Cytoblock preparations for examination ofcervical and other cells

B Dagg, D L S Eustace, X Han, S Money, E Heyderman

JrClin Pathol 1992;45:1122-1123

AbstractThere are a number of antibodies whichmay be of value in the investigation ofcervical smears, effusions, and celis

grown in monolayer culture. The ShandonCytoblock method was used to prepare

discs of such celis suitable both for diag-nosis and for a variety of other tech-niques.

( Clin Pathol 1992;45:1122-1123)

Department ofHistopathology, StThomas's Hospital,London SE1 7EHB DaggX HanS MoneyE HeydermanDepartment ofObstetrics andGynaecology, UMDS,St Thomas's Hospital,London SEl 7EHD L S EustaceCorrespondence to:Dr B Dagg, Department ofHistopathology, Royal FreeHampstead NHS Trust,London NW3 2QGAccepted for publication8 July 1992

Cervical smears stained by the Papanicolaoutechnique have the advantages of speed, sim-plicity, and familiarity. Considerable experi-ence has been built up in their interpretation,and though there are problems with inter- andintra-observer error,' in those countries inwhich screening programmes have been widelyimplemented there has been a considerable fallin incidence of cervical carcinoma and mortal-ity.2 Smears have the disadvantage of beingunique so that if several immunostains are tobe performed there will be considerable differ-ences in cell populations between slides madeat the same time from the same scrape. Theepithelial cells may be obscured by pus in thepresence of severe inflammation, or the smear

may be too thick to interpret.

MethodsThe efficacy of the Shandon Cytoblockmethod (Shandon, Cheshire) for the evalu-ation of cervical cytology was investigated witha view to using the preparations for imunocy-tochemical techniques.

J

Specimens were

9^ 0

VA

0

..

Figure I Normal smear. Most of the squames are cut as non-overlapping six

obtained from 28 unselected patients referredto a colposcopy clinic because of an abnormalsmear result. A Cervex brush (Vernacare,Lancs) which has a gently concave flattenedportion that brushes the external cervix, and apointed centre that samples the upper portionof the endocervical canal, was used.3The brush head was detached by pushing it

off manually and put into unbuffered 10%formalin in a Universal container. The con-tainer was shaken to dislodge as many cells aspossible, and centrifuged at 1000 rpm for threeminutes. The brush floated to the surface andwas discarded, and the pellet was spun down inthe same container at 3000 rpm for a further10 minutes. The formalin was poured off andthe pellet resuspended by vortexing in fourdrops of the gelling reagent (Reagent 2 of theCytoblock kit). The cell disc was prepared bycentrifuging the cells into the embeddingcassette according to the manufacturer'sinstructions. It was then routinely processed toparaffin wax and folded in half before embed-ding to obtain maximum thickness for sec-tioning.

Sections were cut at 4 ,um and were stainedwith haematoxylin and eosin. Unstained sec-

tions were mounted on silane treated slidesand immunostained with our murine mono-clonal antibody TDM39 raised against theXH1 cervical carcinoma cell line.4 A streptavi-din/biotin/peroxidase technique' with a rabbitanti-mouse biotinylated second antibody(Dako, Bucks) was used. Endogenous perox-idase was inhibited as described previously.6

Results

Twenty five out of 28 of the sections containedsufficient cervical epithelial cells for diagnosisand 14 also contained strips of either squa-mous or endocervical cells or both (figs 1, 2and 3). Three contained too few cells fordiagnosis. Four others had not been cut intodeeply enough, but with further experiencethis proportion should decrease. However, we

would prefer to cut deeper levels as required,rather than risk wasting previous material bycutting too far into the block to begin with.One further block contained relatively few cellseven when cut deeper, but as these showedsevere dysplasia the preparation was con-sidered satisfactory. Twelve blocks containedpus, often in large sheets, which might havemade a conventional smear difficult to inter-pret. In the Cytoblock sections sufficient epi-thelial cells for diagnosis were seen between the

*@S sheets of polymorphs (fig 3). There was sat-

isfactory immunostaining with our TDM39ngle cells. antibody (fig 4).

1122

on May 15, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.45.12.1122 on 1 Decem

ber 1992. Dow

nloaded from

Cytoblock preparations for examination of cervical and other cells

iv

A*:4#

art.

&; #̂t -, r 40 rfl.l a

6 -4"e

i _ ^g

. lb.%,1%

I

t... 4

,% I

Figure 2 Sheets ofpolymorphs surround squames showing some reactizenlargement but no dysplasia.

£

A.

0

A.f..t, ..

Figure 3 There is a sheet of koilocytic dysplastic focally keratinising cervical cells acentre. Surrounding this there is a mixture of dysplastic cells, cells showing inflammchange, normal mature cells, and a few polymorphs.

e*. ? wa

# 4

at 4 I. I4

* '-t .0

,. * a

<##,v],,%t . S

4 f

L

s ~~~*Figure 4 Inflamed smear immunostained with DM39 antibody using an ABCperoxidase technique. Two overlapping squames with perinuclear haloes zin the centrepositive, normal squames among the inflammatory cells are negative, while occasiorpolymorphs are weakly positive.

Discussion* Cytoblock preparations are easier to prepare

- than cell pellets using agar or gelatin.7 Both ofthese reagents have to be specially preparedand kept warm. The cells need to be cen-

.* trifuged, the agar or gelatin added, and whenset it is sometimes quite difficult to scrape outthe pellet, and those at the bottom of thecentrifuge tube may be left behind. When theagar or gelatin block is processed and thereforewarmed again, the cells often disperse aroundthe edges, and it is difficult to control the size

. of the block. When Cytoblocks are used the, blocks are of small uniform size, reduced

., i,¶* si quantities ofreagents can be used, and immuno-Z staining is more easily standardised. In a

:t M @ preliminary experiment we used direct cyto-. ,S, *j spin preparations, but each preparation fromf"le * the same scrape was potentially different from

p the next, and only a limited number could beI _-..' -P.' made from each sample. Laboratories with anve nuclear up-to-date Cytospin will already have the

stainless steel clips that hold the special cas-sette and loading funnel. The fixing and gellingreagents, the embedding cassette and Cyto-funnel presently cost 84p a block. However,the method is more satisfactory and reproduci-ble and somewhat less time consuming thanthe preparation of agar or gelatin blocks.Though these cost only IOp a block, they areless satisfactory especially when the sample isscanty.

Cytoblock preparations made by thismethod can be used for haematoxylin andeosin staining and immunostaining in the sameway as normal paraffin wax embedded cellpellets or tissue blocks. To characterise ourcervical carcinoma cell lines and nude mousexenografts, we carried out total genomic DNAin situ hybridisation on Cytoblock sections todistinguish between murine and human cells.8In situ hybridisation may also be used forhuman papillomavirus subtyping. Doublelabelling, to demonstrate a variety of antigensin the same section, can also be performed on

in the these preparations. A prospective study com-zatory paring conventional smears, Cytoblock pre-

parations, colposcopic findings and biopsyspecimens is under way and will be reportedelsewhere.

This work was supported by the Jean Shanks Foundation (BD),Birthright (DLSE), the Dunhill MedicalTrust (XH), and the StThomas's Hospital Research Endowments Fund (SM).

1 Van der Graaf Y, Vooi;s GP, Gaillard HLJ, Go DMDS.Screening errors in cervical cytologic screening. ActaCytol 1987;4:434-8.

2 Day NE. Screening for cancer of the cervix. J EpidemiolCommun Health 1989;43:103-6.

3 Waddell CA, Rollason TP, Amarilli JM, Cullimore J,McConkey CC. The cervex: an ectocervical brush sam-pler. Cytopathology 1990;1:171 -81.

4 Han X, Lyle R, Eustace DLS, et al. XH 1-a new cervicalcarcinoma cell line and xenograft model of invasion,'metastasis' and regression Br Jf Cancer 1991 ;64:645-54.

5 Hsu S-M, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase tech-niques: A comparison between ABC and unlabelledantibody (PAP) procedures. J Histochem Cytochem1981;29:577-85.

6 Heyderman E. Immunoperoxidase technique in histopa-thology: applications, methods, and controls. J Clin Pathol1979;32:971 -8.

L 7 Heyderman E, Ebbs SR, Larkin SE, Brown BME, HainesAMR, Bates T Response of breast carcinoma to endo-crine therapy predicted using immunostained pelletedfine needle aspirates. BrJ Cancer 1989;60:630-3.

8 Obara T, Conti CJ, Baba M, Resau JH, Trifillis A L, Trumpe are BF, Klein-Szanto AJP. Rapid detection of xenotrans-nal planted human tissues using in situ hybridization. Am J

Pathol 1986;122:386-91.

1123

I

.40 4r,Pt'.4p,

0

0 OF"

_M. _ld

I . .11 "..

on May 15, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.45.12.1122 on 1 Decem

ber 1992. Dow

nloaded from