• Embed Doc
  • Copy Link
  • Readcast
  • Collections
  • 1
    CommentGo Back
Download
 
DOI:10.1111/j.1365-2303.2006.00334.x
Heparanase expression: a potential ancillary diagnostic tool fordistinguishing between malignant cells and reactive mesotheliumin body cavity effusions
V. Doviner*, B. Maly*, T. Reinhartz*, I. Vlodavsky
and Y. Sherman*
*Department of Pathology, Hadassah-Hebrew University Hospital, Jerusalem and
Cancer and Vascular Biology ResearchCenter, The Bruce Rappaport Faculty of Medicine, Technion, Haifa, Israel
 Accepted for publication 4 January 2006 
V. Doviner, B. Maly, T. Reinhartz, I. Vlodavsky and Y. Sherman
Heparanase expression: a potential ancillary diagnostic tool for distinguishing betweenmalignant cells and reactive mesothelium in body cavity effusions
Objective:
Heparanase, an endoglycosidase that cleaves heparan sulphate, is frequently expressed incarcinomas and was suggested to play a role in cell invasion and metastasis. We investigated whether heparanaseexpression may serve as a reliable marker to discriminate benign mesothelial cells from malignant cells shed into body cavities.
Methods and results:
Cytological smears of effusions from 51 hospitalized patients were immunostained forheparanase. Strong immunoreactivity was noted in 35 of 40 (88%) carcinoma samples and in all three malignantmesothelioma cases. Only rare (<3%) reactive mesothelial cells were noted showing a faint negligible staining.Specificity was 100%, sensitivity 88%, and positive and negative predictive values were 100% and 89%respectively.
Conclusions:
Our results suggest that heparanase may be of value as a complementary component in adiagnostic panel of markers, contributing to its reliability and accuracy.
Keywords:
serons, effusions, malignant cells, mesothelial cells, heparanase, immunocytochemistry, cytology
Introduction
One of the main challenges of diagnostic cytology isfinding an unequivocal and reliable method fordiscriminating between benign reactive mesothelialcells and malignant cells exfoliating into body cavityfluids. Cells from reactive or hyperplastic mesotheli-um shed from body cavity surfaces in various biolo-gical settings may present a wide range of deviationfrom normal cellular morphology, making it difficult,or even impossible, to distinguish them from malig-nant cells by means of purely cytological criteria.Many diagnostic procedures, such as DNA analysis by means of flow or image cytometry,
1–3
AgNORevaluation,
4,5
in situ
hybridization,
6,7
identification ofK-
ras
mutations by PCR,
8
and detection of humantelomerase in cells from body fluids,
9,10
have beeninvestigated as possible ancillary diagnostic means formeeting the challenge, but none of these has beenwidely accepted as a plausible diagnostic procedure.Likewise, application of various monoclonal antibod-ies aimed at discriminating between the different celltypes is, as of now, of limited value.
11
Most of theantibodies in use are aimed at marking specificantigens in the cells under examination, thus enablingidentification of the origin of cells (mesotheliumversus epithelium), but not their biological nature(i.e. benign proliferating mesothelial cells versusmalignant mesothelioma, or metastatic carcinoma).So far only a few immunohistochemical markers have been suggested as useful in discriminating betweenreactive mesothelial cells and malignant cells. Epithe-lial membrane antigen (EMA), in use as a marker ofepithelial cells, was reported to be expressed inmalignant mesothelioma cells, but not in their react-ive counterparts.
12–14
This observation, however, has
Correspondence:Yoav Sherman, MD, MIAC, Department of Pathology,Hadassah-Hebrew University Hospital, PO Box 12023, Jeru-salem, Israel.Tel.: +972 2 6776536; Fax: +972 2 6426268;E-mail: sherman@md.huji.ac.il
Cytopathology
2007,
18
, 13–19
ª
2006 The AuthorsJournal compilation
ª
2006 Blackwell Publishing Ltd
13
 
not been confirmed in other studies.
15,16
Anothersuggestion, the result of a more extensive qualitativeanalysis of EMA expression, was that only malignantmesothelial cells display membrane staining for thisantigen, whereas such a pattern is not observed inreactive mesothelial cells.
17
Another marker underinvestigation is Sialosyl-Tn, which was found to beexpressed in carcinoma cells from various sites, andproposed to be associated with tumour metastases anddecreased survival.
18,19
This marker has been found to be partially useful in distinguishing reactive meso-thelial cells from carcinoma cells in serous effusionsand was suggested to be of potential value as part of adiagnostic immunohistochemical panel.
20
As none ofthese methods has proven to be unequivocal, there isa genuine need to search for a more specific andreliable biological indicator of malignant cells thatdistinguishes them from reactive mesothelial cells.One attribute of malignant cells that may serve as a basis for a diagnostic discriminating procedure is theirexpression of a host of extracellular matrix (ECM)degrading enzymes, whose activity is thought to beessential for cancer invasion and metastasis.
21,22
Among these enzymes is an endo-
b
-glucuronidase(heparanase) that cleaves heparan sulphate (HS) atspecific intrachain sites.
23,24
HS interacts with ECMmacromolecules such as collagen, laminin and fibro-nectin and with different attachment sites on the cellmembrane, suggesting a key role for this proteoglycanin the self-assembly and insolubility of ECM compo-nents, as well as in cell adhesion and locomotion.
25–27
Cleavage of HS by heparanase was suggested, there-fore, to play an important role in tumour invasion andmetastatic activity.
28
Indeed, various studies have demonstrated a corre-lation between expression of heparanase and themetastatic potential of tumour cells.
23,24
Heparanasewas also reported to be overexpressed in histologicalsamples of various malignant tumours such as oeso-phageal, gastric, colorectal, pancreatic, ovarian andlung carcinomas.
29–36
Some of these studies alsodemonstrated a correlation between heparanaseexpression and poorer prognosis.
29–34
The growing body of data, indicating a differentialand tumour-specific expression of heparanase invarious cancers, led us to examine whether cellularexpression of heparanase in cytological specimensfollows the same pattern, namely that it will bemanifested in malignant cells to a much higher degreethan in reactive mesothelial cells. If so, it could become a useful and reliable diagnostic adjunct tothe task of determining the nature of cells accumu-lating in body cavity effusions.
Materials and methods
We evaluated 51 cytological specimens of benign andmalignant effusions, submitted for cytological exam-ination to the division of cytology (Department ofPathology, Hadassah Medical Center, Jerusalem) fromJanuary 2001 to October 2002. Specimens wereobtained from patients with a previous diagnosis ofcancer, clinical and radiological suspicion of malig-nancy, or non-neoplastic conditions. The above caseswere retrospectively selected on the basis of sufficientdiagnostic material (>100 cells on a slide) for evalu-ation. The distribution of cases according to tumourorigin and clinical diagnosis is presented in Table 1.For immunocytochemical staining we used alcohol-fixed, unstained smears that were processed forimmunocytochemical detection of heparanase asfollows: slides were rinsed in tap water and thenimmersed in phosphate-buffered saline (PBS), pH 7.4,for 3 minutes. The slides were then fixed in coldacetone at
)
20
°
C for 10 minutes. Endogenous per-oxidase activity was blocked by incubation in 3%hydrogen peroxide at room temperature for 10 min-utes, and then the slides were rinsed (
·
2) in tap waterfor 3 minutes. After blocking of nonspecific binding by preincubation with non-immune serum (ABC-kit;Zymed, San Francisco, CA, USA) for 20 minutes, theslides were incubated for 1 hour at room temperaturewith a monoclonal anti-heparanase antibody (HP-92.4) diluted 1 : 4 in PBS. This antibody was kindlyprovided by InSight Pharmaceuiticals Ltd (Rehovot,
Table 1.
Cases grouped according to primary site or disease(determined on the basis of clinical history and/or histolog-ical diagnosis)History/disease
n
Benign (
n
¼
34)Liver disease 4Heart failure 5Unknown aetiology 2Reactive cells in malignant effusions 23Malignant (
n
¼
40)Breast carcinoma 11Lung carcinoma 7Gastrointestinal tract adenocarcinoma 9Ovarian carcinoma 8Other carcinomas 5
V. Doviner et al.
14
Cytopathology
2007,
18
, 13–19
ª
2006 The AuthorsJournal compilation
ª
2006 Blackwell Publishing Ltd
 
Israel). Biotinylated secondary antibody (ABC-kit)was applied for 10 minutes, followed by incubationwith HRP-labelled streptavidin (ABC-kit) for 10 min-utes. After each step, the slides were washed twicewith PBS for 10 minutes, and then reacted for colourdevelopment with Zymed aminoethyl carbozole sub-strate kit (Zymed) followed by counterstaining withMayer’s haematoxylin. Sections from colon carci-noma, which expresses heparanase, were stained inparallel and served as positive controls. For thenegative control, a duplicate slide from each casewas stained, with the primary antibody being substi-tuted for an isotypic mouse immunoglobulin G.Immunocytochemically stained slides were evalu-ated by four independent observers (T.R., B.M., Y.S.and V.D.). Staining in malignant cells was scored in allspecimens diagnosed as either adenocarcinoma ormalignant mesothelioma. Staining of benign meso-thelial cells was similarly scored in both reactivespecimens and malignant effusions with sufficientnumber of reactive cells in the background (
n
¼
23).Staining intensity was graded semiquantitatively asnegative, weak (1+, compared with the intensity of background staining) or strong (++, obvious andunmistakably distinguished from background stain-ing). Positive samples were defined as such only if thestaining intensity was strong. They were divided intotwo groups, one consisting of cases where less than10% of the tumour cells (in malignant effusions) orreactive mesothelial cells (in benign effusions) reac-ted, and the second consisting of cases where morethan 10% of the examined cells demonstrated a strongstaining intensity. In any case of variability in assess-ment among the four observers a consensus wasreached by discussing the slide in question at amultiheaded microscope.
Results
Positive immunoreactivity of heparanase, defined asstrong intensity of cytoplasmatic staining (Figure 1),was detected in 35 of 40 (88%) patients with meta-static carcinoma. Of these, in 26 (65%) cases, morethan 10% of the malignant cells were positive,whereas in the remaining nine (23%) cases, thepositive cells constituted less than 10% of the entirepopulation of malignant cells. Of the various adeno-carcinomas examined, those of gastrointestinal-tractorigin (
n
¼
9) exhibited a perfect score of 100%sensitivity with no negative cases. The other variantspresented positive reactivity rate ranging from 80%(other carcinomas) to 91% (breast). All the samplesfrom cases of malignant mesothelioma manifestedpositive staining of at least some of the malignant cells(Table 2).None of the reactive mesothelial cells in samplesfrom either benign or malignant effusions (defined inthe latter as such by their benign morphology),showed any considerable degree of immunoreactivityto heparanase (Table 2). In some of these samples,rare isolated cells or small clusters of cells exhibitedfaint negligible staining of the same intensity as thenon-specific background staining, and were clearlydistinguishable from the neighbouring malignant cells(Figure 1, arrows).Overall sensitivity for strongly stained cases was88%. The specificity was 100%, positive predictivevalue was 100% and negative predictive value was87% (Table 3). Included in the calculation were 23cases with reactive cells from malignant effusions.
Discussion
The cytological examination of cells suspected of beingneoplastic focuses on the analysis of cytomorpholog-ical features pertinent for defining the biologicalnature of the cells in question and, if possible, theirhistogenesis (general activity and functional differen-tiation respectively).
37
Attempts at defining the natureof cells shed into body cavities are confounded bydifficulties from both aspects. Malignant epithelialcells may often present a benign-looking morphologyidentical to that of reactive mesothelial cells andconversely, reactive proliferating mesothelial cellsmay often show features of atypia barely distinguish-able from those encountered in malignant epithelialcells. Also, the unequivocally malignant cases mayoften pose a problem in discriminating between cellsoriginating from adenocarcinoma versus mesotheli-oma.The various antibodies currently in use as diagnosticadjuncts are directed mainly at helping to define thefunctional differentiation of the cells, namely theirorigin. The logic of this routine stems from theassumption that cells in body cavity fluids that arecharacterized immunocytochemically as epithelial, areinvariably malignant and of metastatic origin. How-ever, cells defined as mesothelial, are either neoplasticoriginating from malignant mesothelioma, or reactive.The battery of immunological markers in useconsists of a group of monoclonal antibodies, aimedat specific targets in the various cell types. Antibodies
Heparanase expression: a potential ancillary diagnostic tool
15
Cytopathology
2007,
18
, 13–19
ª
2006 The AuthorsJournal compilation
ª
2006 Blackwell Publishing Ltd
of 00

Leave a Comment

You must be to leave a comment.
Submit
Characters: ...
11 / 16 / 2010<span class="translation_missing">en_US, this_document_made_it_onto_the</span>Rising List!
You must be to leave a comment.
Submit
Characters: ...