(2002) 16, 335–343  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Apoptosis induction by hypercross-linking of the surface antigen CD5 with anti-CD5 monoclonal antibodies in chronic lymphocytic leukemia DP Cioca and K Kitano

Second Department of Internal Medicine, Shinshu University School of Medicine, Matsumoto, Nagano, Japan

We evaluated cells from 24 patients with B cell chronic lympho- are secreted by malignant B cells and can sequester the MoAb cytic leukemia (B-CLL) to determine apoptosis induced by CD5 away from the malignant cells.11 These difficulties can be hypercross-linking. Following the CD5 hypercross-linking with anti-CD5 monoclonal antibodies (MoAbs), we identified 10 avoided by using MoAbs that recognize antigens more univer- patients where CD5 hypercross-linking induced apoptosis sally expressed on the surface of B cells. One such antigen is (group A) and 14 patients whose cells were resistant to the anti- the surface antigen CD20, against which a humanized anti- CD5 MoAbs (group B). The programmed cell death pathway of CD20 MoAb, Rituximab (Genentech, San Francisco, CA, USA; the cells from patient group A was caspase-3 and poly (ADP- IDEC Pharmaceuticals, San Diego, CA, USA) has been applied ribose) polymerase (PARP)-dependent, involved a reduction of 12 ⌬⌿ to B-CLL patients. However, clinical trials proved only a lim- the mitochondrial transmembrane potential and a down- 13 regulation of the anti-apoptotic Bcl-2, Mcl-1 and iNOS . ited activity of Rituximabin B-CLL. Another target antigen Early activation-associated molecules such as CD25 and CD69 is the surface molecule CD52, Campath-1H being a were expressed at higher levels than in controls after 6 h of humanized anti-CD52 MoAbwhich proved to beeffective in culture with anti-CD5 MoAb. The expression of CD5 and of patients with B-CLL.14 CD72, the ligand for CD5, were significantly lower in group A CD5 is a monomeric 67-kDa class I transmembrane glyco- compared with group B. Anti-CD20 MoAb had similar activity which belongs to the receptor cysteine-rich domain with anti-CD5 MoAb and the combination of the two MoAbs seemed to be additive. In this study, it is suggested that the family of polyanion binding receptors. The surface marker cells from some B-CLL patients can be induced into pro- CD5, which was primary described as a marker, was grammed cell death by CD5 hypercross-linking with anti-CD5 identified on certain B cell tumors and reported in a human MoAbs. leukemia study.15 Subsequently, it was identified on a subset Leukemia (2002) 16, 335–343. DOI: 10.1038/sj/leu/2393 of normal B cells16 designated as the B-1 subset and which is Keywords: apoptosis; B-CLL; CD5 hypercross-linking; caspase-3; found mostly in the peritoneal and pleural cavities.17 monoclonal antibodies Therefore, it seems logical to assess the effects of anti-CD5 MoAbs in B-CLL since expression of CD5 on tumor cell sur- face is a hallmark of the disease, and because previous studies Introduction have shown that CD5 ligation induces apoptosis in normal B cells without affecting T cells.18 B cell chronic lymphocytic leukemia (B-CLL) represents the In this study, we identified a group of B-CLL patients whose quintessential example of a human malignancy in which the cells are induced into programmed cell death by anti-CD5 clonal excess of B cells is caused principally by defects that MoAbs. Our findings may help correlate the clinical hetero- prevent cell turnover due to programmed cell death rather geneity of B-CLL with the developmental stages at which B than by alterations in cell cycle regulation. In the vast majority cells display different responses to various stimuli, and may of patients, B-CLL cells gradually accumulate not because provide a definition of different B-CLL subsets. they are dividing more rapidly than normal, but because they are surviving too long.1 Many factors have been described to determine this abnormal extended life span due to defective Materials and methods apoptosis of the B-CLL cells, among them bcl-2, bax and bcl- 2,3 xL gene over-expression, increased serum levels of inter- Patients feron gamma, interleukin-2, 4, 6 and 13,4 interleukin-8,5 free 6 7 iC3b(the ligand for beta2 integrins), and alpha interferon. The twenty-four patients (12 men and 12 women) enrolled in B-CLL cells have a down-regulated expression of the this study were randomly chosen at hospitals from Nagano, apoptosis-inducer CD95 (Fas), a disrupted CD95-dependent Yamanashi and Saitama prefectures, Japan. The mean age of 8 9 apoptotic pathway, and a perturbed T cell/B cell interaction. the patients was 66.5 years (median 70.5, range 27 to 88 In an attempt to counteract the defective apoptosis, mono- years). B-CLL was diagnosed according to standard clinical clonal antibodies (MoAbs) were used in order to trigger an and laboratory criteria, and at the time of inclusion in our intracellular antiproliferative or apoptotic signal upon binding study, seven patients were in stage 0, seven patients were in 10 with their cognate antigen. The early immunotherapy stage I, two patients were in stage II, two patients were in reagents used targeted the immunoglobulin variable region stage III, and six patients were in stage IV according to Rai (idiotype) of the malignant B cell. However, these antibody classification.19 Other characteristics of the B-CLL patients are reagents proved to be impractical as treatment modalities, summarized in Table 1. because being idiotype-specific, they have to be synthesized individually for each patient, and also because target idiotypes Cell preparation and culture Correspondence: K Kitano, the Second Department of Internal Medi- cine, Shinshu University School of Medicine, 3–1-1 Asahi, Matsu- After informed consent, heparinized samples of peripheral moto, Nagano-ken, 390–8621, Japan; Fax: 81-263-32-9412 blood were obtained from each patient. Peripheral blood Received 14 February 2001; accepted 9 November 2001 mononuclear cells were obtained from the interphase cell Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 336 Table 1 Characteristics of B-CLL patients

Patient Age/Sex Rai Disease WBC CD5/CD19 CD23 CD25 Previous No. stage duration (%) (%) (%) (%) analysis treatment (years)

1 73/F IV 2.3 68 200 91 91.1 86.7 24.3 ND — 2 71/M II 8.0 6810 66 89.4 87.9 44.8 46,XY,+12 CPM 3 78/M I 0.8 30 250 72 92.5 43.3 18.7 47,XY,+22 — 4 58/F IV 6.9 67 100 87 92.9 53.9 24.9 46,XX CPM 5 72/M I 0.2 40 580 53 86.5 34.5 13.0 46,XY,inv(12) — 6 27/M IV 0.4 23 800 92 94.8 69.3 2.4 46,XY — 7 55/M 0 4.0 23 100 83 92.0 72.7 49.6 46,XY — 8 71/M III 0.4 12 900 93 91.3 69.9 38.6 45,X,−Y CHOP 9 71/F IV 8.0 56 800 97 91.9 65.6 56.0 ND — 10 59/F I 4.5 20 400 79 94.9 46.8 31.6 ND CPM 11 64/F 0 9.2 29 500 87 73.5 31.8 12.6 ND CPM 12 45/M I 1.2 19 700 65 88.5 57.0 23.2 46,XY — 13 63/F I 2.9 31 100 81 93.1 95.7 36.8 46,XX,+12 — 14 72/M IV 6.3 15 400 91 82.4 49.4 20.3 ND — 15 66/M 0 2.0 17 590 56 91.9 38.5 25.5 46,XY PSL 16 49/M I 5.5 32 800 79 96.4 51.2 16.3 47,XY,t(1;13),inv(2),del(14) — 17 71/M I 10.1 35 700 79 89.5 61.2 24.5 ND — 18 88/F III 0.2 35 000 87 92.3 60.3 29.8 46,XX,inv(11)(p11;q21) — 19 76/F 0 6.1 31 100 87 92.1 37.6 11.3 ND — 20 45/F 0 0.8 42 500 85 93.3 67.7 41.4 46,XX — 21 72/F 0 8.2 55 660 86 97.9 49.8 32.6 46,XX CPM 22 70/F 0 1.4 13 450 68 96.1 38.6 25.5 46,XX — 23 71/F IV 7.2 13 440 75 94.1 69.4 11.2 ND — 24 70/M II 0.6 15 500 72 85.4 49.8 58.1 46,XY —

WBC, white blood cell count (×106/l); +12, trisomy 12; CPM, cyclophosphamide; CHOP, cyclophosphamide/doxorubicin/ vincristine/prednisolone; PSL, prednisolone; ND, not determined.

layer after Ficoll–Hypaque (Pharmacia, Uppsala, Sweden) Immunofluorescence antibodies and reagents density gradient centrifugation and washed twice in PBS. B- CLL cells were purified by negative selection executed by removal of T and NK cells using immunomagnetic anti-CD2 The murine MoAbs used in this study were phycoerythrin (PE)-

coated microbeads (Dynabeads M-450; Dynal, Oslo, conjugated anti-CD19 (HIB19, IgG1), fluorescein isothiocyan- Norway). Because anti-CD19 and anti-CD20 MoAbs were ate (FITC)-conjugated anti-CD23 (M-L233, IgG1), anti-CD24 20,21 reported to influence the apoptosis phenomenon, a posi- PE (ML5, IgG2), anti-CD25 FITC (M-A251, IgG2), anti-CD69 tive selection using microbeads linked to these antibodies was FITC (FN50, IgG1) and anti-CD72 FITC (J4–117, IgG2), pur- not considered suitable. The monocytes were removed from chased from PharMingen International, peridinin chlorophyll ° the cell suspensions by adherence for 2 h at 37 C to a plastic A protein-conjugated anti-CD19 (SJ25C1, IgG1) from Becton surface (Costar, Cambridge, MA, USA). Control flow cytome- Dickinson, and PE-conjugated anti-CD95 (DX2, IgG1) from try analysis performed on purified B-CLL cells showed that Dako (Glostrup, Denmark). For control staining, corres- + these cells comprised more than 95% CD19 B cells (mean ponding fluorescent isotypes were used: IgG1 from Becton ± 97.2 1.9%). Dickinson and IgG2 from Immunotech. The antibodies used The purified B-CLL cells were cultured at the concentration for intracellular staining were FITC-conjugated mouse anti- × 5 of 5 10 /ml for different periods of time in six-well plates human Bcl-2 oncoprotein (clone 124, IgG1) from Dako, PE- (BD Labware, Franklin Lakes, NJ, USA), in RPMI 1640 conjugated polyclonal rabbit anti-active caspase-3, anti- medium (Nikken BioMedical Laboratory, Kyoto, Japan) sup- human Bax and anti-human Mcl-1 from PharMingen, and the plemented with 10% fetal bovine serum, 2 mML-glutamine corresponding biotinylated swine anti-rabbit coupled with and 1% antibiotics (penicillin 100 IU/ml, 100 ␮g/ml strepto- RPE streptavidin from Dako. The FITC-conjugated anti-

mycin and fungizone, from Bio-Whittaker, Walkersville, MA, inducible nitric oxide synthase (iNOS) (clone 6, IgG2a)was USA). The plates were placed in a 5% CO2–95% air fully purchased from Transduction Laboratories (Lexington, KY, humidified atmosphere at 37°C and incubated until cell har- USA). The apoptosis marker Annexin-V-Fluos FITC-labeled vesting. For the surface hypercross-linking, the wells were pre- was purchased from Boehringer Mannheim, Mannheim, Ger- coated 12 h before at 4°C with 100 ␮g goat anti-mouse IgG many. The counterstaining dyes for cell viability used together (PharMingen International, Tokyo, Japan). B-CLL cells were with Annexin V were propidium iodide (PI) from Calbiochem cultured in the presence of anti-CD5 MoAbs (clone L17F12, (La Jolla, CA, USA), and 7-amino-actinomycin D (7-AAD)

IgG2, from Becton Dickinson Immunocytometry Systems, San from Molecular Probes (Eugene, OR, USA). The FITC-conju- Jose, CA, USA; and clone UCHT2, IgG2, from PharMingen gated rabbit polyclonal anti-PARP cleavage site (214/215) was International) and/or anti-CD20 (clone L27, IgG1, from Becton purchased from BioSource International (Camarillo, CA, USA). Dickinson) at 15 ␮g/ml, also in the presence of a correspond- The transmembrane potentiometric fluorescent marker 3,3Ј-

ing subclass of control antibody (IgG1 from Becton Dickinson, dihexyloxacarbocyanine iodide (DiOC6(3)) was purchased and IgG2 from Immunotech, Coulter, Marseille, France). from Molecular Probes.

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 337 Surface immunophenotype and intracellular flow two groups. The software used was StatView (Abacus cytometric analysis Concepts, Berkeley, CA, USA). Differences were considered statistically significant for P values Ͻ0.05. The Kolmogorov– Cells were extracted from their culture media after 6, 12 and Smirnov statistics were used to assess the flow cytometric 24 h, washed in PBS and stained at 4°C for 15 min with the histograms and the statistical significance of their shifting. appropriate amount of MoAb, washed again in PBS and sub- jected to flow cytometric analysis on a FACScan (Becton Dickinson, Mountain View, CA, USA) equipped with a Results 488 nm blue argon laser. On a forward scatter/side scatter (FSC/SSC) dot-plot, the cell population was electronically Induction of apoptosis by CD5 hypercross-linking gated, to avoid analysis of unwanted debris. 3 × 104 events with anti-CD5 MoAbs were analyzed each time, and results were plotted on logarith- mic scales. The percentage of cells expressing an antigen on After 24 h of culture in the presence of anti-CD5 MoAb, the their surface was calculated using Cell Quest (Becton determination of the apoptotic levels in the highly purified B- Dickinson). The relative level of antigen detected on the B- CLL cells revealed a significant increase of apoptosis in the CLL cell surface was quantified as the mean fluorescence cells from 10 patients, where the apoptotic levels increased intensity ratio (MFIR). MFIR was calculated according to the more than two times compared with the control. Therefore, following formula: the mean fluorescence intensity (MFI) of we divided the patients into two groups, A and B, according cells stained with a specific FITC- or PE-conjugated MoAbdiv- to the pro-apoptotic effect exerted by the anti-CD5 MoAbs ided by the MFI of control cells. For intracellular staining, cells (Table 2). The apoptotic level in group A, determined as phos- were fixed and permeabilized using the two-step IntraStain phatidylserine externalization, was 38.8 ± 11.0%, while that from Dako, according to the procedure recommended by in group B was 9.7 ± 4.9% (P Ͻ 0.01). The spontaneous apop- the manufacturer. totic levels were not different between the two groups: 10.4 ± 5.3% (group A) vs 10.2 ± 4.2% (group B). As shown in Figure 1a, apoptosis, as determined by pho- Apoptosis assays phatidylserine exposure on the outer cellular membrane, was induced on the CD5 hypercross-linked B-CLL cells from The Annexin V fluorescent assay was performed according to patient group A mainly between 6 and 12 h of culture, so sig- the manufacturer’s instructions. Cells were extracted from the nificant increases in apoptotic levels could be detected after culture media and washed with PBS, then 106 cells were 12 h of culture. However, the apoptosis-induction timing was stained with 2 ␮l Annexin V FITC in the incubation buffer not homogeneous, because the cells from one patient were

(10 mM Hepes/NaOH, pH 7.4, 140 mM NaCl, 5 mM CaCl2). induced into apoptosis before 6 h, and also in the cells from In order to discriminate between apoptotic cells and cells another patient apoptosis was induced later, between 12 and which have permeabilized their membranes (dead cells), a 24 h after the hypercross-linking. In the CD5 hypercross- counterstain using 5 ␮g/ml PI or 7-AAD was used. 3 × 104 linked cells from patient group B, apoptosis induction did not events from the double-gated area (gate 1 based on FSC/SSC take place at all (Figure 1b). A second anti-CD5 MoAb used, pattern and gate 2 for CD5+ cells) were analyzed each the clone UCHT2, which recognizes a different CD5 epitope time, results being plotted on a Annexin V/PI logarithmic (demonstrated by blocking experiments where binding of one scattergram. MoAb did not prevent binding of the other MoAb) produced

The DiOC6(3) fluorochrome assay, which reveals the dis- results almost identical with L17F12, indicating that pro- ruption of the mitochondrial transmembrane potential (⌬⌿m), grammed cell death induction was not dependent on the epi- an early and irreversible step of ongoing tope recognized by the anti-CD5 antibody. In the absence of apoptosis, was performed according to Zamzami et al.22 106 the goat anti-mouse hypercross-linker antibody, apoptosis ° cells were incubated with 40 nM DiOC6(3) for 15 min at 37 C, induction was weaker (data not shown), similar to results followed by flow cytometric analysis. In this assay, the apop- obtained with IgM20 and CD20.21 totic cells are identified by their decreased ⌬⌿m These differential apoptotic rates were repetitively repro- low (DiOC6(3) ). duced, the B-CLL cells being obtained twice from the patients The PARP fluorescent assay was performed by fixing and over a period of a few months, demonstrating that the two permeabilizing 106 cells, then staining for 30 min at 4°C with distinct B-CLL populations are maintaining their different pro- 10 ␮l of FITC-conjugated anti-PARP cleavage site (214/215) apoptotic potentials over time. antibody. The caspase-3 fluorescent assay was performed according to the manufacturer’s indications. 106 cells were permeabil- The apoptotic pathway of anti-CD5 MoAb ized and then stained for 20 min at 4°C with 20 ␮l of the PE-conjugated anti-active caspase-3 MoAb, which recognizes As shown in Figure 2a and Table 3, experiments performed on active human caspase-3, as opposed to pro-caspase-3. The B-CLL cells from eight group A patients after CD5 hypercross- following flow cytometric analysis could discriminate linking revealed increased percentages of cells having detect- between cells which have or not exposed the conformational able levels of active caspase-3, compared with the control epitope formed by the cleavage of the pro-caspase-3. samples (4.6 ± 3.8% control vs 28.9 ± 6.7% in CD5 hypercross-linked cells, P Ͻ 0.01). Thus, the anti-CD5 programmed cell death induction pathway involved the Statistical analysis activation of one essential executioner caspase, namely the caspase-3. The paired Student’s t-test was used to determine the statistical Poly (ADP-ribose) polymerase cleavage is considered to be significance of the data obtained and to compare means of one of the classic characteristics of apoptosis. PARP is a 116-

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 338 Table 2 Apoptosis induction by anti-CD5 MoAb

Patient No. Apoptosis (%) Absolute Apoptosis ratio apoptosis increasea anti-CD5 MoAb/control Control Anti-CD5 MoAb

Group A 1 7.2 29.1 21.9 4.04 2 7.4 40.0 32.9 5.40 7 7.1 36.0 28.9 5.07 8 9.1 46.7 37.9 5.13 9 21.2 52.3 31.1 2.46 10 3.8 19.9 16.1 5.23 13 13.2 30.4 17.2 2.30 17 7.2 33.6 26.2 4.66 20 16.7 54.0 37.3 3.23 24 11.0 45.9 34.9 4.17 Group B 3 8.9 11.9 3.0 1.33 4 17.6 17.2 −0.2 0.97 5 11.3 7.6 −3.7 0.67 6 18.1 18.9 0.8 1.04 11 10.4 6.3 −4.1 0.60 12 13.1 13.5 0.4 1.03 14 8.7 12.3 3.8 1.41 15 5.8 3.6 −2.2 0.62 16 3.6 3.3 −0.3 0.91 18 10.1 8.7 −1.4 0.86 19 9.1 10.0 0.9 1.09 21 11.7 10.2 −1.5 0.87 22 9.6 9.8 0.2 1.02 23 4.6 2.6 −2.0 0.57

aAbsolute apoptosis increase is calculated as follows: apoptosis (%) by anti-CD5 Moab minus apoptosis (%) of the control.

kDa nuclear protein which is strongly activated by DNA in intensity at 24 h, and the resulting percentages of apoptotic strand breaks. During apoptosis, caspase-3 and -7 cleave B-CLL cells were higher when the monoclonal antibodies PARP to yield an 85-kDa and a 25-kDa fragment.23 As shown were used together (Figure 3). The mean ± s.d. apoptotic levels in Figure 2band Table3, significant increases of the 85-kDa at 24 h in the seven cases tested were 10.08 ± 6.32% (control), fragment of cleaved PARP, as a hallmark of apoptosis induc- 37.83 ± 12.20% (anti-CD5), 29.66 ± 8.98% (anti-CD20), and tion in anti-CD5 hypercross-linked B-CLL cells, were detected 59.37 ± 11.90% (anti-CD5 + anti-CD20). Thus, it is suggested in the cells from eight group A patients tested (8.1 ± 4.0% that the pro-apoptotic effects of anti-CD5 and anti-CD20 control vs 34.9 ± 9.7% CD5 hypercross-linked, P Ͻ 0.01). MoAbs are additive. In the B-CLL cells from patient group B, Another finding was that the mitochondrial transmembrane where anti-CD5 MoAbs were ineffective, anti-CD20 MoAb- potential ⌬⌿m, which has been shown to constitute an early induced apoptosis still appeared. Moreover, in the two cases and irreversible step of programmed lymphocyte death,22 was from group A in which CD20 hypercross-linking did not pro- already reduced at 6 h after CD5 hypercross-linking. As duced apoptosis induction, anti-CD5 MoAbwas ableto do shown in Figure 3 and Table 3, using the potentiometric this (Figure 4), suggesting that the mechanism of induction of

marker DiOC6(3), the percentage of cells displaying a reduced programmed cell death may be different from that used by ⌬⌿m significantly increased in the CD5 hypercross-linked anti-CD20 MoAb. samples in the eight group A patients tested (14.7 ± 4.6% con- trol vs 43.1 ± 9.5% CD5 hypercross-linked, P Ͻ 0.01). There- fore, another important step of the apoptotic pathway initiated Modulation of the apoptosis-related proteins after by CD5 cross-linking seems to involve the mitochondria. CD5 hypercross-linking Nothing significant happened with caspase-3, PARP and

DiOC6(3) levels in the cells from the 11 group B patients Members of the Bcl-2 family have been demonstrated to be tested (Table 3). directly related to apoptosis induction propensity or resist- ance.24 Therefore, we examined the endogenous levels of Bcl- 2, Bax and Mcl-1 protein expression in B-CLL cells after CD5 The effect of anti-CD20 MoAb on B-CLL cells hypercross-linking with anti-CD5 MoAbs in the B-CLL cells of eight group A patients and 11 group B patients. As shown in Anti-CD20 antibodies have demonstrated apoptosis-inducing Figure 5a, band c, Bcl-2 and Mcl-1 protein expression suf- effects on B-CLL cells,21 therefore we compared the activities fered a significant decrease after CD5 hypercross-linking in B- of anti-CD5 and anti-CD20 MoAbs (clone L27). In our hands, CLL cells from patient group A (Bcl-2 MFIR 26.9 ± 7.7 control clone L27 had a clear in vitro apoptosis-inducing effect in five vs 12.5 ± 6.6 in CD5 hypercross-linked, P Ͻ 0.01, Mcl-1 MFIR of the seven patient group A cases tested. The pro-apoptotic 27.7 ± 7.4 control vs 11.5 ± 5.4, P Ͻ 0.01), while Bax protein effects of anti-CD5 and anti-CD20 MoAbs were comparable expression was significantly increased (MFIR 14.9 ± 10.3 con-

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 339

Figure 2 Characterization of anti-CD5 induced apoptosis. B-CLL cells were cultured for 24 h in the presence of an anti-CD5 MoAbor an isotype-matched mouse Ig (control samples). After the fixation and permeabilization procedure, cells were stained with PE-conjugated anti-active caspase-3 (a), and with FITC-conjugated anti-PARP cleav- age site (214/215)-specific antibody (b). Dotted lines represent the fluorescent-labeled isotype-matched control antibodies, solid lines are the samples examined. A representative FSC/SSC pattern shifting is depicted in (c). Representative for the eight group A patients tested.

Expression of activation-associated molecules following CD5 hypercross-linking

B-CLL cells from both groups A and B were analyzed to verify modulation of CD25, CD69, CD23, CD72, CD95 and CD24 surface antigens after 6 h of culture in the absence (control) or presence of anti-CD5 MoAbs. Following CD5 hypercross- linking, only CD25 and CD69 early activation markers suf- fered a significant up-regulation in both groups. The average Figure 1 Apoptosis in B-CLL cells hypercross-linked with anti- percentage of CD25-expressing cells increased to 53.7 ± CD5 MoAb. B-CLL cells were cultured for different periods in the pres- 15.0% (control 27.9 ± 15.0%, P Ͻ 0.01), and the average ence of anti-CD5 MoAb(clone L17F12) or an isotype-matched mouse percentage of CD69-expressing cells increased to 30.3 ± Ig (control samples), at the concentration of 15 ␮g/ml, strongly hyp- ± Ͻ ercross-linked with goat anti-mouse Ab. After 6, 12 and 24 h, the 16.3% (control 14.7 11.9%, P 0.01). Mean fluorescence apoptotic percentages of the cultured cells were measured by Annexin intensities of CD25 and CD69 suffered a consequent increase V FITC/PI staining and flow cytometric analysis. Data represent the following CD5 hypercross-linking, the CD25 and CD69 histo- mean ± s.d. of the values obtained in the 10 patients with B-CLL from grams being significantly shifted to the right (CD25 MFIR 32.2 group A (a) and the 14 patients from group B (b). ± 6.5 vs 16.7 ± 7.6 control, P Ͻ 0.05, CD69 MFIR 20.9 ± 8.4 vs 11.5 ± 5.7 control, P Ͻ 0.05). Nothing significant happened either with the percentages of CD23-, CD72-, CD95- and trol vs 29.5 ± 6.5, P Ͻ 0.01). Because a functional inducible CD24-expressing cells, or with the mean fluorescence inten- nitric oxide synthase displaying anti-apoptotic activity was sities of these surface antigens on B-CLL cells upon CD5 demonstrated in B-CLL and ,25–27 we also hypercross-linking. examined iNOS expression after CD5 hypercross-linking in eight group A and 11 group B patients, and found a significant decrease of its levels in the B-CLL cells of group A patients Analysis of group characteristics (MFIR 10.8 ± 3.4 control vs 5.1 ± 2.0, P Ͻ 0.01), as shown in Figure 5d. Nothing significant happened with the Statistical analysis of the clinical and biological characteristics expression of these apoptosis-related proteins in the cells from of patient groups A and B revealed a few significant differ- the 11 group B patients tested. ences between them. One of the most important findings was

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 340 Table 3 Summary of differences in apoptotic markers and surface phenotype between group A and B

Control Anti-CD5

Group A Group B P value Group A Group B P value

Annexin V (%) 10.4 ± 5.3 10.2 ± 4.2 NS 38.8 ± 11.0 9.7 ± 4.9 Ͻ0.01 Caspase-3 (%) 4.6 ± 3.8 4.6 ± 2.5 NS 28.9 ± 6.7 4.6 ± 2.6 Ͻ0.01 PARP (%) 8.1 ± 4.0 8.1 ± 3.6 NS 34.9 ± 9.7 8.1 ± 3.9 Ͻ0.01 ± ± ± ± Ͻ DiOC6(3) (%) 14.7 4.6 14.4 4.8 NS 43.1 9.5 14.2 5.0 0.01 Bcl-2 MFIR 26.9 ± 7.7 27.8 ± 6.8 NS 12.5 ± 6.6 28.1 ± 8.1 Ͻ0.01 Bax MFIR 14.9 ± 6.5 14.3 ± 6.3 NS 29.5 ± 10.3 13.9 ± 6.7 Ͻ0.01 Mcl-1 MFIR 27.7 ± 7.0 27.6 ± 7.4 NS 11.5 ± 5.4 28.8 ± 8.0 Ͻ0.01 iNOS MFIR 10.8 ± 3.4 10.7 ± 3.6 NS 5.1 ± 2.0 11.0 ± 3.6 Ͻ0.01 CD25 (%) 40.6 ± 11.7 19.6 ± 9.3 Ͻ0.01 CD23 (%) 70.4 ± 16.0 48.9 ± 12.2 Ͻ0.01 CD5 MFIR 35.0 ± 7.5 44.0 ± 8.8 Ͻ0.05 CD72 MFIR 3.7 ± 1.3 7.3 ± 2.6 Ͻ0.01

MFIR, mean fluorescence intensity ratio; NS, statistically not significant.

Figure 4 The additive effect of anti-CD5 and anti-CD20 MoAbs. B-CLL cells from seven group A patients were cultured in the absence Figure 3 Representative flow cytometric result for anti-CD5 and (control samples) or presence of anti-CD5 (clone L17F12), anti-CD20 anti-CD20 hypercross-linked B-CLL cells. B-CLL cells were cultured (clone L27), or anti-CD5 plus anti-CD20 MoAbs, strongly cross-linked in the presence of an isotype-matched mouse MoAbplus goat anti- with goat anti-mouse Ab. After 24 h the percentage of apoptotic cells mouseAb(a), anti-CD5 (clone L17F12) plus goat anti-mouse Ab(b), in each cell population was quantified by the Annexin V FITC method, Ͻ anti-CD20 (clone L27) plus goat anti-mouse Ab(c), or anti-CD5 plus as described in Materials and methods. *Indicates P 0.05. anti-CD20 plus goat anti-mouse Ab(d). The typical flow cytometric histograms illustrated here were obtained by analyzing the above- named four populations of cells stained with the mitochondrial poten- tial marker DiOC6(3). The percentages shown represent the percent- ± ± ages of apoptotic cells displaying a reduced mitochondrial transmem- group B compared with group A (MFIR 7.3 2.6 vs 3.7 1.3, + + ⌬⌿ low = Ͻ brane potential ( m ). FL1 log DiOC6(3). P 0.01). At the same percentage of CD5 CD19 cells (91.2 ± 2.6% group A vs 90.9 ± 6.5% group B), the CD5 MFIR was significantly higher on B-CLL cells from patient group B than that the B-CLL cells from patient group A had significantly the CD5 MFIR of B-CLL cells from patient group A (44.0 ± 8.8 higher expression of CD25 and CD23 than those from patient vs 35.0 ± 7.5, P Ͻ 0.05). Chromosomal analysis showed that group B, both on percentages of CD25- and CD23-expressing two patients from group A had trisomy 12, but none from cells (40.6 ± 11.7% vs 19.6 ± 9.3% for CD25, P Ͻ 0.05, and group B presented this abnormality. The differences between 70.4 ± 16.0% vs 48.9 ± 12.2% for CD23, P Ͻ 0.05), also on the two patient groups regarding age (64.9 ± 9.3 years group MFIR (23.9 ± 5.7 vs 12.3 ± 5.4 for CD25, P Ͻ 0.05, and 56.8 A vs 64.9 ± 15.6 years group B), WBC counts (31301 ± 19807 ± 16.4 vs 35.9 ± 10.2 for CD23, P Ͻ 0.05). Another significant × 106/l group A vs 30384 ± 15784 × 106/l group B), lympho- finding was that at the same percentage of CD72+ cells (97.4 cyte percentages ( 82.6 ± 9.5% group A vs 77.5 ± 12.9% ± 2.3% group A and 97.1 ± 1.8% group B), the CD72 mean group B) and Rai stage (1.8 ± 1.5 group A vs 1.6 ± 1.7 group fluorescence intensity was higher on the B-CLL cells from B) were under the threshold of statistical significance.

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 341

Figure 5 Modulation of apoptosis-related proteins. After 24 h of culture in the presence of anti-CD5 MoAbs or an isotype-matched mouse Ig (control samples), B-CLL cells were fixed and permeabilized as stated in Materials and methods, then stained with FITC-conjugated anti-Bcl- 2 (a), PE-conjugated anti-Bax (b) and anti-Mcl-1 (c), and FITC-conjugated anti-iNOS (d). The dotted lines represent the fluorescent-labeled isotype-matched controls, the solid lines are the samples examined. Representative for the eight group A patients tested.

Discussion distribution between groups A and B matches the previous findings of Gibson et al37 and Barnett and Reilly,38 that B- Our results indicating that CD5 hypercross-linking with anti- CLL patients with more than 50% CD23-positive cells in the CD5 MoAbs results in programmed cell death induction in B- peripheral blood have a significantly higher mean level of CLL cells from a group of patients are at variance with those CD25-positive cells (13.8 ± 10.7% Gibson et al,49± 29% of Cantor et al,28 who described an anti-apoptotic effect of Barnett and Reilly, 40.6 ± 11.9% in patient group A in our CD5 cross-linking in bovine CD5+ B cells and no effect in present report) than do patients with less than 50% CD23- CD5+ B cells from animals with persistent (a positive cells (5.1 ± 4.2% Gibson et al,23± 20% Barnett and non-neoplastic expansion of CD5+ B cell population induced Reilly, 19.6 ± 9.3% in patient group B in our present report). by bovine leukemia virus). However, CD5+ cells from persist- The higher activation state of the B-CLL cells from patient ently lymphocytotic cattle are different from human B-CLL group A also matches some previous observations that the CD5+ cells, and also human normal CD5+ cells have a differ- state of activation of malignant B-1 cells was correlated with ent behavior compared with bovine CD5+ B cells, CD5 lig- increased apoptosis.39 ation resulting in apoptosis induction in human CD5+ B The second significant difference found by us between B- cells.18 CLL cells from patient groups A and B concerned the CD5 The cross-linking of the CD5 antigen by immobilized anti- and CD72 MFIR values, which were found to be significantly CD5 MoAbs on human T cells was shown to induce func- higher on cells from patient group B. CD72 has been defined tional interleukin-2 receptors,29 and also ligation of the CD5 as the CD5 ligand on the B cell membrane,40 and its co-modu- molecule with anti-CD5 MoAbon human T cells was demon- lation with CD5 has been demonstrated.41 One hypothesis strated to induce expression of the early activation antigen which could explain the decreased CD72 MFIR values is that CD69.30 Moreover, on B-CLL cells, Cerutti et al31 demon- reduced surface CD72 expression follows by co-modulation strated that anti-CD5 MoAbs significantly up-regulate ␣ CD5 MFIR decrease upon activation. Indeed, Jamin et al42 (CD25), ␤, and ␥ subunits of interleukin-2 receptor, and also found a decline of CD5 MFI on activated T cells which up-regulate the early activation antigen CD69.31 Our results expressed CD25. Another argument sustaining this hypothesis confirm these findings, the increases in percentages of acti- is the finding that on B cells, the occupancy of CD72 vated CD25- and CD69-positive cells being complementary enhances CD23 expression,43 thus suggesting that on B-CLL with the increases in surface expression. cells from patient group A, the CD72 surface antigen is bound Programmed cell death induction in B-CLL cells following to a counterstructure and subsequently its expression is down- cell activation resembles the activation-induced cell death regulated. The significance of the lower CD5 and CD72 MFIR (AICD) of T cells. The term AICD was initially used by Ashwell on B-CLL cells from patient group A could indicate a higher et al32 to describe the T cell receptor-induced death of T cell level of CD5–CD72 interaction, with reciprocal down- hybridomas, and later extended to nontransformed human T modulation. lymphocytes33 and splenic T cells.34 The term AICD was also Another potential ligand identified for the CD5 antigen is 44 applied to describe B cell receptor-mediated apoptosis on ton- the B cell surface immunoglobulin VH framework region. sillar B cells and Burkitt cell lines,35 and the death CD5 was identified to be associated with the human B cell of normal B cells promoted by hypercross-linking of their receptor complex45 and to negatively regulate antigen recep- surface immunoglobulins.36 tor-mediated signaling. CD5 appears to regulate the early sig- The two main significant differences found between B-CLL naling events induced by B cell receptor cross-linking in B-1 cells from patient groups A and B were the higher percentages cells46 by setting a threshold which may play a role in protect- of CD25- and CD23-positive cells in group A and the dimin- ing cells from antigen receptor firing either spontaneously or ished CD5 and CD72 MFIR (indicating a lower CD5 and by weak cross-reactive interactions with antigens. In this CD72 surface expression) on cells from patient group A, com- respect, our finding of a higher activation state of B-CLL cells pared with patient group B cells. The CD23 and CD25 antigen from patient group A suggests that indeed the CD5 hypercross-

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 342 linking removes the CD5 threshold of negative regulation on dichotomy and to establish possible B-CLL therapeutic B cell receptor and the activated cells become able to enter approaches using anti-CD5 MoAbs. programmed cell death. It is also suggested that the regional interaction of surface immunoglobulin with CD5 could pro- vide a autostimulatory signal necessary for B cell survival.47 Acknowledgements This weak but persistent survival signal provided to the B-CLL cell by the continual ‘tickling’ of CD5 could be disrupted by We thank Professor Kendo Kiyosawa for valuable advice and anti-CD5 MoAbs, which have a higher relative avidity,44 thus Dr Fumihiro Ishida for constructive discussion. We also thank opening the pathway of programmed cell death. Death signals Mr Susumu Ito for technical assistance and Drs Mayumi Ueno, triggered by B cell receptor ligation are dependent on caspase- Shigetaka Shimodaira, Shojiro Takagi, Naohiko Chiba, Hiroshi 3,48 pointing to our finding of caspase-3 involvement in anti- Morishita, Shigeki Seki, Naoaki Ichikawa, Toshiro Ito, Hikaru CD5-induced apoptosis. Moreover, we describe the preceding Kobayashi and Hiroshi Saito for providing patient samples. step of the death pathway in anti-CD5-induced apoptosis, the disruption of the mitochondrial potential, which was also 49 found in B cell receptor-induced apoptosis. All these support References the hypothesis that anti-CD5 MoAbs disrupt the weak but essential CD5–Ig framework interaction and induce pro- 1 Reed JC. Molecular biology of chronic lymphocytic leukemia. grammed cell death via the same mitochondria–caspase path- Semin Oncol 1998; 25: 11–18. way. 2 Gottardi D, Alfarano A, De Leo M, Stacchini A, Aragno M, Rigo The findings that cells from two of our group A B-CLL A, Veneri D, Zanotti R, Pizzolo G, Caligaris-Cappio F. In leu- kaemic CD5+ B cells the expression of Bcl-2 gene family is shifted patients were induced into programmed cell death by anti- toward protection from apoptosis. Br J Haematol 1996; 94: CD5 but not by anti-CD20 MoAb and that anti-CD20 MoAb 612–618. could induce apoptosis in the cells from group B patients sug- 3 Klein A, Miera O, Bauer O, Golfier S, Schriever F. Chemosensiti- gest different apoptosis induction pathways. Signaling through vity of B cell chronic lymphocytic leukemia and correlated CD72 was described as having an anti-apoptotic activity in B expression of proteins regulating apoptosis, cell cycle and DNA cells, the anti-immunoglobulin-induced death of mouse repair. Leukemia 2000; 14: 40–46. 50 4 Tangye SG, Raison RL. Human cytokines supress apoptosis of leu- spleen B cells being blocked by CD72 signaling. However, kaemic CD5+ cells and preserve expression of bcl-2. Immunol important differences exist between mouse spleen B cells and Cell Biol 1997; 75: 127–135. the human B-CLL cells we used in our experiments. One 5 Molica S, Vitelli G, Levato D, Levato L, Datillo A, Gandolfo GM. hypothesis which could explain apoptosis induction by CD5 Clinico–biological implications of increased serum levels of hypercross-linking is that the anti-CD5 MoAbs could disrupt interleukin-8 in B-cell chronic lymphocytic leukemia. Haema- the CD5–CD72 interaction and promote cell death. If this is tology 1999; 84: 208–211. 6 Plate JM, Long BW, Kelkar SB. Role of beta2 integrins in the pre- the case, it would explain our finding of selective apoptosis vention of apoptosis induction in chronic lymphocytic leukemia induction in the B-CLL group, in which both CD5 and CD72 B cells. Leukemia 2000; 14: 34–39. had lower detectable expression, possibly because of a higher 7 Mainou-Fowler T, Prentice AG. Modulation of apoptosis with interaction level with its counterstructure. It seems possible cytokines in B-cell chronic lymphocytic leukemia. Leuk Lym- that in our extensive cross-linking conditions, where a supra- phoma 1996; 21: 369–377. optimal signal was delivered, both the CD5-IgV framework 8 Laytragoon-Lewin N, Duhony E, Bai XF, Mellstedt H. Downregul- H ation of the CD95 receptor and defect CD40-mediated signal region and CD5–CD72 interactions are perturbed, resulting in transduction in B-chronic lymphocytic leukemia cells. Eur J programmed cell death induction. Haematol 1998; 61: 266–271. The findings that cells from two of our group A B-CLL 9 Kneitz C, Goller M, Wilhelm M, Mehringer C, Wohlleben G, patients were induced into programmed cell death by anti- Schimpl A, Tony HP. Inhibition of T cell/B cell interaction by B- CD5 but not by anti-CD20 MoAb and that anti-CD20 MoAb CLL cells. Leukemia 1999; 13: 98–104. could induce apoptosis in the cells from group B patients sug- 10 Multani PS, Grossbard ML. -based therapies for hematologic malignancies. J Clin Oncol 1998; 16: 3691–3710. gest different apoptosis induction pathways and have potential 11 Meeker TC, Lowder J, Maloney DG, Miller R, Thielelmans K, therapeutic implications, considering that a subgroup of B-CLL Warnke R, Levy R. A clinical trial of anti-idiotype therapy for B patients may be sensitive to anti-CD5 MoAb therapy even if cell malignancy. Blood 1985; 73: 1349–1363. the anti-CD20 therapy is ineffective. Finally, regarding the 12 McLaughlin P, White CA, Grillo-Lopez AJ. Clinical status and opti- additive pro-apoptotic effect of anti-CD5 and anti-CD20 mal use of Rituximabfor B-cell . Oncology 1998; 12: MoAbs, the possibility of a combined MoAb therapy has to 1769–1777. 13 Nguyen DT, Amess JA, Doughty H, Hendry L, Diamond LW. be further considered. IDEC-C2B8 anti-CD20 (rituximab) immunotherapy in patients In conclusion, the results of this study demonstrate the with low-grade non-Hodgkin’s lymphoma and lymphoprolifer- existence of two different groups of B-CLL patients, whose ative disorders: evaluation of response on 48 patients. Eur J leukemic cells have a different behavior after the extensive Haematol 1999; 62: 76–82. cross-linking of their surface antigen CD5 with anti-CD5 14 Osterborg A, Fassas AS, Anagnostopoulos A, Dyer MS, Catovsky MoAbs. Cells from both patient groups suffered an early acti- D, Mellstedt H. Humanized CD52 monoclonal antibody Cam- path-1H as first-line treatment in chronic lymphocytic leukemia. vation phenomenon, and following that, cells from patient Br J Haematol 1996; 93: 151–153. group A entered programmed cell death. The apoptosis induc- 15 Boumsell L, Bernard A, Lepage V, Degos L, Lemerle J, Dausset J. tion was caspase-3 and PARP-dependent and involved down- Some chronic lymphocytic leukemia cell bearing surface immuno- regulation of the anti-apoptotic Bcl-2, Mcl-1 and iNOS pro- globulins share determinants with T cells. Eur J Immunol 1978; 8: teins, together with an up-regulation of the pro-apoptotic Bax 900–904. protein. Further research is necessary to define the molecular 16 Caligaris-Cappio F, Gobbi M, Bofill M, Janossy G. Infrequent nor- mal B lymphocytes express features of B-chronic lymphocytic leu- mechanisms of apoptosis induction after CD5 hypercross-link- kemia. J Exp Med 1982; 155: 623–628. ing, the specific features which distinguish the two B-CLL 17 Kantor AB, Herzenberg LA. Origin of murine B cell lineages. Annu groups, to clarify the pathophysiological significance of this Rev Immunol 1993; 11: 501–539.

Leukemia Apoptosis by anti-CD5 in CLL DP Cioca and K Kitano 343 18 Pers J-O, Jamin C, Le Corre R, Lydyard PM, Youinou P. Ligation 34 Radvanyi LG, Mills GB, Miller RG. Religation of the T cell receptor of CD5 on resting B cells, but not on resting T cells, results in after primary activation of mature T cells inhibits proliferation and apoptosis. Eur J Immunol 1998; 28: 4170–4176. induces apoptotic cell death. J Immunol 1993; 150: 5704–5715. 19 Rai KR, Sawitsky A, Cronkite EP, Chanana AD, Levy RN, Pasternak 35 Daniel PT, Oettinger U, Mapara MY, Bommert K, Bargou R, BS. Clinical staging of chronic lymphocytic leukemia. Blood 1975; Dorken B. Activation and activation-induced death of human ton- 46: 219–234. sillar B cells and cells: lack of CD95 (Fas/APO- 20 Chaouchi N, Vazquez A, Galanaud P, Leprince C. B cell antigen 1) ligand expression and function. Eur J Immunol 1997; 27: receptor-mediated apoptosis. Importance of accessory molecules 1029–1034. CD19 and CD22, and of surface IgM cross-linking. J Immunol 36 Parry SL, Hasbold J, Holman M, Klaus GGB. Hypercrosslinking 1995; 54: 3096–3104. surface IgM or IgD receptors on mature B cells induces apoptosis 21 Shan D, Ledbetter JA, Press OW. Apoptosis of malignant B cells that is reversed by costimulation with IL-4 and anti-CD40. J Immu- by ligation of CD20 with monoclonal antibodies. Blood 1998; 91: nol 1994; 152: 2821–2829. 1644–1652. 37 Gibson J, Neville S, Joshua D, Kronenberg H. CD23 expression 22 Zamzami N, Marchetti P, Castedo M, Zanin C, Vayssiere JC, Petit in CLL. Br J Haematol 1989; 72: 598. PX, Kroemer G. Reduction in mitochondrial potential constitutes 38 Barnett D, Reilly JT. Lack of corelation between cell surface acti- an early irreversible step of programmed lymphocyte death in vation antigen expression and clinical stage in CLL. Br J Haematol vivo. J Exp Med 1995; 181: 1661–1672. 1989; 73: 572. 23 Kaufmann SH, Desnoyers S, Ottaviano Y, Davidson NE, Poirier 39 Peng B, Raveche E. Apoptosis induction in CD5+ (Ly1+) malignant GG. Specific proteolytic cleavage of poly (ADP-ribose) poly- B cells. Leukemia 1993; 7: 789–794. merase: an early marker of chemotherapy-induced apoptosis. Can- 40 Van de Velde H, von Koegen I, Luo W, Parnes JR, Thielemans K. cer Res 1993; 53: 3976–3985. The B-cell surface protein CD72/Lyb-2 is the ligand for CD5. Nat- 24 Yang E, Korsmeyer SJ. Molecular thanatopsis: a discourse on the ure 1991; 351: 662–665. BCL2 family and cell death. Blood 1996; 88: 386–401. 41 Jamin C, Le Corre R, Pers JO, Dueymes M, Lydyard PM, Youinou 25 Zhao H, Dugas N, Mathiot C, Delmer A, Dugas B, Sigaux F, Kolb P. Modulation of CD72 by ligation of B cell receptor complex J-P. B-cell chronic lymphocytic leukemia cells express a functional molecules on CD5+ cells. Int Immunol 1997; 9: 1001–1009. inducible nitric oxide synthase displaying anti-apoptotic activity. 42 Jamin C, Lamour A, Pennec L, Hirn M, Le Goff P, Youinou P. Blood 1998; 92: 1031–1043. Expression of CD5 and CD72 on T and B cell subsets in rheuma- 26 KolbJP. Mechanisms involved in the pro- and anti-apoptotic role toid arthritis and Sjogren’s syndrome. Clin Exp Immunol 1993; 92: of NO in human leukemia. Leukemia 2000; 14: 1685–1694. 245–250. 27 Roman V, Zhao H, Fourneau JM, Marconi A, Dugas N, Dugas B, 43 Katira A, Kamal M, Gordon J. Occupancy of CD72 (the CD5 Sigaux F, KolbJP. Expression of a functional induciblenitric oxide counterstructure) enhances interleukin-4-dependent CD23 synthase in hairy cell leukaemia and ESKOL cell line. Leukemia expression in resting B lymphocytes. Immunology 1992; 76: 2000; 14: 696–705. 422–426. 28 Cantor GH, Pritchard SM, Dequiedt F, Willems L, Kettmann R, 44 Pospisil R, Fitts MG, Mage RG. CD5 is a potential selecting ligand Davis WC. CD5 is dissociated from the B-cell receptor in B cells for B cell surface immunoglobulin framework region sequences. from bovine leukemia virus-infected, persistently lymphocytotic J Exp Med 1996; 184: 1279–1284. cattle: consequences to B-cell receptor-mediated apoptosis. J Virol 2001; 75: 1689–1696. 45 Lankester AC, van Schijndel GM, Cordell JL, van Noesel CJ, van 29 Verwilghen J, Vandesande R, Vandenberghe P, Ceuppens JL. Lier RA. CD5 is associated with the human B cell antigen receptor Crosslinking of the CD5 antigen on human T cells induces func- complex. Eur J Immunol 1994; 24: 812–816. tional IL2 receptors. Cell Immunol 1990; 131: 109–119. 46 Sen G, Bikah G, Venkataraman C, Bondada S. Negative regulation 30 Vandenberghe P, Verwilghen J, Van Vaeck F, Ceuppens JL. Lig- of antigen receptor-mediated signaling by constitutive association ation of the CD5 or CD28 molecules on resting human T cells of CD5 with the SHP-1 protein tyrosine phosphatase in B-1 B cells. induces expression of the early activation antigen CD69 by a cal- Eur J Immunol 1999; 29: 3319–3328. cium- and tyrosine kinase-dependent mechanism. Immunology 47 Pospisil R, Mage RG. CD5 and other superantigens as ‘ticklers’ of 1993; 78: 210–217. the B-cell receptor. Immunol Today 1998; 19: 106–108. 31 Cerutti A, Trentin L, Zambello R, Sancetta R, Milani A, Tassinari 48 Chen W, Wang HG, Srinivasula S, Alnemri ES, Cooper ND. B C, Adami F, Agostini C, Semenzato G. The CD5/CD72 receptor cell apoptosis triggered by antigen receptor ligation proceeds via system is coexpressed with several functionally relevant counter- a novel caspase-dependent pathway. J Immunol 1999; 163: structures on human B cells and delivers a critical signaling 2483–2491. activity. J Immunol 1996; 157: 1854–1862. 49 Berard M, Mondiere P, Casamayor-Palleja M, Hennino A, Bella 32 Ashwell JD, Longo DL, Bridges SH. T-cell tumor elimination as a C, Defrance T. Mitochondria connects the antigen receptor to result of T-cell receptor-mediated activation. Science 1987; 237: effector caspases during B cell receptor-induced apoptosis in nor- 61–64. mal human B cells. J Immunol 1999; 163: 4655–4662. 33 Ucker DS, Meyers J, Obermiller PS. Activation-driven T cell death. 50 Nomura T, Han H, Howard MC, Yagita H, Yakura H, Honjo T, II. Quantitative differences alone distinguish stimuli triggering Tsubata T. Antigen receptor-mediated B cell death is blocked by nontransformed T cell proliferation or death. J Immunol 1992; signaling via CD72 or treatment with dextran sulfate and is defec- 149: 1583–1592. tive in autoimmunity-prone mice. Int Immunol 1996; 8: 867–875.

Leukemia