28Clin Pathol 1994;47:248-252

A new monoclonal antibody (3A5) that recognisesa fixative resistant epitope on tissue macrophagesand mornocytes

E H Jaspars, E Bloemena, P Bonnet, R J Scheper, E Kaiserling, C J L M Meijer

AbstractAims-To develop a monoclonal antibodyspecific for hnman macrophages in rou-tinely processed material.Methods-The monoclonal antibody wasderived from a mouse popliteal lymphnode after subcutaneous immunisation inthe footpad with fragments of humanspleen depleted of lymphocytes and ery-throcytes.Results-3A5 is a monoclonal antibodyreactive with macrophages, monocytes,and histiocytes in routinely processed(formalin fixed, paraffin wax embedded)human tissue specimens. Unlike the wellknown panmacrophage marker KPl(CD68), neither dendritic cells (inter-digitating cells, Langerhans' cells, andmicroglia) nor myeloid, lymphoid, orepithelial cells stained with 3A5.Conclusion-As the staining pattern of3A5 is restricted, compared with othermacrophage markers and the recognisedepitope survives common fixation andembedding procedures, 3A5 is a valuablemarker for histiocytes and macrophagesin routine diagnostic applications.

7 Clin Pathol 1994;47:248-252)

Department ofPathology, FreeUniversity Hospital,Amsterdam, TheNetherlandsE H JasparsE BloemenaP BonnetR J ScheperC J L M MeijerDepartment of SpecialHistopathology andCytopathology,Institute ofPathology,University ofTfibingen, GermanyE KaiserlingCorrespondence to:E H Jaspars, Department ofPathology, Free UniversityHospital, De Boelelaan1117, 1007 MBAmsterdam, TheNetherlands.

Accepted for publication14 October 1993

Macrophages are important to the immunesystem: they are the main effector cells in theinnate immune response and are also indis-pensable in adaptive immunity because theyprocess antigen, secrete soluble factors, andguide lymphocytes in their essential recircula-tion."1 Unlike the numerous antigens on Tand B lymphocytes which have been identifiedover the past few years, specific markers forhuman macrophages are rare and not yet wellcharacterised. An important reason for this isthat the mononuclear phagocyte system(MPS) is morphologically and functionallydiverse, and the differentiation state does notalways correlate with the phenotypic appear-ance nor with the function of the cell.

Several antibodies have been raised againstmonocytes/macrophages, but most ofthem donot recognise all the cells or are not lineagespecific. Recently, CD68 was introduced bythe Fourth International Workshop onLeucocyte Differentiation Antigens as a 110kilodalton antigen, reactive with MPS cells.'Ever since, several anti-CD68 monoclonalantibodies have been described which differsubtly in their immunohistochemical stainingpatterns, possibly because of variations in gly-

cosylation of the CD68 protein in differentcell types.45 The functional aspects of CD68are still unclarified, mainly due to its localisa-tion in lysosomal granules,6 which hindersfunctional studies with antibodies.

Besides the need for lineage specific anti-bodies in fundamental research, pan-macrophage markers are also of great value indiagnostic pathology. Because of the greatmorphological heterogeneity in macrophagecells, a clear distinction from other cells canbe extremely difficult, especially when themacrophages have undergone malignanttransformation. Among the expanding panelof anti-CD68 monoclonal antibodies KP1 isone of the most widely used in routine diag-nostic pathology, merely because it stains cellson paraffin wax embedded tissues.7 However,besides macrophages this monoclonal anti-body also reacts with myeloid cells,8 "plasma-cytoid" T lymphocytes,9 and epithelial cells incertain disorders.'°01

MethodsPRODUCTION OF MONOCLONAL ANTIBODIESA spleen specimen, surgically removedbecause of a vena lienalis thrombosis, wasdepleted of lymphocytes and erythrocytes bycutting and squeezing it through a nylongauze. The remaining substance was frag-mented by a Polytron blender, suspended inphosphate buffered saline (PBS), emulsifiedin Freund's Complete Adjuvant (1 in 1), andinjected subcutaneously (three times) in thefootpads of Balb/C mice (10lOIfootpad).Four days after the last injection, lymphocytesfrom the popliteal lymph nodes were fusedwith Sp20-Agl4 mouse myeloma cells, fol-lowing standard techniques, as describedbefore."3 The supematant fluids were initiallyscreened on cryostat sections ofhuman spleenand tonsil for immunohistochemical reactivitywith macrophages. Further testing was per-formed on paraffin wax embedded material.The selected hybrid (3A5) was subclonedthree times.

For double staining (see below), the follow-ing monoclonal antibodies were used: theanti-CD68 KP1, commercially obtained fromDako reagents (Copenhagen, Denmark), andPG-Mi, kindly provided by Dr B Falini(Perugia, Italy).

CELLS AND CELL LENESRoutine bone marrow smears taken fordiagnostic purposes were obtained fromthe Pathology Department of Westeinde

248

on June 25, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.47.3.248 on 1 M

arch 1994. Dow

nloaded from

http://jcp.bmj.com/

A new monoclonal antibody for use on tissue macrophages and monocytes

Hospital, The Hague, The Netherlands. Theslides were air-dried and stored at - 20°C.Before use they were fixed in acetone for 10minutes. Peripheral blood mononuclear cellswere prepared from anticoagulated bloodby Ficoll-Hypaque density centrifugation.Monocytes were purified by counterflowcentrifugation elutruation as describedbefore. 14THP-1 (a monocytic leukaemia cell line),

U-937 (a B cell lymphoma cell line withmonocytic characteristics), K-562 (an erythro-leukaemia cell line with pluripotential differ-entiation abilities), and HL60 (a myeloid cellline) were cultured in RPMI 1640 containing10% fetal calf serum, supplemented with 4mM glutamin, 10 U/ml penicillin, and 0-1mg/ml streptomycin.

Peripheral blood monocytes and cell lineswere stimulated for 0 5, 1, or 2 hours inRPMI medium containing 100-500 U/mlinterferon y, 10-40,ug/ml lipopolysaccharide(LPS), or 5-20 nM phorbol myristate acetate(PMA). Immunocytochemical staining of thecells was performed on cytospin preparations,which were air-dried, fixed in acetone, andstored at - 20°C.

FIXATION PROCEDURESMost of the tissues were obtained from surgicalprocedures in our own clinic (SurgicalDepartment, Free University Hospital,Amsterdam, The Netherlands). Some of thespecific histiocytic lesions were kindly pro-vided by Professor H Kerl (Gratz,Switzerland) and Dr M Santucci (Florence,Italy). Part of the tissue was snap frozen andstored in liquid nitrogen. Frozen sections (4,um) were cut, mounted on poly-L-lysinecoated glass slides, dried and fixed in acetonefor 10 minutes before performing immunohis-tochemistry. Other part of the tissues werefixed in Sensofix (Sensomed, Beuningen, TheNetherlands), sublimate buffered formalin orformalin, and embedded in paraffin wax.Sections (4,m) were cut and stored at 4°C.Before immunostaining, the slides weredewaxed in xylol and rehydrated in a gradedseries of ethanol.

IMMUNOPEROXIDASE STAININGFor immunohistochemistry, a standard two-step immunoperoxidase staining procedurewas performed as described before.'5 Toimprove the staining results of 3A5, the for-malin fixed tissues were preincubated with 10mM citrate buffer (pH 6) for 5 minutes at 720Watts in a microwave oven. Frozen sectionsand material fixed in Sensofix or sublimatebuffered formalin needed no additionaltreatment. Isotype analysis was performedusing isotype specific peroxidase conjugates(Serotec, Oxford, England).

For double staining the slides were firststained with the first monoclonal antibody bythe two-step immunoperoxidase procedureusing 3f3 diamino benzidine tetrahydrochlo-ride as a substrate; the second monoclonalantibody was visualised by the alkaline phos-phatase method, essentially as described.'5 16

ResultsThe hybridoma supernatant fluids were ini-tially screened on frozen sections of humantonsil tissue and selected for specific reactivitywith tissue macrophages. The selected anti-body (3A5) also stained macrophages inparaffin wax embedded material and provedto be of the IgG2B subclass.

IMMUNOHISTOCHEMICAL REACTIVITY OF 3A5ON TISSUE SECTIONSNormal tissuesIn lymph nodes, tonsils, and mucosa-associ-ated tissue (MALT), 3A5 strongly stained thestarry sky macrophages in germinal centres(fig 1), but also macrophages in T cell areasand in the sinuses. Follicular dendritic cells(FDC) and interdigitating cells (IDC) did notstain; neither did lymphocytes, myeloid cells,and other cell types. The white pulp of thespleen showed a comparable staining patternwith the other lymphoid tissues; red pulpmacrophages were also positive. Furthermore,3A5 reacted with most macrophages and histi-ocytes in all other tissues investigated (table1). Remarkably, cells from the Langerhans-interdigitating cell series and microgliashowed no reactivity.

Haemopoietic cellsIn peripheral blood smears monocytes showedintracytoplasmatic staining with 3A5. Somereactivity was also observed within basophilicgranulocytes, but no other cell types werepositive (table 1). The immunocytochemicalstaining of peripheral blood monocytes wasgranular and rather weak. Expression of the3A5 defined antigen could not be upregulatedby exposing the cells to various concentrationsof monocyte activating agents, like PMA,interferon y, or LPS.

In bone marrow smears only macrophagesand mast cells stained with 3A5.Megakaryocytes and myeloid cells did not,nor did other blood forming elements.Monocyte precursor cells could not bedetected by the antibody.

Cell linesOnly the THP-1 monocytic line showed some

Table 1 Immunoreactivity of3A5 on normal tissues

Lymphoid tissue:Tonsil/lymph node Starry sky macrophages,

interfollicular macrophages, sinusmacrophages

Spleen Macrophages in red and white pulpThymus Medullary and cortical macrophagesBone marrow Macrophages, basophilic

granulocytes and mast cells. Noother haemopoietic cells

Non-lymphoid tissue:LungKidneyLiver

Skin

PancreasGutTestisUterusThyroidBrain

Alveolar macrophagesInterstitial macrophagesKupffer cells, periportal

macrophagesDermal macrophages (no

Langerhans' cells)Interstitial macrophagesLamina propria macrophagesInterstitial macrophagesInterstitial macrophagesInterstitial macrophagesPerivascular macrophages (no

microglia)

249

on June 25, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.47.3.248 on 1 M

arch 1994. Dow

nloaded from

http://jcp.bmj.com/

20aspars, Bloemena, Bonnet, Scheper, Kaiserling, Meijer

-A.:

:i~4r: i,

"

4:

Figure 1by 3A5.

Starry sky macrophages in a germinal centre ofa reactive lymph node, stained

a AhAMt ; .. s v

;d ' tiA+-93:' W ,- *$S;§*t- r iA~~~~~~~, * 9'.-,

A new monoclonal antibody for use on tissue macrophages and monocytes

with the monoclonal antibody (fig 4).However, the tumour cells in histiocytosis Xand in the different variants of malignantfibrous histiocytoma lacked the antigen.Mast cell disorders-Reactive and neoplasticmast cells in almost all cases (urticaria pig-mentosa, cutaneous mastocytoma, systemicmastocytosis and malignant mastocytosis)stained, though the staining was much weakerthan in macrophages and histiocytes.Epithelial disorders-In sialadenitis, adeno-

* lymphoma of the salivary gland and hyper-plastic polyps of the colon, 3A5 stained onlymacrophages and histiocytes. In contrast,KP1 showed reactivity with epithelial cells in

* these disorders, as has recently beendescribed.10

Figure 4 A giant cell tumour of the synovium showing strong reactivity with 3A5 in thegiant cels.

Figure S Immunohistochemical analysis of interfollicular cells in a lymph node reactivewith 3A5 (brown) and KP1 (blue). In this experiment the slides were first incubated with3A5. Note the variable double staining ofsome of the cells.

a;i'a'.N~~~~~~~J-43

Figure 6 Immunohistochemical double staining of macrophages in a reactive lymph node,first incubated with 3A5 (brown), followed by PGM-1 (blue). G = germinal centre,I = interfollicular area.

IMMUNOHISTOCHEMICAL DOUBLE STAININGTo compare 3A5 with other macrophagemarkers-KPl and PG-Mi-double stainingwas done on tonsil tissue sections. When 3A5was added as the first monoclonal antibody,followed by KP1, the different macrophagecells showed diverse staining patterns. Instarry sky macrophages very little or no KP1reactivity remained; interfollicular macro-phages showed a variable, but clearly dis-cernible, two-coloured intracytoplasmaticstaining (fig 5). In this area some cells withdendritic morphology were only positive forKP1. When the slides were first incubatedwith KP1, no further reactivity with 3A5 wasobserved.

Double staining analysis with 3A5 and PG-Ml, adding 3A5 first, showed that PG-Mlalso stained some cells with dendritic mor-phology in the paracortex which were 3A5negative, although less than KP1 (fig 6).However, no double staining cells weredetected. When PG-Mi was added first inthese experiments, no further 3A5 reactivitycould be detected.

Discussion3A5 is a highly specific monoclonal antibodyfor the macrophage lineage, as has beenshown in an extensive immunohistochemicalanalysis of normal and diseased human tis-sues. Besides basophilic granulocytes andmast cells, it does not cross-react with othercell types. The detection of basophilic granu-locytes and their tissue derivatives by 3A5,however, is not very surprising, as it hasrecently been shown that these cells are veryclosely related to the monocytic lineage.'7According to its tissue distribution, the 3A5defined antigen is very likely to belong to theCD68 cluster. Compared with KP1, one ofthe most widely used anti-CD68 monoclonalantibodies,7 3A5 shows a far more restrictedstaining pattern, excluding myeloid cells, den-dritic cells (IDCs, Langerhans' cells, andmicroglia), lymphoid cells and epithelial cells.In immunohistochemical double stainingexperiments 3A5 and KP1 hinder each otherfrom binding, suggesting that the monoclonalantibodies react with partly overlapping epi-topes. Because 3A5 cannot block all KP1

I'4. t.. _k

4 .e

._

411,., It.

it ,

251

.,FAI

II'qk.

.o41

.-A

.I -1.

I

.j

ii .,

'i.

on June 25, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.47.3.248 on 1 M

arch 1994. Dow

nloaded from

http://jcp.bmj.com/

22aspars, Bloemena, Bonnet, Scheper, Kaiserling, Meijer

reactivity in this sequential staining method, itseems that part of the KP1 antigen lacks aspecific structure that is necessary for 3A5recognition. Apparently, these KP1 positive,3A5 negative molecules are variably presentwithin the different macrophages of lymphoidtissue. As the 3A5 defined epitope is moststrongly expressed in the more differentiated cellsand its expression is not influenced by variousactivating agents, 3A5 may detect a differ-entiation rather than an activation molecule.

In immunohistochemical staining patterns3A5 shows remarkable similarity with themacrophage-restricted anti-CD68 mono-clonal antibody PG-Ml.12 Double stainingexperiments on tonsil and lymph node sec-tions, however, indicated a small populationof paracortical cells with dendritic morphol-ogy, not reactive with 3A5 and still PG-Mipositive. In contrast to 3A5, PG-Mi does notstain normal or reactive mast cells.'8 There-fore, both monoclonal antibodies apparentlydo not recognise exactly the same epitope.

Like KP1 and PG-Mi, the 3A5 definedepitope is resistant to routine fixation proce-dures, which makes it very useful for diagnos-tic pathologists. Moreover, 3A5 does not reactwith undifferentiated tumour cells other thanthose of histiocytic origin. The fact that histio-cytosis X cells did not react can be explainedby the fact that these tumours are derivedfrom Langerhans' cells, which are also nega-tive with the monoclonal antibody. No reac-tivity was observed in tumour cells of thedifferent types of malignant fibrous histio-cytoma. In spite of their name, however, thetrue origin of these tumours still remainshighly controversial. 19 20 Our findings supportthe idea that (at least part of) the malignantfibrous histiocytomas show no histiocytic dif-ferentiation.

These data led us to conclude that 3A5 isnot only a highly specific marker for macro-phages, useful in discriminating monocytesand macrophages from other cells of themononuclear phagocyte system, such asmyeloid cells and dendritic cells, but that itmay also be of great value in routine diagnos-tics to distinguish histiocytes from tumour cellsand for typing undifferentiated malignancies.We thank Professor H Kerl and Dr M. Santucci for providingparaffin wax sections of histiocytic disorders, Dr B Falini forthe gift of PG-Mi, Miss PA den Otter for expert technicalassistance and Dr P van der Valk for helpful comments.

1 Unanue ER, Allen PM. The basis for the immunoregula-tory role of macrophages and other accessory cells.Science 1987;236:551-7.

2 Van den Berg TK, Breve JJP, Damoiseaux JGMC, et al.Sialoadhesin on macrophages: its identification as a lym-phocyte adhesion molecule. Jf Exp Med 1992;176:647-55.

3 Micklem KJ Cordell JL, Rigney EM, et al. A macrophage-associated antigen defined by five mAb. In: Knapp W,Leucocyte Typing IV. White cell differentiation antigens.Oxford: Oxford University Press, 1989:843-6.

4 Pulford KAF, Sipos A, Cordell JL, et al. Distribution ofthe CD68 macrophage/myeloid associated antigen. IntImmunol 1990;2:973-80.

5 Micklem KJ, Rigney EM, Cordell JL, et al. A humanmacrophage-associated antigen (CD68) detected by sixdifferent monoclonal antibodies. Br J7 Haematol 1989;73:6-11.

6 Saito N, Pulford KAF, Breton-Gorius J, et al.Ultrastructural localization of the CD68 macrophage-associated antigen in human blood neutrophils andmonocytes. Am J Pathol 1991;139:1053-9.

7 Pulford KAF, Rigney EM, Micklem KJ, et al. KP1: a newmonoclonal antibody that detects a monocyte/macrophage associated antigen in routinely processedtissue sections. Jf Clin Pathol 1989;42:414-2 1.

8 Warnke RA, Pulford KAF, Pallesen G, et al. Diagnosisof myelomonocytic and macrophage neoplasms inroutinely processed tissue biopsies with monoclonalantibody KP1. AmJ Pathol 1989;135:1089-95.

9 Facchetti F, De Wolf-Peeters C, Mason DY, et al.Plasmacytoid T cells: immunohistological evidence fortheir monocyte/macrophage origin. Am Jf Pathol 1988;133:15-21.

10 Bolz S, Ruck P, Schaumberg-Lever G, et al.Immunoelektronenmikroskopische Befimde an Zellendes Makrophagensystems mit den MonoklonalenAntikorpern KP1 (CD68) und KiMlP. Verh Dtsch GesPathol 1991;75:381.

11 Doussis IA, Gatter KC, Mason DY. CD68 reactivityof non-macrophage derived tumours in cytologicalspecimens. Jf Clin Pathol 1993;46:334-6.

12 Fallini B, Flenghi L, Pileri SA, et al. PG-Mi: a newmonoclonal antibody directed against a fixative resistentepitope on the macrophage-restricted form of the CD68molecule. Am J Pathol 1993;142: 1359-72.

13 Kohler G, Milstein C. Continuous cultures of fused cellssecreting antibodies of predicted specificity. Nature1975;256:495-7.

14 Figdor CG, Bont WS, Touw I, et al. Isolation of function-ally different human monocytes by counterflow centrifu-gation elutriation. Blood 1982;60:46-53.

15 Jaspars LH, Van der Linden JC, Scheffer GL, et al.Monoclonal antibody 4C7 recognizes an endothelialbasement membrane component that is selectivelyexpressed in capillaries of lymphoid follicles. Jf Pathol1993;170:121-8.

16 Mullink H, Henzen-Logmans SC, Alons-van KordelaarJJM, et al. Simultaneous immunoenzyme staining ofvimentin and cytokeratins with monoclonal antibodies asan aid in the differential diagnosis ofmalignant mesothe-lioma from pulmonary adenocarcinoma. Virchows Arch(Cell Pathol) 1986;52:55-65.

17 Horny HP, Schaumburg-Lever G, Bolz S, et al. Use ofmonoclonal antibody KP1 for identifying normal andneoplastic human mast cells. Jf Clin Pathol 1990;43:719-22.

18 Horny HP, Ruck P, Xiao JC, Kaiserling E.Immunoreactivity of normal and neoplastic human tis-sue mast cells with macrophage-associated antibodieswith special reference to the recently developed mono-clonal antibody PG-Mi. Hum Pathol 1993;24:355-8.

19 Woods GS, Beckstead JH, Turner RR, et al. Malignantfibrous histiocytoma tumor cells resemble fibroblasts.Am J7 Surg Pathol 1986;10:323-35.

20 Roholl PJ, Kleyne J, Van Unnik JAM. Characterization oftumour cells in malignant fibrous histiocytomas andother soft tissue tumours in comparison with malignanthistiocytes. II. Immunoperoxidase study on cryostatsections. Am J Pathol 1985;121:269-74.

252

on June 25, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.47.3.248 on 1 M

arch 1994. Dow

nloaded from

http://jcp.bmj.com/