CLINICAL AND IMMUNOPATHOLOGY 46, 177- 185 (1988)

Clinical and Immunological Findings in Large B-Cell Chronic Lymphocytic

ALBERTO~RFAO,* MARCOSGONZALEZ,*JESUSFERNANDO SANMIGUEL,* MARIA CONSUELOCA~ZO,* PURIFICACIONGALINDo,t MARIADOLORESCABALLERO,SRAMIROJIMENEZ,§ AND ANTONIO LOPEZBORRASCA*

*Servicio de Hematologia, Hospital Vniversitario, Salamanca; fVnidad docente de Bioestadistica, Facultad de Biologia, Vniversidad de Salamanca; SServicio de Hematologia, Hospital Virgen Blanca, Leon; $Servicio de Hematologia, Hospital Virgen de la Vega, Salamanca, Spain

In order to define the characteristics of B-CLL cases in which the predominant cell population is composed of large , we studied 97 patients with B-CLL, com- paring the cell morphological features with the clinical and biological findings and the immunological phenotype of the proliferating cells. Multivariant analysis showed that there were three significantly different morphological groups: Typical CLL, large lym- phocyte CLL (LLL), and CLL with prolymphocytes (CLL/PL). The LLL group showed a greater incidence of lymphadenopathies (P < 0.05) and higher percentages of both p,+6+ cells (P < 0.01) and Fmc/7 + cells (P < 0.001) than in typical CLL. The main differences between LLL and CLL/PL were the peripheral count and the percentage of Fmc/7+ cells (P < 0.002)-both higher in the CLL/PL group-and the percentage of mouse rosette-forming cells (P < O.Ol)-lower in CLU PL. Further studies including functional assays and survival analyses could contribute to elucidating whether these groups are different entities or a single disease with marked heterogeneity. 0 1988 Academic Press. Inc.

INTRODUCTION For many years the term chronic lymphocytic leukemia (CLL) has included several different lymphoproliferative disorders such as B-CLL, prolymphocytic- leukemia (PLL), and (HCL) (1, 2). Currently, these diseases are well differentiated from B-CLL. However, within this latter group of patients there continues to be considerable clinical and biological heterogeneity (3, 4). Accordingly, although in most cases the morphological appearance of the cells is that of typical small lymphocytes with clumped chromatin and scanty cytoplasm, it is not uncommon to find cases in which the predominant cell population is represented by large lymphocytes with abundant cytoplasm and mature clumped chromatin without a nucleolus (5, 6). These findings pose the question of consid- ering whether one is dealing with a single disease with a certain heterogeneity or whether there might be different disorders. In this sense, recently Melo er al. (7) suggested the possibility that some cases of B-CLL in prolymphocytic transfor- mation might correspond to a clinical, morphological, and immunological entity different from both B-CLL and B-PLL, whereas other cases would correspond to the intermediate form described by Enno et a/. (8). Immunological markers consistitute an additional tool for elucidating these

177 0090-1229/88 $1.50 Copyright 0 1988 by Academic Press. Inc. All nghts of reproduction in any form reserved. 178 ORFAO ET AL. aspects since they permit the phenotypic distinction between cells according to their stage of antigenic maturation. Ln order to define the characteristics of B-CLL cases in which the predominant cell population is composed of large lymphocytes, we studied peripheral blood (PB) cells of 97 patients with B-CLL, comparing the morphological features with the clinical and biological findings and the immunological phenotype of the prolif- erating cells. Four patients with B-PLL were also included in the analysis.

MATERIAL AND METHODS One hundred and one patients were studied between June 1984 and January 1986. Ninety-seven were B-CLL and four were B-PLL. Cases in which lymph node biopsy findings revealed non-Hodgkin’s lymphoma were excluded from the study. The diagnostic criteria for B-CLL were those proposed by RAI et al. (9). The B-CLL cases were classitied according to Rai’s staging and Binet’s staging (9. 10). The immunological markers and the morphological assessment were per- formed in all cases at the moment of diagnosis, prior to chemotherapy. The clinical and hematological data analyzed included age, sex, presence of lymphadenopathy in two or more anatomical regions, size of spleen, presence of anemia (hemoglobin < 10 g/dl) and/or thrombopenia (< 100 x IO9 platelets/liter), PB lymphocyte count, percentage of bone marrow (BM) lymphocyte infiltration, and the pattern of BM involvement (11). Morphology. Cell morphology was studied by light microscopy in PB films stained with May-Grunwald-Giemsa. According to the criteria of Melo et ctl. (7), with slight modifications, seven different types of lymphoid cells were estab- lished (Table 1). A total of 200 lymphoid cells were counted in each slide by one of us after a random sample of 40 cases had been independently assessed by two of the authors to test for the reproducibility of the criteria employed: simulta- neously, the diameter of the lymphocytes was measured with an occular microm- eter, counting 100 cells per patient. Immunological markers. Mononuclear cells from PB were obtained by Ficoll- Hypaque (Pharmacia Fine Chemicals) density gradient centrifugation and tested for (a) spontaneous rosette-forming cells with AET-treated sheep erythrocytes (SRFC) (12); (b) spontaneous rosette-forming cells with mouse erythrocytes (MRFC) (13); (c) surface immunoglobulins (sIg) were detected by direct immuno- fluorescence with rabbit F(ab’), antisera conjugated with fluorescein isothio- cyanate (FITC) polyvalent to human Igs and specific to human heavy and light chains (Behring and Nordic Lab); and (d) surface antigens assessed by indirect immunofluorescence with live monoclonal antibodies (McAb), GRBl (anti-class II MHC antigens) (14) Bl (anti-p35 antigen or CD20) (lS>, Fmc-7 (which recog- nizes a subpopulation of normal PB B lymphocytes and the majority of cells in B-PLL and HCL) (16), Fmc-8/Fmc-56 (anti-p24 antigen or CD9) (17), and Leu-1 (anti-p67 antigen or CD5) (18). According to the intensity, four degrees of fluores- cence were separated: undetectable, weak, moderate, and strong. The percentage of MRFC, sIg, Fmc-7, and Leul) positive cells was calculated from the total of B lymphocytes after removing the percentage of SRFC-T cells. Each case was con- TABLE 1 5: CRITERIA USEDFORTHEMORPHOLOGICALDEFINITIONOFLYMPHOID CELLS ii Cell Chromatin Amount of Other ? features Cell size pattern Nucleolus cytoplasm features 8 Small lymphocyte <2 erythrocytes Clumped Absent Scanty (high N/C) F: Large lymphocyte >2 erythrocytes Clumped Absent or very small Relatively abundant G and poorly defined (intermediate N/C) 5 Prolymphocyte Usually >2 Clumped Large, vesicular, Usually abundant erythrocytes prominent (intermediate N/C) ij Cleft lymphocyte >2 erythrocytes Clumped Absent Scanty (high N/C) Cleft nucleus 2 Lymphoplasmocyte >2 erythrocytes Clumped Absent or very small Abundant (low N/C) Slightly excentric $ nucleus 24 Basophilic 8 cytoplasm and halo perinuclear z3 Immunoblast >3 erythrocytes Finely dispersed Large, prominent, Very abundant (low Strongly basophilic usually >I N/C) cytoplasm L Granular >2 erythrocytes Clumped Absent Usually abundant Large azurophilic 2 lymphocyte (low N/C) cytoplasm granules

z W 180 ORFAO ET AL. sidered positive for a marker when more than 10% of the cells displayed positivity for that marker. Statistical methods. To estimate the significance between means, the Mann- Whitney U or Student c test was used. A contrast for dichotomous variables was employed to estimate the significance between those parameters which were evaluated as presence/absence or positive/negative. To analyze whether the groups established were indeed different and, if so, which were the variables that most contribute to such a differentiation, the simultaneous HJ-Biplot technique (19, 20) was employed. The variables concerning clinical features were consid- ered as discontinuous variables, while the markers were considered both as con- tinuous and discontinuous, since they were introduced in two different analyses both as percentages of positive cells and as positive/negative cases. Low-range approximation was done according to the techniques proposed by Golub and Reinsch (21).

RESULTS According to the morphology of lymphoid cells the 97 B-CLL patients were distributed as follows (Table 2): (a) typical CLL, 41 cases; (b) mixed CLL with small/large lymphocytes (CLL/LL), 28; (c) large lymphocyte CLL (LLL), 11; (d) CLL with prolymphocytes (CLL/PL), 7; and (e) CLL with cleft lymphocytes (Cleft LL), 10. No case of mixed large lymphocytes/prolymphocytes was de- tected. Mean lymphocytic diameters were significantly larger (P < 0.001) in the cases of LLL (11.3 -+ 0.8 km) than in those of typical CLL (7.7 -+ 1.2 pm). None of the cases had to be changed with respect to initial classification, in which erythrocyte size was employed as the comparative reference, after the measure- ment of lymphocytic diameter. The clinical and biological characteristics of the different groups are shown in Table 3. The LLL patients showed a significantly lower incidence of anemia and/ or thrombopenia (P < 0.01) with respect to the typical CLL. The finding of en- larged lymph nodes in two or more anatomical regions was more common in LLL (90%) than in the typical CLL cases (63%) (P < 0.05). All except one LLL patient were in stage I or II while 41% of typical CLL cases were in advanced stages, III/IV. No significant differences were observed between typical CLL and LLL with respect to sex, age, spleen involvement, and PB lymphocyte count (Table 3 and Fig. 2).

TABLE 2 B-CLL: GROUPSOFPATIENTSACCORDINGTOCYTOMORPHOLOGICALTYPES Groups Morphological criteria No. ofcases Typical CLL X35% small lymphocytes 41 CLLILL > 10% and =GO% large lymphocytes 28 LLL >SO% large lymphocytes 11 CLLIPL > 10% and ~55% prolymphocytes 7 Cleft LL >20% cleft lymphocytesO 10 a In no case was the percentage of cleft lymphocytes 30%. LARGE B-CELL CHRONIC LYMPHOCYTIC LEUKEMIA 181 TABLE 3 CLINICAL/HEMATOLCXXAL FEATURESOF THE MORPHOLOGICALGROUPS Feature TypicalCLL CLLlLL LLL CLL/PL Cleft LL PLL Lymph nodec 63%” 70% 90% 80% 67% 25% Spleen 43% 52% 40% 60% 67% 100% Anemia and/or thrombocytopeniad 41% 25% 10% 20% 25% 100% PB lymphocytosise 57 2 536 64 +- 81 41 t 27 81 2 50 57 f 65 140 e 14 % BM involvementf 76 i 16 61 k 21 63 2 17 73 2 26 79 t 13 85 e 13 Cases >70% BM lymphocytesg 65% 50% 22% 67% 80% 100% B Results expressed as percentage of positive cases. b Results expressed as means k SD. c Typical CLL vs LLL (P < 0.05). d Typical CLL vs LLL (P < 0.01). E Typical CLL vs CLL/PL (P < 0.05) and LLL vs CLL/PL (P < 0.05). f Typical CLL vs LLL (P < 0.05). * Typical CLL vs LLL (P < 0.01) and LLL vs CLL/PL (P < 0.01).

The patients with mixed morphology between these two forms (CLL/LL) dis- played intermediate values regarding BM , hemoglobin, platelet counts, and enlarged lymph nodes (Table 3). The group of patients with the inter- mediate form between typical CLL and PLL (CLL/PL) was characterized by a larger spleen and higher PB lymphocyte counts. Although lymphoid infiltration in BM was similar to that of the typical CLL group, there was a lower incidence of anemia and/or thrombopenia. Cleft LL exhibited a clinical spectrum similar to that of typical CLL. No significant differences were found among the morpholog- ical groups with respect to the pattern of BM infiltration. The immunological markers showed 15% sIg-negative cases in the whole series. However, when considering the different morphological groups it was ob- served that all LLL cases were positive for sIg with a weak-moderate intensity and a l.~+6+ predominant phenotype (80% of the cases). Conversely, in 11 out of 34 typical CLL cases sIg was not detected, l.~+ being the predominant isotype among the positive cases (Fig. 1). In the group of mixed CLL/LL 96% of the cases were sIg+ with a slight preponderance of l.~+6+cases. The expression of the Fmc/7 antigen was significantly higher (P < 0.001) in LLL and CLL/LL than in typical CLL. In these three groups of patients there were no significant differ- ences either in the proportion of MRFC or in the expression of GRBl, CD20, CD9, and CD5 antigens, although in the LLL cases the latter three markers dis- played slightly higher reactivity (Table 4 and Fig. 2). The CLL/PL group differed from the previous groups in the greater intensity of sIg (moderate/strong) with a more differentiated heavy chain isotype: p,+6+, 6+, l.~+&+o+ (Fig. 1). Reactivity with Fmc/7 was also significantly stronger than that of both LLL (P < 0.002) and typical CLL (P < 0.001). Furthermore, there was a significantly lower proportion of MRFC with respect to LLL and typical CLL (P 182 ORFAO ET AL.

__/ ‘Typical CL1 B.d Other % of slg+cells phenotypes ,/ Morphological groups

1 slg phenotype (heavy and light chains) FIG. I. Type of slg in the different morphological subgroups. n.d., sIg heavy and light chains not detected.

< 0.01) (Table 4 and Fig. 2). The Cleft LL cases expressed a phenotype similar to that of typical CLL (Table 4). The four patients diagnosed as B-PLL displayed more than 70% of prolymphocytes in PB and their clinical characteristics are shown in Table 3. The immunological phenotype was that of mature B lympho- cytes: sIg+ + +, Fmc/7+, MRFC- (Table 4).

DISCUSSION The morphological heterogeneity of B-CLL has been widely documented and a large lymphocyte with abundant cytoplasm and clumped chromatin without a nu- cleolous occasionally constitutes the predominant cell population in this group of (5, 6). Recently, attempts have been made to establish a relationship between morphological findings and the immunological phenotype in B lympho- proliferative disorders (7, 22). However, in CLL cases in which the proliferating cells are large lymphocytes no studies have been conducted to establish their clinical, biological, and immunological characteristics. Morphologically there are two main features which differentiate small from large lymphocytes: cell size and the amount of cytoplasm. Our results showed an excellent correlation between the measurement of lymphocyte diameter in mi- crometers and the morphological classification employed, based on lymphocyte size in relation to that of erythrocytes, confirming the reproducibility of this score method (7). On analyzing whether the morphological categories established really repre- sented different groups of patients, regarding other clinical and immunological characteristics apart from the morphological ones, it was confirmed that the LLL LARGE B-CELL CHRONIC LYMPHOCYTIC LEUKEMIA 183

TABLE 4 IMMIJNOLOGICALMARKERSOFTHEMORPHOLOGICALGROUPS

Typical Marker CLL CLLiLL LLL CLLiPL Cleft LL PLL slg” 44 i- 36 56 rfr 26 73 e 29 71 2 32 58 e 34 83 ” 7 (72%) (96%) (100%) (100%) (80%) (100%) MRFCb 58 2 28 53 t 26 62 2 18 43 2 33 49 2 21 426 (90%) (93%) (100%) (71%) (90%) (25%) B, 67 t 25 63 ? 25 82 2 23 80 t 28 57 2 30 94 + 2 (100%) (100%) (100%) (100%) (100%) (100%) GRB, 82 ” 18 80 rt 20 84 t 16 81 2 18 70 2 23 91 -t 1 (100%) (100%) (100%) ( loo%) (100%) Uoo%) Fmc-7’ 226 12 ” 23 13 r 14 27 T 23 3c5 72 2 22 (6%) (18%) (18%) (67%) (10%) (100%) Leu- I 40 t 31 43 2 29 45 2 36 38 k 30 49 2 32 31 2 27 (72%) (79%) (89%) (83%) (78%) (67%) Fmc-8156 33 r 30 31 2 28 36 + 31 38 2 32 40 k 35 31 * 37 (64%) (67%) (80%) (67%) (67%) (50%) Note. Results expressed both as means f SD of positive cells and as percentage of positive cases (> 10% positive cells). ” Typical CLL vs LLL (P < 0.001) and typical CLL vs CLL/PL (P < 0.02). b Typical CLL vs CLL/PL (P < 0.01) and typical LLL vs CLL/PL (P < 0.01). c Typical CLL vs LLL (P < 0.001). typical CLL vs CLUPL (P < 0.001). and LLL vs CLLiPL (P < 0.002). cases formed a group which was significantly different from typical CLL, CLL/ PL, and classical PLL. In the LLL cases all the patients were sIg+ with a pre- dominant ~+a+ phenotype, in contrast to the typical CLL cases that frequently displayed a negative or weak p.+ sIg phenotype. Both findings suggest that the lymphocyte in LLL is a more differentiated cell. This fact would also be sup- ported by the significant increase in this group of patients in the expression of Fmc-7, a marker associated with an advanced stage of maturation (16, 23). The higher incidence of anemia and/or thrombopenia in typical CLL could be related to the greater BM infiltration found in this group. LLL also appeared to be clearly differentiated from both CLL/PL and PLL. The main differences between them were the greater extent of PB lymphocytosis and the more mature phenotype (Fmc-7+, MRFC-) observed in the latter two groups. The CLL/PL patients showed an intermediate phenotype between typical CLL and PLL, confirming other previous findings (7, 24). Accordingly, in a hypothetical scheme of B-cell differentiation, the large lymphocyte would be a more differentiated cell than the small lymphocyte, in turn, preceding the prolymphocyte. Nevertheless, in our study we found no cases of intermediate forms of large lymphocytes/prolympho- cytes but did find intermediate forms of CLL/LL and CLL/PL. These findings, together with the different pattern of infiltration observed in LLL cases, being predominantly in lymph nodes instead of spleen and/or BM (as is the rule in PLL and typical CLL, respectively), suggest that the large lymphocytes, rather than an intermediate stage in the B-cell differentiation between the small lymphocytes 184 ORFAO ET AL.

T

Fmc-7 0 / i /

s maemia and/or thrombopenia L n E.M. Lymphocytosis>70% m A n j Typical CLLJ Lymph Node Slg / [Cleft LLI 0 Leo-1 “\ l

M&C

FIG. 2. HJ-biplot multivariant analysis: clinical and surface characterization of the morphological groups. and the prolymphocytes, could represent a with a different peripheral tro- phism. In summary, in this series we have seen that the cases of LLL are not only morphologically different from typical CLL, CLL/PL, and classical PLL, but also expressed some different phenotypic, clinical, and biological characteristics. The group of cases with mixed small/large lymphocytes (CLL/LL) exhibited clin- ical and immunological features between typical CLL and LLL, suggesting that this group could be a transitional form. Analysis of the clinical and prognostic significance of these morphological subgroups together with functional studies of these different cell populations could contribute to elucidating whether one is dealing with different clinical entities or whether there is a single disease with important heterogeneity in its morphological, clinical, and immunological mani- festations.

ACKNOWLEDGMENTS

We are grateful to Dr. H. Zola for the gift of the FMC monoclonal antibodies used in this study and th J. V. Melo for critical review of the paper. This work has been partially supported by a grant from the Educational Council of Castilla-Leon, Spain. A. Orfao was supported by a grant from the instituto National de Investigacao Cientifica.

REFERENCES

1. Bouroncle, B. A., Blood 53, 412, 1979. 2. Galton, D. A. G., Catovsky, D., and Wiltshaw, E., Cancer 42, 901, 1978. 3. Galton, D. A. G., In “Recent Advances in Haematology, 3” (A. V. Hoffbrand, Ed.), pp. 183-206. Churchill-Livingstone, London, 1982. LARGE B-CELL CHRONIC LYMPHOCYTIC LEUKEMIA 185 4. Oifao, A., Gonzalez, M., San Miguel, J. F., Galindo, I?, Moro, M. J., Jimenez Galindo, R., Ba- scones, C., and Lopez Borrasca, A., Sangre 31, 550, 1987. 5. Woessner, S., Lafuente Rodes, R., and Sans-Sabrafen, J., Sangre 22, 871, 1977. 6. Zacharski, L. R., and Linman, J. W., J. Natl. Cancer Inst. 48, 51, 1972. 7. Melo, J. V., Catovsky, D., and Galton, D. A. G., Brit. J. Haematol. 63, 377, 1986. 8. Enno, A., Catovsky, D., Obrien, M., Cherchi, M.. Kumaran, T. 0.. and Galton, D. A. G., Brjt. J. Haematol. 41, 9, 1979. 9. Rai, K. R., Sawitsky, A., Cronkite, E. P., Chanama, A. D.. Levy, R. N., and Pasternack, B. S., Blood 46, 219, 1975. 10. Binet, J. L., Catovsky, D., Chandra, Dighiero, G.. Montserrat, E., Rai, K. R., and Sawisthy, D., Brit. J. Haematol. 48, 365, 1981. 11. Hemandez-Nieto, L., Montserrat, E., Muncunill, J., and Rozman, C., Lancer 1, 1269, 1977. 12. Kaplan, M. E., and Clark, C., J. Immunol. Methods 5, 131, 1974. 13. Statlopoulos, G., and Elliot, E. U., Lancer i, 600, 1974. 14. Ruiz-Cabello, E, Gonzalez, M., Lopez, M. A., Cabrera, A., San Miguel, J. F., and Garrido, F., Sangre 30, 190, 1985. 15. Nadler, L. M., Ritz, J., Hardy, R., and Pesando, J. M., J. C/in. Invest. 67, 134, 1981. 16. Catovsky, D., Cherchi, M., Brooks, D., Bradley, J., and Zola, H., Blood 58, 406, 1981. 17. Zola, H., McNamara, P. J., Moore, H. A., Smart, I. J., Brooks, D. A., Beckman, I. G. R., and Bradley, J., Clin. Exp. Immunol. 52, 655, 1983. 18. Knowles, D. M., Halper, I?, Azzo. W., and Wang, C., Cancer 52, 1369. 1983. 19. Gabriel, K. R., Biometrika 58, 453, 1971. 20. Galindo, F?, Questtio 10, 13, 1986. 21. Golub, G. H.. and Reinsch, C., Numer Math. 14, 403, 1970. 22. Ottolander, G. J., Schuit, H., Waayer, J., Huibregtsen, L.. Hijmans. W.. and Jansen, J., C/in. Immunol. Immunopathol. 35, 92, 1985. 23. Zola, H.. Moore, H. A., and Hohmann, A., J. Immunol. 133, 321, 1984. 24. Melo. J. V., Catovsky, D., and Galton, D. A. G., Brit. J. Huematol. 64, 77, 1986. Received March 16. 1987; accepted with revision July 4, 1987