(2002) 16, 1691–1698  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Surface membrane antigen expression changes induced in vitro by exogenous growth factors in chronic lymphocytic leukemia cells J Vilpo1, J Hulkkonen2, M Hurme2 and L Vilpo1

1Department of Clinical Chemistry, University of Tampere Medical School and Laboratory Centre of Tampere University Hospital, Tampere, Finland; and 2Department of Microbiology and Immunology, University of Tampere, Medical School and Laboratory Centre of Tampere University Hospital, Tampere, Finland

The factors determining the growth and survival of cells in B necrosis factor superfamily members, such as chronic lymphocytic leukemia (CLL) have remained poorly CD30,10 CD4011–13 and CD95.14–16 understood. We investigated the effects of optimal mitogen combinations (OMCs) on the expression of 26 surface mem- In addition to the data accumulated concerningtheir bio- brane antigens among 33 CLL patients. The seven OMCs used chemical actions and cellular functions, some interesting were selected after pre-testing 14 combinations of (1) S. aureus clinical aspects of several membrane antigens are emerging. Cowan I (SAC), (2) interleukin-2 (IL-2), (3) tumor necrosis factor The marked heterogeneity in the clinical course of patients alpha (TNF-␣) and (4) 12-O-tetradecanoylphorbol 13-acetate with apparently similar features can be explained by variable (TPA; also known as phorbol 12-myristate 13-acetate or PMA). expression of regulatory molecules. This, on the other hand, In flow cytometry we revealed that OMCs induced statistically highly significant upregulation of the expression of CD5, should provide novel ways to intervene in the clinical course CD11c, CD19, CD22, CD23, CD25, CD38, CD40, CD45, CD45RO, of CLL progression. All these aspects point to the importance CD95, CD126, CD130 and FMC7, and downregulation of CD20 of membrane signals mediated by a large number of mol- and CD124 expression. Interestingly, the expression of CD27, ecules. The aim of the present work was to investigate whether CD45RA, CD79b, CD80, CD122 and that of the immunoglobulin the expression of membrane antigens can be altered. The ␬ ␭ ␦ ␮ superfamily members CD21, Ig- , Ig- , Ig- and Ig- were challenging molecules were selected according to their not significantly affected under similar conditions. The expression of several antigens was co-regulated, suggesting capacity to induce maximal proliferation in vitro. For this pur- common regulatory pathways. These antigens include pose, the most effective mitogen combination for each patient CD11c/CD5, CD11c/CD22, CD11c/CD126, CD11c/FMC7 as well was chosen from a panel of 14 combinations of four B-CLL as CD27/CD45, CD27/CD45RA and CD27/CD79b. Upregulation cell mitogens, as previously described.17–19 of surface antigen expression, induced by OMCs, should be applicable in therapy in vitro and in vivo, and in nega- tive stem cell selection for autotransplantation. Furthermore, the current strategy to enhance cell surface antigen expression Materials and methods may be a versatile tool to raise humoral and cell-mediated host defense against CLL cells. Upregulation of mediating positive growth signals (eg CD25, CD40) and negative signals Patients and samples or apoptosis (eg CD95) may be used to sensitize cells to chemotherapy and programmed cell death. Clinical specimens were obtained after informed consent from Leukemia (2002) 16, 1691–1698. doi:10.1038/sj.leu.2402583 Keywords: chronic lymphocytic leukemia; cluster of differentiation; 33 CLL patients referred to the CLL out-patient clinic at Tam- mitogen stimulation; surface membrane antigens pere University Hospital. Diagnosis and staging were based on standard clinical, morphological and immunophenotyping criteria. All patients had a CD19+/CD5+/CD23+ immunophen- Introduction otype, and mature CLL, CLL/mix or CLL/PL morphology. Most patients had chromosomal aberrations, but none indicating 20 Chronic lymphocytic leukemia (CLL) is a clonal neo- chronic lymphoid other than typical CLL. The plasm of small, resting, long-lived B cells. Despite recent clinical hematological data of the 33 CLL patients are illus- advances in understandingof the genetics, 1 biology,2 treat- trated in Table 1. ment,3 and clinical behavior4 of CLL, there is still no estab- Peripheral blood mononuclear cells were isolated from hep- lished cure for the disease, and its progression and outcome arinized (Noparin; Novo Nordik, Dagsvaerd, Denmark) blood are highly unpredictable. New strategies are required. samples by centrifugation over a Lymphoprep layer CLL cells co-express low levels of surface membrane (Nycomed, Oslo, Norway) at a density of 1.077 g/ml. The immunoglobulins (SmIgs) and the CD5 molecule. The domi- cells were washed twice with phosphate-buffered saline (PBS, nant immunophenotype is easily detectable by flow cytome- pH 7.4) and once with medium consistingof RPMI 1640 (20 try: weak SmIg, positive CD5, positive CD23, negative FMC7, mM Hepes; ICN Biochemicals, Costa Mesa, CA, USA), 10% weak/negative CD22, and negative CD79b.5,6 heat-inactivated fetal calf serum (Gibco BRL, Paisley, UK), 2 mM L-glutamine (Gibco BRL) and antibiotics (Gibco BRL, Many membrane molecules are known to influence B-CLL ␮ tumor cells through their effects on the mechanisms regulating penicillin, 50 U/ml, streptomycin, 50 g/ml). Cell counting cell proliferation and apoptosis. These include, for instance, was performed by usingTechnicon H1, H2 or H3 analyzers IL-2 receptor components (CD25, CD122 and CD132),7 IL-4 (Bayer Diagnostica, Tarrytown, NY, USA). Density gradient receptor (CD124),8,9 IL-6 receptor (CD126),9 as well as tumor centrifugation yields – without other purification steps – have been good (range 29–92%, mean 61.7%, s.d. 16.2%, n = 36), as previously described. This information is necessary for interpretation of the results.21 The proportion of Correspondence: J Vilpo, Department of Clinical Chemistry, Tampere University Hospital, PO Box 2000, FIN-33521 Tampere, Finland; Fax: plus polyclonal T and B was 1–13%, indicating 358 3 247 4048 that 87–99% of the isolated cells represented the leukemic Received 15 October 2001; accepted 9 April 2002 population. Surface antigen changes in CLL J Vilpo et al 1692 Table 1 Clinical and hematological data at the time of sampling of 33 patients with chronic lymphocytic leukemia

Patient No. Age Sex Binetb Blood FABc Progression Therapy OMCd n = 33a lymphocytes (×109/l)

1 54 m A 63 CLL S None 10 2 64 m A 46 CLL T LP 11 3 72 m B 84 CLL F None 8 4 69 m B 92 CLL F L 10 5 61 f A 86 CLL S None 6 6 50 f A 130 CLL S None 10 8 58 m C 310 CLL/PL T LP, 6×COP, 4 7×CHOP 9 68 m B 67 CLL S None 11 10 68 f A 51 CLL I None 8 11 68 f C 88 CLL T L 4 12 64 m C 132 CLL T LP 4 13 68 m C 114 CLL F None 10 14 66 m C 188 CLL/PL T LP, 7×COP, 10 10×CHOP 15 57 m A 63 CLL F None 10 16 68 m A 81 CLL I None 11 17 57 f A 120 CLL F None 6 18 73 f C 79 CLL/mix F None 10 19 55 m A 36 CLL/PL F None 10 20 53 m B 69 CLL/PL I None 10 21 73 f B 142 CLL/mix I None 10 22 55 f B 59 CLL/PL I None 6 23 62 m B 59 CLL S None 10 24 79 m B 97 CLL/mix T 4×CHOP, LP 8 26 48 m A 93 CLL S None 6 27 67 m A 67 CLL I None 6 28 67 m C 178 CLL F/T LP 11 29 78 f B 224 CLL F None 8 30 69 m A 69 CLL F None 11 32 70 m A 134 CLL F None 10 33 57 m A 68 CLL S None 12 35 76 f C 206 CLL F None 12 36 78 m A 155 CLL S None 11 37 63 m C 340 CLL/PL F None 14

aCases 7, 25, 31 and 34 were not available for this investigation. bBinet et al.38 cBennett et al.39 dOMC; optimal mitogen combination selected from 14 combinations;17,19 4 = SAC (0.005%) + IL-2 (10 U/ml), 6 = TPA10 (10 ng/ml) + IL-2 (10 U/ml), 8 = TPA1 (1 ng/ml) + IL-2 (10 U/ml), 10 = TNF-␣ (10 ng/ml) + IL-2 (10 U/ml), 11 = TNF-␣ (10 ng/ml) + TPA1 (1 ng/ml), 12 = TNF- ␣ (10 ng/ml) + TPA1 (1 ng/ml) + IL-2 (10 U/ml), 14 = TNF-␣ (10 ng/ml) + TPA10 (10 ng/ml) + IL-2 (10 U/ml). IL-2, interleukin-2; SAC, Staphylococcus aureus Cowan 1; TPA, 12-O-tetradecanoylphorbol 13-acetate; TNF-␣, recombinant tumor necrosis factor alpha; S (slow), blood lymphocytes increased less than 20% within a year; F (fast, doubling time 1 year or less; I (intermediate) between S and F; T (therapy), chemotherapy given, natural disease progression non-evaluable; L, Leukeran (chlorambucil); P, predniso(lo)ne; CH(O)P, cyclophosphamide-hydroxydaunorubicin (doxomycin)-oncovine (vincristine)-prednisolone.

Proliferation and activation of CLL cells in suspension nucleic acids were precipitated with 0.2 M perchloric acid cultures and the radioactivity in the macromolecular and nucleic acid fraction was counted. Stimulation indices were Mononuclear cells were incubated in 200 ␮l aliquots (2 × 106 calculated by dividingthe day 4 tracer incorporation values in cells/ml) in RPMI-1640 medium (vide supra)at+37°Cina mitogen-stimulated cultures by the corresponding day 4 tracer

humidified atmosphere containing5% CO 2. Mitogens were incorporation values in unstimulated control cultures. Indices used accordingto Table 1: these were IL-2 (Boehringer- were calculated for activation (protein synthesis or leucine Mannheim, Mannheim, Germany); TPA (Sigma Chemical, St incorporation values). An optimal mitogen combination Louis, MO, USA); SAC (Pansorbin Cells; Calbiochem-Novabi- (OMC) was selected for every patient on the basis of the ochem, La Jolla, CA, USA) and TNF-␣ (Immunex, Los Angeles, mitogen combination inducing the highest rate of thymidine CA, USA). Proliferation activity and total protein synthesis of incorporation in 4 day cultures. the cells were measured as the rate of incorporation of (methyl-3H)-thymidine and (U-14C)-L-leucine, respectively. The specific activities were: 3H-dThd, 70 Ci/mmol; 14C-Leu, Immunophenotyping 35 mCi/mmol or 3.5 mCi/mmol. The tracers were purchased from Amersham (Amersham, Bucks, UK). At 72 h of culture, Immunophenotypingwas performed with cryopreserved cells 0.1 ␮Ci of 3H-dThd per well and 0.1 or 0.01 ␮Ci of 14C-Leu by usingflow cytometry (FACSCalibur, Becton Dickinson per well were added. After a further 24 h, the proteins and Immunocytometry Systems, Mountain View, CA, USA) and

Leukemia Surface antigen changes in CLL J Vilpo et al 1693 commercial mouse monoclonal with respective membrane was analyzed in a histogram data display using a immunoglobulin isotype controls (Table 2). The expression of logarithmic scale. The number of CD45-positive cells was antigens was determined after 4 days in vitro stimulation with always Ͼ98% (99.79 ± 0.35). The mean proportion of CD14- an OMC (time point D4) and compared with that of cells ana- positive cells in an analyzed gate was 0.77% (s.d. 1.11, range lyzed immediately after thawing(D0). Stainingwas performed 0.05–6.82). We used geometric mean fluorescence (GMF) as accordingto the suppliers’ instructions. For cell surface a surrogate measure for antigen expression as proposed else- immunoglobulin staining, the cells were first incubated in PBS where (Imperial Cancer Research Fund FACS Laboratory at (pH 7.4) at 37°C for 30 min to remove heterophilic antibodies. http://lif.icnet.uk/axp/facs/davies/stats.html). The background The buffers used in all the steps did not contain serum. GMF values of pertinent isotype control antibodies were sub- The forward scatter channel was checked prior to every tracted to obtain the final net GMF figures. analysis usingunlabeled calibration particles (SPHERO Cali- There were some drop-outs from the analyses, as indicated bration Particles, blank 6.5–8 ␮m; Pharmingen, San Diego, by the numbers in Figure 1. Firstly, SAC particles adhered very CA, USA). Instrument linearity was checked usingcommercial tightly to CLL cells in 4-day cultures. Owing to the strong standard reagents (Immuno-Brite Standards Kit, Coulter, Hia- bindingof labeled antibodies to the SAC particles, flow cyto- leah, FL, USA) prior to the study. Fluorescence compensations metric determination of membrane antigens in SAC-stimulated were set prior to every analysis usingcalibration beads cultures was impossible. Secondly, in some cases the mitog- (caliBRITE; Becton Dickinson) and FACSComp software ens induced very strong aggregation of cells, making reason- (Becton Dickinson). Instrument calibration was additionally able surface membrane antigen determinations impossible. checked usingImmunobrite level II standard beads (Coulter). On the basis of the data obtained, the interassay variability in this study was 6.5% for the FL1 channel and 7.6% for the Statistical treatment FL2 channel. The percentages of cells were calculated on the basis of Statistical analyses were performed usinga statistical package: data obtained from two-color immunofluorescence dot blots STATISTICA for Windows (ver. Win 5.1D, 1996, StatSoft, (results not shown). Analyses were carried out usingCellQU- Tulsa, OK, USA). EST Software (Becton Dickinson). Quadrant markers were set relative to negative immunoglobulin isotype controls in a such way that 99% (99.05 ± 0.04) of the cells labeled with control Results antibodies were located in the left lower quadrant. When esti- matingthe reactivity of known antigenson the surface Cell surface antigen expression changes induced by of malignant B cells, the proportion of CD2 cells was taken optimal mitogen combinations into account. The intensity of antigen expression in the cell A statistically significant difference (P Ͻ 0.001, Wilcoxon matched pairs test; GMF on D4 minus GMF on D0) was noted Table 2 Monoclonal antibodies used with 16 of the 26 surface antigens when expression of unstimulated cells was compared with that of cells stimulated Antigen and Label Isotype Clone for 4 days ex vivo. In most cases the OMC induced enhance- sourcea ment of expression (Figure 1). Only CD20 and CD124 were B ␥ CD2 FITC 2a Leu-5b repressed. Individual changes were also noted in regard to ten B ␥ CD20 FITC 1 Leu16 other antigens, but no common up- or downregulation trends B ␥ CD45RA FITC 1 Leu-18 B were demonstrable (Table 3). CD45RO PE ␥2a Leu-45RO B ␥ CD122 PE 1 Anti-IL-2R-p75 B ␥ CD19 PE 1 sJ25C1 B CD5 FITC ␥2a Leu-1 Association of membrane antigen expression changes B ␥ CD23 PE 1 Leu-20 with proliferation and activation FMC7I FITC ␮ FMC7 B ␥ CD22 FITC 2b Leu-14 Proliferation and activation results are not shown in detail, but B ␥ CD38 PE 1 Leu-17 17,19 I similar results have been published elsewhere. Prolifer- CD124 PE ␥1 S4-56C9 I ␥ ation per se was not significantly associated with antigen CD126 PE 1 M91 D ␥ CD11c FITC 1 KB90 expression changes (Table 4). Surprisingly, a negative corre- D ␥ CD21 FITC 1 1F8 lation was observed concerningCD5 and CD11c, showing D ␥ CD79b FITC 1 SN8 that suppressed proliferation was associated with strongupre- ␥P ␥ FITC 1 G18-145 gulation of these antigens. The protein synthesis-based acti- ␦P ␥ FITC 2a IA6-2 P vation index, assessed by leucine incorporation, correlated ␮ PE ␥1 G20-127 P ␥ with the expression changes of several antigens (Table 4). CD95 FITC 1 DX2 P ␥ CD40 FITC 1 5C3 Interestingly, the correlation was negative in some cases. P ␥ CD27 FITC 1 M-T271 (See also discussion concerningproblems with multiple P ␥ CD130 PE 1 VC041 correlations, below.) B ␥ CD25 FITC 1 Anti-IL-2R B ␥ CD80 PE 1 Anti-B7 B CD45/CD14 FITC/PE ␥1/␥2b Anti-Hle-1/Leu-3M B Correlations between the expression changes of ␬/␭ FITC/PE ␥1/␥1 TB28-2/1-155-2 different antigens aB, Becton Dickinson; D, Dako; I, Immunotech; P, Pharmingen. For more accurate description of antigen proteins, see eg We also analyzed whether possible co-regulation of changes http://www.ncbi.nlm.nih.gov/PROW. in the expression of different antigens took place. To this end,

Leukemia Surface antigen changes in CLL J Vilpo et al 1694

Figure 1 Effect of optimal mitogen combinations in 4-day cultures on the surface membrane expression (geometric mean fluorescence) of 16 antigens showing a statistically significant change (P Ͻ 0.001, as analyzed by usingWilcoxon matched pairs test). The numbers of eligible CLL patients are given above each chart. Point, median; whiskers, min and max; rectangles, 25–75%. Note that CD124 and CD20 were downregulated by the OMCs.

the order correlations amongthe 26 antigenswere ana- relationships, in vitro, between the rate of CLL cell prolifer- lyzed. Owingto the greatnumber of comparisons, we have ation, total protein synthesis and the expression of 26 cell sur- chosen to illustrate only the most significant correlations face membrane antigens. The growth promoting mechanisms (Figure 2). Interestingly, CD11c and CD27 were co-regulated of these modulators in normal B cells are gradually emerging with four and three other antigens, respectively. A negative, and similar signaling pathways are also operative in B-CLL, albeit non-linear correlation, was observed with the changes although not necessarily in an identical manner as in normal of CD23 and CD95 expression. B cells (for a review, see Meinhardt et al2). In this study we selected seven mitogen combinations from a set of 14. This set was previously formed on the basis of four known B-CLL Kinetics of induction of surface membrane antigens cell mitogens tested in this laboratory.17–19 Several novel fea- tures concerningsurface membrane antigenexpression The analysis of the kinetics of some example antigens demon- emerged from the present analyses. The current data clearly strated that additional induction may take place after day 4 demonstrated that many different types of membrane antigen (Figure 3). are inducible or repressible under certain conditions. Information about cell surface membrane antigen regulation in CLL is very limited. In a recent investigation it was reported Discussion that several CLL cell antigens, including CD58, CD80, CD86, and MHC class I molecules, can be upregulated by CpG-oli- In this study we examined the effects of four known CLL cell gonucleotides.22 Another immunostimulant, the guanosine ␣ growth promoters (SAC, IL-2, TPA, TNF- ) on the inter- analogloxoribine, effectively upregulatedCLL surface anti-

Leukemia Surface antigen changes in CLL J Vilpo et al 1695 Table 3 OMC stimulation-associated changes of surface mem- proteins discussed here). The importance of CD5 in CLL brane antigens of B-CLL cells in vitro immunophenotypingis established. However, as recently reviewed, the role of CD5 as an independent lineage marker Surface Geometric mean fluorescence for normal polyclonal human CD19+/CD5+ has remained antigen unresolved.25 Another possibility exists that CD5 is an acti- Day 0 Day 4 vation marker. The present investigation demonstrated that Median Min–Max Median Min–Max this is the case in CLL cells; a positive correlation was observed between OMC-induced activation (protein CD21 15 0.1–57 15 4–48 CD27 29 11–99 22 2–163 synthesis) and enhancement of the expression of CD5. A simi- CD45RA 105 9–643 119 8–909 lar correlation has been previously noted between the release CD79b 8 0.01–22 7 2–35 of a known activation marker, ie beta-2-microglobulin, and CD80 0 0–22 0 0–628 protein synthesis in OMC-stimulated CLL cells.19 The negative CD122 0 0–1 0 0–21 correlation of CD5 with proliferation, as demonstrated here, ␬ Ig- 5 0–65 11 0–72 suggests that CD5 intensity cannot be used as a proliferation Ig-␭ 1.6 0–244 4.0 0.01–834 Ig-␦ 4.3 0.01–20.4 1.7 0–14 marker for CLL cells. Ig-␮ 7.1 0.01–84 13 0.01–358 We also demonstrated a variable but in some cases remark- able enhancement of CD11c expression in OMC-treated CLL CLL cells were incubated in vitro for 4 days in the presence of an cells. CD11c is an alpha X chain of the CD11c/CD18 optimal mitogen combination, whereafter surface antigen adhesion molecule, which mediates leukocyte spread on to expression was determined by using flow cytometry and fluor- protein-coated surfaces. Low levels of expression have been escent antibodies as described in Materials and methods. The ‘Day recorded in CLL cells.26–28 Higher intensities have been asso- 0’ cells were assessed without stimulation. ciated with early-stage disease,27 while cells with 11q deletion carry significantly lower amounts of CD11c.28 It remains to be resolved why low expression of CD11c, a molecule mediating Table 4 Relationship of proliferation (thymidine incorporation) leukocyte spread, is paradoxically associated with 11q and total activation (leucine incorporation) with the OMC-induced 1 antigen expression change as assessed by using Spearman’s rank order deletion, a marker of poor prognosis of CLL. The current correlation. Only statistically significant results (P Ͻ 0.05) are illus- approach provides a means to upregulate CD11c expression trated and possibly to affect the biological behavior of CLL cells. We also demonstrated that the expression of CD5, CD22 (an End Antigen R P adhesion and signaling molecule), CD126 (IL-6 receptor) and point FMC7 (currently unknown function) are co-modulated, with CD11c, by OMCs. There is a possibility that this simply Proliferationa CD5 −0.436 0.0161 reflects a common ‘non-specific’ activation, since the − CD11c 0.372 0.0372 enhancement of expression of all these antigens was associa- CD45RA 0.354 0.0433 ted with the enhancement of total protein synthesis. Similar Activationb CD5 0.524 0.0029 co-regulation was observed as regards CD27 (tumor necrosis CD11c 0.528 0.0019 − factor receptor family), with CD45, CD45RA and CD79b. CD19 0.404 0.0269 Although CD27 is known to mediate a co-stimulatory signal CD22 0.6889 Ͻ0.0001 CD45RA −0.4702 0.0058 of B cell activation, more precise interpretation of its role in CD45RO 0.4410 0.0147 OMC-stimulated B-CLL cells requires that its basic cellular CD79b −0.4731 0.0062 functions become better characterized. CD80 −0.3656 0.0470 CD19 is a B lymphocyte development, activation and differ- CD122 0.4262 0.0188 entiation regulator. Considerable enhancement of this B cell FMC7 0.3700 0.0341 co-receptor member by OMCs is in contrast to the regulation

a 3 of its complementation partner CD21 and other B cell recep- Proliferation was measured as H-thymidine incorporation tor-associated molecules or SmIgtypes examined here. The (dpm/culture well). bActivation was measured as leucine incorporation index: leucine overall activation of CLL cells does not explain this exception, incorporation in cultures treated with an OMC for 4 days per leucine since the protein synthesis rate was inversely correlated to incorporation in unstimulated control cultures. CD19 expression enhancement. The significance of CD19 in CLL as well as its increased expression in OMC-treated cells, as shown here, remains to be determined. gens such as CD22, CD23, CD25, CD38 and CD54, while CD20 and CD124 were the only OMC-downregulated anti- CD5, CD11c, CD20, HLA-DR, and mIg-␭ showed little or no gens in this investigation. The precise functions of CD20 are change.23 Furthermore, it has been shown previously that the currently unknown. The structure of CD20, however, indi- CD40-ligand (CD40L) CD154 increases the expression of cates that it may be involved in the regulation of B cell acti- CD40, CD54, CD69, CD70, CD80 and CD95, whereas the vation and proliferation. Hence, modulation of these putative protein kinase C modulator bryostatin-1 increased expression functions by OMCs is possible. It is conceivable that similar of CD40, CD54, CD69 and CD95.24 To our knowledge, our downregulation of CD20 expression may also take place in OMC strategy provides, for the first time, tools to upregulate vivo and should be taken into account in anti-CD20 antibody CD5, CD11c, CD19, CD22, CD23, CD25, CD38,CD40, therapy and in anti-CD20-based in vitro purging and negative CD45, CD45RO, CD95, CD126, CD130 and FMC7, and selection of stem cell preparations processed for autotrans- downregulate CD20 and CD124 in CLL cells. plantation. CD5 is a signal transducing molecule from the scavenger CD23 belongs to the sialoadhesins of immunoglobulin receptor superfamily (see http://www.ncbi.nlm.nih.gov/ supergene family products. It functions as an IgE synthesis PROV/guide for upgraded information about this and other regulator and triggers cytokine release by human monocytes;

Leukemia Surface antigen changes in CLL J Vilpo et al 1696

Figure 2 Spearman rank order correlations between the expression changes of different surface membrane antigens (P р 0.0004). A negative correlation between the changes in CD23 vs CD95 is illustrated in the last panel. Correlation coefficient R, significance (P) and the number of eligible patients are given above each panel.

10 and 11 (Table 1), ie TPA + IL-2, TNF␣ + IL-2 and TNF␣ + TPA. This demonstrates that physiological cytokines and pro- tein kinase C modulators may be involved in the regulation of CD38 expression. IL-2 and TNF␣ have thoroughly been investigated in B-CLL and their role in disease progression is evident, as recently reviewed.32 The upregulation of CD38 by TNF␣, IL-2, or both, as well as by agents modulating protein kinase C activity, may explain the high expression of CD38 in progressive forms of CLL. CD40 is a member of the tumor necrosis factor receptor superfamily gene products, which have functions in B lym- phocyte growth, differentiation and isotype-switching as well as in rescue signaling from apoptosis in cells. Native B-CLL cells are unable to function effectively as anti- gen-presenting cells, and they fail to induce a relevant T cell immune response. Therefore, the CD40 pathway in B-CLL Figure 3 Kinetics of induction on days 0–6 of four example anti- cells has been actively investigated. Expression of both CD40 gens of two patients. and CD154 (CD40L) may have an autocrine or paracrine function in cell growth (reviewed by Orsini and Foa32). The TNF␣, IL-1, IL-2 and GM-CSF. CD23 ligands include CD21 most effective upregulators of CD40 were OMC combinations + ␣ + and CD11c, the latter beingremarkably upregulatedby 6, 8 and 10 (Table 1), containingTPA IL-2 and TNF IL- OMCs. This may indicate collaboration of CD11c and CD23 2. It has been demonstrated that the serum of patients with B- 33 in OMC-treated B-CLL cells. CD23 is unstable on CLL cells; CLL contains increased levels of CD40L and that the in vivo, it is rapidly cleaved from the surface into a stable CD40/CD40L system may be effective in inhibitingthe apop- 34 ␣ form, sCD23 (reviewed by Molica29). Serum sCD23 is an totic response to fludarabine. As discussed above, TNF and established risk factor in CLL. Remarkable enhancement of IL-2 are thought be involved in CLL progression32 and the CD23 expression was induced by OMCs in vitro. It is conceiv- present results imply that upregulation of their signaling cas- able that similar cytokine modulation may mark progressive cades may be a mediator of proliferation and resistance to disease in vivo. apoptosis. It is also noteworthy that the expression of IL-2 CD38 has multiple functions, such as NAD glycohydrolase, receptor subunits, as exemplified by the alpha subunit (CD25) ADP-ribosyl cyclase, cyclic ADP ribose hydrolase and base- is enhanced by OMCs. This kind of receptor amplification exchange activities. Particular interest in this molecule in CLL may further potentiate the anti-apoptotic (via CD40) and is based on its possible value as a marker of poor prog- growth promoting (via CD38?) effects of IL-2. nosis30,31 and its association with the naive or unmutated Remarkable upregulation of CD95 was induced by OMCs maturation stage of the disease.30 In some cases, remarkable in most of the patients. CD95 (APO-1 or ) is a trans- enhancement of CD38 expression was induced by an OMC. membrane glycoprotein that belongs to the TNF receptor The most effective combinations in this regard were Nos 8, superfamily. It is thought to be one of the principal mediators

Leukemia Surface antigen changes in CLL J Vilpo et al 1697 of apoptosis-inducingsignals.In addition to CD40L and Acknowledgements bryostatin-1 (see above), upregulation of CD95 by IFN␣, IFN␥ and SAC has been observed in other laboratories, as recently We thank Leena Pankko and Merja Suoranta for their techni- reviewed.32 B-CLL cells may be resistant to CD95-mediated cal assistance. This work was supported by grants from the apoptosis,32 but this certainly is a sum effect of many regu- Medical Research Fund of Tampere University Hospital, and lators of apoptotic pathways. Hence, it remains to be investi- the Finnish Cancer Organization. gated if the considerable upregulation of CD95 by OMCs can be used to kill B-CLL cells. A against FMC7 detects an antigen References that is present in types of chronic other than typical B-CLL (egB-PLL, HCL, leukemic non-Hodgkin’s 1Do¨hner H, Stilgenbauer S, Benner A, Leupolt E, Krober A, Bul- 5,6 linger L, Dohner K, Bentz M, Lichter P. Genomic aberrations and ). It has recently been proposed that FMC7 anti- survival in chronic lymphocytic leukemia. N Engl J Med 2000; gen expression on normal and malignant B cells can be pre- 343: 1910–1916. dicted by the expression of CD20.35 This may be explained 2 Meinhardt G, Wendtner C, Hallek M. Molecular pathogenesis of by the recent findingthat FMC7 antibody detects a confor- chronic lymphocytic leukemia: factors and signaling pathways mational epitope on the CD20 molecule.36 The current data regulating cell growth and survival. J Mol Med 1999; 77: 282–293. 3 Rai K, Peterson B, Appelbaum F, Kolitz J, Elias L, Sepherd L, Hines demonstrated a good correlation between CD22 and FMC7 = = J, Threatte G, Larson R, Cheson B, Schiffer C. Fludarabine com- expression changes (R 0.59, P 0.0007, Spearman’s rank pared with chlorambucil as primary therapy for chronic lympho- order correlation) and the epitope bound by the FMC7 anti- cytic leukemia. N Engl J Med 2000; 343: 1750–1757. body may be closer to other antigens co-regulated with 4 Kipps T. Chronic lymphocytic leukemia. Curr Opin Hematol CD11c, ie CD5 and CD126, in addition to CD22, than to 2000; 7: 223–232. CD20. No correlation was seen between FMC7 and CD20 (R 5 Moreau EJ, Matutes E, A’Hern RP, Morilla AM, Morilla RM, =− = Owusu-Ankomah KA, Seon BK, Catovsky D. Improvement of the 0.06, P 0.76). This indicates very different regulation of chronic lymphocytic leukemia scoringsystem with the mono- epitopes detected by FMC7 and CD20 antibodies under the clonal antibody SN8 (CD79b). Am J Clin Pathol 1997; 108: OMC-stimulation used in this work. 378–382. Although B prolymphocytic leukemia (B-PLL) can be dis- 6 Matutes E, Owusu-Ankomah K, Morilla R, Garcia Marco J, Houli- han A, Que TH, Catovsky D. The immunological profile of B-cell tinguished as a separate independent clinical and immunolog- 37 disorders and proposal of a scoringsystem for the diagnosisof ical entity, many features are similar to those of B-CLL, in CLL. Leukemia 1994; 8: 1640–1645. particular, atypical and transformed cases of B-CLL. It is inter- 7 de Totero D, di Celle F, Cignettyi A, Foa R. The IL-2 receptor esting to note that several B-CLL antigens can be upregulated, complex: expression and function on normal and leukemic B cells. as shown here, so that the immunophenotype starts to Leukemia 1995; 9: 1425–1431. 8 Fluckiger AC, Briere F, Zurawski G, Bridon JM, Banchereau J. IL- resemble that of B-PLL. These antigens include FMC7, CD22, 37 13 has only a subset of IL-4-like activities on B chronic lympho- CD79b, CD25, CD11c and CD38, indicatingthat OMC-like cytic leukaemia cells. Immunology 1994; 83: 397–403. factors may be particularly operative in vivo in B-PLL. This is 9 van Kooten C, Rensink I, Aarden L, van Oers R. Effect of IL-4 and in accordance with the fast clinical progression of that disease. IL-6 on the proliferation and differentiation of B-chronic lympho- In this study we used multiple comparison procedures, cytic leukemia cells. Leukemia 1993; 7: 618–624. 10 Trentin L, Zambello R, Sancetta R, Facco M, Cerutti A, Perin A, which have an internal tendency to reveal non-existingcorre- Siviero M, Basso U, Bortolin M, Adami F, Agostini C, Semenzato lations. On the other hand, there are very few possibilities to G. B lymphocytes from patients with chronic lymphoproliferative accurately determine the P values in this kind of situation. disorders are equipped with different costimulatory molecules. One robust and very conservative approach is based on Bon- Cancer Res 1997; 57: 4940–4947. ferroni-type inequality, generally used with variance analysis, 11 Fluckiger AC, Rossi JF, Bussel A, Bryon P, Banchereau J, Defrance T. Responsiveness of chronic lymphocytic leukemia B cells acti- in which each P value inside a comparison family is multi- vated via surface Igs or CD40 to B-cell tropic factors. Blood 1992; plied by the number of comparisons. Takingthis into con- 80: 3173–3181. sideration, only the results with the smallest P values are 12 Crawford DH, Catovsky D. In vitro activation of leukaemic B cells presented. by interleukin-4 and antibodies to CD40. Immunology 1993; 80: To conclude, our most prominent findings are significant 40–44. 13 Buhmann R, Nolte A, Westhaus D, Emmerich B, Hallek M. CD40- up- or downregulation of a number of ‘new’ B-CLL surface activated B-cell chronic lymphocytic leukemia cells for tumor antigens by using individually selected optimal mitogen com- immunotherapy: stimulation of allogeneic versus autologous T binations. Furthermore, novel co-regulation of several anti- cells generates different types of effector cells. Blood 1999; 93: gens was observed. These molecules are important in cell sig- 1992–2002. naling, growth promotion and apoptosis induction or 14 Panayiotidis P, Ganeshaguru K, Foroni L, Hoffbrand AV. Expression and function of the FAS antigen in B chronic lympho- prevention, for example. Our results introduce new aspects to cytic leukemia and . Leukemia 1995; 9: strategies in B-CLL research: (1) understanding and inter- 1227–1232. vention in proliferation and apoptosis by means of regulation 15 Kamihira S, Yamada Y, Hirakata Y, Tsuruda K, Sugahara K, via immunomodulators, (2) enhancingthe expression of inter- Tomonaga M, Maeda T, Tsukasaki K, Atogami S, Kobayashi N. estingcandidate antigensfor in vitro antibody purging and in Quantitative characterization and potential function of membrane Fas/APO-1 (CD95) receptors on leukaemic cells from chronic B vivo antibody therapy as well as for negative selection of stem and T lymphoid leukaemias. Br J Haematol 1997; 99: 858–865. cells for autotransplantation. This should ultimately lead to a 16 Tsuruda K, Yamada Y, Hirakata Y, Sugahara K, Maeda T, Atogami better understandingof the clinical behavior of CLL as well S, Tomonaga M, Kamihira S. Qualitative and quantitative charac- as to the development of novel treatment strategies. Both are terization of Fas (APO-1/CD95) on leukemic cells derived from urgently needed in CLL. patients with B-cell neoplasms. Leukemia Res 1999; 23: 159–166. 17 Larramendy M, Siitonen S, Zhu Y, Hurme M, Vilpo L, Vilpo J, Knuutila S. Optimized mitogen stimulation induces proliferation of neoplastic B-cells in chronic lymphocytic leukemia: signifi-

Leukemia Surface antigen changes in CLL J Vilpo et al 1698 cance for cytogenetic analysis. Cytogenet Cell Genet 1998; 82: Cheson B (ed). Chronic Lymphoid Leukemias. Marcel Dekker: 215–221. New York, 2001, pp 231–260. 18 Vilpo J, Vilpo L. Short term growth potential of crypreserved B 30 Damle RN, Wasil T, Fais F, Ghiotto F, Valetto A, Allen SL, Buch- chronic lymphocytic leukemia cells in suspension and in semiso- binder A, Budman D, Dittmar K, Kolitz J, Lichtman SM, Schulman lid medium. Leukemia 2001; 15: 992–994. P, Vinciguerra VP, Rai KR, Ferrarini M, Chiorazzi N. Ig V gene 19 Vilpo J, Vilpo L, Hurme M, Vuorinen P. Induction of beta-2- mutation status and CD38 expression as novel prognostic indi- microglobulin release in vitro by chronic lymphocytic leukaemia cators in chronic lymphocytic leukemia. Blood 1999; 94: 1840– cells: relation to total protein synthesis. Leukemia Res 1999; 23: 1847. 913–920. 31 Ibrahim S, KeatingM, Do KA, O’Brien S, Huh YO, Jilani I, Lerner 20 Koski T, Karhu R, Visakorpi T, Vilpo L, Knuutila S, Vilpo J. Com- S, Kantarjian HM, Albitar M. CD38 expression as an important plex chromosomal aberrations in chronic lymphocytic leukemia prognostic factor in B-cell chronic lymphocytic leukemia. Blood are associated with cellular drugand irradiation resistance. Eur J 2001; 98: 181–186. Haematol 2000; 65: 32–39. 32 Orsini E, Foa R. Cytokines and regulatory molecules in the patho- 21 Vilpo J, Vilpo L, Hulkkonen J, Lankinen M, Kuusela P, Hurme genesis and clinical course of B-cell chronic lympohocytic leuke- M. Non-specific bindingcompromises the purification yields of mia. In: Cheson B (ed). Chronic Lymphoid Leukemias. Marcel leukemic B-cells in chronic lymphocytic leukemia: prevention by Dekker: New York, 2001, pp 127–159. collagen coating. Eur J Haematol 1998; 60: 65–67. 33 Younes A, Snell V, Consoli U, Clodi K, Zhao S, Palmer JL, Thomas 22 Decker T, Schneller F, Sparwasser T, Tretter T, Lipford GB, EK, Armitage RJ, Andreeff M. Elevated levels of biologically active Wagner H, Peschel C. Immunostimulatory CpG-oligonucleotides soluble CD40 ligand in the serum of patients with chronic lym- cause proliferation, cytokine production, and an immunogenic phocytic leukaemia. Br J Haematol 1998; 100: 135–141. phenotype in chronic lymphocytic leukemia B cells. Blood 2000; 34 Romano MF, Lamberti A, Tassone P, Alfinito F, Costantini S, Chiu- 95: 999–1006. razzi F, Defrance T, Bonelli P, Tuccillo F, Turco MC, Venuta S. 23 Goodman MG, Spinosa JC, Saven A, Piro LD, Wormsley S. New Triggering of CD40 antigen inhibits fludarabine-induced apoptosis perspectives on the approach to chronic lymphocytic leukemia. in B chronic lymphocytic leukemia cells. Blood 1998; 92:990– Leuk Lymphoma 1996; 22: 1–10. 995. 24 Kitada S, Zapata JM, Andreeff M, Reed JC. Bryostatin and CD40- 35 Hubl W, Iturraspe J, Braylan RC. FMC7 antigen expression on nor- ligand enhance apoptosis resistance and induce expression of cell mal and malignant B-cells can be predicted by expression of survival in B-cell chronic lymphocytic leukaemia. Br J CD20. Cytometry 1998; 34: 71–74. Haematol 1999; 106: 995–1004. 36 Serke S, Schwaner I, Yordanova M, Szczepek A, Huhn D. Mono- 25 Ghia P, Caligaris-Cappio F. The origin and nature of chronic lym- clonal antibody FMC7 detects a conformational epitope on the phocytic leukemia lymphocyte. In: Cheson B (ed). Chronic CD20 molecule: evidence from phenotypingafter rituxan therapy Lymphoid Leukemias. Marcel Dekker: New York, 2001, pp 63– and transfectant cell analyses. Cytometry 2001; 46: 98–104. 80. 37 Matutes E, Brito-Babapulle V, Yullie M, Catovsky D. Prolympho- 26 Lucio PJ, Faria MT, Pinto AM, da Silva MR, Correia Junior ME, da cytic leukemia of B- and T-cell types. In: Cheson B (ed). Chronic Costa RJ, Parreira AB. Expression of adhesion molecules in chronic Lymphoid Leukemias. Marcel Dekker: New York, 2001, B-cell lymphoproliferative disorders. Haematologica 1998; 83: pp 525–565. 104–111. 38 Binet J, Auquier A, Dighiero G, Chastang C, Piguet H, Goasguen 27 Marotta G, Raspadori D, Sestigiani C, Scalia G, Bigazzi C, Lauria J, Vaugier G, Potron G, Colona O, Oberling F, Thomas M, Tcher- F. Expression of the CD11c antigen in B-cell chronic lymphopro- nia G, Jacquillat C, Boivin P, Lesty C, Duault M, Monconduit M, liferative disorders. Leuk Lymphoma 2000; 37: 145–149. Belabbes S, Gremy F. A new prognostic classification of chronic 28 Sembries S, Pahl H, Stilgenbauer S, Do¨hner H, Schriever F. lymphocytic leukemia derived from a multivariate survival analy- Reduced expression of adhesion molecules and cell signaling sis. Cancer 1981; 48: 198–206. receptors by chronic lymphocytic leukemia cells with 11q 39 Bennett J, Catovsky D, Daniel M, Flandrin G, Galton D, Gralnick deletion. Blood 1999; 93: 624–631. H, Sultan C. Proposals for the classification of chronic (mature) B 29 Molica S. Prognostic factors in chronic lymphocytic leukemia. In: and T lymphoid leukaemias. J Clin Pathol 1989; 42: 567–584.

Leukemia