Lymphotoxin-␣ Is an Autocrine Growth Factor for Chronic Lymphocytic Leukemia B Cells P Kulmburg, M Radke and W Digel

Home , Lymphotoxin, Lymphotoxin alpha, Lymphotoxin beta

Leukemia (1998) 12, 493–498  1998 Stockton Press All rights reserved 0887-6924/98 $12.00 http://www.stockton-press.co.uk/leu Lymphotoxin-␣ is an autocrine growth factor for chronic lymphocytic leukemia B cells P Kulmburg, M Radke and W Digel

Department of Internal Medicine I (Haematology/Oncology), Albert-Ludwigs-University of Freiburg, Freiburg, Germany

Lymphotoxin-alpha (LT), also called TNF-beta, which belongs patches, although the thymus appears to be normal. Within to the ‘TNF family’ was originally isolated from a lymphoblas- the white pulp of the spleen, there is a failure of normal segre- toid cell line. LT enhances the proliferation of activated B cells and augments B cell proliferation induced by IL-2. It functions gation of B and T cells, and lymphocytes positive for immuno- as an autocrine growth factor for EBV-infected B cell lines and globulin M are present in increased numbers in both the has been implicated in the pathogenesis of B cell malignancies. spleen and peripheral blood.9 Moreover, LT has been found We tested the expression of LT mRNA in B-CLL and found that to be an autocrine growth factor for EBV-infected B cell LT was expressed in highly purified leukemic cells in 11 out of lines.10 In this report we analyzed the potential role of LT in 11 patients examined. Regulation of expression of LT mRNA is CLL, a low-grade B cell malignancy. We found that the malig- aberrant in B-CLL cells, since LT mRNA expression was not detected in fresh peripheral blood mononuclear cells or B cells nant B-CLL cells from all 11 patients investigated constitut- identified in seven out of seven normal individuals. In addition, ively produced LT. Proliferation of B-CLL cells is stimulated LT mRNA expression was detected for up to 6 days in purified by LT and inhibited by antisense LT oligonucleotides. More- unstimulated in vitro cultures of B-CLL cells. Glucocorticos- over, it appears that LT prevents apoptosis of the malignant teroids, that have been effectively used in the treatment of B-CLL cells. lymphoid malignancies, were added to the cultures and abrogated the LT mRNA expression after an incubation time of 12 h. Addition of recombinant LT to cultures increased proliferation of B-CLL cells while proliferation of these cells was inhibited Materials and methods by antisense oligonucleotides against LT mRNA. B-CLL cells cultured with LT antisense oligonucleotides (asLT) as well as Cells glucocorticoid-treated cells showed reduced viability and a DNA fragmentation ladder characteristic of apoptosis suggest- Peripheral blood samples were obtained from 11 patients ing a relationship between down-regulation of LT mRNA expression and the induction of apoptosis. These studies sup- fulfilling diagnostic and immunophenotypic criteria for com- port the role of LT in the growth regulation and development mon B cell CLL at the Medical Center of the University of of B-CLL cells. Freiburg. The medium age of the patients was 61, ranging Keywords: CLL; lymphotoxin; autocrine growth factor; apoptosis from 51 to 77 years. The clinical stage of the untreated patients ranged from Rai stage 1 to stage 4. Eight patients were male and three female. Blood was anticoagulated with hep- Introduction arin. Mononuclear cells were isolated by Ficoll–Hypaque (Pharmacia, Freiburg, Germany) density-gradient centrifug- Chronic lymphocytic leukemia (CLL), the most common adult ation. Plastic adherent cells were removed by incubation of leukemia observed in Western societies, is a malignancy of the cell suspension on Petri culture dishes (Nunc, Wiesbaden, relatively mature B cells. Patients are characterized by high Germany) for 90 min at 37°C. The resulting cell population numbers of circulating clonal B cells, resulting from the trans- was depleted of T-cells by rosetting for 60 min at 4°C with formation and clonal expansion of a single B lymphocyte neuraminidase-treated sheep RBCs. Non-rosetting cells were within the CD5+ subset.1,2 Many cytokines, including TNF-␣, harvested from the interface of a second Ficoll gradient. The IL-2, and IL-7 have been suggested to be involved in the number of residual monocytes and T cells were determined induction or regulation of B-CLL proliferation.3–5 Lympho- by immunofluorescence with CD14 and CD3. All leukemic B toxin-alpha (LT) is a soluble product of activated lymphocytes cell populations used in our experiments had less than 1% + + that was first defined by its cytotoxic activity against fibro- CD3 and CD14 cells. blasts.6 LT is now known to be produced by activated type I CD4+ T helper lymphocytes, CD8+ lymphocytes, and certain B-lymphoblastoid cell lines.7 The gene encoding human LT is Cell surface markers and reagents located on chromosome 6, upstream of the evolutionarily related gene encoding TNF within the major histocompat- The mAbs used to determine immunophenotype were CD3, ibility complex. LT and TNF are structurally related molecules CD4, CD5, CD8, CD10, CD14, CD19, all of which were and share many common biological features.8 Studies of natu- obtained from Becton Dickinson (Heidelberg, Germany). ral functions of LT by production of an LT-deficient mouse PCA-1 (CD27) and antibodies against cytoplasmic immuno- strain suggest an essential role for LT in the normal develop- globulins and surface immunoglobulin anti-␬, anti-␭, anti-␥ ment of peripheral lymphoid organs.9 Mice deficient in LT and anti-␦ originated from Coulter (Krefeld, Germany). 12-O- have no morphologically detectable lymph nodes or Peyer’s tetradecanoylphorbol-13-acetate (TPA; Sigma Chemicals, Munich, Germany), which has been shown to prevent apoptosis of B-CLL cells,11 was used at a final concentration − of 2 × 10 8 M. Hydrocortisone was used at a concentration of Correspondence: W Digel, Department of Internal Medicine I, Frei- ␮ burg University Hospital, Hugstetter Strasse 55, D-79106 Freiburg, 1 M (Sigma). Highly purified recombinant LT was obtained Germany; Fax: 49 761 270 3206 from British Bio-Technology (Oxon, UK, specific activity 1.0– Received 7 January 1997; accepted 27 November 1997 2.0 × 107 units/mg protein). The Quantikine LT assay was Lymphotoxin-␣ is an autocrine growth factor for CLL B cells P Kulmburg et al 494 Table 1 Effect of antisense LT oligonucleotides on 3H-TdR incor- (Gibco BRL, Eggenstein, Germany) using random hexamers. poration in neoplastic B cells Generally, 5 ␮l of cDNA were used as a template for amplification in the PCR. Primers for LT were 5′-ATG Case None SenseLT AntisenseLT ACA CCT GAA CGT CTC TTC-3′ and 5′-C GAA GGC TCC AAA GAA GAC AGT ACT-3′, the expected size of the ± ± ± 1 6628 420 5856 620 210 30 PCR product was 610 bp, and restriction enzyme PstI pro- 2 2250 ± 250 2187 ± 320 725 ± 25 3 892 ± 320 790 ± 110 230 ± 55 duced fragments of 477 and 133 bp. Primers were purchased 4 3600 ± 320 3059 ± 420 880 ± 84 from Clontech Laboratories (Palo Alto, CA, USA). Primers for 5 390 ± 52 424 ± 62 140 ± 33 LT-beta were 5′-CTC TAC TGT CTC GTC GGC TAC C-3′ and 5′-GCT TTC CAT TCT TTA TTT TCA TCA-3′; the expected size B cells (3 × 106 cells/ml) from five patients with CLL were incubated of the PCR product was 466 bp. Standard PCR was performed with senseLT oligonucleotides and antisenseLT oligonucleotides with Ampli Taq DNA polymerase (Perkin Elmer Cetus, Nor- (15 ␮M) for 5 days. Proliferation rate was determined by pulse with walk, CT, USA). After 35 cycles of PCR, 10 ␮l of each 50 ␮l 3 ␮ H-TdR (1 Ci/well) during the last 16 h of incubation. Data illustrate sample was subjected to separation by agarose gel electro- a mean c.p.m. ± s.d. from triplicate wells. phoresis and stained with ethidium bromide.

obtained from R&D Systems (Minneapolis, MN, USA). This specific LT assay has a high sensitivity with a minimum detect- Analysis of apoptosis able dose of 7 pg/ml. The LT sense oligodeoxynucleotide used Apoptosis was studied by light microscopy and by DNA gel in cell cultures was 5′-CCATGACACCACCTG-3′, where the electrophoresis. Cytocentrifuge preparations were stained ATG corresponds to the start of translation. The LT antisense was 5′-CAGGTGGTGTCATGG-3′ (complimentary to the with Wright–Giemsa for light microscopy analysis of × 6 sense oligonucleotide). Both were purified by FPLC and used apoptosis. 2 10 cells were assayed for DNA fragmentation after 48 h of culture as previously described.11 Briefly, mono- at concentrations of 15 ␮M, which were added to the cultures nuclear cells were centrifuged at 500 g for 10 min. The cell every 24 h. pellet was lysed by adding 0.5 ml of hypotonic lysing buffer The rabbit anti-human lymphotoxin antibody for FACS was obtained from Genzyme (Cambridge, MA, USA). (10 mmol/l Tris, pH 7.4; 1 mmol/l EDTA; 0.2% Triton X-100) followed by gentle mixing. The lysates were immediately centrifuged at 11 000 g for 10 min to separate intact from frag- Isolation of cell subsets mented chromatin. Supernatants and pellets were precipitated in separate microfuge tubes for 16–48 h at −20°C in 50% iso- Leukemic cells were incubated on ice for 30 min with PE- labeled anti-CD5 and with FITC-labeled anti-CD19, washed and analyzed in a FACScan flow cytometer (Becton Dickinson). In some experiments the double-labeled CD5+, CD19+ or CD5− and CD19+ cells were sorted on a FACStar Plus instrument (Becton Dickinson) to a more than 99% purity.

Isolation of RNA

Polyadenylated RNA was purified using the QuickPrep Micro mRNA purification kit (Pharmacia). The RNA from 106 cells was then dissolved in 30 ␮l water, and 10 ␮l were used for cDNA synthesis. The first-strand cDNA synthesis was carried out according to the SuperScript preamplification system

Table 2 Induction of LT by TPA in cultured B-CLL cells

Case LT(pg/ml)

None TPA

1 0 126 2 0 392 3 0 184 4 0 295 Figure 1 LT PCR product from CLL patients’ cells before and after 5 0 172 restriction enzyme digestion with PstI. Lane 1, molecular weight 6 0 235 marker; lane 2, LT PCR product of cDNA template synthesized from 7 0 119 RNA isolated from B-CLL cells. Lane 3, same as lane 2, after PstI digestion. Lanes 4 and 5 represent LT-PCR products before and after restric- B cells (3 × 106) cells/ml from seven patients with CLL were incu- tion enzyme digestion of a second patient. The expected size of the − bated with TPA (2 × 10 8 M) for 2 days. LT was measured in the LT-PCR product was 610 bp, while the fragments were 477 bp and supernatant by ELISA. 133 bp. Lymphotoxin-␣ is an autocrine growth factor for CLL B cells P Kulmburg et al 495 propanol and 0.5 mol/l NaCl. The precipitates were collected LT Assay by centrifugation at 11 000 g for 10 min, air dried, resus- pended in TE buffer and heated at 55°C for 10 min. Electro- LT secretion by B-CLL cells was determined in culture super- phoresis was continued for 3 h at 20 V. Subsequently, the gels natants using a Quantakine ELISA from R&D Systems, Wies- were incubated in TBE buffer containing DNase-free RNase A baden, Germany. The sensitivity of the test was Ͻ16 pg/ml. at 37°C for 3 h to digest RNA, stained with ethidium bromide, washed, and photographed. In this DNA fragmentation assay, intact chromatin and oligonucleosomal fragment are separ- Results ated by centrifugation. Therefore, in the absence of apoptosis, the gel lane with the corresponding supernatant DNA LT mRNA expression in neoplastic B cells from CLL appears empty.11 patients

LT cytokine expression was examined by PCR to determine whether LT was consistently expressed in cells from patients Culture conditions and proliferation assay with CLL. This analysis showed that LT was expressed in cells from 11 out of 11 CLL patients tested. The LT PCR product migrated in agarose gel electrophoresis to the approximate The rate of DNA synthesis of CLL cells was measured by 3H- size predicted for the mRNA-derived template (610 bp; TdR uptake. 6 × 105 cells were seeded in triplicate in ﬂat-bot- Figure 1, lane 2). Digestion of the LT PCR product with the tom microtiter wells (Nunc) in a total volume of 200 ␮l RPMI restriction enzyme PstI revealed two fragments as expected 1640 (Gibco, Karlsruhe, Germany) supplemented with 10% from the cDNA sequence (477 bp and 133 bp; Figure 1, lane heat-inactivated fetal calf serum. Cultures were incubated up 3). To rule out contaminating DNA as a source of template, ° to 144 h at 37 C, 5% CO2 and 98% humidity, and were we performed PCR using RNA without prior reverse transcrip- pulsed with 1 ␮Ci 3H-TdR (6.7 Ci/mmol, Amersham, tion. No LT product was evident in the absence of a cDNA Braunschweig, Germany) 7 h before cells were harvested (Cell template. A PCR product derived from genomic DNA would harvester; Scatron, Lier, Norway). Incorporated radioactivity be expected to consist of 943 bp since the reverse and forward was measured by liquid scintillation counting (Packard Tricarb primers for LT span two introns of genomic sequence con- 1500, Frankfurt, Germany). sisting of 86 bp and 247 bp. Thus, LT was expressed in cells

Figure 2 Suppression of LT mRNA expression by hydrocortisone. The LT mRNA expression was suppressed 12 h after addition of hydrocortisone (HC; 1 ␮M)toin vitro cultured B-CLL cells (3 × 106 cells/ml). Without any additives, LT mRNA could be detected up to 6 days (144 h) in cultures (A, actin; LT, lymphotoxin; HC, hydrocortisone). C1 represents a PCR control without RT, C2 represents a water control. Similar results were obtained with B-CLL cells from three other patients. Lymphotoxin-␣ is an autocrine growth factor for CLL B cells P Kulmburg et al 496 from CLL patients. No LT mRNA was detected in mononuclear hydrocortisone (Figure 2). The same results were obtained in cells or isolated B cells from seven healthy individuals repeated experiments in three different patients (data not although similar amounts of beta-actin mRNA could be ampli- shown). fied in parallel experiments. Dual-color fluorescence and cell sorting were used to determine whether the leukemic cells themselves expressed LT. The double-positive CD5+, CD19+ Reduction of DNA synthesis and induction of cells were clearly identified as a major population and were apoptosis by antisense LT sorted to a purity of more than 99%. The cells were then washed and processed for the isolation of RNA. PCR demon- strated that only the subset consisting of the leukemic popu- Addition of recombinant LT to purified B-CLL cells suggested lation expressed LT mRNA (data not shown). In addition, we that B-CLL cells can be growth stimulated by exogenous LT examined the expression of LT-beta by PCR and it was not (Figure 3). To test further the function of endogenously pro- detected in any of the eight CLL cases tested. duced LT on the DNA synthesis, we used antisense LT oligonucleotides. Daily addition of antisense LT oligonucleotides (15 ␮M) to the cultures reduced the DNA synthesis as meas- 3 Suppression of the LT mRNA expression by ured by H-TdR incorporation (Figure 3 and Table 1). Addition hydrocortisone of LT and antisense LT oligonucleotides in conjunction with the B-CLL cultures reversed the inhibited DNA synthesis caused by antisense LT oligonucleotides (Figure 3). However, B-CLL cells underwent rapidly programmed cell death after sense LT (15 ␮M) oligonucleotide did not significantly influ- addition of hydrocortisone to the cell cultures.12 Under the ence DNA synthesis (Table 1). Furthermore, B-CLL cells cul- assumption that LT is an essential endogenous growth factor tured with antisense LT oligonucleotides showed character- for the neoplastic B cells, we then analyzed whether hydro- istic features of apoptosis. After 48 h, 30% of cells showed cortisone abrogated the LT mRNA expression. Using RT/PCR morphologic evidence of apoptosis such as nuclear fragmen- amplification we showed that LT mRNA expression was tation and loss of cell volume. At the same time, we observed absent in the hydrocortisone-treated cultures after an incu- degradation of DNA into fragments (Figure 4). bation time of 12 h, whereas LT mRNA expression could be detected in the unstimulated B-CLL cultures for up to 6 days (Figure 2). Similar amounts of beta-actin mRNA were ampli- fied in parallel experiments in the cultures with or without

Figure 4 Endogenously produced LT inhibits DNA fragmentation Figure 3 Reduction of DNA synthesis by addition of antisense LT in B-CLL cells. The cells (2 × 106/ml) were cultured for 48 h with oligonucleotides to cultures. 3H-TdR incorporation of B-CLL cells medium + 10% FCS, sense LT oligonucleotides (15 ␮M, sLT), antisense (3 × 106 cells/ml) after addition of sense LT (15 ␮M, sLT), antisense LT LT oligonucleotides (15 ␮M, asLT), and hydrocortisone (1 ␮M, HC). (15 ␮M, asLT), LT (10 ng/ml), hydrocortisone (1 ␮M), and the combi- The DNA ladder was performed as described in Ref. 11. Intact nation of LT and asLT. Similar results were obtained in repeated chromatin and oligonucleosomal fragments are separated by centri- experiments in three different patients. Data are mean c.p.m. ± s.d. fugation. Therefore, in the absence of apoptosis, the gel lane with the from triplicate wells. corresponding supernatant DNA appears empty. Lymphotoxin-␣ is an autocrine growth factor for CLL B cells P Kulmburg et al 497 Secretion of LT after stimulation duction of a growth factor by B-CLL cells has also been shown for IL-7.5,19 TNF-␣/LT binding proteins are greatly increased The supernatants from B-CLL cultures were examined for the in the sera of B-CLL patients.20 Our data show that the malig- ability of B-CLL cells to constitutively secrete LT. In contrast nant cell clone secretes these soluble TNF/LT-receptors (Digel to our results with PCR, LT could not be detected by ELISA in et al, unpublished results). the supernatant of unstimulated cultures in the seven patients Endogenous production of LT might trigger release of sol- examined over a culture period of 1–3 days. Intracellularly, uble LT receptors by the malignant cells. These soluble recep- LT has been observed reacting at low concentrations, there- tors bind and neutralize TNF-␣ and LT, and thus, could com- fore, the level of LT secreted by B-CLL may be too low to pete with cell-bound receptors for the ligand. This could measure by ELISA. Thus, we investigated the possibility that partially explain why LT levels in sera of patients and culture B-CLL cells produce LT after stimulation. Phorbolesther TPA, supernatants appear to be low and LT could not be detected which has been shown to induce LT and prevent apoptosis of in the supernatant without stimulation. − B-CLL cells,11 was used at a final concentration of 2 × 10 8 M B-CLL cells express both types of TNF receptors.21 Mem- to induce LT secretion by the B-CLL cells. TPA induced up to brane-bound LT heterotrimers could not be detected on the 392 pg/ml LT in B-CLL cultures (3 × 106 cells/ml) from seven B-CLL cells. Expression of LT-beta is a prerequisite in the for- of seven patients after an optimal incubation time of 48 h mation of LT heterotrimers on the cell surface.22 Furthermore, (Table 2). Furthermore, we tested for membrane-bound LT we could not detect LT-beta in any of the cases tested. It seems heterotrimers by using FACS and Ab against LT, but no mem- unlikely that B-CLL express lymphotoxin-beta-specific brane-bound LT was detected on B-CLL cells of the eight receptors. cases tested. Programmed cell death appears to be regulated by cytokines acting as survival signals.23 B-CLL cells cultured with asLT showed a DNA ladder characteristic of apoptosis. Thus, Discussion autocrine secretion of LT might delay apoptosis and extended survival of malignant clones similar to what has been shown The production of LT is restricted to lymphocytes or cells for IFN-␥.24 Glucocorticoid hormones, commonly used to derived from lymphoid progenitors.7 EBV-infected lympho- treat patients with B-CLL, abrogated the LT mRNA expression blastoid cell lines produce LT constitutively and the auto- and induced apoptosis. A possible mechanism that could nomously produced LT serves as a growth factor for EBV- explain this finding is abrogation of LT mRNA expression by infected B cells stimulating their growth in an autocrine glucocorticoids. Factor deprivation resulting in apoptosis has fashion.10 In this report, we have examined the role of LT in been well documented in other systems, such as hematopo- B-CLL cells that had been isolated from 11 patients. We found ietic colony-stimulating factors.25 that the mRNA for LT was expressed in the leukemic B cells In summary, our data document autocrine aberrant of all patients. In contrast, LT mRNA could not be detected in secretion of LT by B-CLL cells and suggest that autocrine LT mononuclear cells or B cells from healthy individuals. The production could result in clonal expansion of B-CLL cells by detection of LT in the supernatant after stimulation indicated prevention of programmed death.26 Suppression of an that LT mRNA is actually translated into protein by the leu- endogenous cytokine pathway might be an important tool for kemic B cells. Our data suggest that LT may be produced by treatment of low-grade lymphomas in the future. the malignant lymphoma clone and act as an autocrine growth factor stimulating proliferation of B-CLL growth. Simi- lar findings have been previously reported for TNF-␣.13 In References contrast to TNF-␣ secretion, LT mRNA expression appears to be observed in every single case examined and its spon- 1 Gale RB, Foon KA. Chronic lymphocytic leukemia. Ann Intern taneous production appears to be rather low. This suggests Med 1985; 103: 101–120. that LT may act in an autocrine fashion in B-CLL cells. More- 2 Dighiero G, Travade P, Chevret S, Fenaux P, Chastang C, Binet JL and The French Cooperative Group on CLL. B cell chronic lym- over, our data show that antisense LT oligonucleotides inhibit phocytic leukemia: present status and future directions. Blood B-CLL proliferation in all five patients examined. Importantly, 1991; 78: 1901–1914. this blocking can be reversed by exogenous LT, suggesting 3 Fluckiger AC, Rossi JF, Bussel A, Bryon P, Banchereau J, Defrance that LT is a growth factor for B-CLL cells independent of TNF- T. Responsiveness of chronic lymphocytic leukemia B cells acti- ␣. LT and TNF-␣ share a 30% sequence homology and pos- vated via surface Igs or CD40 to B-cell tropic factors. Blood 1992; 80: 3173–3181. sess many overlapping biological activities.14 Also, TNF-␣ and ¨ 4 Digel W, Stefanic M, Schoniger W, Buck C, Raghavachar A, Frick- 15 ␣ LT use common receptors. In most cases, TNF- has been hofen N, Heimpel H, Porzsolt F. Tumor necrosis factor induces 4,13 shown to stimulate proliferation of malignant B cells, but proliferation of neoplastic B cells from chronic lymphocytic leuke- rarely to inhibit B cell growth.16 TNF-␣ and LT have divergent mia. Blood 1989; 73: 1242–1246. effects on proliferation of hairy cell leukemia B cells, with 5 Digel W, Schmid M, Heil G, Conrad P, Gillis S, Porzsolt F. Human TNF-␣ being stimulatory while LT is inhibitory.17 In contrast, interleukin-7 induces proliferation of neoplastic cells from chronic LT promoted growth of lymphoblastoid B cell lines while TNF- lymphocytic leukemia and acute leukemias. Blood, 1991; 78: 753–759. ␣ 10 partially reversed the stimulatory effect of LT. 6 Granger GA, Williams TW. Lymphocyte cytotoxicity in vitro: acti- Since LT expression was only detected in leukemic B cells vation and release of a cytotoxic factor. Nature 1968; 218: but not in B cells from normal donors, LT expression might 1253–1254. be related to leukemic transformation. The mechanisms 7 Kehrl JH, Alvarez-Mon M, Delsing GA, Fauci AS. Lymphotoxin is underlying the aberrant lymphoproliferative stimulus in an important T cell-derived growth factor for human B-cells. chronic leukemias, such as B-CLL, remain unclear. A critical Science 1987; 238: 1144–1146. 8 Paul NL, Ruddle HN. Lymphotoxin. Annu Rev Immunol 1988; 6: event in the process of B cell immortalization may be the 407–438. initiation of an autocrine loop, where cells produce growth 9 De Togni P, Goellner J, Ruddle HN, Streeter PR, Fick A, Maria- ¨ factors that support their own proliferation.18 Aberrant pro- thasan S, Smith SC, Carlson R, Shornick LP, Strauss-Schonenberger Lymphotoxin-␣ is an autocrine growth factor for CLL B cells P Kulmburg et al 498 J, Russell JH, Karr R, Chaplin DD. Abnormal development of per- 18 Gordon J, Ley SC, Melamed MD, English LS, Hughes-Jones NC. ipheral lymphoid organs in mice deficient in lymphotoxin. Science Soluble factor requirements for the auto-stimulatory growth of B 1994; 264: 703–707. lymphoblasts immortalized by Epstein–Barr virus. J Exp Med 1984; 10 Esrov BZ, Kurzrock R, Pocsik E, Pathak S, Kantarjian HM, Zipf TF, 159: 1554–1559. Harris D, Talpaz M, Aggarwal BB. Lymphotoxin is an autocrine 19 Frishman BJ, Long B, Knospe W, Gregory S, Plate J. Genes for growth factor for Epstein-Barr virus infected B cell lines. J Exp Med interleukin 7 are transcribed in leukemic cell subsets of individuals 1993; 177: 763–774. with chronic lymphocytic leukemia. J Exp Med 1993; 177: 11 McConkey DJ, Aguilar-Santelises M, Hartzell P, Eriksson I, 955–964. Mellstedt H, Orrenius S, Jondal M. Induction of DNA fragmen- 20 Digel W, Porzsolt F, Schmid M, Herrmann F, Lesslauer W, Brock- tation in chronic B-lymphocytic leukemia cells. J Immunol 1992; haus M. High levels of circulating soluble receptors for tumor 146: 1072–1076. necrosis factor in hairy cell leukemia and type B chronic lymphocytic leukemia. J Clin Invest 1992; 89: 1690–1693. 12 Manabe A, Coustan-Smith E, Behm FG, Raimondi SC, Campana ¨ D. Bone marrow-derived stromal cells prevent apoptotic cell 21 Digel W, Schoniger W, Stefanic M, Janssen H, Buck C, Schmid death in B lineage acute lymphoblastic leukemia. Blood 1992; 79: M, Raghavachar A, Porzsolt F. Receptors for tumor necrosis factor 2370–2377. on neoplastic B cells from chronic lymphocytic leukemia are expressed in vitro but not in vivo. Blood 1990; 76: 1607–1613. 13 Cordingly FT, Hoffbrand AV, Heslop HE, Tumer M, Blanche A, 22 VanArsdale T, VanArsdale S, Force W, Walter B, Mosialos G, Kieff Reittie JE, Vyakamam A, Meager A, Brenner MK. Tumour necrosis E, Reed J, Ware C. Lymphotoxin-beta receptor signaling complex: factor as an autocrine tumour growth factor for chronic B cell role of tumor necrosis factor receptor-associated factor 3 recruit- malignancies. Lancet 1989; 1: 969–971. ment in cell death and activation of nuclear factor kappaB. Proc 14 Gray PW, Aggarwal BB, Benton CV, Bringman TS, Henzel WJ, Natl Acad Sci USA 1997; 94: 2460–2465. Jarrett JA, Leung DW, Moffat B, Ng P, Svedersky LP et al. Cloning 23 Raff MC. Social control on cell survival and cell death. Nature and expression of cDNA for human lymphotoxin, a lymphokine 1992; 356: 397–400. with tumor necrosis activity. Nature 1984; 312: 721–724. 24 Buschle M, Campana D, Carding SR, Richard C, Hoffbrand AV, 15 Loetscher H, Steinmetz M, Lesslauer W. Tumor necrosis factor: Brenner MK. Interferon-gamma inhibits apoptotic cell death in B receptors and inhibitors. Cancer Cells 1991; 3: 221–226. cell chronic lymphocytic leukemia. J Exp Med 1993; 177: 213– 16 Janssen O, Kabelitz D. Tumor necrosis factor selectively inhibits 218. activation of human B cells by Epstein–Barr virus. J Immunol 1988; 25 Williams GT, Smith CA, Spooncer E, Dexter TM, Taylor DR. 140: 125–130. ¨ Haematopoietic colony stimulating factors promote cell survival 17 Buck C, Digel W, Schoniger W, Stefanic M, Raghavachar An, by suppressing apoptosis. Nature 1990; 343: 76–79. Heimpel H, Porzsolt F. Tumor necrosis factor-␣ but not lympho- 26 Hoffbrand AV, Panayiotidis P, Reittie J, Ganesshaguru K. toxin stimulates the growth of tumor cells in hairy cell leukemia. Autocrine and paracrine growth loops in chronic lymphocytic leu- Leukemia 1990; 4: 431–434. kemia. Semin Hematol 1993; 30: 306–317.