J Clin Pathol 1988;41:510-515

Monoclonal antibody EBM/1 1: high cellularspecificity for human macrophagesP M A KELLY, E BLISS, J A MORTON, J BURNS, J O'D McGEE

From the University of Oxford, Nuffield Department ofPathology, John Radeliffe Hospital, Oxford

SUMMARY A monoclonal antibody, EBM/11, was raised against isolated human lung macrophages.Immunohistochemically this antibody reacted with freshly isolated lung macrophages and bloodmonocytes, mononuclear cells (presumptive macrophages) in sections of lung, skin, stomach, smalland large bowel, pancreas, spleen, tonsil, placenta, liver, gall bladder, heart, thyroid, pituitary, brain,and peritubular and mesangial cells in kidney. Microglial cells and osteoclasts also labelled withEBM/1 1. The antibody reacted with cytoplasmic structures rather than with cell membranes. Theepitope recognised by EBM/l I was present on four polypeptides (of 120, 70, 64 and 22 kilodaltons). Itdid not react with any other cell type in the tissues screened except the epithelium of renal proximaltubules.

This antibody may be useful in identifying and elucidating the function of macrophages inpathological processes.

There is a need for reliable and sensitive reagents forthe identification ofhuman macrophages, particularlyin studies of lymphoproliferative disease, the immuneresponse, inflammation and neoplasia. To this end a-1-antitrypsin (al-AT),' lysozyme,2 peanut agglutinin(anti-T lectin),3 and non-specific esterase4 have beenused as specific markers for macrophages. All of these,however, are found in cells which are manifestly notmacrophages. For example, a1-AT is present not onlyin macrophages but also in polymorphs5 and inhepatocytes.6 Several monoclonal antibodies tomacrophages/monocytes have been produced to over-come the shortcomings of histochemical methods formacrophage/monocyte identification.7 Most of thesehave been used in studies of isolated monocytes but itis clear that many of them label cells other thanmacrophages/monocytes.9'The aim of this study was to produce monoclonal

antibodies to macrophages and to determine theirspecificity for this cell type. The long term objectivewas to use these reagents with high specificity formacrophages to elucidate macrophage functions inphysiological and pathological states.

Material and methods

PREPARATION OF MACROPHAGESSpecimens of fresh human lung resected for carcinomawere obtained from the Churchill Hospital, Oxford.

Accepted for publication 25 November 1987

510

and the East Birmingham Hospital. Lungs A-eretransported to this laboratory in supplementedLeibovitz medium (L15). The supplements consistedof 10 mmol/l 1-glutamine, 100 ji/ml each of benzvl-penicillin and streptomycin, and gentamycin 80 li ml(Warrick, UK). Unless otherwise specified. all cellculture reagents were obtained from Gibco Biocult(UK).About 50 g of macroscopically normal lung tissue

distant from the tumour was minced thoroughly withscissors with as little L-15 medium as was practicable.The disaggregated lung tissue was mixed with an equalvolume of L-15 medium, degasscd carefully in a 1000ml beaker in a vacuum desiccator at 760 torr after amethod similar to that ofMason et al.7 The componentcells were isolated by squeezing the mince throughsterile surgical gauze, then by filtering the eluted cellsin L- 15 medium through a steel tea strainer of largepore size. This cell isolation procedure was repeatedonce more on the disaggregated tissue, and the cellsfrom both isolates were pooled. The mince wassubsequently extracted similarly with L- 15 mediumadditionally supplemented with 6 mmol/l edetic acid(Sigma, UK), 5% fetal calf serum, 10 mmol/l lig-nocaine hydrochloride (Astra, UK), and enough 01mmol/l sodium bicarbonate (BDH, UK) to bring thepH to 70. Cells extracted with both media wereimmediately centrifuged at 120 g for 10 min in 50 mlFalcon tubes. Both cell pellets were pooled, resuspen-ded in fresh L-1 5 medium containing antibiotics onlyplaced on a Ficoll-Hypaque gradient (1077 g), andcentrifuged at 400 g for 20 minutes. The cells which

on August 13, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.41.5.510 on 1 May 1988. D

ownloaded from

High cellular specificity ofEBMIII for macrophagesconcentrated at the interface were removed, washed,and resuspended at 106 cells/ml in L-1 5 medium, andplaced in 80 cm2 culture flasks (Nunc, Denmark) in 80ml of RPM1 1640 medium supplemented with 5%fetal calf serum, benzylpenicillin (100 p/ml), strep-tomycin (100 i/ml), and gentamycin (80 p/ml). Afterincubation in a 5% carbon dioxide and 95% oxygenatmosphere for one hour at 37°C the adherent mono-layer was agitated and washed gently with two changesof the same medium, thus affording virtually puremacrophage monolayers. Macrophages were deta-ched from the plastic with a sterile rubber policeman.With very fresh lung specimens, a total yield of6 x 106 adherent macrophages/50 g of lung tissue wastypical, though with lung samples of up to four or fivehours old (transported in L-1 5 medium) the totalyields were typically reduced to 3 x 106 adherentmacrophages/50 g lung. These macrophages werejudged viable according to their ability to excludetrypan blue.

MONOCLONAL ANTIBODY PRODUCTION AND

SCREENINGAbout 107 purified macrophages in 0 2 ml Hanks'sbalanced salt solution (BSS) were mixed with an equalvolume of complete Freund's adjuvant and injectedsubcutaneously into Balb C mice. The animalsreceived three intraperitoneal booster doses of 107 cellsin 0 2 ml of BSS, without adjuvant, at two weekintervals. Three days after the final injection separatedmouse spleen cells (about I o8 cells) were filteredthrough a sterile stainless steel strainer and fused with107 mouse myeloma cells (NSI) using 50% polyeth-ylene 4000 (BDH, UK) in RPMI 1640 medium; thecell suspension was agitated for 90 seconds andincubated for a further five minutes at 22°C." Afterimmediate washing with RPM 1 1640 containinghypoxanthine, methotrexate, and thymidine (HAT)the cells were resuspended in 50 ml of RPM 11640containing HAT and 10% fetal calf serum, and platedinto Costar wells." This method afforded between 300and 400 hybrid colonies. Two hundred of thesecolonies were separated, subcloned, grown and frozenin liquid nitrogen.As a primary screen for macrophage reactive

antibodies culture supernatants were tested by anindirect immunoperoxidase procedure'2 on cryostatsections of human lung and tonsil fixed at - 20°C inacetone for 10 minutes. The cells from those super-natants, which stained lung or tonsil macrophages, orboth, were definitively cloned by dilution cloningusing either Terasaki microtitre plates or round-bot-tom (300 p1) 96 well plates with mouse spleen feedercells at 106/ml." The supernatants from these cloneswere tested immunohistochemically for antimacro-phage specificity, on a second more extensive organscreen. This consisted of acetone fixed (- 20°C for 20

511

minutes) cryostat sections of normal lung, tonsil,lymph node, spleen, liver, pancreas, stomach, smalland large bowel, gall bladder, kidney, heart, brain,pituitary, thyroid, placenta, and skin. Cytocentrifugedbuffy coat cells from human peripheral blood andisolated lung macrophages fixed in acetone were alsoincluded in this secondary screen. All tissues and cellsused in this secondary screen were obtained fresh atbiopsy, or from necropsy organs within 24 hours ofdeath.

Immunohistochemical controls consisted of tissuesor cells treated with culture medium alone, orirrelevant monoclonal antibodies to Mallory'sbodies'" (JMBI, JMB2), followed by rabbit antimouseIg conjugated with peroxidase (Dako, UK), or, in afew instances, rabbit antimouse followed by alkalinephosphatase-mouse antialkaline phosphatase com-plex. The conditions for immunohistochemicalanalysis are as described,'2 except that after theapplication of primary antibody (or control antibody)sections or cells for immunoperoxidase staining weretreated for 30 minutes with methanol containing 1%H202 (v/v) at 22°C and washed in PBS before incuba-tion with second antibody. In some tissues the reactionof EBM/1 1 was detected using the alkaline phos-phatase-antialkaline phosphatase method using astandard procedure.'2The immunoglobulin class of interesting mono-

clonal antibodies was determined by the Ouchterlonymethod using subclass specific antisera (Dako, UK).

WESTERN BLOTSLung macrophages were isolated and maintained forfour to 24 hours in L-1 5 medium supplemented withantibiotics (see above) and 10% heat inactivated fetalcalf serum. Cultures were washed twice in Hanks'sBSS, harvested with a rubber policeman, and themacrophages recovered by centrifugation at 1500 g forfive minutes. The cell pellet was dispersed in aminimum volume of a buffer solution containingTriton-X 100 (0 1%), 5 mmol/l edetic acid, 0-01 mmol/lTris-hydrochloric acid, 2 mmol/l phenylmethylsul-phonfluoride, 2 5 mmol/l iodoacetamide of pH 7 4.The cell suspension was sonicated for 30 seconds andcentrifuged at 1500 g for 15 minutes. Ten volumes ofabsolute ethanol were added to the supernatant,incubated at 22°C for five minutes, and theprecipitated protein recovered by centrifugation(12 000 g for five minutes). The pellet was dissolved at80°C for five minutes in 5% sodium dodecyl sulphate(SDS) containing 2% mercaptoethanol. Proteins wereelectrophoresed on 7-5% polyacrylamide gels in anelectrophoresis buffer containing 80 mmol/l Tris, 50mmol/l boric acid, and 0 1% SDS at pH 8 5; molecularweight standards were run on each gel. Proteins weretransferred (in an electric field for 16 hours) tonitrocellulose filters (0.45 pm pore size) using a buffer

on August 13, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.41.5.510 on 1 May 1988. D

ownloaded from

Kelly, Bliss, Morton, Burns, McGee

.- t

I

I .0.:: v~~00

;r...41, '

.:6'# _:".6

',:

..I

s{. o..

. s..b b

.i*, .\- 54,..

.,4.',XI4f .SX B/4 $*1-

Fig I Macrophages in alveolar air spaces and some alveolarlining cells are EBMIJJ positive. EBMIJ1 reaction iscytoplasmic and granular. EBM/JI reactivity was visualisedby indirect immunoperoxidase histochemistry (unlessotherwise stated). Sections are counterstained withhaematoxylin.

,''* w^ 4.

,.i'X,X~~~~~~~1m7Fig 3 Lymph nodes: interdigitating reticulum cells, sinusmacrophages, and germinal centre macrophages are ERM/JJpositive. (APAAPprocedure.)

containing 25 mmol/l Tris, 0-2 mol/l glycine, 0-25%Tween 20 (v/v), and 20% methanol (v/v) at pH 8-3.Filters were washed for 60 minutes in PBS containing0-5% Tween 20 (v/v) (PBS/Tween), incubated at 22°Cfor two hours in EBM/I I (ascites) diluted 1/200 in thesame buffer containing 0-1% human AB serum, andwashed three times for 15 minutes in PBS/Tween.Filters were incubated at 22°C for 60 minutes in goatantimouse Ig (Dako, UK), diluted 1/20 in PBScontaining 3% bovine serum albumin (v/v), andwashed three times in PBS/Tween. A complex ofmouse monoclonal antibody to alkaline phosphataseand alkaline phosphatase (Dako, UK) was added tothe filters for 60 minutes at 22°C. After washing in PBS(four lots of 10 minutes) the labelled bands werevisualised by incubating filters in PBS containing nitroblue tetrazolium and 5- bromo4chloro-3 indolylphosphate for about 20 minutes at 22°C.'4 Molecularweights of EBM/1 1 containing polypeptides were

Fig 2 Hepatic Kupifer cells and mononuclear cells in portaltracts are EBMI/ I positive. Reaction product is granular.Brown granules in hepatocytes are lipofuchsin.

-4.-...'.; ,4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4

Fig 4 Macrophages in gastric lamina propria are EBM/I IIpositive.

calculated from the migration distances of the markerproteins.

Results

Twenty monoclonal antibodies in the primary screenreacted with pulmonary and tonsillar macrophages.Of these, 18 reacted with other cell types in themultiorgan screen; in general, these antibodies reactednot only with macrophages but also with stratifiedsquamous epithelia, other epithelia or polymorphsand lymphocytes, or a combination. Of the remainingtwo antibodies, one, designated EBM/ 11, reacted withalmost all macrophages and monocytes, and the otheronly with lung macrophages. The characteristics oftheother antibody will be described elsewhere.EBM/ 1I reacted with isolated alveolar macro-

phages and with the same cells in intact lung (fig 1),Kupffer cells in liver (fig 2), and littoral cells in splenic

512

.,w t.. i

:. IT i:I

t'

4

on August 13, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.41.5.510 on 1 May 1988. D

ownloaded from

High cellular specificity ofEBMIJJ for macrophages

f:IL . .X

'S^.

Si

¾',tE *s -

Fig 5 Epidermal Langerhans' cells and mononuclear cells inupper dermis are EBMIJI positive (APAAP procedure.)

J, !

0.

'9@

I 4%,~~~~I

@6- M?,l 4b . SAFig 6 Renalperitubular mononuclear cells are EBMIJIpositive.

Fig 7 All monocytes in this buffy coat preparationfromperipheral blood are EBMIH positive. All other white cellsare EBM/IJ I negative.

red pulp. Interdigitating cells ofT zones in tonsil andlymph node (fig 3) also reacted but not dendriticreticulum cells. Mononuclear cells (with a macrophagemorphology) in the lamina propria of stomach, small,and large bowel were also EBM/l positive (fig 4).Some, but not all, Langerhans' cells of skin stainedstrongly (fig 5). EBM/l also stained mononuclearcells in connective tissue of skin (fig 5) tonsil, portaltracts of liver, gall bladder, placenta, heart andbetween muscle bundles in the alimentary tract.Mononuclear cells around thyroid follicles andrenal tubules (fig 6) mesangial cells in glomeruli,and occasional nonendocrine mononuclear cells in thepituitary also reacted with EBM/I 1. Microglial cells inbrain also labelled with EBM/I1 L" Osteoclasts,'6multinucleate giant cells in giant cell myocarditis,'7megakaryocytes and some mononuclear cells in bonemarrow also stained.'61'

In buffy coat preparations the antibody stainedmonocytes but no other white cells (fig 7).The antibody also stained stellate mononuclear cells

(presumptive macrophages) in the stroma of invasivebreast cancers'9 but not the neoplastic epithelial cells

Fig 8 Invasive ductal cancer ofbreast. Malignant cells areEBMI I1 negative. Many mononuclear stromal cells areEBMIJ I positive.

Table Distribution ofEBMI II positive cells in normaltissues

Organ Cell types

Lung Alveolar macrophagesLiver Kupffer cells, portal tract mononuclear

cells2'Gall bladder MacrophagesSkin Dermal macrophages; Langerhans' cellsBrain Microilial cells; perivascular mononuclear

cellsKidney Mononuclear cells around renal tubules;

mesangial cells in glomeruliSpleen Macrophages and dendritic cells in red and

white pulp"Lymph nodes Interdigitating reticulum cells; sinusoidal

and germinal centre macrophagesTonsil Macrophages (B and T cell zones)Stomach Macrophages in villi and mononuclearIleum cells in circular and longitudinalColon J muscle coatsHeart Mononuclear cells between muscle bandsPancreasStromal mononuclear cellsThyroid Mononuclear cells around folliclesPituitary Nonendocrine mononuclear cellsBuffy coat Monocytes onlyBone Osteoclasts,'6 mononuclear cells,'6

megakaryocytes"

*9,

4

4

40%i. .. 4

513"qw-

to .,.

ft. I

lixt Iwl

A %b "

t

... f

-%. A16 a

on August 13, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.41.5.510 on 1 May 1988. D

ownloaded from

514

Fig 9 Western blot ofisolated lung macrophage proteins.EBM/I 1 reacts withfour polypeptides ofrelative molecularmass 120, 70, 64 and 22 kilodaltons.

(fig 8). Unlike some other monoclonal antibodies tomacrophages, EBM/11 stains intracytoplasmic gran-ular structures (figs 1 and 2) but not macrophage cellmembranes. The cells with which EBM/ 11 reacted arelisted in the table.EBM/ 1I gave a weak reaction with the epithelium

of proximal convoluted tubules of kidney.'8 In somefreeze dried liver preparations hepatocytes give a weakfocal reaction which was distinguishable from lipo-fuchsin granules. Not all normal or diseased liversshowed this phenomenon and this is being studiedfurther.EBM/I 1 reacted with an epitope on four polypep-

tides from lung macrophages on Western blots (fig 9).These have relative molecular masses of 120, 70, 64and 22 kilodaltons. The immunoglobulin subclass ofEBM/l is IgGI .

Discussion

EBM/I1 reacts with all cells included in the classic

Kelly, Bliss, Morton, Burns, McGeemononuclear/phagocyte system such as Kupffer cells,sinusoidal cells of spleen, and alveolar macrophages.It also labels cells with a macrophage morphologyaround renal tubules and thyroid follicles, non-endocrine cells of the pituitary, and Langerhans' cellsof the epidermis; all of these cells have now beenincluded in the mononuclear/phagocyte system asdefined by their reactivity with F4/80 in mouse tissue.20In man, there is phenotypic evidence that Langerhans'cells belong to the mononuclear/phagocyte system.2 22Franklin et alestimated that 90% of Langerhans' cellsstained with EBM/l 1 ifHLA-DR positivity is taken asthe standard.'8 Interestingly microglial cells werepositive with EBM/ 1.15 Del Rio Hortego assumedthat microglial were part of the mononuclear/phagocyte system.23 Mouse microglia also reacts withF4/80.'520 Some glomerular mesangial cells reactedwith EBM/l 1; there is corroborative evidence thatsome mesangial cells are part of the mononuclear/phagocyte system.24 EBM/ 1I has been shown to reactwith osteoclasts,'6 megakaryocytes, and marrowmononuclear cells.'618 These three cell types thereforehave a phenotypic similarity to macrophages.Additionally, all of them react with other antibodieswhich have broad specificity for macrophages.8 Ath-anasou discusses the association between osteoclasts,megakaryocytes, and mononuclear phagocyte systemcells.25The molecular and cellular nature of the antigen

with which EBM/I 1 reacts is now the subject ofinvestigation. The data from Western blotting indicatethat it reacts with four major polypeptides fromisolated lung macrophages of relative molecularweights 120, 70, 64 and 22 kd. These polypeptidesdiffer in mass from those identified by otherantimacrophage monoclonals. The functions of thesepolypeptides are not yet known. EBM/ 1I also differsfrom most other monoclonal antibodies to macro-phages in staining intracytoplasmic structures whichhave a granular appearance (fig 7) on light micros-copy; many other antimacrophage monoclonals reactwith cell membrane components.7 Preliminary ultra-structural evidence indicates that the antibody mayreact with rough or smooth endoplasmic reticulum; itdoes not react with mitochondria, Golgi apparatus, orlysosomes (PMA Kelly, I Ferguson, and J O'DMcGee, unpublished observations). The similaritiesand differences of EBM/ 11 reactivity compared withother macrophage monoclonals is discussed at lengthby Franklin et al.'8 At the Third International Work-shop on Human Leucocyte Differentiation Antigens,EBM/ 1I was assigned to group 12, the mixedantimacrophage group. Within this group it isincluded in subgroup 4, which includes thoseantibodies with broad specificity.8 Unlike many of theantibodies in this group, it shows strong reactivity with

on August 13, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.41.5.510 on 1 May 1988. D

ownloaded from

High cellular specificityfor macrophages with EBMIJJ 515

peripheral blood monocytes.The screening procedure used here for the genera-

tion of EBM/1 1 is instructive. Although 20 mono-clonal antibodies were produced which reacted withthe primary immunogen (lung macrophages), only oneof these showed high cellular specificity for cells of themononuclear/phagocyte system. This emphasises theimportance of screening monoclonal antibodiesagaitst large numbers of intact tissues before claimsare madefor cellular specificity. We have shown in thispaper that EBM/ 1I labels known or presumptive cellsof the human mononuclear/phagocyte system in nor-mal tlssues with a high degree of specificity.

It has also been shown to label multinucleate cells ingiant cell myocarditis'7 and in giant cell tumour ofbone,26 the tumour cells in a single case of malignanthistiocytosis (unpublished observations), and alsopresumptive macrophages in human breast cancer(fig 8).' There has been controversy about the natureof these cells in these various diseases."26 We believethat the demonstration of the shared relatively specificEBM/1 1 positive immunophenotype between thesecells and normal mononuclear/phagocyte system cellshas contributed to the understanding of these diseasesand that this antibody will assist in the study of otherdisease processes in man.

This work was supported by a grant from the CancerResearch Campaign (J O'D McGee). PMAK holds atravelling studentship from the National University ofIreland. We thank Miss Lesley Watts for typing themanuscript.

Refereomes

1 Isaacson P, Jones DB, Millward-Sadler GH, Judd MA, Payne S.Alpha-l-antitrypsin in human macrophages. J Clin Pathol198 1;34:982-90.

2 Mason DY, Taylor CR. The distribution of muramidase(lysozyme) in human tissues. J Clin Pathol 1975;28:1488-503.

3 Howard DR, Batsakis JG. Peanut agglutinin: a new marker fortissue histiocytes. Am J Clin Pathol 1982;77:401-8.

4 Yam LT, Li CY, Crosby WH. Cytochemical identification ofmonocytes and granulocytes. Am J Clin Pathol 1971;55:283-90.

5 Benitez-Bibiesca L, Frere-Horta R. Immunofluorescent localisa-tion of alpha-I -antitrypsin in human polymorphonuclearleucocytes. Life Sci 1978;21:99-104.

6 Triger DR, Millward-Sadler GM, Czaykowski AA, et al. Alpha-l-antitrypsin deficiency and liver disease in adults. Q J Med1976;45:351-72.

7 Mason R, Austyn J, Brodsky F, Gordon S. Monoclonalantimacrophage antibodies: human pulmonary macrophagesexpress HLA-DR (Ia-like) antigens in culture. Am Rev RespirDis 1982;125:586-93.

8 Hogg N, Horton MA. Myeloid antigens: new and previouslydefined clusters. In: McMichael AN, Ed., Leucocyte typing I11.Oxford: Oxford University Press, 1987:576-629.

9 Breard J, Reinherz EL, Kung PC, Godstein G, Schlossman SF. A

monoclonal antibody reactive with human peripheral bloodmonocytes. J Immunol 1980;124:1943-8.

10 Hogg N, Slusarenko M, Cohen J, Reiser J. Monoclonal antibodywith specificity for monocytes and neurons. Cell 1981;24:875-84.

11 Bishop CE. A miniaturised single-step method of cell cloning.J Immunol Methods 1981;46:53-62.

12 Gatter KC, Falini B, Mason DY. The use of monoclonalantibodies in histopathological diagnosis. In: Anthony P,MacSween RNM, eds. Recent advances in histopathology. No.12. Edinburgh: Churchill-Livingstone, 1984:35-67.

13 McGee JO'D, Morton JA, Barbatis C, et al. Monoclonalantibodies to Mallory bodies/intermediate filaments and HLA(Class 1) antigens in human liver disease. In: McMichael AJ,Fabre JW, eds. Monoclonal antibodies in clinical medicine.London: Academic Press, 1982.

14 Chan VT-W, Fleming KA, McGee JO'D. Detection of sub-picogram quantities of specific DNA sequences on blothybridization with biotinylated probes. Nucleic Acids Res1985;13:8083-91.

15 Esiri MM, McGee JO'D. A monoclonal antibody to macrophages(EBM/ 1 ) labels macrophages and microglial cells in humanbrain. J Clin Pathol 1986;39:615-21.

16 Athanasou NA, Heryet A, Quinn J, Gatter KC, Mason DY,McGee JO'D. Osteoclasts contain macrophage and mega-karyocyte antigens. J Pathol 1986;150:239-46.

17 Theaker JM, Gatter KC, Heryet A, Evans DJ, McGee JO'D.Giant cell myocarditis: evidence for the macrophage origin ofthe giant cells. J Clin Pathol 1985;38:160-4.

18 Franklin WA, Pulford K, Falini B, et al. Immunohistologicalanalysis of human mononuclear phagocytes using monoclonalantibodies. Lab Invest 1986;54:322-35.

19 Kelly PMA, Davison RS, Bliss E, McGee JO'D. Macrophages inhuman breast diseases. A quantitative immunohistologicalstudy. Br J Cancer (in press).

20 Lee S-H, Starkey P, Gordon S. Quantitative analysis of totalmacrophage content in adult mouse tissues. J Exp Med1985;161:475-89.

21 Wood GS, Turner RR, Shiurba RA, Eng L, Warnke RA. Humandendritic cells and macrophages. Am J Pathol 1985;119:73-82.

22 Crocker J. Reticulum cells and related structures in lymph nodes:their properties and roles in antigen processing. In: ThompsonR, ed. Recent advances in clinical immunology. No. 4, Edinburgh:Churchill Livingstone, 1987:19-43.

23 Del Rio Hortego P. In: Penfield's cytology andcellular pathology ofthe nervous system. Vol 2., New York: Hoeber, 1932:483-534.

24 Michielsen P, Creemers J. The structure and function of theglomerular mesangium. In: Dalton AJ, Haguenan F, eds.Ultrastructure of the kidney. New York: Academic Press,1967:57.

25 Athanasou NA, Quinn J, McGee JO'D. Immunocytochemicalanalysis of the human osteoclast: phenotypic relationship toother marrow derived cells. Bone and Mineral 1988;3:317-33.

26 Athanasou NA, Bliss E, Gatter KC, Heryet A, Woods CG, McGeeJO'D. An immunohistological study of giant cell tumour ofbone: evidence for an osteoclast origin ofthe giant cells. J Pathol1985;147: 153-9.

27 Barbatis C, Kelly P, Greveson J, Heryet A, McGee JO'D.Immunocytochemical analysis of HLA Class II (DR) antigensin human liver disease. J Clin Pathol 1987;40:879-84.

Requests for reprints to: Professor J O'D McGee, NuffieldDepartment of Pathology, John Radcliffe Hospital, Level 1,Headington, Oxford OX3 9DU, England.

on August 13, 2021 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.41.5.510 on 1 May 1988. D

ownloaded from