm that CK expression is modulated in transitional cell carcinomas of various grades, and that it may be possible to identify the grade of the tumor based on the CK expression observed using immunochemical staining. All three re ports also identify CK 18 as being expressed in greater amounts as the tumor grade increases. Moll et al. [13] and Cintorino et al. [3] both described the novel expres sion of CK 17 in transitional cell carcinomas. Moll et al. [13] presented data that shows CK 13 expression de
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
Patient
Cytokeratin Expression in Bladder Cytopreparations
V
Fig. 4. Immunostained cytopreparations of a cathetcrized blad der specimen from a 39-year-old male (patient 7). See table 2 for diagnosis. The following immunostaining pattern is typical of inva sive, nonpapillary transitional cell carcinomas, a A small, malignant cell is nonreactive with antibody 1091 (anti-CK 18). b A urothelial cell reacts weakly with antibody K4.62 (anti-CK 19). c A cluster of small, malignant cells stain intensely with antibody K8.12 (anti-CK 13). Indirect immunoperoxidase, hematoxylin counterstain. X 40.
increased numbers of CK 13-positive cells indicate that a carcinoma in situ could be present in addition to a low grade (I—II) papillary transitional cell carcinoma (see also patient 3, table 2). Patient 11: the lack of cells posi tive for CK 18, the increased numbers of CK 13-positive cells, and the decreased numbers of CK 19-positive cells suggest, that a nonpapillary invasive carcinoma may be present in this patient (see also patient 7, table 2). In the hope that these predictions are substantiated, these pa tients’ histories are being followed for the histopathological results of a tissue biopsy.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
creasing as the tumor grade increases. Reedy et al. [ 16] described a similar pattern of CK 19 expression in blad der tumors induced in rats. In the present study using immunostained cytoprepa rations and histological sections of normal human blad der. the following general conclusions were drawn: All urothelial cells were positive with the antibody K.8.13 (CKs 5, 7 and 8) (fig. 2d). Cells that were not positive are not urothelial cells (fig. 3a. d). Basal urothelial cells stained positive for CK 13 (Sigma K8.12) (fig. 2b, 3c, 4c). Superficial urothelial cells stained positive for CK 18 (TR 1091) (fig. 2a, 3a). The degree of transitional cell carcinoma differentia tion is positively associated with the percentage of cells positive for CK 18 (TR 1091) and negatively associated with the percentage of cells positive for CK 13 (Sigma K8.12) (table 2). Specimens from patients with well-differentiated pa pillary transitional cell carcinomas were characterized by a substantial percentage of cells positive for CK 18 (TR 1091) (fig. 2a), and a low percentage of cells positive for CK 13 (Sigma K8.12) (fig. 3c). Specimens from pa tients with less well-differentiated papillary transitional cell carcinomas were characterized by fewer cells posi tive for CK 18 and more cells positive for CK 13 (fig. 4). Cytology specimens from the patient in our study with invasive transitional cell carcinoma were character ized by a decrease (43% compared to 100%) in the per centage of cells expressing CK 19 (Sigma K4.62) (pa tient 7, table 2). In studies with rat bladder tumors in duced in vivo, we observed that as malignant potential increased, the expression of CK 19 decreased [16]. The staining pattern of patient 7 is consistent with these results (43% of cells positive for CK 19) and is supported by histopathology (infiltrating carcinoma). Using information based on the data presented in table 2, the following predictions of tumors present in patients 8-11 (table 3) are made: Patient 8: the absence of cells positive for CK 18, the increased numbers of CK 13-positive cells, and the decreased numbers of CK 19positive cells, suggest that a nonpapillary invasive carci noma is present in this patient (see also patient 7, table 2). Patient 9: the CK staining pattern of this patient indicates that a low grade (I—II) papillary transitional cell carcinoma could be present in this patient (see also patients 1 and 2. table 2). Subsequent to analysis, the bladder tumor of this patient was evaluated by histopa thology, and was diagnosed as a papillary transitional cell carcinoma, grade I (no invasion). Patient 10: the
309
310
Reedy/Colombo-Burke/Resau
Table 3. Comparison of immunochemical staining results with cytology diagnosis alone Patient
Age/sex
Method
Immunochemical staining percent positive cells/500 cells
Cytology screening diagnosis
1091
K8.12
K8.13
K4.62
8
53/F
wash
atypical transitional cells present
0
40
100
25
9'
43/F
cath
atypical transitional cells present
73
54
100
100
10
66/M
cath
suspicious for malignancy: significantly atypical transitional cells are present
90
100
100
100
11
90/M
void
positive for malignancy: malignant cells are present which represent a transitional cell carcinoma
0
50
100
30
1 Patient 9 was evaluated by histopathology following this analysis and was diagnosed as having a papillary transitional cell carcinoma, grade I (no invasion).
References 1 Baker. W.; White, R.; Rossitto. P.; Min. B.; Cardiff. R.: Quanti tative analysis of keratin 18 in the urine of patients with bladder cancer. J. Urol. 140: 436-439(1988). 2 Basta, M.; Attallah, A.; Seddek. M.; El-Mohamady, H.; Al-Hilaly, E.; Atwaan, N.; Ghoneiwm, M.: Cytokeratin shedding in urine: A biological marker for bladder cancer? Br. J. Urol. 61: 116-121 (1988). 3 Cintorino, M.; Vecchio, M.; Bugnoli, M.; Petracca, R.: Cytoker atin pattern in normal and pathological bladder urothelium: Immunohistochemical investigation using monoclonal anti bodies. J. Urol. 139: 428-432 (1988). 4 Dalquen, P.; Bittel, D.; Gudat, F.; Overbeck. J.; Heitz, P.: Com bined immunoreaction and Papanicolaou’s stain on cytological smears. Pathol. Res. Pract. 181: 50-54 (1986). 5 Frost, J.: The cell in health and disease. An evaluation of cellular morphologic expression of biologic behavior; in Wied, G.; von Hamm, E.; Koss, L.; Reagan, J. (eds): Monographs in Clinical Cytology; 2nd ed., vol. 2 (Karger, Basel, 1986). 6 Gigi-Leitner. O.; Geiger, B.: Antigenic interrelationship between the 40-kilodalton cytokeratin polypeptide and desmoplakins. Cell Motil. Cytoskeleton 6: 628-639 (1986). 7 Huszar. M.; Gigi-Leitner. O.; Moll. R.: Franke. W.; Geiger, B.: Monoclonal antibodies to various acidic (type I) cytokeratins of stratified epithelia: Selective markers for stratification and squamous cell carcinomas. Differentiation 31: 141-153 (1986). 8 Kern, W.; Urinary tract; in Keebler. C.; Reagan. J. (eds): A Man ual of Cytotechnology; 6th ed. (ASCP Press, Chicago 1983). 9 Korowska, I.; Zipfel, S.: The potential usefulness of monoclonal antibodies in the determination of histologic types of lung cancer in cytologic preparations. Acta Cytol. 32: 675-679 (1988). 10 Koss, L.; Wersto, R.; Simmons, D.; Deitch, D.; Herz, F.; Freed, S.: Predictive value of DNA measurements in bladder washings: Comparison of flow cytometry, image cytomorphometry, and
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
In this study, we have demonstrated that immunostaining of bladder cytology specimens for CKs enhances the cytopathology diagnosis such that a prediction can be made of the tumor’s malignant potential. Specimens that have malignant cells present should be stained with these antibodies to aid in the determination of malignant potential and differentiation of the transitional cell car cinoma present. The results based on immunostained cytopreparations correlate well with the surgical pathol ogy diagnosis. This aids in the evaluation of the need for a biopsy of the bladder mucosa. We also have demonstrated that the use of this panel of monoclonal anti-CK antibodies is reproducible and standardizable. Unlike more expensive assays of malig nancy, such as the nude mouse [20, 21] or amnion inva sion [11], immunostaining of cytology specimens can be performed quickly, reliably, and inexpensively in any cytology laboratory. Moreover, this technique is not lim ited to bladder cytopreparations. Other immunostaining panels can be developed to aid the diagnosis of tumors from other organ systems. In summary, we have demonstrated that: (1) urothe lial cells can be distinguished from nonurothelial cells (e.g. squamous cells and renal tubule cells) in immuno stained cytopreparations; (2) specific urothelial cell types can be identified in immunostained cytopreparations; (3) CK staining of cytopreparations can be correlated with the histopatholoy of bladder tumors in humans, and (4) immunostaining of cytology specimens is an aid in predicting the malignant potential of bladder tumors.
Cytokeratin Expression in Bladder Cytopreparations
18 Travis, W.; Wold, L.: Immunoperoxidase staining of fine needle aspiration specimens previously stained by the Papanicolaou technique. Acta Cytol. 31: 517-520 (1987). 19 Trump, B.; Valigorsjy, J.; Jones, R.; Mergner, W.; Garcia. J.: Cowley, R.: The application of electron microscopy and cellular biochemistry to the autopsy. Observations on cellular changes in human shock. Hum. Pathol. 6: 499-516 (1975). 20 Valerio, M.; Fineman. E.; Bowman. R.; Harris. C ; Stoner, G.; Autrup, H.; Trump, B.; McDowell, E.; Jones, R.: Long-term sur vival of normal adult human tissues as xenografts in congenitally athymie nude mice. J. Natl. Cancer. Inst. 5: 849-858 (1981). 21 Yoakum, G.; Lechner, J.; Gabrielson, E.; Korba, B.; Malan-Shibley, L.; Willey, J.: Valerio, M.: Shamsuddin, A.; Trump, B.: Harris, C.: Transformation of human bronchial epithelial cells transfected by Harvey-ras oncogene. Science 227: 1174-1179 (1985).
Received: May 9, 1990 Accepted: June 13, 1990 Dr. James H. Resau Department of Pathology School of Medicine University of Maryland at Baltimore 10 S. Pine Street Baltimore, MD 21201 (USA)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
cytology in patients with a past history of urothelial tumors. Cancer 64: 916-924 (1989). 11 Liotta, L.; Lee, C.; Morakis, D.: New method for preparing large surfaces of intact human basement membrane for tumor inva sion studies. Cancer Lett. 11: 141-152 (1980). 12 Menard. S.; Rilke. F.; Torre, G.; Mariani-Costantini, R.; Rcgazzoni, R.; Tagliabue, E.; Alasio. L.; Colnaghi. M.: Sensitivity enhancement of the cytologic detection of cancer cells in effu sions by monoclonal antibodies. Am. J. Clin. Pathol. 83: 571576 (1985). 13 Moll, R.; Achtstatter, T.: Becht, E.; Balcarova-Stander, J.; Ittensohn, M.; Franke, W.: Cytokeratins in normal and malignant transitional epithelium: Maintenance of expression of urothelium differentiation features in transitional cell carcinomas and bladder carcinoma cell culture lines. Am. J. Pathol. 132: 123— 144 (1988). 14 Ramaekers, F.; Huysmans, A.: Moesker. O.; Schaart, G.; Her man, C.; Vooijs, P.: Cytokeratin expression during neoplastic progression of human transitional cell carcinomas as detected by monoclonal and a polyclonal antibody. Lab. Invest. 52: 31-38 (1985). 15 Ramaekers. F.; Mosker, O.; Huysmans, A.; Schaart, G.; Westerhof, G.; Wagenaar, S.; Herman, C.; Vooijs, G.: Intermediate fila ment proteins in the study of tumor heterogeneity: An in-depth study of tumors of the urinary and respiratory tracts. Ann. N.Y. Acad. Sci. 455: 614-634 (1985). 16 Reedy, E.: Heatfield, B.; Trump. B.; Resau, J.: Correlation of cytokeratin patterns with histopathology during neoplastic pro gression in the rat urinary bladder. Pathobiology 58: 15-27 (1990). 17 Segasothy, M.; Tseng, M.; Birch, D.; Fairley, K.; Kincaid-Smith, P.: Immunocytologic dissection of the urine sediment using monoclonal antibodies. Am. J. Clin. Pathol. 90: 691-696 (1988).
311
Pathobiology 1990;58:304-311
Discrimination of Cell Types in Primary Transitional Cell Carcinoma by Monoclonal Anti-Cytokeratin Antibodies1,2 Edward A. Reedya 3, Kristin L. Colombo-Burkeb, James H. ResaiC 1 Department of Pathology, School of Medicine, University of Maryland at Baltimore, and bUniversity of Maryland Medical System, Department of Anatomic Pathology, Baltimore, Md., USA
Key Words. Cytokeratin • Bladder • Transitional cell carcinoma • Cytology • Immunocytochemistry • Antibodies • Differentiation Abstract. The morphologic discrimination between low and high grade malignant tumor cells arising in the uri nary bladder is a major diagnostic problem in cytopathology. Using immunochemical peroxidase staining of cytokeratins (CKs) of human bladder exfoliative cytology specimens, we have been able to disciminate between transi tional cell carcinoma cells, atypical cells and normal bladder cells. Commercially available monoclonal antibodies used in this study were: anti-CK 13 (Sigma K8.12), anti-CK 5, 7 and 8 (Sigma K8.13), anti-CK 19 (Sigma K4.62), and anti-CK 18 (Transformation Res. 1091). When normal bladder cells are stained with these antibodies, CK 18 is specific for superficial cells; CK 13 is specific for the basal type cells and CKs 5, 7, 8 and 19 are expressed by all urothelial cell types. Four cases diagnosed by cytopathological criteria as 'atypical' and 7 diagnosed as ‘positive’ (various grades) were used in this study. Cytologic diagnosis was confirmed by histopathology in 7 cases. Tissue was not available for histopathology in 4 cases. Malignant cells with a ‘basal’ or ‘immature’ phenotype preferentially stained with CK 13 and were associated with increased metastatic or malignant potential. Patients with higher grade tumors had more cells positive for CK 13 than did patients with atypical or lower grade malignancies. Patients with well-differentiated, low grade tumors predominantly shed cells that expressed CK 18 and CK 19. High grade, inva sive bladder lesions were characterized by many cells expressing CK 13, and fewer cells expressing CK 19. The combination of diagnosis by morphologic criteria on Papanicolaou-stained specimens with immunochemical char acterization of the same cells for CKs facilitate an accurate diagnosis of bladder lesions.
Exfoliative cytology of bladder washings and voidings has been shown to be an effective means of monitoring patients for bladder carcinoma [10]. Cytopathological techniques have been enhanced by applying mono- and polyclonal antibodies to cytologic preparations from var ious body sites to detect specific lesions or cell types [4, 9, 12, 18]. Segasothy et al. [ 17] have advocated the use of monoclonal antibodies to identify different cell types found in urine to aid in the diagnosis of nephritis. Normal human urothelium is known to be a ‘com plex’ epithelium, characterized by ‘simple’ and ‘squa mous’ epithelium cytokeratin (CK) subtypes [13]. The transitional epithelium of the urinary bladder (urothe
lium) is composed of several specialized cell types. Pre vious studies have shown that different cell layers of the urothelium express various CK subtypes which can be detected and discriminated from each other by immuno chemical and biochemical methods [13, 14, 16]. Bladder cytology from washings, voidings and cathe terizations is an effective means of evaluating bladder tumors. Immunocytochemical techniques can be easily applied to all these types of specimens. Since alcohol fix1 Supported in part by funding from the National Disease Re source Interchange (NDRI). 2 A portion of these results were presented in abstract format at the 1989 American Society of Cytology Annual Meeting. 3 Submitted by E.A.R. to the University of Maryland as partial fulfillment for the degree of doctor of philosophy.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
Introduction
Cytokeratin Expression in Bladder Cytopreparations
Materials and Methods Specimens All cytology and surgical pathology specimens used in this study were obtained from the University of Maryland Hospital, Depart ment of Pathology. The protocols were reviewed and approved by the Human Volunteers Research Committee at the University of Maryland. Ethanol-fixed, Papanicolaou (Pap)-stained cytoprepara tions were examined and characterized by standard criteria [5, 8], Abnormal diagnoses were correlated with surgical specimens from 7 of the 11 cases, and the cytology diagnosis and the surgical pathol ogy diagnosis were compared. Normal human bladder was obtained at immediate autopsy [19] within 1 h of somatic death from a 19-year-old white male with no previous history of urinary tract disease. Urinary bladder mucosa was fixed in 100% (absolute) ethanol, and processed and sectioned for light microscopy. Histologic sections were evaluated microscop ically by HE staining. Next, serial 5-p.m paraffin sections were cut and stained using the immunohistochemical technique described below. Stains Used Pap Stain. The Pap stain is the standard cytological multichromatic stain made up of aqueous hematoxylin and ethanolic counter stains. Specimens included in this study were stained in the follow ing manner: Gill’s hematoxylin, 90 s; tap water rinse until clear; 95% ETOH. 60s; OG-6, 2 min; 95% ETOH. 60s; modified EA, 5 min; 95% ETOH, two changes, 60s each; 100% ETOH. two changes, 60 s each; xylene, two changes, 3 min each. Specimens were initially screened and diagnosed using this stain.
Immunochemical Stains. Monoclonal mouse anti-CK antibodies were used in an indirect immunochemical procedure for both tissue sections and cytopreparations. When tissue sections were stained, the tissue was deparaffinized with three changes of xylene and the xylene was removed with three changes of 100% ethanol. For immunostaining, both tissue sections and cytology specimens were treated with 3% H 2 O 2 for 20 min, taken to water, then rinsed in PBS (pH 7.4) for 5 min. Nonspecific protein binding was blocked on tis sue sections by incubating with normal nonimmune rabbit serum (Dako Corp., Carpinteria. Calif.) for 15 min. The primary anti bodies used were K8.12 (anti-CK 13), K8.I3 (anti-CK 5, 7 and 8). K4.62 (anti-CK 19) (Sigma Chemical Corp., St. Louis, Mo.), and 1091 (anti-CK 18) (Transformation Research, Framingham, Mass.). All primary antibodies were used at a dilution of 1:20 in PBS except 1091, which was 1:100, and all were incubated for 2 h at room tem perature. All primary antibody reactivities were determined by Western blot analysis using normal human urothelial CKs [see 16]. Specimens were washed with 50 mM Tris-HCl buffer (pH 7.6) be tween antibody incubation steps. Rabbit anti-mouse IgG horserad ish peroxidase-conjugated secondary antibody (Dako) was used at a 1:80 dilution in PBS and incubated for 30 min at room temperature. CK filaments were localized by incubating for 10 min with 0.05% (w/v) 3,3'-diaminobenzidine (Sigma) in Tris buffer freshly prepared and filtered with 0.03% (v/v) H;02. Specimen Evaluation The cytopathology diagnosis was performed on Pap-stained specimens with criteria described by Frost [5] and Kem [8]. An ‘atypical’ diagnosis was based on the presence of slightly to moder ately irregular and enlarged nuclei with hyperchromasia and in creased chromatin granularity. A ‘suspicious’ diagnosis was made when nuclei exhibited coarser chromatin granules, marked hyper chromasia, an increased nuclcar/cytoplasmic (N/C) ratio, and when an abundant number of atypical cells or cell groups were present. A specimen was judged ‘positive’ for malignancy if the cells present exhibited severely atypical nuclear morphology, a very high N/C ratio, and many atypical cells or cell groups, or numerous mitotic figures (fig. 4c). Immunostained specimens were evaluated by counting 500 cells from each slide on each case. Urothelial cells were scored as staining positive or negative, and the percent positive cells were determined. The immunostaining results, cytology diagnosis and surgical pathol ogy diagnosis were compared and correlated.
Results Normal human urothelium is composed of three dis tinct cell types: large, often multinucleated superficial (umbrella) cells at the lumenal surface; intermediate cells, and small basal cells with dense nuclei lying on the basal lamina (fig. 1). In ethanol-fixed paraffin sections of normal human bladder, only the superficial cells stain positive with the antibody 1091 (fig. 2a, table 1). Based on Western blot ting results with this antibody [16], superficial cells
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
ation is used predominantly for cytological preparations, retrospective studies are possible and appropriate for immunocytochemical staining. Prior Papanicolaou staining does not interfere with antibody binding [18]. Our laboratory previously has shown in a rat bladder tumor model that, as bladder carcinomas progress, the developing tumors express unique CK patterns; and CK expression was positively associated with the invasive potential of these induced primary bladder tumors [16]. This report describes how this data, applied to human transitional cell carcinomas, aids in the diagnosis and treatment of the patient. In this study, cytopreparations (direct smear or cytospin) of bladder voidings, washings and catheterizations were immunochemically stained with a panel of four commercially available monoclonal anti-CK antibodies and correlated with the cytopathology diagnosis. These results were then compared to the histopathologic diag nosis. Using the panel of antibodies utilized by Reedy et al. [ 16], different urothelial cell types could be identified in immunostained cytopreparations. The cell types iden tified in the immunostained cytopreprations were corre lated with the histopathologic diagnosis.
305
306
Reedy/Colombo-Burke/Resau
Table 1. CK subtypes and antibody reactivities in normal human bladder CK
Antibody
Histological location
5. 7. 8 19 13 18
Sigma K8.13 Sigma K4.62 Sigma K8.12 TR 1091
all layers all layers basal superficial
Fig. 1. Smear preparation of normal human urinary bladder from a 19-year-old white male, 2 h postmortem, with no history of urinary tract disease. Numbered arrowheads indicate the different cell types present: large superficial cells are often multinucleated (arrowhead I), intermediate cells (arrowhead 2), and small basal cells with dense nuclei (arrowhead 3). Pap stain. X 40.
Fig. 2. Immunostained ethanol-fixed normal human bladder mucosa from a 19-year-old white male, 2 h postmortem, with no history of urinary tract disease, a Antibody 1091 (anti-CK 18) reacts specifically and intensely with the superficial cells of the urothelium. b Antibody K8.12 (anti-CK 13) reacts specifically with the
d basal and some intermediate ceils of the urothelium. c Antibody K4.62 (anti-CK 19) reacts with all cell layers of the urothelium. d Antibody K8.13 (anti-CK 5, 7 and 8) reacts intensely with all cell layers of the urothelium. Indirect immunoperoxidase, hematoxylin counterstain. X 40.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
b
307
Cytokeratin Expression in Bladder Cytopreparations
express CK 18, while other cells in the urothelium ex press CK 18 but in amounts below the levels of detection with immunohistochemical techniques [13]. Patients with a papillary carcinoma of the bladder had 49% (pa tient 5) to 85% (patient 2) of cells in the urine positive for CK 18 (table 2). In normal bladder, these cells corre spond to those found on the lumenal surface of the blad der (fig. 2a, 3a, table 1). When antibody K8.12 was used to stain ethanol-fixed paraffin-embedded sections of normal human bladder, only the basal layers of cells were reactive (fig. 2b, table 1). Therefore, based on the reported reactivities of this antibody [7, 16], these cells express CK 13. In immunostained cytopreparations, the percentage of cells posi tive for basal CK (CK 13) ranged from 34% (patient 3) to 100% (patients 4 and 7). Generally, the number of
positive cells increased as the histologic grade (malignant potential) increased (patients 1-3, 5-7). Patients 4 and 7 had 100% of cells positive for basal CKs. By histopathology, patient 4 had a carcinoma in situ in addition to a papillary transitional cell carcinoma. Patient 7 was diag nosed to have a nonpapillary transitional cell carcinoma by histopathology (table 2). Previous reports have established the reactivity of the antibody K4.62 as specific for CK 19 [6. 16], Immuno histochemical results of normal bladder stained with the antibody K4.62 indicate that all cell layers in the urothe lium express CK 19 (fig. 2c, table 1). Of patients whose surgical diagnosis was known, 6 of 7 had 100% of the cells counted as positive for CK 19. The 1 patient (No. 7) that did not have all cells positive was diagnosed to have an invasive transitional cell carcinoma (table 2).
b
Fig. 3. Immunostained cytopreparations of a bladder washing from a 66-year-old male (patient 1). See table 2 for diagnoses. The following immunostaining pattern is typical of low grade, noninvasive, papillary transitional cell carcinoma, a Antibody 1091 (anti-CK 18) intensely stains the large, multinucleated superficial (‘umbrella’) cells, while the squamous cells (arrows) are unreactive. b A group of small cells resembling basal cells are nonreac-
tive with the antibody 1091. c Cells resembling superficial cells are unstained with the antibody K.8.12 (arrows) (anti-CK 13), while similar appearing cells and a group of small basal-like cells stain intensely, d A urothelial cell is positive when stained with the antibody K4.62 (anti-CK 19), but squamous cells (arrows) are negative. Indirect immunoperoxidase, hematoxylin counterstain. X 40.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
d
Reedy/Colombo-Burke/Resau
308
Table 2. Comparison of immunochemical staining results with cytology and surgical diagnosis Age/sex
Method
Cytology screening diagnosis
Surgical pathology diagnosis
Immunochemical staining percent positive cells/500 cells 1091
K.8.12
K8.I3
K.4.62
1
66/M
wash
atypical transitional cells present
grade II papillary transitional cell carcinoma, no invasion
83
56
100
100
2
67/M
wash
atypical transitional cells present
grade II papillary transitional cell carcinoma, no invasion
85
63
100
100
3
63/M
cath
suspicious for malignancy: grade I papillary transitional groups of atypical cell carcinoma, no invasion transitional cells are present
67
34
100
100
4
67/M
wash
positive for malignancy: cells are present which represent a transitional cell carcinoma
flat transitional cell carcinoma in situ and grade I papillary transitional cell carcinoma, no invasion
69
100
100
100
5
73/M
wash
positive for malignancy: cells are present which represent a transitional cell carcinoma
grade II papillary transitional cell carcinoma, no invasion
49
53
100
100
6
71/M
void
positive for malignancy: cells are present which represent a transitional cell carcinoma
grade II/III papillary transitional cell carcinoma, no invasion
51
62
100
100
7
39/M
cath
positive for malignancy: cells are present which represent a transitional cell carcinoma
grade III infiltrating transitional cell carcinoma with squamous features, smooth muscle invasion
50
100
100
43
Based on immunohistopathology, all cell layers of the urothelium stain strongly with antibody K.8.13 in etha nol-fixed paraffin sections (fig. 2d, table 1). Squamous cells found in cytopreparations stain weakly or not at all (fig. 3a, b). Only cells that displayed a urothelial pheno type and stained strongly with K8.13 were included in this data. Cells that possessed a squamous morphology (flat with pyknotic nucleus) or were negative for K.8.13 were excluded from the cells counted because they were npt considered urothelial cells. Therefore, all cells (100%) included in this study’s data stained positive with the antibody K8.13 (table 2).
Discussion Previous studies have attempted to identify CKs as markers for transitional cell carcinoma. Baker et al. [1] correlated the relative amount of CK 18 in transitional
cell carcinoma using a radioimmunoassay (RIA) system. Basta et al. [2] used an ELISA system to detect the pres ence of keratin in the urine of patients with transitional cell carcinoma. Both groups identified CK as a marker of urothelial malignancy but did not specifically identify cancer cells. More definite results regarding the expression of spe cific CKs in bladder carcinomas have been presented by Ramaekers et al. [14, 15], Cintorino et al. [3], and Moll et al. [13], Generally, all the reports confirm that CK expression is modulated in transitional cell carcinomas of various grades, and that it may be possible to identify the grade of the tumor based on the CK expression observed using immunochemical staining. All three re ports also identify CK 18 as being expressed in greater amounts as the tumor grade increases. Moll et al. [13] and Cintorino et al. [3] both described the novel expres sion of CK 17 in transitional cell carcinomas. Moll et al. [13] presented data that shows CK 13 expression de
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
Patient
Cytokeratin Expression in Bladder Cytopreparations
V
Fig. 4. Immunostained cytopreparations of a cathetcrized blad der specimen from a 39-year-old male (patient 7). See table 2 for diagnosis. The following immunostaining pattern is typical of inva sive, nonpapillary transitional cell carcinomas, a A small, malignant cell is nonreactive with antibody 1091 (anti-CK 18). b A urothelial cell reacts weakly with antibody K4.62 (anti-CK 19). c A cluster of small, malignant cells stain intensely with antibody K8.12 (anti-CK 13). Indirect immunoperoxidase, hematoxylin counterstain. X 40.
increased numbers of CK 13-positive cells indicate that a carcinoma in situ could be present in addition to a low grade (I—II) papillary transitional cell carcinoma (see also patient 3, table 2). Patient 11: the lack of cells posi tive for CK 18, the increased numbers of CK 13-positive cells, and the decreased numbers of CK 19-positive cells suggest, that a nonpapillary invasive carcinoma may be present in this patient (see also patient 7, table 2). In the hope that these predictions are substantiated, these pa tients’ histories are being followed for the histopathological results of a tissue biopsy.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
creasing as the tumor grade increases. Reedy et al. [ 16] described a similar pattern of CK 19 expression in blad der tumors induced in rats. In the present study using immunostained cytoprepa rations and histological sections of normal human blad der. the following general conclusions were drawn: All urothelial cells were positive with the antibody K.8.13 (CKs 5, 7 and 8) (fig. 2d). Cells that were not positive are not urothelial cells (fig. 3a. d). Basal urothelial cells stained positive for CK 13 (Sigma K8.12) (fig. 2b, 3c, 4c). Superficial urothelial cells stained positive for CK 18 (TR 1091) (fig. 2a, 3a). The degree of transitional cell carcinoma differentia tion is positively associated with the percentage of cells positive for CK 18 (TR 1091) and negatively associated with the percentage of cells positive for CK 13 (Sigma K8.12) (table 2). Specimens from patients with well-differentiated pa pillary transitional cell carcinomas were characterized by a substantial percentage of cells positive for CK 18 (TR 1091) (fig. 2a), and a low percentage of cells positive for CK 13 (Sigma K8.12) (fig. 3c). Specimens from pa tients with less well-differentiated papillary transitional cell carcinomas were characterized by fewer cells posi tive for CK 18 and more cells positive for CK 13 (fig. 4). Cytology specimens from the patient in our study with invasive transitional cell carcinoma were character ized by a decrease (43% compared to 100%) in the per centage of cells expressing CK 19 (Sigma K4.62) (pa tient 7, table 2). In studies with rat bladder tumors in duced in vivo, we observed that as malignant potential increased, the expression of CK 19 decreased [16]. The staining pattern of patient 7 is consistent with these results (43% of cells positive for CK 19) and is supported by histopathology (infiltrating carcinoma). Using information based on the data presented in table 2, the following predictions of tumors present in patients 8-11 (table 3) are made: Patient 8: the absence of cells positive for CK 18, the increased numbers of CK 13-positive cells, and the decreased numbers of CK 19positive cells, suggest that a nonpapillary invasive carci noma is present in this patient (see also patient 7, table 2). Patient 9: the CK staining pattern of this patient indicates that a low grade (I—II) papillary transitional cell carcinoma could be present in this patient (see also patients 1 and 2. table 2). Subsequent to analysis, the bladder tumor of this patient was evaluated by histopa thology, and was diagnosed as a papillary transitional cell carcinoma, grade I (no invasion). Patient 10: the
309
310
Reedy/Colombo-Burke/Resau
Table 3. Comparison of immunochemical staining results with cytology diagnosis alone Patient
Age/sex
Method
Immunochemical staining percent positive cells/500 cells
Cytology screening diagnosis
1091
K8.12
K8.13
K4.62
8
53/F
wash
atypical transitional cells present
0
40
100
25
9'
43/F
cath
atypical transitional cells present
73
54
100
100
10
66/M
cath
suspicious for malignancy: significantly atypical transitional cells are present
90
100
100
100
11
90/M
void
positive for malignancy: malignant cells are present which represent a transitional cell carcinoma
0
50
100
30
1 Patient 9 was evaluated by histopathology following this analysis and was diagnosed as having a papillary transitional cell carcinoma, grade I (no invasion).
References 1 Baker. W.; White, R.; Rossitto. P.; Min. B.; Cardiff. R.: Quanti tative analysis of keratin 18 in the urine of patients with bladder cancer. J. Urol. 140: 436-439(1988). 2 Basta, M.; Attallah, A.; Seddek. M.; El-Mohamady, H.; Al-Hilaly, E.; Atwaan, N.; Ghoneiwm, M.: Cytokeratin shedding in urine: A biological marker for bladder cancer? Br. J. Urol. 61: 116-121 (1988). 3 Cintorino, M.; Vecchio, M.; Bugnoli, M.; Petracca, R.: Cytoker atin pattern in normal and pathological bladder urothelium: Immunohistochemical investigation using monoclonal anti bodies. J. Urol. 139: 428-432 (1988). 4 Dalquen, P.; Bittel, D.; Gudat, F.; Overbeck. J.; Heitz, P.: Com bined immunoreaction and Papanicolaou’s stain on cytological smears. Pathol. Res. Pract. 181: 50-54 (1986). 5 Frost, J.: The cell in health and disease. An evaluation of cellular morphologic expression of biologic behavior; in Wied, G.; von Hamm, E.; Koss, L.; Reagan, J. (eds): Monographs in Clinical Cytology; 2nd ed., vol. 2 (Karger, Basel, 1986). 6 Gigi-Leitner. O.; Geiger, B.: Antigenic interrelationship between the 40-kilodalton cytokeratin polypeptide and desmoplakins. Cell Motil. Cytoskeleton 6: 628-639 (1986). 7 Huszar. M.; Gigi-Leitner. O.; Moll. R.: Franke. W.; Geiger, B.: Monoclonal antibodies to various acidic (type I) cytokeratins of stratified epithelia: Selective markers for stratification and squamous cell carcinomas. Differentiation 31: 141-153 (1986). 8 Kern, W.; Urinary tract; in Keebler. C.; Reagan. J. (eds): A Man ual of Cytotechnology; 6th ed. (ASCP Press, Chicago 1983). 9 Korowska, I.; Zipfel, S.: The potential usefulness of monoclonal antibodies in the determination of histologic types of lung cancer in cytologic preparations. Acta Cytol. 32: 675-679 (1988). 10 Koss, L.; Wersto, R.; Simmons, D.; Deitch, D.; Herz, F.; Freed, S.: Predictive value of DNA measurements in bladder washings: Comparison of flow cytometry, image cytomorphometry, and
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
In this study, we have demonstrated that immunostaining of bladder cytology specimens for CKs enhances the cytopathology diagnosis such that a prediction can be made of the tumor’s malignant potential. Specimens that have malignant cells present should be stained with these antibodies to aid in the determination of malignant potential and differentiation of the transitional cell car cinoma present. The results based on immunostained cytopreparations correlate well with the surgical pathol ogy diagnosis. This aids in the evaluation of the need for a biopsy of the bladder mucosa. We also have demonstrated that the use of this panel of monoclonal anti-CK antibodies is reproducible and standardizable. Unlike more expensive assays of malig nancy, such as the nude mouse [20, 21] or amnion inva sion [11], immunostaining of cytology specimens can be performed quickly, reliably, and inexpensively in any cytology laboratory. Moreover, this technique is not lim ited to bladder cytopreparations. Other immunostaining panels can be developed to aid the diagnosis of tumors from other organ systems. In summary, we have demonstrated that: (1) urothe lial cells can be distinguished from nonurothelial cells (e.g. squamous cells and renal tubule cells) in immuno stained cytopreparations; (2) specific urothelial cell types can be identified in immunostained cytopreparations; (3) CK staining of cytopreparations can be correlated with the histopatholoy of bladder tumors in humans, and (4) immunostaining of cytology specimens is an aid in predicting the malignant potential of bladder tumors.
Cytokeratin Expression in Bladder Cytopreparations
18 Travis, W.; Wold, L.: Immunoperoxidase staining of fine needle aspiration specimens previously stained by the Papanicolaou technique. Acta Cytol. 31: 517-520 (1987). 19 Trump, B.; Valigorsjy, J.; Jones, R.; Mergner, W.; Garcia. J.: Cowley, R.: The application of electron microscopy and cellular biochemistry to the autopsy. Observations on cellular changes in human shock. Hum. Pathol. 6: 499-516 (1975). 20 Valerio, M.; Fineman. E.; Bowman. R.; Harris. C ; Stoner, G.; Autrup, H.; Trump, B.; McDowell, E.; Jones, R.: Long-term sur vival of normal adult human tissues as xenografts in congenitally athymie nude mice. J. Natl. Cancer. Inst. 5: 849-858 (1981). 21 Yoakum, G.; Lechner, J.; Gabrielson, E.; Korba, B.; Malan-Shibley, L.; Willey, J.: Valerio, M.: Shamsuddin, A.; Trump, B.: Harris, C.: Transformation of human bronchial epithelial cells transfected by Harvey-ras oncogene. Science 227: 1174-1179 (1985).
Received: May 9, 1990 Accepted: June 13, 1990 Dr. James H. Resau Department of Pathology School of Medicine University of Maryland at Baltimore 10 S. Pine Street Baltimore, MD 21201 (USA)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 8/4/2018 9:22:20 PM
cytology in patients with a past history of urothelial tumors. Cancer 64: 916-924 (1989). 11 Liotta, L.; Lee, C.; Morakis, D.: New method for preparing large surfaces of intact human basement membrane for tumor inva sion studies. Cancer Lett. 11: 141-152 (1980). 12 Menard. S.; Rilke. F.; Torre, G.; Mariani-Costantini, R.; Rcgazzoni, R.; Tagliabue, E.; Alasio. L.; Colnaghi. M.: Sensitivity enhancement of the cytologic detection of cancer cells in effu sions by monoclonal antibodies. Am. J. Clin. Pathol. 83: 571576 (1985). 13 Moll, R.; Achtstatter, T.: Becht, E.; Balcarova-Stander, J.; Ittensohn, M.; Franke, W.: Cytokeratins in normal and malignant transitional epithelium: Maintenance of expression of urothelium differentiation features in transitional cell carcinomas and bladder carcinoma cell culture lines. Am. J. Pathol. 132: 123— 144 (1988). 14 Ramaekers, F.; Huysmans, A.: Moesker. O.; Schaart, G.; Her man, C.; Vooijs, P.: Cytokeratin expression during neoplastic progression of human transitional cell carcinomas as detected by monoclonal and a polyclonal antibody. Lab. Invest. 52: 31-38 (1985). 15 Ramaekers. F.; Mosker, O.; Huysmans, A.; Schaart, G.; Westerhof, G.; Wagenaar, S.; Herman, C.; Vooijs, G.: Intermediate fila ment proteins in the study of tumor heterogeneity: An in-depth study of tumors of the urinary and respiratory tracts. Ann. N.Y. Acad. Sci. 455: 614-634 (1985). 16 Reedy, E.: Heatfield, B.; Trump. B.; Resau, J.: Correlation of cytokeratin patterns with histopathology during neoplastic pro gression in the rat urinary bladder. Pathobiology 58: 15-27 (1990). 17 Segasothy, M.; Tseng, M.; Birch, D.; Fairley, K.; Kincaid-Smith, P.: Immunocytologic dissection of the urine sediment using monoclonal antibodies. Am. J. Clin. Pathol. 90: 691-696 (1988).
311
