CD137 and CD137 ligand constitutively coexpressed on human T and B leukemia cells signal proliferation and survival

Abstract

CD137, a member of the tumor necrosis factor receptor family, provides expansion and survival signal to T cells. Its ligand, CD137L, in addition to its ability to costimulate T cells, signals back into antigen presenting cells promoting their activation and differentiation. Recently, CD137 has been proposed as a therapeutic target to improve and sustain anticancer immune response. Several activated T leukemia and B lymphoma cell lines expressed CD137 or CD137L, respectively, and soluble CD137L has been found in sera of leukemia patients. However, the functionality and role of these costimulatory molecules in hematologic malignancies are until now unknown. Interestingly, we observed constitutive CD137 and CD137L coexpression on both human T and B leukemia cell lines. The constitutive CD137 expression on unstimulated T or B leukemia cells presents some differences compared to CD137 expressed on PMA/ionomycin-activated T leukemia cells. Surprisingly, in spite of the low expression level, both tumor CD137 and CD137L molecules signaled in T and B leukemia cells inducing proliferation and prolonging survival. In addition, CD137/CD137L system ligation opposed the anticancer drug cytotoxic effects, reducing the apoptotic DNA fragmentation and stimulating proliferation of doxorubicin-escaped leukemia cells. Although the role of leukemia CD137/CD137L system in vivo is unknown, these data suggest that these costimulatory molecules might confer an advantage to hematologic tumors promoting survival, sustaining cellular growth and contributing to drug resistance. © 2003 Wiley-Liss, Inc.

CD137/CD137 ligand (L), members of tumor necrosis factor (TNF) receptor/ligand families, regulate cell activation and proliferation of immune system.1, 2 CD137 receptor delivers potent costimulatory signals to activated T lymphocytes maintaining cell division, facilitating differentiation into effector and memory cells and acting as an important survival factor.3, 4, 5 In addition, triggering of 4-1BB, the murine counterpart of human CD137, in effector CD8/NK cells is able in in vivo murine models to eradicate large well-established poorly immunogenic tumors, conferring also a long-lasting protection against tumors.6, 7, 8 Although CD137 is mainly involved in regulating activated T lymphocyte functions, expression has also been found in activated B cells, monocytes/macrophages, dendritic cells, eosinophils and neutrophils.9, 10, 11, 12

CD137L, the physiologic ligand of CD137, is mainly expressed by antigen presenting cells, including mature dendritic cells, monocytes/macrophages and B cells. CD137L, in addition to its ability to costimulate T cells by triggering CD137 receptor, also signals back into antigen presenting cells inducing proliferation, prolonging survival and enhancing secretion of proinflammatory cytokines.13, 14, 15, 16

However, CD137L expressed by activated T lymphocytes, different from antigen presenting cells, inhibits proliferation and promotes apoptosis,17, 18 indicating CD137L's ability to affect opposite signaling pathways depending on the cell type involved.

CD137 expression is inducible in several human T leukemia cell lines and constitutive CD137L expression has been reported for many human B lymphomas. The molecular characteristics of CD137/CD137L expressed on tumor cells are comparable to those expressed on normal immune cells.19, 20, 21 In addition, high amounts of soluble CD137L have been found in sera of patients with non-Hodgkin's lymphoma, myelodysplastic syndrome or acute myeloid leukemia.21 However, the meaning of CD137/CD137L expression in leukemia cells is unknown. Considering that some members of TNF receptor family are upregulated in human hematologic malignancies and are able to trigger responses favoring tumor progression,22, 23 we studied CD137/CD137L molecular and functional characteristics in human T and B leukemia cell lines.

MATERIAL AND METHODS

Human cell lines

The T-cell lines derived from acute leukemia, CCRF-CEM and Jurkat, and the B lymphoma cell lines, Raji and IM-9, were obtained from ATCC (Manassas, VA). These cells were grown in RPMI 1640 medium containing 2 mM glutamine, 10 mM HEPES, 1 mM sodium pyruvate and 10% heat-inactivated fetal bovine serum (FBS).

Reagents

The soluble fusion proteins and the soluble fusion protein FITC-labeled human CD137L-mCD8 and human CD137-hIg were purchased from Ancell (Bayport, MN). Monoclonal antihuman CD137 antibody was purchased from Ancell. Polyclonal rabbit antihuman CD137L and antihuman CD137 antibodies and FITC-labeled sheep antirabbit immunoglobulin were from Chemicon International (Temecula, CA). Polyclonal rabbit antihuman CD8 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). FITC-labeled goat antimouse IgG from Sigma (St. Louis, MO).

RNA extraction and RT-PCR

Total RNA was extracted from CEM and Raji cells using the Nucleospin RNA II kit (Macherey-Nagel, Duren, Germany) according to the manufacturer's instructions. RT-PCR technique was performed using the 1-tube, 2-enzyme system kit Promega's Access RT-PCR System (Promega, Madison, WI), transcribing and then amplifying 5 μg of total RNA for each sample. For detection of human CD137 and CD137L cDNA, we used previously published primers; in particular, for CD137L, sense primer corresponds to nucleotides 484–513 and antisense primer is complementary to nucleotides 926–949;15, 24 for CD137, sense primer corresponds to nucleotides 128–149 and antisense primer is complementary to nucleotides 810–831.9, 10 After denaturation at 94°C for 2 min, 42 PCR cycles were performed, each consisting of a denaturation step (94°C, 1 min), an annealing step (60°C, 2 min) and an elongation step (72°C, 2 min). For RT-PCR-negative controls, parallel samples were assayed without adding reverse transcriptase. The PCR products (expected size 465 bp for human CD137L, 703 bp for human CD137) were separated by electrophoresis on 2% agarose gel and visualized by staining with ethidium bromide. Amplification of human β-actin or G3PDH genes was used as a control.

Flow cytometry

Leukemia cells (500,000 cells/sample) were analyzed on a FACScan (Becton-Dickinson, Mountain View, CA). Gates were set on live cells only based on forward-side scatter profiles and in some experiments using propidium iodide as a marker of dead or apoptotic cells. For the detection of human CD137, cells were incubated with 10 μg/ml FITC-labeled fusion protein CD137L-mCD8 (1 hr at 4°C); 1 μg/ml monoclonal anti-CD137 antibody (1 hr at 4°C) and, after washing, FITC-labeled antimouse IgG (1:150; 30 min at 4°C); polyclonal rabbit anti-CD137 antibody (1:200; 1 hr at 4°C) and, after washing, FITC-labeled sheep antirabbit immunoglobulin (1:200; 30 min at 4°C). In some experiments, all these reagents were preincubated for 15 min at room temperature with 20 μg/ml CD137-hIg before being added to activated CEM cells. For the detection of human CD137L, cells were incubated with 10 μg/ml FITC-labeled fusion protein CD137-hIg (1 hr at 4°C); polyclonal rabbit anti-CD137L antibody (1:200; 1 hr at 4°C) and, after washing, FITC-labeled sheep antirabbit immunoglobulin (1:200; 30 min at 4°C).

Cell growth and survival

CEM and Raji cells were plated on 24-well tissue culture plates at 150,000 cells/well in RPMI medium at 1% FBS and 300,000 cells/well in RPMI medium at 1.5% FBS, respectively. The reagents polyclonal rabbit anti-CD137, anti-CD137L or anti-CD8 antibodies (1:100–1:10,000), the fusion protein CD137L-mCD8 (1–4 μg/ml), the fusion protein CD137-hIg (2–4 μg/ml) were added and left in the culture medium for the entire experiment length (24–168 hr). After the appropriate times of culture, cells were counted in a hemacytometer counting chamber. Dead and live cells were distinguished by Trypan blue exclusion. Each condition was performed in triplicate. In another set of experiments, all T and B leukemia cells were plated on 96-well tissue culture plates at 30,000 cells/well in 200 μl of RPMI medium at 1% FBS in the presence of polyclonal anti-CD137, anti-CD137L or anti-CD8 antibodies (1:1,000) and CD137-hIg previously attached on the plastic well (2 μg/ml in PBS at 4°C overnight). Each condition was performed in quadruplicate. The plates were incubated at 37°C for 6 days. At the end of culture, 20 μl of Alamar blue (Biosource International, Camarillo, CA) were added to each well and the plates were further incubated for an additional 5 hr. The chemical reduction of Alamar blue in the growth medium is a fluorometric growth indicator based on detection of metabolic activity. Fluorescence was monitored in a multilabel counter Victor 1420 (Wallak, Turku, Finland) at 530 nm excitation wavelength and 590 nm emission wavelength.

Cytotoxicity assays

Leukemia cells were exposed to doxorubicin (Adriablastina, Pharmacia Upjohn, Milan, Italy) or camptothecin (Sigma) for 90 min, washed twice in RPMI medium, then plated at the concentration of 80,000 cells/well in 200 μL RPMI medium containing FBS 1% in the presence of polyclonal anti-CD137, anti-CD137L or anti-CD8 antibodies (1:1,000), soluble CD137L-mCD8 (2 μg/ml) or only medium. After 4 days of culture, cellular viability was measured using Alamar blue assay. Percentage of cell survival was calculated by comparing the cell growth of unpulsed cells.

Photometric enzyme immunoassay for qualitative and quantitative in vitro determination of cytoplasmic histone-associated DNA fragments after induced cell death

After apoptosis induction, cells (10,000 cells/well) were lysed in 200 μl of lysis buffer and 20 μl assayed in the photometric enzyme immunoassay for the qualitative and quantitative in vitro determination of cytoplasmic histone-associated DNA fragments (Cell Death detection ELISA^PLUS, Roche Applied Science, Mannheim, Germany) according to the manufacturer's instructions. Each condition was performed in quadruplicate.

RESULTS

Constitutive CD137 expression on T and B leukemia cells

In flow cytometric studies, surface CD137 protein was inducible in CEM and Jurkat leukemic T cells treated with 10 ng/ml PMA plus 1 μM ionomycin for 48 hr (Fig. 1a). Activated cells were stained by specific monoclonal or polyclonal antibodies as well as by the FITC-CD137L-mCD8 fusion protein. The specificity of all these reagents was demonstrated by the complete inhibition of their binding to activated CEM cells following neutralization with the soluble fusion protein CD137-hIg (Fig. 1a). Surprisingly, unstimulated T and B leukemia cell lines express membrane CD137, being recognized by the polyclonal anti-CD137 antibody or the FITC-CD137L-mCD8 (Fig. 1b). The constitutive CD137 expression on T leukemia cells, especially on CEM cells, was lower in terms of binding sites and percentage of positive cells compared to activated cells. However, no binding on unstimulated cells was observed with the monoclonal anti-CD137 antibody that also scarcely recognized the activated Jurkat cells.

Figure 1. FACS analysis of CD137 expression on T and B leukemia cells. CEM and Jurkat cells activated with PMA and ionomycin for 48 hr (a) or unstimulated CEM, Jurkat, Raji and IM-9 cells (b) were stained with polyclonal anti-CD137 antibody (1:200) plus FITC-labeled antirabbit Ig; 10 μg/ml FITC-labeled fusion protein CD137L-mCD8; or 1 μg/ml monoclonal anti-CD137 antibody plus FITC-labeled antimouse IgG. All these reagents were preincubated with 20 μg/ml CD137-hIg for 15 min at room temperature before being added to activated CEM cells (a). The samples were analyzed on a FACScan. Shaded histograms represent the proper background fluorescence level of each condition.

CD137 surface expression kinetic studies indicate similar profiles between polyclonal and monoclonal anti-CD137 antibodies on activated CEM cells with the only difference in capability of the polyclonal antibody to recognize also unstimulated cells (Fig. 2).

Figure 2. FACS analysis of CD137 kinetic expression on CEM cells stimulated with PMA and ionomycin. CEM cells were stained with polyclonal anti-CD137 antibody (1:200) plus FITC-labeled antirabbit Ig (a) or 1 μg/ml monoclonal anti-CD137 antibody plus FITC-labeled antimouse IgG (b) and then analyzed on a FACScan. Shaded histograms represent CEM cells stained only with a secondary FITC-labeled antirabbit or antimouse Ig. Open histograms represent unstimulated CEM or CEM stimulated with PMA and ionomycin for 5, 24 and 48 hr as indicated.

In addition, transcription of CD137 mRNA has been found in activated CEM and Jurkat cells but also in unstimulated CEM, Jurkat and Raji cells (Fig. 3). The level of CD137 mRNA expression differed among the cell lines, being extremely high for activated CEM cells, moderate for Raji and weak for unstimulated CEM (Fig. 3a). In activated CEM cells, the amplified product of 703 bp was so abundant it overflowed during the electrophoresis run. Differences in the amount of amplified products were also observed between activated and unstimulated Jurkat cells (Fig. 3b). In RT-PCR studies, besides the expected products of 703 bp, 2 additional bands of 450 and 545 bp were visualized in PMA/ionomycin-activated CEM and Jurkat cells. Likely they correspond to soluble forms of CD137 generated by differential splicing as previously described.25

Figure 3. CD137 and CD137L mRNA expression in Raji, CEM and Jurkat cells. RT-PCR analysis of human CD137, CD137L, β-actin and G3PDH mRNA expression was performed on total RNA extracted (a) from unstimulated CEM cells (lane 1), CEM cells stimulated with 10 ng/ml PMA plus 1 μg/ml ionomycin for 24 hr (lane 2), and unstimulated Raji cells (lane 3); and (b) from unstimulated Jurkat cell (lane 1) or Jurkat cell stimulated with 10 ng/ml PMA plus 1 μg/ml ionomycin for 24 hr (lane 2). The specific amplification products were separated by 2% agarose gel electrophoresis and visualized by staining with ethidium bromide.

Constitutive CD137L expression on T and B leukemia cells

Surface CD137L protein expression, on unstimulated Raji and IM-9 B lymphoma cells, has been detected by specific staining with the polyclonal anti-CD137L antibody or the fusion protein CD137-hIg in FACS analysis (Fig. 4). Interestingly, CEM and Jurkat leukemic T cells, although on a subset of their cell population, also expressed membrane CD137L protein (Fig. 4). Activation of T leukemia cells with PMA and ionomycin for 48 hr reduced the number of cells expressing membrane CD137L molecules (Fig. 4).

Figure 4. FACS analysis of CD137L expression on T and B leukemia cells. Unstimulated Raji, IM-9, CEM and Jurkat cells and PMA plus ionomycin-activated CEM and Jurkat cells were stained with polyclonal anti-CD137L antibody (1:200) plus FITC-labeled antirabbit Ig or 10 μg/ml FITC-labeled fusion protein CD137-hIg. The samples were analyzed on a FACScan. Shaded histograms represent the proper background fluorescence level of each condition.

In addition, mRNA encoding CD137L was observed by RT-PCR studies in unstimulated Raji, Jurkat and CEM cells (Fig. 3). Activation of CEM cells increased the level of CD137L mRNA transcripts (Fig. 3a), while activation of Jurkat cells did not determine significant modulation in CD137L gene expression (Fig. 3b).

CD137/CD137L ligation enhances proliferation and survival in both T and B leukemia cells

CD137 crosslinking mediated by the polyclonal anti-CD137 antibody or CD137L-mCD8 enhanced both CEM and Raji proliferation as indicated by the significant increase of living cells after 96 hr of culture (Fig. 5). In addition, in CEM cells, a profound effect on cell death delay was observed after 168 hr of culture. Similar proliferative and survival signals were also observed by CD137L crosslinking obtained with the polyclonal anti-CD137L antibody (Fig. 5).

Figure 5. Effects on CEM and Raji cell growth and survival induced by CD137 and CD137L ligation. CEM and Raji cells, plated at 150,000 and 300,000 cells/well, respectively, were cultured for the entire length of experiment in the presence of anti-CD137, anti-CD137L, anti-CD8 antibodies (1:1,000) or 2 μg/ml CD137L-mCD8. At the indicated time points, cells were counted in a hemacytometer chamber. Dead and live cells were distinguished by Trypan blue exclusion. Data represent the mean of 5 independent experiments. Triple asterisk, p < 0.001 vs. control medium; one-way ANOVA, Tukey test.

Anti-CD137 or anti-CD137L antibodies were effective in a dilution range from 1:500 to 1:10,000, while CD137L-mCD8 was effective at concentrations of 1–4 μg/ml (data not shown). The CD137L-mCD8 spontaneous aggregation in multimeric forms (as evidenced by HPLC GPC analysis; data not shown) may be responsible for its functionality even in soluble forms. Anti-CD8 antibody, used as isotype control for rabbit serum, was unable to modulate CEM and Raji proliferation (Fig. 5).

CD137-hIg, although binding CD137L expressed on Raji and CEM cells, did not stimulate proliferation when given in soluble form. Consequently, this property, likely due to its incapacity to aggregate in solution in a grade superior to the intrinsic protein dimerization (data not shown), was used to verify the specificity of CD137L-mCD8 and anti-CD137 antibody proliferative effects. As shown in Figure 6, the responses induced by CD137L-mCD8 (2 μg/ml) were almost completely inhibited and those triggered by the anti-CD137 antibody (1:5,000) were reduced by about 50% in CEM cells and 70% in Raji cells when these molecules were pretreated for 30 min at room temperature with 4 μg/ml soluble CD137-hIg before being added to the cultures. These data confirm a real involvement of CD137 receptors in CEM and Raji cell growth.

Figure 6. Neutralizing effects of soluble CD137-hIg protein on the anti-CD137 antibody and CD137L-mCD8-induced cell growth and survival in human CEM and Raji cells. CEM and Raji cells, plated at 150,000 and 300,000 cells/well, respectively, were cultured in the presence of anti-CD137 antibody (1:5,000) or 2 μg/ml fusion protein CD137L-mCD8. In the indicated conditions, the stimuli were preincubated for 30 min at room temperature with 4 μg/ml soluble CD137-hIg fusion protein before being added to the cultures. After 96 hr of culture, cells were counted in a hemacytometer chamber. Dead and live cells were distinguished by Trypan blue exclusion. Data represent the mean of 3 independent experiments. Double asterisk, p < 0.01; triple asterisk, p < 0.001 vs. the appropriate condition without soluble CD137-hIg fusion protein; one-way ANOVA, Tukey test.

The fluorimetric Alamar blue assay showed that in CEM, Jurkat, Raji and IM-9 leukemic cells, there was a significant increase in active metabolic cells in response to anti-CD137 and anti-CD137L antibodies as well as to CD137-hIg previously attached to the plastic well after 6 days of culture (Fig. 7). Therefore, the fusion protein CD137-hIg was able to generate a signal exclusively when inducing CD137L clustering.

Figure 7. CD137/CD137L-mediated cell growth and survival in T and B leukemia cells. CEM, Jurkat, IM-9 and Raji cells, plated on 96-well tissue culture plates at 30,000 cells/well, were cultured in the presence of anti-CD137, anti-CD137L or anti-CD8 antibodies (1:1,000) or CD137-hIg (2 μg/ml; previously attached on the plastic well) for 144 hr. Cell growth was measured as chemical reduction of Alamar blue. Data represent the mean of arbitrary fluorescence units of quadruplicate wells ± SEM of 1 representative experiment out of 3. Asterisk, p < 0.05; double asterisk, p < 0.01; triple asterisk, p < 0.001 vs. control medium; one-way ANOVA, Tukey test.

The simultaneous triggering of CD137 and CD137L expressed by Raji and CEM cells determined a significant additive effect on proliferation/survival (Fig. 8). The combination between anti-CD137 and anti-CD137L antibodies (1:5,000) was significantly different from the single antibody concentration (p < 0.001), indicating that the pathways turned on by this receptor/ligand pair are not antagonistic but generate the same final effects.

Figure 8. Effects of anti-CD137 and anti-CD137L antibody combination on CEM and Raji cell growth and survival. CEM and Raji cells, plated at 150,000 and 300,000 cells/well, respectively, were cultured in the presence of anti-CD137 or anti-CD137L antibodies (1:2,500 or 1:5,000) and the combination of anti-CD137 plus anti-CD137L antibodies (1:5,000). Raji cells were counted at 96 hr while CEM cells were counted at 168 hr in a hemacytometer chamber. Dead and live cells were distinguished by Trypan blue exclusion. Data represent the mean of 3 independent experiments. Asterisk, p < 0.05; double asterisk, p < 0.01; triple asterisk, p < 0.001 vs. control medium; one-way ANOVA, Tukey test.

Enhancement of cell viability and proliferating capacity by CD137/CD137L ligation in T and B leukemia cells exposed to anticancer drugs

To investigate whether tumor CD137 and CD137L act as survival factors after exposure to DNA-damaging agents such as anticancer drugs, leukemia cells were pulsed for 90 min with doxorubicin (150–250 nM for CEM cells and 250 nM for Raji, Jurkat and IM-9 cells) or camptothecin (1 μM for Raji cells). The cells were then plated in the presence of anti-CD137, anti-CD137L, anti-CD8 antibodies or the soluble CD137L-mCD8. Ligation of both CD137 and CD137L significantly enhanced the cell viability/proliferation in all T and B leukemia cells after 96 hr of culture (Fig. 9). To determine whether CD137/CD137L triggering, besides its ability to increase cell proliferation in the anticancer drug-escaped cells, also inhibited apoptosis, we measured the cytoplasmic histone-associated DNA fragments with the specific cell death detection ELISA kit. Triggering of both CD137 and CD137L significantly reduced the DNA fragmentation induced by 90-min doxurubicin pulse in CEM cells cultured for 24 hr and in Raji and Jurkat cells cultured for 48 hr (Fig. 10a and b). Similar protective effects were also observed when Raji cells, stimulated overnight through CD137/CD137L ligation, were exposed to 1 μM camptothecin for 4 hr (Fig. 10c).

Figure 9. Triggering of CD137/CD137L promotes cell growth and survival in anticancer drug-treated T and B leukemia cells. Leukemic cells were incubated at 37°C for 90 min with 150–250 nM doxorubicin for CEM cells (a), with 250 nM doxorubicin for Raji, Jurkat and IM-9 cells (b and c) and with 1 μM camptothecin for Raji cells (c). After the pulse, cells were washed and then plated at the concentration of 80,000 cells/well in the presence of anti-CD137, anti-CD137L or anti-CD8 antibodies (1:1,000) or of 2 μg/ml CD137L-mCD8 as indicated. Unpulsed cells were also plated and used as the basal control growth. The plates were incubated at 37°C for 96 hr and the metabolic activity of live cells was measured as chemical reduction of Alamar blue. Data represent the mean of percent survival obtained from 3 independent experiments. Asterisk, p < 0.05; double asterisk, p < 0.01; triple asterisk, p < 0.001 vs. control medium; one-way ANOVA, Tukey test.

Figure 10. Protection by CD137/CD137L ligation on apoptotic DNA fragmentation induced by anticancer drugs in T and B leukemia cells. Leukemic cells (10,000 cells/sample) were lysed and the cell lysates were assayed in a specific quantitative photometric enzyme immunoassay to determine cytoplasmic histone-associated DNA fragments induced by cell death. In (a) and (b), cells were incubated at 37°C for 90 min with doxorubicin, 200 nM for CEM cells, 250 nM for Raji cells and 250–500 nM for Jurkat cells. After the pulse, cells were washed and plated in the presence of anti-CD137, anti-CD137L or anti-CD8 antibodies (1:1,000) for CEM and Raji cells or 2 μg/ml CD137L-mCD8 and 2 μg/ml CD137-hIg, previously attached to plastic well, for Jurkat cells. CEM cells were then cultured for 24 hr and Raji and Jurkat cells for 48 hr. In (c), Raji cells were pretreated or not with polyclonal anti-CD137, anti-CD137L antibody (1:1,000) or 2 μg/ml CD137L-mCD8 or 2 μg/ml CD137-hIg, previously attached to plastic well, overnight. Then 1 μM camptothecin was added to the cell culture for 4 hr. All data are presented as the absorbance value subtracted from the specific background and are the mean of triplicate well ± SEM of 1 representative experiment out of 3. Double asterisk, p < 0.01; triple asterisk, p < 0.001 vs. control medium; one-way ANOVA, Tukey test.

DISCUSSION

The costimulatory molecules, CD137 and CD137L, are important activating and survival factors for T lymphocytes and antigen presenting cells. Through bidirectional signals, this receptor/ligand pair is essential for induction, amplification and maintenance of immune response.4, 8, 13 Leukemia cells have maintained expression of these molecules being CD137-inducible in T leukemia cell lines and CD137L constitutively expressed by B lymphomas.18, 19, 20 Now we report a constitutive coexpression of CD137 and CD137L molecules on both T and B leukemia cells. Interestingly, tumor CD137 and CD137L molecules can be indifferently used by leukemia cells to transduce proliferative and antiapoptotic signals.

Accumulating data indicate that CD137 is not only a costimulatory molecule restricted to T lymphocytes but that almost all adaptive and innate immune cells as well as nonimmune cells provide this receptor.9, 10, 11, 12 However, all the molecules reported as CD137 may not have identical molecular structures and features. It is well known that several members of the TNF receptor/ligand superfamily show a multiplicity of isoforms and multimeric aggregation states26 and much evidence suggests a variety of CD137 receptors, too. Differences in protein molecular weight and recognition by specific monoclonal anti-CD137 antibodies between CD137 expressed in transfected SF-1 insect cells and activated T lymphocytes have been reported.20 CD137 protein appears in monomeric and dimeric forms20 and has glicosylation sites.9 In addition, several CD137 mRNA transcripts are expressed in activated T lymphocytes and further additional isoforms have been described in human activated chondrocytes.27 Another open question is the requirement of cell activation for CD137 membrane expression. In almost all cells expressing the receptor, including lymphocytes and leukemic T cells,19, 20 activation seems to be mandatory but recently constitutive CD137 expression on neutrophils, monocytes and dendritic cells has been reported.11, 12, 28 However, conflicting data are present in the literature.11, 29 Now we report that CD137 is constitutively expressed by T and B leukemia cells and activation of leukemic T cells is associated with an increase in receptor-binding sites and appearance of a peculiar epitope. In unstimulated T and B leukemia cells, the RT-PCR studies revealed the transcription of a consistent region of CD137 gene, including those codifying for the extracellular and transmembrane domains. Membrane CD137 protein was recognized by CD137L-mCD8 and the polyclonal anti-CD137 antibody. These reagents were highly specific for human CD137 molecule as demonstrated by the complete inhibition of their binding to activated CEM cells when neutralized with CD137-hIg. However, exclusively activated T leukemia cells, strongly for CEM cells and weakly for Jurkat cells, were recognized by the monoclonal anti-CD137 antibody. Kinetic studies suggest that polyclonal and monoclonal anti-CD137 antibodies mark the same molecular structures in activated CEM cells in view of the similar identification in binding site amount modification following activation, but the 2 antibodies recognize different epitopes on CD137 protein. In fact, the polyclonal anti-CD137 antibody was not able to prevent the monoclonal anti-CD137 antibody binding toward the recombinant protein CD137-hIg in an ELISA assay as well as toward CD137 expressed on activated CEM in FACS analysis and vice versa (data not shown). Therefore, we assume that the epitope bound by the monoclonal anti-CD137 antibody is associated with a peculiar CD137 glicosylation or, hidden in membrane of unstimulated cells, becomes recognizable only following conformational protein changes due to activation. The data we obtained indicate the variability of CD137 receptor(s) responsible for the mutability in its detection, likely explaining the contradicting data on CD137 expression present in the literature.

Constitutive CD137L expression was found not only on B lymphomas as widely described in the literature,19, 21 but also in a subset of T leukemia CEM and Jurkat cell population. CD137L protein identification was done not only with specific antibodies but also through binding of the physiologic ligand CD137 protein. Finding of specific CD137L mRNA in activated or unstimulated CEM and Jurkat cells confirms CD137L expression in T leukemia cells. Although not described previously, these data are not dissonant from the CD137L detectability observed on T lymphocytes.17 In addition, the CD137L membrane protein decrease after PMA/ionomycin activation of CEM and Jurkat cells may be due to CD137L shedding, a phenomenon already described for activated T lymphocytes.17 Activation of shedding or transcription of soluble CD137L isoforms could therefore explain the unexpected increase in CD137L mRNA transcripts observed in activated CEM cells.

Of extreme interest is the ability of the constitutive tumor CD137 and CD137L molecules to transduce signals in spite of their low expression level and peculiarities. Similar to the physiologic responses triggered in activated T lymphocytes, tumor CD137 promotes proliferation and survival in T as well as in B leukemia cells. If the CD137L-induction of positive signals for B lymphoma cell growth could be easily supposed, it was unexpected in T leukemia cells representing a substantial difference between normal and tumor T cells. In fact, while in T lymphocytes CD137L triggering induces apoptosis17 and inhibits anti-CD3 antibody-mediated proliferation,18 in leukemic T cells CD137L has been adapted to be used as a growth promoter. In addition, the proliferative/survival signals mediated by both CD137 and CD137L go toward the same directions since no reciprocal inhibition but at least additive effects were observed when they were triggered simultaneously. Ligation of both CD137 and CD137L enhanced cell viability in T and B leukemia exposed to anticancer drugs such as doxorubicin and camptothecin. Increase in cell proliferation and apoptosis inhibition was able to reduce the cytotoxic effects turned on by these drugs. Moreover, it has been reported that subcytotoxic concentrations of doxorubicin and bleomycin enhances CD137 expression in activated CEM cells,30 amplifying consequently the antiapoptotic signals generated through this receptor. Therefore, coexpression of CD137/CD137L receptor ligand pair may contribute to resistance to cytotoxic chemotherapy in leukemia cells. Besides, another member of TNF receptor superfamily, CD40, showed protective antiapoptotic effects in non-Hodgkin's lymphoma cell lines exposed to doxorubicin.31 Although the role of leukemia CD137/CD137L system in vivo is unknown, the high amounts of soluble CD137L in sera of patients with hematologic malignancies21 may support an involvement of these molecules in such diseases. In addition, the CD137/CD137L receptor ligand pair might act as a tumor growth promoter and survival factor not restricted to hematologic tumors. Functional CD137L expression in colon carcinoma cell lines,24 CD137 expression on lung carcinoma and hepatoma cells9 and enhanced frequency of CD137-expressing blood vessels in malignant tumors32 have been described. In addition, a downregulation of 4-1BB, together with other inhibitors of apoptosis, has been observed in mouse mammary adenocarcinoma cells after therapeutic treatment with protein kinase inhibitor or inhibition of nuclear factor kB activation.33

In conclusion, the coexpression of CD137/CD137L receptor/ligand pair may confer an advantage to both T and B leukemias promoting survival, sustaining cellular growth and contributing to drug resistance.