J Clin Pathol 1990;43:334-336
Immunohistochemical detection of neuroblastomain frozen sections of bone marrow trephine biopsyspecimens
MM Reid, A J Malcolm, A G McGuckin
AbstractFrozen sections from bone marrowtrephine biopsy specimens from childrenwith disseminated neuroblastoma werestained using the monoclonal antibodyUJ13A. An immunoalkaline phosphatasetechnique was the preferred stainingmethod. Acceptable histological detailwas obtained from this material *anddeposits of tumour cells detected. Sqmeapparently fibrous tissue was also stainedby this antibody. The results show thatthis immunohistological approach isfeasible and provide encouragement forits addition to the range of investigationscurrently available for assessing themarrows of children with this disease.
When assessing children with disseminatedneuroblastoma, histological examination ofbone marrow trephine biopsy specimens ismore rewarding than either standard cyto-logical or immunocytochemical methods ofexamining aspirated material.' The interpreta-tion of conventional histological preparations,however, remains uncertain in many cases.2The ability to detect neuroectodermal antigensin the marrow cores of such patients mayimprove interpretation of their specimens.This report describes an immunohistologicaltechnique using frozen sections of marrowcores stained with the monoclonal antibodyUJ13A and shows that such an approach isfeasible.
Department ofHaematology, RoyalVictoria Infirmary,Newcastle upon TyneA J MalcolmAG McGuckinMM ReidDepartment ofPathology, UniversityofNewcastle uponTyneMM ReidCorrespondence to:DrMM Reid, Departmentof Haematology, RoyalVictoria Infirmary,Newcastle upon Tyne NE14LPAccepted for publication8 November 1989
MethodsFresh iliac crest trephine biopsy specimens,taken with a paediatric Jamshidi needle as partof routine staging or re-evaluation, were
divided into two portions. One was processedfor conventional histological examination, the
other embedded in a glycerol based compound(OCT, Miles Scientific), rapidly frozen usingisopentane cooled in liquid nitrogen, then
stored at - 70°C until required. The biopsyspecimens studied were obtained from four
patients with disseminated neuroblastoma at
diagnosis two to four months after startingtreatment. All patients were known to have
UJ13A positive tumour cells.
SECTIONING AND FIXATION
Sections, 7 um thick, were cut using a cryostat(Reichert-Jung Frigocut) at a temperaturebetween -15 and - 25°C. A variety of section
adhesives was tested, including aminopropyl-triethoxysilane (APES), poly-L-lysine, andchrome alum gelatin. This last was found to bemost useful in our hands and was usedsubsequently at all times. It was particularlynecessary to ensure adherence of sections ofhypoplastic marrow, with relatively normalarchitecture, which proved the most fragile.Sections were dried overnight at room tem-perature, fixed in acetone for 10 minutes, airdried and washed in TRIS buffered saline(TBS), pH 7-6, for two consecutive five minutewashes, followed by incubation in non-immune rabbit serum for 10 minutes.
IMMUNOSTAININGEarly experiments compared the indirectimmunoperoxidase with the comparableimmunoalkaline phosphatase technique. In theimmunoperoxidase technique heavy haemo-siderin deposits in many specimens appearedbrown and could be confused with positivelystained cells. Native peroxidase activity ofhaemopoietic cells was not always easily over-come. The ability to block endogenous alkalnephosphatase activity increased contrast. Theauthors also preferred the colour reactionproduct in the immunoalkaline phosphatasemethod. Immunoalkaline phosphatase wastherefore the preferred option.The primary antibody was UJ13A (kindly
donated by Dr J Kemshead, Tumour Im-munology Laboratory, Institute of ChildHealth). It is a murine monoclonal antibodyraised against fetal brain and recognises the celladhesion molecule N-CAM (personal com-munication, Dr J Kemshead) present onneuroectodermal cells. It was diluted with non-immune rabbit serum in TBS. Optimal work-ing dilutions were established for each batch ofantibody.
Secondary antiobody was an alkaline phos-phatase conjugated rabbit anti-mouse antibody(Dako, No D314). It was used at a final workingdilution of 1/10 in 20% non-immune rabbitserum in TBS.
Sections were incubated with UJ13A for 30minutes at room temperature in a humidenvironment. A test section was simul-taneously incubated with rabbit serum omit-ting primary antisera, which acted as a negativecontrol. A section from a primary tumourbiopsy specimen was used as a positive control.Sections were washed for 10 minutes in TBS.The secondary antibody was applied similarlyand washed in TBS. Alkaline phosphatase wasvisualised using naphthol AS-TR phosphate in
334
on Novem
ber 19, 2020 by guest. Protected by copyright.
http://jcp.bmj.com
/J C
lin Pathol: first published as 10.1136/jcp.43.4.334 on 1 A
pril 1990. Dow
nloaded from
Immunohistochemical detection of neuroblastoma
4,. ..r
'* T
*,*. .I,.,.
-~ ~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*w
..
11
Figure 2 Positive tumour cells, normal haemopoieticcells and negative tumour cells. (Immunoalkalinephosphatase.)
Figure I Positive tumour deposit after treatment. Note strongly positive eosinophilsandfaint background positivity. (Immunoperoxidase.)
a standard cytochemical technique. Endo-genous alkalinie phosphatase activity wasblocked with levamisole incorporated into theenzyme development medium. Sections werecounterstained with Mayer's haematoxylin,mounted in an aqueous mountant, and viewedwith a standard light microscope.
ResultsAlthough the preferred immunostaining tech-nique was immunoalkaline phosphatase, anexample of successful immunoperoxidasestaining is shown in fig 1. Acceptable histo-logical detail is essential for this technique to beworth pursuing. Examples of variable cellu-larity and architecture are shown in figs 2 to 4.
Figure 2 shows UJ13A positive tumour cells,normal marrow cells, and at least one deposit ofUJ13A negative tumour in a biopsy specimentaken four months after starting aggressivechemotherapy. Similar histological appear-
-
A~~~~~~~
Figure 3 Hypoplastic marrow with several small deposits of tumour cells aftertreatment. (Immunoalkaline phosphatase.)
ances had been found in the conventionallyhandled biopsy material. Figure 3 shows multi-ple small deposits of residual tumour in hypo-plastic marrow two months after startingtreatment. Figure 4 shows UJ13A positivetissue in streaming whorls of apparentlyfibrous material in marrow from an intensivelytreated patient. Within it are some cells withlarge, oval, or elongated nuclei whose identityis unclear. Similar appearances can be found atdiagnosis (data not shown). Figure 5 shows anumber of UJ13A positive osteoblasts alongthe edge of growing bone. Golden coloureddeposits of haemosiderin are shown in figs 3and 4, clearly contrasting with the red reactionproduct of the immunoalkaline phosphatasetechnique.
DiscussionFormalin fixed, paraffin wax embeddedtrephine biopsy specimens from children withneuroblastoma can be examined by immuno-histochemical techniques using, for example,antisera against neurone specific enolase.3Formalin fixation and decalcification pro-cedures can mask or even destroy many cellantigens, thus restricting the range of inves-tigations which can be carried out. UJ13Areacts strongly with most neuroblastomas, butthe antigen or epitope which it recognises isrendered undetectable by formalin.4 Frozensections or some alternative fixative5 arerequired for its use in histological sections.
Frozen, undecalcified bone marrow is anotoriously difficult tissue from which toobtain untraumatised sections which adherewell to the slide. Hypoplastic marrow withrelatively normal architecture provided thegreatest challenge. Nevertheless, with applica-tion and the readiness to persevere, acceptablehistological detail can be obtained, examples ofwhich are contained in this report. The labourintensive nature of the work, however, whencompared with handling paraffin wax em-bedded, decalcified material, is a drawback.The high level of non-specific peroxidase
activity in haemopoietic cells and thepreference by the authors for the colourreaction in the immunoalkaline phosphatasetechnique were the major factors affecting thechoice of immunostaining technique. Onlysimple, two stage immunoperoxidase andimmunoalkaline phosphatase techniques were
335
on Novem
ber 19, 2020 by guest. Protected by copyright.
http://jcp.bmj.com
/J C
lin Pathol: first published as 10.1136/jcp.43.4.334 on 1 A
pril 1990. Dow
nloaded from
Reid, Malcolm, McGuckin
Figure 4 Positivematerial within "fibrous"tissue, after treatment.(Immunoalkalinephosphatase.)
studied. The more complex peroxidase-anti-peroxidase (PAP) or alkaline phosphatase-anti-alkaline phosphatase (APAAP) techniquesmight have given both a stronger visual signaland reduced non-specific background staining.UJ13A stained frankly primitive cells in
marrow at diagnosis (data not shown). Infiltra-tion at diagnosis is usually easy to detect, andconventional histological, cytological, andimmunocytochemical stains are much simplerto perform. The use of this antibody in frozensections, however, draws attention to smalltumour deposits in hypocellular or fibroticmarrow after treatment has been started andalso shows that antigen negative tumour cells
Figure 5 Positive osteoblasts. (Immunoalkaline phosphatase.)
may coexist with positive cells in the samebiopsy specimen. The presence of immuno-positive material in hypocellular "fibrous"tissue both at diagnosis (data not shown) and,in some cases, after treatment, is worth noting.This material may be akin to that found in moredifferentiated or partially treated primaryneuroblastomas. It is possible that some of the"fibrous" material detected by conventionalhistological techniques is a product of neuro-ectodermal cells rather than reactive bonemarrow fibroblasts.
Osteoblasts often stained with UJ13A, con-firming previous observations by us' andothers,5 but their individual morphologicalappearances and restriction to the edge of bonetrabeculae usually makes their distinction fromtumour simple. This contrasts with the poten-tial problem of interpreting immunofluores-cence studies of cell suspensions. Osteoblastsare easily aspirated from the marrow of chil-dren, often as groups or clumps, and in inten-sively treated children may constitute a sub-stantial proportion of aspirated cells.
In summary, this report shows that, despitethe technical difficulties, it is possible to obtainuseful information from immunohistochemicalstudies of frozen bone marrow in this disease.These techniques could be extended to othermonoclonal antibodies. Comparison of a rangeof immunohistochemical techniques usingfrozen and formalin fixed material with con-ventional histological examination in a largeseries of children may lead to a clearer pictureof the importance of the various histologicalabnormalities of the marrow in children withdisseminated neuroblastoma.
A G McGuckin is supported by a grant from the North ofEngland Children's Cancer Research Fund.
1 Carey PJ, Thomas L, Buckle G, Reid MM. Immunocyto-chemical examination of bone marrow in disseminatedneuroblastoma. J Clin Pathol 1990;43:9-12.
2 ReidMM, Hamilton PJ. Histology ofneuroblastoma involv-ing bone marrow: the problem of detecting residualtumour after initiation of chemotherapy. Br J Haematol1988;40:487-90.
3 Ironside JW, Shiach CR, Cuthbert AC, Robinson EA.Histology of neuroblastoma involving bone marrow. Br JHaematol 1989;71:565-6.
4 Carey PJ, Malcolm AJ, Pearson ADJ, Kemshead JT, ReidMM. Immunocytochemical detection of tumours ofneuroectodermal origin. J Clin Pathol 1988;41:586-9.
5 Oppedal BR, Storm Mathisen I, Kemshead JT, BrandtzaegP. Bone marrow examination in neuroblastoma patients; amorphological immunocyto- and immunohistochemicalstudy. Hum Pathol 1989;20:800-5.
336
on Novem
ber 19, 2020 by guest. Protected by copyright.
http://jcp.bmj.com
/J C
lin Pathol: first published as 10.1136/jcp.43.4.334 on 1 A
pril 1990. Dow
nloaded from
