J Clin Pathol 1994;47:43-47

Expression of aminopeptidase-n (CD 13) innormal tissues and malignant neoplasms ofepithelial and lymphoid origin

J Dixon, L Kaklamanis, H Turley, I D Hickson, R D Leek, A L Harris, K C Gatter

AbstractAims-To provide a detailed knowledgeof the distribution of the CD13 molecule,also known as the protease aminopepti-dase-N, on both normal tissues andmalignant neoplasms of epithelial andlymphoid origin.Methods -CD13 antigen was examinedby immunocytochemistry, using arecently produced antibody (VS5E)alongside a commercially available anti-CD13 monoclonal antibody. The VS5Erecognising CD13 was produced byimmunising a doxorubicin resistantbreast cancer cell line (MCF-7-Adr). Astriking feature of this antibody was thatit stained the doxorubicin resistant cellsbut not the parental cell line. Both anti-bodies were tested on a broad range ofnormal tissues and three common typesof epithelial malignancy (colon n = 28,lung n = 30, breast n = 35), and 12 casesof Hodgkin's and 52 of non-Hodgkin'slymphomas.Results-CD13 was expressed on manytissue and cell types outside thehaematopoietic system. In particular itwas present on breast epithelium and in20% (seven of 35) of breast carcinomas,but absent in normal and neoplasticcolonic and bronchial tissues and lym-phomas.Conclusions-This study provides notonly detailed information about theexpression of the CD13 antigen, but alsoraises the important possibility thatCD13 expression may correlate withdrug resistance in breast carcinomas.

(7 Clin Pathol 1994;47;43-47)

Nuffield DepartmentofPathology, JohnRadcliffe Hospital,University of Oxford,Oxford OX3 9DUJ DixonL KaklamanisRD LeekK C GatterICRF Laboratories,Institute ofMolecularMedicine, JohnRadcliffe HospitalH TurleyI D HicksonA L HarrisCorrespondence to:Dr K C Gatter

Accepted for publication18 August 1993

The myeloid cell surface antigen CD 13, alsoknown as the protease aminopeptidase-N, is a

150 kilodalton surface glycoprotein1 2

encoded by a gene located on the long arm ofchromosome 15 at bands q25-26. This 967amino acid integral membrane protein has a

large extracellular carboxy-terminal domaincontaining a pentapeptide motif, characteris-tic of members of the zinc binding metallo-protease superfamily. The amino terminalintracellular hydrophobic segment can poten-tially act both as a signal for membrane inser-tion and as a stable membrane spanningsegment.2

Interest in the CD13 antigen arose when itwas discovered that myeloid cell linesexpressed it. Although CD 13 is absent from

the surface of normal B and T lymphocytesand most lymphocytic malignancies, it is nowwell established that increased CD13 expres-sion is detected in acute myeloblasticleukaemia (AML) and in some cases of acutelymphoblastic leukaemia (ALL). A high levelof CD13 expression in these two types ofneoplasm is correlated with a poor prognosis,in that there is a reduced success rate ininducing remission as well as a shortenedsurvival time, particularly in patients withA_L.3CD 13 expression is not restricted to

myeloid cells and related neoplasms. Non-haematopoietic cells such as fibroblasts, osteo-clasts, and the epithelium that line the renalproximal tubules, intestine, and bile ductcanaliculi are also CD 13 positive.4 CD 13expression has also been detected in hepato-cellular, gall bladder, and renal carcinoma celllines.56 Mesenchymal tumours of smoothmuscle, fibrous, synovial, osteogenic, andperipheral nerve sheath origin are also CD13positive.7 Reports of CD1 3 expression in a fewcases of neuroblastoma, fallopian tube carci-noma, and in some ovarian and cervical carci-noma cell lines have also been made.8CD 13 is identical with human aminopepti-

dase-N,9 a zinc binding protease that binds tothe membrane through an N-terminal seg-ment. It is expressed in many tissues andmainly on the brush border membrane of thesmall intestine, on proximal tubules, onsynaptic membranes of the central nervoussystem, and on the surface of granulocytesand macrophages.1013

Recently we produced monoclonal anti-bodies using the cell line MCF-7-AdR whichis a doxorubicin resistant breast cancer cellline. One monoclonal antibody, VS5E, wasselected as staining only the doxorubicinresistant cells but not the parental cell line.Characterisation of VS5E showed that itrecognised CD13, raising the possibility thatthis antigen might be involved in some aspectof drug resistance. To investigate this it isnecessary to have a detailed knowledge of thedistribution of CD13 on both normal andneoplastic tissues. The present study wasdesigned to provide this on a range of normaltissues and in five types of neoplasms.

MethodsFresh tissue samples from 13 normal tissues(adrenal, brain, breast, bronchus, colon, kid-ney, liver, nasal mucosa, pancreas, skin,testes, tongue and tonsil) and from five typesof neoplastic tissue (12 cases of Hodgkin's

43

on August 4, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.47.1.43 on 1 January 1994. Dow

nloaded from

Dixon, Kaklamanis, Turley, Hickson, Leek, Harris, Gatter

Table 1 Monoclonal antibodies, VS5E and WM47, bindCD13 epitopes

Transfected cells VSSE WM47

COSS + CD13 construct + pSV2neo + +COS + pSV2 neo - -

205

disease, 52 non-Hodgkin's lymphomas, 35breast, 30 lung and 28 colon carcinomas)

117 were used. Different types and grades oflymphoma, such as chronic lymphocytic

97 leukaemia, and B or T large cell lymphomas,and three types of lung neoplasm: squamouscell carcinoma; adenocarcinoma; and small

67 cell carcinoma were included.

VS5E was produced as detailed below.

IMMUNISATION WITH ANTIGENThe source of antigen was a membrane prepa-

- 45 ration of MCF-7-AdR cells. About 4 x 105cells were injected intraperitoneally into a

Balb/c mouse, together with Freund's com-

plete adjuvant. Eleven days later the mouse

Figure 1 Immunoblot ofa membrane preparationfrom MCF7AdR cells.LaneA shows that VSSErecognises a band ofmolecular weight 150kilodaltons. Lane B is anegative control stainedwith an irrelevant antibody(anti-bromodeoxyuridine).

was given a booster injection of about1-6 x 106 cells, with non-complete adjuvant.The booster was repeated again seven dayslater.

Three days after the last injection cell sus-

pensions were prepared from the spleen ofthe mouse and were fused with cells from themyeloma cell line NS1, using polyethyleneglycol to induce cell fusion, as described byPulford et al.14

BIOCHEMICAL ANALYSIS OF THE VS5E EPITOPECell cultures (MCF-7-AdR) were washedtwice with phosphate buffered saline (PBS)and then harvested with 0-25% trypsin and0-02% EDTA. The cell suspension was

washed in culture medium and centrifugedfor 5 minutes at 1000 rpm at 4°C. The cellswere washed twice in PBS and centrifuged fora final spin at 2000 rpm for 10 minutes at4°C. The cell pellet was solubilised with an

equal volume of 1% (v/v) Triton X-100(Sigma) in 10 mM TRIS-HCI (pH 8) for20 minutes on ice, to yield integral mem-

brane and associated proteins. This was thenspun at 200 rpm for 15 minutes at 4°C, thesupernatant fluid collected, and used eitherimmediately for analysis by (PAGE) polyacry-lamide gel electrophoresis or stored as mem-

brane extract at -70°C.

SDS-PAGE AND IMMUNOBLOTTINGSodium dodecyl sulphide (SDS) PAGE was

performed as follows. Samples were sus-

pended in buffer containing Dithiothreitoland boiled for 2 minutes, allowed to cool,and loaded on to a gel (5 5% stacking gel,6-5% running gel). Molecular weight markerswere run at the same time. Proteins weretransferred electrophoretically using a

Transblot (Biorad) apparatus to nitrocellu-lose membranes and the protein detected byapplying the antibody followed by a rabbit-anti mouse peroxidase conjugate (Dako) andvisualised using diaminobenzidine with metalion enhancement (fig 1).WM47 (anti CD 13) was produced by

Favaloro et al 8 (commercially supplied byDako).To test the reactivity of the two monoclonal

antibodies used in this project to the CD13antigen, COS cells (monkey kidney fibrob-lasts) were either transfected with a cDNAencoding CD13, together with the selectablemarker pSV2 neo, or with pSV2 neo alone,and the cells selected for G4 18 resistance.Transfectants were then exposed to the VS5Eantibody and the goat anti-mouse fluoresceinisothiocyanate (FITC) conjugate. The cellswere separated on fluorescence activated cellsorting (FACS) and the 10% highest fluoresc-ing cells were collected.The immunohistochemical staining was

performed using the alkaline phosphatase-anti-alkaline phosphatase method (APAAP),as has been described before. Cryostat tissuesections (6 [tm thick) were fixed in acetonefor 10 minutes at room temperature, left todry overnight, and either stained immediatelyor stored at -20'C until required. The pri-mary antibodies were added to the dry sec-

tions in a moist chamber for 30 minutes.Anti-mouse immunoglobulins (Dako code Z259 at 1 in 50 dilution with TBS and APAAPcomplex (Dako code D 651 at 1 in 50 dilu-tion in TBS) were added and incubated for30 minutes for each one. This step was

repeated twice (the second and third incuba-tions for 10 minutes only) to enhance theintensity of the final staining. The alkalinephosphatase substrate was applied afterwardsfor 15-20 minutes, and the sections, afterbeing washed with TBS and tap water, werecounterstained with haematoxylin andmounted on a suitable aqueous mountingmedium. All antibody incubations were

:Figure 2 VSSEimmunostaining of (A)COS cells transfected withthe full length cDNA ofCD13 with the selectablemarkerpSV2 neo and (B)with the selectable markerpSV2 neo alone.

A B

44

Jm:

on August 4, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.47.1.43 on 1 January 1994. Dow

nloaded from

CD13 expression in normal and malignant tissues

Table 2 Expression ofCD13 in non-haematopoietictissues

Adrenal + Reticular connective tissue- Endothelium- Epithelium

Brain + Endothelium of cerebral capillariesBreast + Epithelium of some secretory ducts

+ Basement membrane of secretory ducts+ Interlobular connective tissue

Bronchus + Epithelium of secretory ducts ofbronchial glands

+ Basement membrane- Respiratory epithelium

Colon + Macrophages- Mucosal epithelium

Kidney + Epithelium of proximal convolutedtubule ducts

+ Distal convoluted tubulesLiver + Epithelium of bile ducts

+ Bile canaliculiNasal + Epithelium of secretory ductsmucosa + Connective tissue stroma

+ Smooth muscle- Endothelium

Pancreas + Epithelium of exocrine ducts- Islets

Small + Epitheliumintestine + MacrophagesSkin + Basement membrane

+ Macrophages- Endothelium- Epithelium

Testis + Leydig cells- Spermatozoa-Sertoli cells

Tongue + Basement membrane+ Stroma

Tonsil + Basement membrane+ macrophages

followed by two one minute washings inTBS.

ResultsTRANSFECTED CELLSThe reactivity of VS5E and WM47 to COScells either transfected with a full lengthcDNA encoding CD13 (positive control), orwith the pCDM8 vector (negative control), isshown in table 1 and fig 2. These controls

were performed to ensure that the two mono-clonal antibodies used in the project reactedspecifically with the CD13 antigen. As thetable indicates, VS5E and WM47 react onlywith the COS cells containing the CD13 con-struct.

NORMAL TISSUESThe distribution of CD13 in 13 different nor-mal tissues was investigated using VS5E andWM47. The two antibodies displayed identi-cal reactivities in all tissues. The results aresummarised in table 2. CD13 was expresseduniversally by macrophages, fibroblasts,smooth muscle cells and stromal cells. CD13was absent from the keratinised squamousepithelium of the skin and the pseudo-stratified respiratory epithelium of thebronchus, but was found on both the secre-tory epithelium of the glands and ducts of theexocrine pancreas, nasal polyp, hepatic bileducts, bronchial glands, breast, small intes-tine and the reabsorptive epithelium of therenal proximal convoluted tubules. CD13was expressed by exocrine glandular secretoryepithelium, but the extent of the expressionvaried. At least 80% of the exocrine glandsand ducts of the pancreas, mucous glands ofthe nasal mucosa and the respiratory tract,and the hepatic bile ducts expressed CD13.The antigen was completely absent fromcolonic epithelium.CD13 expression did not occur on the

endocrine secretory cells of the islets ofLangerhans in the pancreas or on the Leydigcells in the testes.CD13 was not expressed by any endothe-

lium in systemic capillaries, but was clearlypresent on the luminal side of cerebral capil-laries (fig 3).

4a

Figure 3 Staining ofnormal tissues with anti-CD13 antibody VSSE: (A) breast; (B) kidney; (C) pancreas; (D) brain.

45

%J6

on August 4, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.47.1.43 on 1 January 1994. Dow

nloaded from

Dixon, Kaklamanis, Turley, Hickson, Leek, Harris, Gatter

Table 3 Expression of CD13 in different neoplasms

No of cases positiveTissue with VSSE and WM47

Hodgkin's disease 0/12Non-Hodgldn's lymphomas 0/52Breast 7/35Colon 0/28Lung 0/30

NEOPLASMSTo investigate whether CD 13 was expressedby neoplasms other than the well charac-terised leukaemic cell lines, five differentmalignant neoplasms were studied, three ofwhich were of non-haematopoietic origin.CD 13 was expressed by breast neoplasms inseven of 35 (20%) of the cases but was notexpressed by Hodgkin's disease cells andnon-Hodgkin's lymphomas, colon, or lungcarcinomas. In most of the positive breastcases every tumour cell expressed CD13, butin a few cases CD13 was expressed focally(table 3) (fig 4).

DiscussionCD 13 was expressed in many different nor-mal tissues. It was found on macrophages,stromal cells, smooth muscle cells and fibrob-lasts within basement membranes and con-nective tissue throughout all the tissuesexamined. The wide distribution of thisenzyme suggests that it performs functionsthat may be site dependent.'0 It has been sug-gested that in the pancreas and liver CD13,which is identical with aminopeptidase-N, anectoenzyme with a metalloprotease activity,may be essential for the breakdown of thetripeptide glutathione, which is present inhigh concentrations. CD 13 was expressed bythe reabsorptive epithelium of the proximalconvoluted tubules of the kidney, but not bythe absorptive epithelial lining of the colonicmucosa. In a recent study, however,'5aminopeptidase-N expression was detected inthe normal colonic epithelium and in 30% ofthe colonic carcinomas by immunofluores-cence. A possible explanation could be thedifference in epitope specificity of the anti-bodies used.

It has been reported that CD 13 isexpressed in the small intestine, where it is

thought to have an important role in peptidedigestion. Intestinal brush border aminopep-tidase-N constitutes 8% of the total proteinand is thought to play a part in the final stageof peptide digestion, with a preference forneutral amino acids. In this particular casethe enzyme has a nutritional protein process-ing role.'6 CD13 is the major receptor for dif-ferent types of human coronaviruses whichcause upper respiratory tract and intestinalinfections.'7 '8CD13 was expressed by the endothelium

of the cerebral capillaries, but not by theendothelium of the systemic capillaries.Whether this finding reflects morphologicaland functional differences between capillariesof the systemic and cerebral circulations isnot known. What is known, however, is thatCD 13 collaborates with CD 10 (neutralendopeptidase) to inactivate bioactive pep-tides on synaptic membranes of the centralnervous system, including enkephalins andendorphins that are involved in signal trans-duction at the cell membrane.'9 20

For patients with ALL, survival studiesshow that expression of CD13 is positivelycorrelated with a reduced success rate ininducing remission and a shortened survival.The production of a new monoclonal CD13antibody (VS5E) aroused interest in CD 13 asa potential marker of drug resistance. VS5Ewas produced using a membrane preparationof MCF-7 AdR, a breast cancer deriveddoxorubicin resistant cell line. This newmonoclonal antibody stained MCF-7 AdRbut not the parental cell line (unpublishedobservations). The most interesting finding ofthis study is that while most breast carcino-mas do not express CD13, a few (20%) arepositive, presumably reflecting the positivityof normal breast epithelium. It would beinteresting to determine whether this subsetof CD13 positive breast carcinomas exhibitan altered response in vivo to chemothera-peutic agents.We are now undertaking a large study of

breast cancer cases to confirm these findingsand to attempt to elucidate whether there is acorrelation between CD13 expression and aresponse to treatment and ultimately toprognosis.This work was supported by the Imperial Cancer ResearchFund.

Figure 4 Staining ofbreast cancer with anti-CD13 antibody VSSEshowing (A) positive caseand (B) negative case.

46

1- A-i-.:.i

...o "f -,.-.w .1. '- 4

-,M&lw 4 C.'

91 1.

on August 4, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.47.1.43 on 1 January 1994. Dow

nloaded from

CD13 expression in normal and malignant tissues

1 Look AT, Peiper SC, Rebentisch MB, et al. Transfer andexpression of the gene encoding a human myeloid mem-brane antigen (gp 150). J7 Clin Invest 1985;75:569-79.

2 Sakai K, Hattori T, Sagawa K, Yokoyama M, TakatsukiK. Biochemical and functional characterization ofMCS-2 antigen (CD 13) on myeloid leukaemic cells andpolymorphonuclear leukocytes. Cancer Res 1987;47:5572-6.

3 Sobol RE, Mick R, Royston I, et al. Clinical importance ofmyeloid antigen expression in adult acute lymphoblasticleukaemia. N Engl Jf Med 1987;316:1111-17.

4 Hogg N, Horton MA. Myeloid antigens: new and previ-ously defined clusters. In: McMichael AJ, ed. Leukocytetyping III. Proceedings of the Third International Workshopon Human Leukocyte Differentiation Antigens. Oxford:Oxford University Press, 1987: 576-621.

5 Look AT, Ashmun RA, Shapiro LH, et al. Report on theCD1 3 (aminopeptidase N) cluster workshop. In: KnappW, Dorken B, Gilks WR, eds. Leukocyte typing IV. Whitecell differentiation antigens. Oxford: Oxford UniversityPress, 1989:784-7.

6 Finstad CL, Cordon Cardo C, et al. Specificity analysis ofmouse monoclonal antibodies defining cell surface anti-gens of human renal cancer. Proc Nail Acad Sci USA1985;82:2955-9.

7 Mechtersheimer G, Moller P. Expression of aminopepti-dase- N (CD13) In mesenchymal tumours. Am Jf Pathol1990;137:1215-22.

8 Favaloro EJ, Bradstock CF, Kabral A, Grimsley P, ZowtyjH, Zola H. Further characterization of human myeloidantigens (gp 160,95; gpl50; gp67): Investigation of epi-topic heterogeneity and non haemopoietic distributionusing panels of monoclonal antibodies belonging toCD11b, CD13 and CD33. Br J7 Haematol 1988;69:163-71.

9 Look AT, Ashmun RA, Shapiro LH, Peiper SC. Humanmyeloid plasma membrane glycoprotein CD13 (gp 150)is identical to aminopeptidase N. J Clin Invest 1989;68:932-41.

10 Kenny AJ, Maroux M. Topology of microvillar membrane

hydrolases of kidney and intestine. Physiol Rev 1982;62:91-128.

11 Matsas R, Stephenson SL, Hryszko J, Kenny AJ, TurnerAJ. The metabolism of neuropeptides. Phase separationof synaptic membrane preparations with Triton X-1 14reveals the presence of aminopeptidase N. Biochem J1985;231:445-9.

12 Noren 0, Sjostrom H, Danielsen EM, Cowell GM,Skovbjerg H. The enzymes of the enterocyte plasmamembrane. In: Desnuelle P, ed. Molecular and cellularbasis of digestion. Amsterdam: Elsevier/North Holland,1986:135-67.

13 Semenza G. Anchoring and biosynthesis of stalked brushborder membrane proteins: Glycosidases and peptidasesof enterocytes and renal tubuli. Ann Rev Cell Biol1986;2:255-313.

14 Pulford K, Micklem K, Thomas J, Jones M, Mason DY.A 72-kD B cell-associated surface glycoproteinexpressed at high levels in hairy cell leukaemia andplasma cell neoplasms. Clin Exp Immunol 1991 ;85:429-35.

15 Quaroni A, Nichols BL, Quaroni E, et al. Expression anddifferent polarity of aminopeptidase N in normal humancolonic mucosa and colonic tumours. Int J Cancer1992;51:404-1 1.

16 McDonald JK, Barrett AJ. Mammalian proteases: glossaryand bibliography. In: Exopeptidases. Vol 2. New York:Academic Press, 1986:98-117.

17 Yeager CL, Ashmun RA, Williams RK, et al. Humanaminopeptidase N is a receptor for human coronavirus229E. Nature 1992;357:420-2.

18 Delmas B, Gelfi J, L'Haridon R, et al. Aminopeptidase Nis a major receptor for the enteropathogenic coronavirusTGEV. Nature 1992;357:417-20.

19 Ashmun RA, Look AT. Metalloprotease activity ofCD 13/Aminopeptidase N on the surface of humanmyeloid cells. Blood 1989;75:462-9.

20 Turner AJ, Hooper NM, Kenny AJ. Metabolism of neu-ropeptides. In. Kenny AJ, Turner AJ, eds. Mammalianectoenzymes. New York: Elsevier Science, 1987:211-42.

47

on August 4, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.47.1.43 on 1 January 1994. Dow

nloaded from