6 JClin Pathol 1992;45:660-663

Papers

Expression of Ki-67 nuclear antigen in B and Tcell lymphoproliferative disorders

N de Melo, E Matutes, I Cordone, R Morilla, D Catovksy

AbstractAims: To determine whether the prolifera-tion rates of tumour cells may relate toprognosis and reflect disease activity.Methods: Blood mononuclear cells from155 patients with B cell (n = 120) orT cell(n = 35) chronic lymphoproliferative dis-orders were tested with the monoclonalantibody Ki-67 by indirect immunoperox-idase or immunoalkaline phosphatasetechniques. B cell diseases includedchronic lymphocytic leukaemia (CLL),CLL in prolymphocytic transformation(CLLIPL), prolymphocytic leukaemia(B-PLL) and non-Hodgkin's lymphoma(B-NHL) in leukaemic phase. The T celldiseases comprised large granular lym-phocyte (LGL) leukaemia, T-PLL, andT-NHL.Results: These showed significantly higherproportions of Ki-67 positive cells inT cell(11.2%) than in B cell (2-90/o) disorders(p < 0.001). The highest values werefound in NHL of both B and T cell types,particularly when low grade diseasetransformed to high grade. The lowestpercentages of Ki-67 positive cells werefound in CLL (1.4%) and LGL leukaemia(1.7%); intermediate values were seen in BPLL (3 3%) andT PLL (5.8%).Conclusions: There is a positive correla-tion between prognosis and proliferationrates in chronic B and T cell lym-phoproliferative disorders. Estimation ofKi-67 in circulating leukaemic cells couldbe used to determine prognosis in lowgrade malignancies.

Academic Departmentof Haematology andCytogenetics, TheRoyal MarsdenHospital and Instituteof Cancer Research,Fulham Road, LondonSW3 6JJN de MeloE MatutesI CordoneR MorillaD CatovskyCorrespondence to:Professor D CatovskyAccepted for publication31 January 1992

Chronic lymphoproliferative disorders com-

prise a heterogeneous group of diseasesresulting from the clonal proliferation ofmature B andT lymphoid cells. Although theymay present problems of differential diagnosis,they have distinct clinical, morphological, andevolutionary features. Previous studies on cellproliferation in these diseases have been doneusing several methods,' but it is not yetknown whether kinetic studies on peripheralblood cells reflect the growth fraction proper-ties of the tumour. Studies of cell proliferationrates with the monoclonal antibody Ki-67 havemainly been applied to lymph nodes of non-

Hodgkin's lymphoma (NHL). However, fewstudies have been performed on peripheralblood6 '- and diseases such as prolymphocytic

leukaemia and some NHL in leukaemic phasehave not been investigated.The purpose of this study was to investigate

the proliferation rate of a group of B and Tmature lymphoproliferative disorders using themonoclonal antibody Ki-67 on the circulatingcells by means of immunocytochemistry. Wealso compared the results obtained withimmunoperoxidase and alkaline phosphataseanti-alkaline phosphatase (APAAP) and inves-tigated if reactivity with Ki-67 is a feature ofsome of these disorders.

MethodsA series of 155 patients with mature B and Tcell malignancies were studied. The diagnosiswas based on clinical features, cell morphol-ogy, immunophenotype, and in some cases byhistological and ultrastructural analysis.

Peripheral blood mononuclear cell suspen-sions were obtained by density gradient cellseparation using Lymphoprep (Nycomed,Oslo, Norway) and cytospin sides were pre-pared at a concentration of 2 x 106 cells/ml(Cytospin II, Shandon Products Ltd, UK).Immunophenotyping was carried out on cellsuspension by indirect immunofluorescenceusing a panel of monoclonal antibodies againstB andT cell antigens and fluorescein conjugat-ed (FITC) goat anti-mouse immunoglobulinsF(ab)2 fragment (Cappel, West Chester, Eng-land) as a second layer and analysed on aFACScan flow cytometer (Becton-Dickinson,Mountain View, California).

Determination of growth fraction wasperformed on cytospins using the monoclonalantibody Ki-67 (Dako, High Wycombe,England) using immunoperoxidase andalkaline phosphatase anti-alkaline phos-phatase (APAAP) techniques. Briefly, cyto-spins wrapped in foil paper and stored at- 20°C were thawed and fixed in pure acetonefor 10 minutes. The cells were incubated withthe monoclonal antibody Ki-67, followed bysequential 30 minute incubations with perox-idase labelled rabbit anti-mouse immunoglo-bulins and swine anti-rabbit immunoglobulins(Dako, High Wycombe). The developing solu-tion contained 30 mg of 3,3'-diaminobenzi-dine (Sigma, Poole) and hydrogen peroxidasein phosphate buffered saline (PBS). The slideswere counterstained with haematoxylin-S(BDH) and mounted witb DPX (BDH). Forthe APAAP method, rabbit anti-mouse immu-noglobulins and APAAP complexes (Dako,

66()

on March 19, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.45.8.660 on 1 August 1992. D

ownloaded from

Expression of Ki-67 nuclear antigen in B and T cell lyniphoproliferative disorders

Table 1 Expression of Ki-67 in B cell diseases

% of positive cellsNo of _ Mean (SEM) of Mean (SEM) of

Disease cases 0 0 2-2-0 > 2-0 all cases positive cases

CLL 70 22 29 19 1-4 (02) 2-0 (02)CLIIPL 11 5 2 5 1-8 (0 6) 3-3 (0-7)B-PLL 9 5 1 3 3-3 (2-0) 7-4 (3-4)B-NHL 29 7 11 11 6-3 (2-2) 8-3 (2 8)IBT* 1 0 0 0 17 8%

*Immunoblastic transformation: patient with Richter's syndrome and peripheral blood diseasewith large cells (fig 1).

HighWycombe) were used as second and thirdlayers. The reaction was developed in a medi-um containing naphthol As-MX phosphate(Sigma), levamisole, dimethylformamide, andFast-Red Salt TR (Sigma, Poole). Subse-quently, the slides were counterstained withhaematoxylin-S (BDH) and mounted withglycergel (Dako). The percentage of Ki-67positive cells was estimated by light microsco-py (oil immersion) by counting 500 cells ineach sample. Only those cells with nuclearstaining were considered positive.The statistical analysis was performed on a

Macintosh microcomputer with a Stat View512 + software program.

ResultsImmunophenotyping disclosed 120 patientswith B lymphoid disease: 70 with chroniclymphocytic leukaemia (CLL), one with CLLin immunoblastic transformation, 11 withCLL with increased numbers of prolympho-cytes (CLL/PL), nine with prolymphocyticleukaemia (B-PLL) and 29 with non-Hodg-kin's lymphoma (B-NHL) in leukaemic phase.The remaining 35 were T cell disorders: 10large granular lymphocyte LGL leukaemia,

Figure I Nuclear expression of Ki-67 in the large cells of a patient with B cell CLL intransformation (immunoperoxidase).

two T cell leukaemia in immunoblastictransformation, 13 T-PLL and 10 T-NHL inleukaemic phase.The results of Ki-67 expression were com-

pared by both methods in circulating cells from64 patients with B-CLL. The number of caseswith Ki-67 positive cells and the mean ofpercentage of the Ki-67 positive cells wereslightly higher with immunoperoxidase thanwith APAAP, but this difference was notsignificant. The immunoperoxidase methodshowed, overall, a better morphological defin-ition of the positive cells than APAAP. Becauseof this, we subsequently evaluated the reac-tivity of Ki-67 in all the haematological malig-nancies using immunoperoxidase.

Results in the patients with B cell diseasesare shown in table 1. The highest proliferationrates were observed in a patient with a previousdiagnosis of CLL who developed Richter'ssyndrome (fig 1) and in cases of B-NHL; CLLcases had low mean values. Intermediatevalues were observed in CLL/PL and B-PLL.The B-NHL group comprised five patientswith intermediate (mantle zone) lymphoma,six patients with high grade (large cell) NHL,and 18 with low grade malignancy, all inleukaemic phase. The highest percentages ofKi-67 positive cells were observed in patientswith high grade malignancy (range 0 8-57%).These results agree with those of similarstudies carried out on tissue sections. ". 13Thehighest value for Ki-67 in NHL group was seenin a lymphoplasmacytic lymphoma whichtransformed to large cell lymphoma (fig 2).AmongT cell diseases the highest expression

of Ki-67 was observed in two patients withmature T cell leukaemia in immunoblastictransformation and in T-NHL. The lattergroup comprised one patient with peripheralTcell lymphoma, five patients with S&ary syn-drome and four with adult T cell leukaemia/lymphoma. These patients had high Ki67values: 15-4%, 540%, and 6-5-60%, respec-tively. The lowest values were observed inpatients with LGL leukaemia. Patients withT-PLL had intermediate values between thesetwo groups (table 2).

Overall, the results in the cases of T celldisease were higher than those in B celldiseases; this difference was significant (table3). It should be noted that within each group ofroughly equivalent diseases, such as CLL andLGL leukaemia, B and T PLL, and B and TNHL, those with a T cell phenotype alwaysexpressed higher Ki-67 values (fig 3).

DiscussionThere was a highly significant difference in theproportion of circulating Ki-67 positive cellsbetween the two main groups of B and Tmature lymphoproliferative disorders. Ourfindings also show that there is a range ofproliferation rates detected on circulating cellsin each group of disorders which may correlatewith prognosis.

Several approaches have been used to esti-mate proliferation rates in different humantumours. Usually they are estimated from the

661

on March 19, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.45.8.660 on 1 August 1992. D

ownloaded from

de Melo, Matutes, Cordone, Morilla, Catovsky

Figure 2 Peripheral blood cells of a patient with a low grade B ceUl NHL in leukaemicphase, now in transformation, to a large ceU lymphoma showing expression of Ki-67(immunoperoxidase).

number of tumour cells in mitosis and enteringDNA synthesis by measuring the uptake ofradioactively labelled DNA precursors.2 14More recently, the monoclonal antibody Ki-67has been used to assess the growth fraction.This reagent detects a nuclear antigen presenton chromsomes in all phases of mitosis andwithin the nucleolus of proliferating cells in allphases of the cell cycle but not in restingcells.' 6B-CLL is a disease with a variable pattern of

survival with the best prognostic factors beingdisease staging,'7 age, white cell count andsex.'8 In general this disease showed the lowestproliferation rates. These results are similar tothose of other cell kinetic studies performed onperipheral blood cells.6 CLL/PL has featuresintermediate between those of CLL andB-PLL,'9 and is associated with disease pro-gression and short survival.20 This group hadslightly higher proliferation rates than CLLand this may correlate with clinical behaviour.In the groups of mature B cell leukaemias

Table 2 Expression of Ki-67 in T cell diseases

% ofpositive cellsNo of Mean (SEM) of Mean (SEM) of

Disease cases 0 0-2-2-0 > 2-0 all cases positive cases

LGL 10 4 2 4 1-7 (0-6) 2-9 (0 6)T-PLL 13 2 0 11 5-8 (1-4) 6-9 (1-5)T-NHL 10 0 0 10 21-2 (5-3) 21-2 (5-3)IBT* 2 0 0 2 44-3 (6-7) 44-3 (6 7)

*Immunoblastic transformation of low grade T cell leukaemia.

Table 3 Comparison of Ki-67 expression in B and T cell disorders

No of Mean (SEM) of Mean (SEM) ofGroup No of cases positive cases all cases positive cases

B 120 81 2-9 (0-6) 4-3 (0 8)T 35 29 11-2 (2 5) 13-6 (2-8)

p < 0-001.

B-PLL has the poorest prognosis and is usuallyunresponsive to treatments which are effectivein CLL.2' In B-PLL we found higher Ki-67values than in CLL and CLL/PL.B-NHL with leukaemic signs are associated

with a variable clinical course and these hadhigher Ki-67 values than CLL, CLL/PL, andB-PLL. Previous studies on histological sec-tions have shown a strong correlation betweenthe proportion of Ki-67 positive cells andhistological grade in NHL." '3 Other studieshave shown, however, that there is often a widerange of proliferation rates within each histo-logical subtype, including high rates in lowgrade lymphomas. 13 22 In series of low gradeNHL Hall et al found a relatively high expres-sion of Ki-67 in patients with a shorter survivalthan those with low Ki-67 expression.'3 Weobserved the highest expression of Ki-67 inpatients with B-NHL with low grade disease intransformation to high grade NHL. Our studysuggests that the determination of Ki-67 inperipheral blood can be useful for detectingand monitoring patients with low grade diseasewho have a progressive leukaemic phase andwho may be undergoing transformation to highgrade disease.Within the various types of mature T cell

leukaemias LGL leukaemia often has a benignclinical course, whereas other types have arelatively aggressive course.23 Thus the low Ki-67 expression observed in LGL leukaemiaseems to correlate with a good prognosis,unlike T-PLL, a distinct disease entity with amedian survival of six months.24 The expres-sion of Ki-67 in this group has not been

25r-

O B cases

0 T cases T20

C)

co

0

151-

101-

5

0LCLUILGL PLLL NHL

IJ

Figure 3 Comparison between B and Tlymphoproliferative disorders. Mean and SEM values ofKi-67 positive cells.

662

.....................................

on March 19, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.45.8.660 on 1 August 1992. D

ownloaded from

Expression of Ki-67 nuclear antigen in B and T cell lymphoproliferative disorders

reported previously and our study suggests thatthe high Ki-67 values correlate with the poorprognosis of T-PLL. The high proportion ofKi-67 positive cells found in all but one of thepatients with adultT cell leukaemia/lymphomaconfirms the findings of other studies on

peripheral blood" 0 and also correlates withthe poor outlook of this disease which has a

median survival of five to seven months.Although some of the patients with Sezarysyndrome also had high Ki-67 values, becausethis disease has a variable course furtherstudies are needed to correlate this finding withprognosis.

Overall, more Ki-67 positive cells were

found in T cell than B cell disorders, againcorrelating with the worse prognosis of patientswith T cell leukaemia or T-NHL in leukaemicphase, except LGL leukaemia. The determina-tion of Ki-67 expression in circulating cellsseems to be an informative and simple tool forassessing disease activity. And it might predictthe clinical course in low grade disorders suchas CLL and may allow sequential studies inperipheral blood samples. Our results are

similar to findings in acute lymphoblasticleukaemia which also showed a correlationbetween DNA labelling index, prognosis, andcell lineage,25 26 with T-ALL having a worse

prognosis than common ALL.

This work was supported in part by Fundacao Pro-sangueHemocentro de Sao Paulo, Brasil (NM) and The CancerResearch Campaign.

We are grateful to Dr Stephen Jacobs for his assistance in thestatistical analysis.

1 Cooper EH, Peckham MJ, Millard RE, Hamlin IME,Gerard-Marchant R. Cell proliferation in human malig-nant lymphomas. Analysis of labelling index. EurJ7 Cancer1968;4:287-96.

2 Theml H, Trepel F, Schick P, Kaboth W, Begemann H.Kinetics of lymphocytes in chronic lymphocytic leukemia:studies using continuous H-thymidine infusion in twopatients. Blood 1973;42:623-36.

3 Costa A, Bonadonna G, Villa E, Valagussa P, Silvestrini R.Labeling index as a prognostic marker in non-Hodgkin'slymphomas. JNCI 1981;66:1-5.

4 Habeshaw JA, Lister TA, Stansfeld AG. Correlation oftransferrin receptor expression with histological class andoutcome in non-Hodgkin's lymphoma. Lancet1983;i:498-500.

5 Akerman M, Brandt L, Johnson A, Olsson H. Mitoticactivity in non-Hodgkin's lymphoma. Relation to the Kielclassification and to prognosis. Br Cancer1987;55:219-23.

6 Scott CS, RamsdenW, Limbert HJ, Master PS, Roberts BE.Membrane transferrin receptor (TfR) and nuclear pro-liferation-associated Ki-67 expression in hemopoieticmalignancies. Leukemia 1988;2:438-42.

7 Falini B, Canino S, Sacchi S, et al. Immunocytochemicalevaluation of the percentage of proliferating cells inpathological bone marrow and peripheral blood sampleswith the Ki-67 and anti-bromo-deoxyuridine monoclonalantibodies. BrJ Haematol 1988;69:311-20.

8 Shirono K, Hattori T, Takatsuki K. Prognostic usefulness ofKi-67 antigen expression in adultT-cell leukaemia. Lancet1989;ii: 1044.

9 Drach J, Gattringer C, Glassl H, Schwarting R, Stein H,Huber H. Simultaneous flow cytometric analysis ofsurface markers and nuclear Ki-67 antigen in leukemiaand lymphoma. Cytometry 1989;10:743-9.

10 Yamada Y, Murata K, Kamihira S, et al. Prognosticsignificance of the proportion of Ki-67 positive cells inadultT-cell leukemia. Cancer 1991;67:2605-09.

11 Grogan MT, Lippman SM, Spier CM, et al. Independentprognostic significance of a nuclear proliferation antigenin diffuse large cell lymphomas as determined by themonoclonal antibody Ki-67. Blood 1988;71: 1157-60.

12 Gerdes J, Dallenbach F, Lennert K. Growth fractions inmalignant non-Hodgkin's lymphomas (NHL) as deter-mined in situ with the monoclonal antibody Ki-67.Haematol Oncol 1984a;2:365-71.

13 Hall PA, Richards MA, Gregory WM, d'Ardenne AJ, ListerTA, Stansfeld AG. The prognostic value of Ki-67immunostaining in non-Hodgkin's lymphoma. J Pathol1985;154:223-35.

14 Kvaloy S, Godal T, Marton PF, Steen H, Brennhovd IO,Abrahamsen AF. Spontaneous (3H)-Thymidine uptake inhistological subgroups of human B-cell lymphomas.ScandJ Haematol 1981;26:221 -34.

15 Gerdes J, Schwab U, Lemke H, Stein H. Production of amouse monoclonal antibody reactive with a humannuclear antigen associated with cell proliferation. Int JCancer 1983;31:13-20.

16 Gerdes J, Lemke H, Baisch H, Wacker HH, Schwab U,Stein H. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the mono-clonal antibody Ki-67. J Immunol 1984b;133:1710-15.

17 Binet JL, Auquier A, Dighiero G, et al. A new prognosticclassification of lymphocytic leukemia derived from amultivariate survival analysis. Cancer 1981 ;48: 198-206.

18 Catovsky D, Fooks J, Richards S. Prognostic factors inchronic lymphocytic leukaemia: the importance of age,sex and response to treatment in survival. Br J Haematol1989;72:141 -9.

19 Melo JV, Catovsky D, Galton DAG. The relationshipbetween chronic lymphocytic leukaemia and prolympho-cytic leukaemia. I. Clinical and laboratory features of 300patients and characterisation of an intermediate group. BrJ Haematol 1986;63:377-87.

20 Melo JV, Catovsky D, Gregory WM, Galton DAG. Therelationship between chronic lymphocytic leukaemia andprolymphocytic leukaemia. IV. Analysis of survival andprognostic features. Br J Haematol 1987;65:23-9.

21 Catovsky D. Prolymphocytic leukaemia. Nouv Rev FrHematol 1982;24:343-7.

22 Kvaloy S, Marton PF, Kaalhus JH, Foss-Abrahamsen A,Godal T. 3H-thymidine uptake in B-cell lymphomas-Relationship to treatment response and survival. Scand JHaematol 1985;34:429-35.

23 Catovsky D, Matutes E. Leukemias of mature T cells.Prolymphocytic and large granular lymphocytic leukemia.In: Knowles D, ed. Neoplastic hematopathology Baltimore:Williams & Wilkins (in press).

24 Matutes E, Brito-Babapulle V, Swansbury J, et al. Clinicaland laboratory features of 78 cases of T-prolymphocyticleukemia. Blood 1991;78:3269-74.

25 Suarez C, Miller DR, Steinherz PG, Melamed MM,Andreeff M. DNA and RNA determination in 111 casesof childhood acute lymphoblastic leukaemia (ALL) byflow cytometry: correlation of FAB classification withDNA stemline and proliferation. Br J Haematol1985;60:677-86.

26 Look AT, Melvin SL, Williams DL, et al. Aneuploidy andpercentage of S-phase cells determined by flow cytometrycorrelate with cell phenotype in childhood acute leuke-mia, Blood 1982;60:959-67.

663

on March 19, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.45.8.660 on 1 August 1992. D

ownloaded from