Malignant and Reactive Cells from Human Lymphomas Frequently

Malignant and reactive cells from human lymphomas frequently express Fas ligand but display a different sensitivity to Fas-mediated apoptosis L Xerri1,2, E Devilard1,2, J Hassoun1, P Haddad3 and F Birg2

1Department of Pathology, Institut Paoli-Calmettes; 2INSERM U 119; and 3Immunotech SA, Marseille, France

Fas ligand (FasL) is capable of inducing apoptosis of lymphoid to transduce an apoptotic signal.11 gld mice harbour a point cells by cross-linking with its natural receptor, Fas. We aimed mutation in the extracellular domain of FasL;12–14 such a to investigate the possible role of the Fas/FasL-mediated apoptosis in the development of human lymphomas. FasL mutation abolishes the binding between FasL and its receptor. mRNA was detected by reverse transcriptase-polymerase chain Mice carrying a homozygous mutation for either Ipr or gld reaction in 38 out of 63 lymphoma biopsy specimens represen- develop similar phenotypes, characterized by massive lym- tative of various subtypes of non-Hodgkin’s lymphoma (NHL) phadenopathy, splenomegaly, B cell activation and hyper- and Hodgkin’s disease. FasL was co-expressed with Fas mRNA gammaglobulinemia.11–14 The human counterpart of Ipr mice in most cases. Flow cytometry (FACS) analysis showed a bright is the recently described auto-immune lymphoproliferative FasL staining in 31% to up to 75% of the total cell population syndrome, caused by dominant Fas mutations.15,16 from 14 out of 16 samples; the presence of the FasL protein 17,18 was confirmed by Western blotting. Dual-color FACS analysis As previously reported, Fas is expressed by malignant showed that FasL was expressed by T cells in B-NHLs and T- cells in various lymphoid malignancies, such as B cell non- NHLs. A significant percentage of B cells in various B-NHLs Hodgkin’s lymphomas (NHLs) derived from germinal centers, also stained positively for FasL. Freshly separated neoplastic ؉ ؊ Hodgkin’s disease (HD), anaplastic large cell lymphomas and B cells from three FasL and one FasL B-NHLs displayed a peripheral T cell lymphomas PTL). Neoplastic lymphoid cells relative resistance to Fas-mediated apoptosis, when compared expressing high Fas levels can be sensitive to Fas-mediated to reactive T cells isolated from the same tissue samples. In 5,6 contrast, the sensitivity to Fas-mediated killing of the T cells apoptosis. The most striking example is the observation that (isolated from two FasL؉ T-NHLs was not uniform. These data a single injection of anti-Fas monoclonal antibodies (mAb show that (1) FasL is expressed in both neoplastic and reactive into nu/nu or SCID mice carrying lymphoma xenotransplants, cells from a significant proportion of lymphoma cases, and (2) 19 ؉ ؉ induces tumor eradication. The Fas/FasL system is also that the intratumoral FasL /Fas reactive T cells are more involved in chemotherapy-induced apoptosis.20 Taken sensitive to Fas-induced apoptosis than the neoplastic -FasL؉/Fas؉ malignant B cells. A putative defect in the Fas/FasL together, these observations suggest that the intra-tumoral lev pathway may thus favor the development of malignant B cell els of FasL, the natural Fas activator, may be a crucial para- populations. meter for the regulation of in vivo lymphoma growth. Keywords: human lymphoma; Fas expression; Fas ligand Expression of FasL in human lymphomas has only been expression; apoptosis; RT-PCR; cell sorting reported so far in the particular case of natural killer-type malignant cells, which express FasL on their surface and release it in the circulation, thereby contributing to specific Introduction tissue alterations in those patients.21 To our knowledge, a comprehensive analysis of the status of FasL expression in Fas/CD95 and its ligand (FasL) are respectively members of fresh human lymphoma tissues has not yet been reported. One two superfamilies of complementary receptors and ligands may hypothesize that FasL production by the malignant cell which play important roles in the regulation of the immune component might result in the elimination of Fas-positive system (see Ref. 1 for review). Fas, also called APO-12,3 is a reactive T cells, and could potentially favor tumor growth. type I membrane protein of 45 kilodaltons (kDa) belonging to Another question to address is the putative interaction the tumor necrosis factor receptor (TNFR) superfamily.4 between reactive T cells expressing FasL and neoplastic B Among these receptors, Fas and the p55 TNFR only share a cells expressing Fas, which might antagonize tumor develop- 70 aminoacid intracellular ‘death domain’ which transduces ment. Because of the role of Fas in receiving and transducing signals for cell death.4 Fas mediates apoptosis when cross- the FasL-induced apoptotic signal, the respective sensitivities linked with agonistic anti-Fas antibodies.5,6 FasL is a 40 kDa of B cell and T cell populations from lymphoma tissues to Fas- type II membrane protein expressed as a membrane-bound induced apoptosis were also investigated. This report shows form, which is then proteolytically processed into a soluble that FasL is expressed in both the neoplastic and reactive cell cytokine retaining biological activity.7–10 population from a significant proportion of lymphoma cases, There is accumulating evidence from Ipr and gld mice mod- and that Fas-positive malignant B cells are less sensitive to els that Fas/FasL-mediated apoptosis is a crucial regulatory Fas-mediated apoptosis than intratumoral Fas-positive reactive process which limits immune response and provides a safe- T cells. guard against emergence of auto-reactive lymphocytes.11–14 Ipr mutant mice harbor either an early transposable element or a mutation in the Fas gene, which either severely reduce Materials and methods the expression of a full size Fas transcript, or abolish its ability Tissue sampling and study design

A series of 63 tumor biopsy specimens representative of the Correspondence: L Xerri, De´partement de Pathologie, and INSERM U 119, Institut Paoli-Calmettes, 232 Boulevard de Sainte Marguerite, BP various subtypes of human lymphomas, including NHL and 156, 13273 Marseille Cedex 9, France HD, was retrospectively analyzed for the expression of FasL Received 13 January 1997; accepted 1 July 1997 and Fas mRNAs using reverse transcriptase-polymerase chain Fas ligand in human lymphomas L Xerri et al 1869 reaction (RT-PCR). Tumors were obtained from untreated Table 1 RT-PCR analysis of Fas and FasL expresssion in NHL patients. The distribution of cases according to the revised and HD European–American classification of lymphoid neoplasms (REAL)22 is detailed in Table 1. A portion of each sample was Sample Type FasL Fas submitted to conventional histopathological processing and standard immunophenotyping; the other part was snap frozen B cell NHLs − ° 6300 Follicular (Low grade) + 0 in liquid nitrogen and stored at 80 C until use. The presence 6461 0 0 of lymphoma lesions was checked on frozen sections prior to 6603 0 + storage. A control panel of benign reactive lymphoid tissues 6627 +++ + (n=3) was simultaneously analyzed. Fas expression was 6634 +++ evaluated simultaneously by RT-PCR and immunohistochemi- 6631 ++ + stry. In addition, fresh lymphoma cells from a representative 4098 Lymphocytic/plasmacytic 0 0 panel of 16 tumors were analyzed for Fas and FasL expression 4189 (Low grade) 0 + by flow cytometry (FACS) and Western blotting. Six out of 4806 ++ + these 15 samples were analyzed for their sensitivity to 5131 +++ + Fas-mediated apopotosis on separated B cell and T cell 6262 ++ 0 populations. 5145 ++ 0 5148 ++ +++ 5185 + +++ 5305 0 0 RT-PCR analysis of FasL and Fas expression 5331 Mantle zone 0 0 Total RNAs were prepared from frozen tissues by lysis in 5509 + 0 guanidium isothiocyanate. Reverse transcription was perfor- 5623 0 0 med as described previously,23 using 1 ␮g of poly(A)+ RNA ␮ 4355 Diffuse large cell +++ ++ from control cells, and 1 g of total RNA from tissues. One 4370 (Intermediate/High grade) ++ ++ quarter of the cDNA preparation was used for FasL amplifi- 4451 +++ cation using a 20-mere sense oligonucleotide (5′-CTA CAG 4622 0 ++ GAC TGA GAA GAA GT-3′) upstream of the initiation codon, 4624 ++ ++ and a 22-mere antisense oligonucleotide (5′-CAA CAT TCT 4670 ++ ++ ++ ++ CGG TGC CTG TAA C-3′) downstream of the termination 4842 9 4849 ++ + codon. Oligonucleotide primers for Fas were chosen so as to 5133 ++ ++ amplify and distinguish the mature mRNAs encoding sFas and 5184 0 0 Fas, respectively (sense primer, 5′-GAC CCA GAA TAC CAA 5196 +++ 0 GTG CAG ATG TA-3′; antisense primer, 5′-CTG TTT CAG 5366 0 + GAT TTA AGG TTG GAG ATT-3′).24 Integrity of RNA samples 5595 + 0 was checked by electrophoresis of an aliquot on denaturing 5201 Burkitt’s + 0 agarose/formaldehyde gels and by performing control ampli- ␤ ′ 5254 0 0 fication for human -actin (sense primer, 5 -TAC CAC TGG 5413 0 0 CAT CGT GAT GAC T-3′; antisense primer, 5′-TCC TTC TGC ATC CTG TCG GCA AT-3′).25 The cDNA samples were mixed T cell NHLs with 50 ␮l of a PCR mixture containing reaction buffer and 4938 Anaplastic +++ + 2.5 U of Taq polymerase (Perkin Elmer Cetus, Foster City, CA, 5403 0 5635 ++ USA). The PCR cycle profile for FasL was the following: denat- + ° 5641 0 uration (94 C for 1 min (2 min for the first cycle)), annealing 5693 0 0 (55°C for 2 min), and extension (72°C for 3 min (10 min for the last cycle)). Thirty cycles of amplification were performed. 4566 Pleomorphic ++ ++ PCR products were visualized in ethidium bromide-stained 5203 +++ ++ 1.5% agarose gels. PCR amplification of Fas was performed 5628 0 0 5694 + 0 under the same conditions, except for the annealing tempera- 6264 +++ 0 ture (60°C instead of 55°C). 4725 AIL-like ++ ++

Southern blot analysis 4854 Lymphoblastic + 0 5333 0 0

Gel electrophoresis-separated PCR products obtained after 5225 Mycosis F. 0 0 FasL and Fas amplification were transferred to Hybond N+ membranes (Amersham, Buckinghamshire, UK) as rec- Hodgkin’s disease ommended by the manufacturer. Membranes were prehy- 3582 NS 0 + bridized for 2 h at 45°Cin5×SSC containing 0.1% laurylsar- 3729 NS 0 0 ++ ++ cosin, 0.02% SDS and 0.5% blocking reagent (Boehringer 3833 NS 3993 MC 0 ++ Mannheim, Mannheim, Germany), and hybridized overnight + ° 4088 NS 0 at 45 C in the same buffer containing a digoxigenin-labeled 4180 NS 0 + internal oligonucleotide, respectively a 18-mere (5′-CAA ATA 4255 NS ++ ++ GGC CAC CCC AGT-3′, spanning the 3′ end of exon 2 and 4434 NS +++ ++ the 5′ end of exon 3) for FasL, and a 24-mere (5′-ACA TGT Fas ligand in human lymphomas L Xerri et al 1870 Table 1 Continued CA, USA), washed in PBS, and then incubated with a FITC- conjugated anti-rabbit mAb (Jackson Immunoresearch, West 5117 NS ++ + Grove, PA, USA) before FACS analysis. Positive controls for 5157 MC +++FasL expression were Jurkat cells activated with PHA 5211 MC ++ ␮ 5213 NS ++ +++ (10 g/ml) for 2 h. Non-activated Jurkat cells were used as a 5231 MC ++ +++ negative control. 5234 NS 0 + The immunophenotype of fresh lymphoma cells were 5300 NS 0 + determined using a panel of conjugated mAbs (Immunotech and Dakopatts, Glostrup, Denmark) directed against differ- AIL, Angio immunoblastic lymphadenopathy; NS, nodular scleros- entation antigens specific for B cells (CD10, CD19, CD20, ing; MC, mixed cellularity. CD21, CD22, CD23, CD24, ␬ and ␭ immunoglobulin light chains, ␦, ␮, ␥, and ␣ immunoglobulin heavy chains), T cells (CD2, CD3, CD4, CD5, CD7, CD8) and Reed– CAT GAA CCC ATG TTT GCA-3′ (nucleotide positions: 2118– Sternberg/activated cells (CD30). For double-color flow cyto- 2142)) for Fas. Oligonucleotide probes were labeled with a 3′ metric analysis, cells were labeled with fluorescently tagged tailing kit (Boehringer Mannheim) according to the supplier’s mAbs recognizing the cell surface molecules, Fas, FasL, CD3 recommendations. After hybridization, filters were washed and CD19 for 30 min at 4°C. They were then washed and twice for 10 min at room temperature in 2 × SSC containing fixed with 1% paraformaldehyde in PBS before FACS analysis. 0.1% SDS, and twice for 10 min at 55°C in 0.1 × SSC contain- A negative control labeling using an isotype-matched FITC- ing 0.1% SDS, then underwent revelation using a chemilumi- conjugated antibody was simultaneously performed. nescence detection kit (Boehringer Mannheim).

Western blot (WB) analysis Purification of B and T cells from fresh lymphoma tissues Frozen tissues crushed in liquid nitrogen, or cell suspensions, were thrown in lysis buffer (10 mM Tris HCl buffer pH 7.4, B cells and T cells from fresh tissue samples, obtained by teas- containing 1% SDS and 1 mM sodium vanadate) and then ing, were separated with magnetic beads (Immunotech) conju- heated in a microwave oven for 10–15 s. Proteins in lysates gated with anti-CD19 and anti-CD3 mouse IgG mAbs. Pur- were separated by 7.5% SDS-PAGE and transferred to Immob- ified populations were then submitted to WB, FACS analysis ilon-P membranes (Millipore, Medford, MA, USA) as rec- and/or apoptotic tests. Isolated B cells were uniformly Ͼ90% ommended by the manufacturer. FasL was detected with a + + + CD20 and Ͻ10% CD2 . Isolated T cells were Ͼ90% CD2 mouse IgG3 mAb, raised against a 18 kDa polypeptide corre- + and Ͻ10% CD19 . The monoclonality of isolated malignant sponding to amino acids 116–277 of human FasL B cell populations was checked by the exclusive detection of (Transduction Laboratories, Lexington, KY, USA). Immunore- ␬ or ␭ light chains (Dakopatts). Sensitivity of purified cells to active material was visualized by enhanced chemilumi- Fas-mediated apoptosis was determined as described below. nescence (Amersham). Jurkat cells stimulated with phytohem- agglutinin A (PHA, 10 ␮g/ml, Sigma, St Louis, MO, USA) served as a positive control. Unstimulated Jurkat cells were used as a negative control. Assessment of Fas-mediated apoptosis

Immunohistochemical analysis Sensitivity of lymphoma cells to Fas-mediated apoptosis was determined in six NHL cases by treatment with the agonistic Fas immunodetection on frozen sections was performed as anti-Fas 7C11 IgM mAb (Immunotech). Purified or total cell previously described18 using the CH-11 mAb (Immunotech, populations were immediately resuspended in RPMI medium Marseille, France). Reactive lymph nodes were used as containing 20% FCS, then incubated with 7C11 (0.05– positive controls. 0.5 ␮g/106 cells) for 1–72 h. The percentage of apoptotic cells was measured by FACS analysis using the anti-7A6 mAb called APO2.7 (Immunotech) as recently described.26 FACS analysis Efficiencies of both 7C11 apoptotic induction and APO2.7 immunodetection were simultaneously controlled on Fas- Surgically removed lymph nodes were immediately pro- expressing Jurkat cells in all the experiments. Peripheral blood cessed. Fresh lymphoma cells were obtained by teasing, lymphocytes (PBL) from three healthy volunteers were also washed, and resuspended in RPMI medium containing 20% used as positive controls for 7C11 induction and APO2.7 fetal calf serum (FCS). For FACS analysis of Fas expression, immunodetection. 7A6 expression was compared with DNA lymphoma cells were incubated with a fluorescein isothiocy- fragmentation detected by the TUNEL method (Mebstain kit; anate (FITC)-conjugated non-apoptosis inducing anti-Fas mAb Immunotech) in case 6631; the results of both methods were (UB2 clone, Immunotech) for 30 min at 4°C. Cells were then found to be correlated (data not shown) as previously demon- washed in phosphate-buffered saline (PBS), fixed with 1% par- strated.26 Negative controls were performed by 7A6 detection aformaldehyde and analyzed on a FACS Star IV flow cyto- on cell populations incubated without 7C11. The relative meter (Becton Dickinson, San Jose, CA, USA). sensitivity/resistance to 7C11-induced cell death of the neo- For detection of surface FasL expression, cells were incu- plastic cell populations in each sample was defined by com- bated for 45 min with affinity-purified rabbit anti-human FasL parison with the reactive cell component isolated from the polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, same tissue sample. Fas ligand in human lymphomas L Xerri et al 1871 Results FasL, as examplified in Figure 3. Fas expression was detected in parallel with FasL expression in most cases (Table 2). FasL and Fas mRNAs are frequently coexpressed in lymphoma tissues WB analysis detects the FasL protein in both The results of RT-PCR experiments are illustrated in Figure 1, neoplastic and reactive cell populations and summarized in Table 1. RT-PCR analysis was repeated at least twice for each sample. The relative differences in ampli- To confirm and extend the RT-PCR and flow cytometry data, fication levels were reproducibly observed in the various WB analysis was performed using a different anti-FasL mAb. experiments, indicating that they reflect differences in mRNA In lysates from total lymphoma tissues, the expected 40 kDa concentrations, rather than random variations. The relative band was detected in the FACS-positive samples (examplified levels of FasL RNA expression could thus be estimated, after in Figure 4), whereas FACS-negative samples gave no signal normalization with control RT-PCR for actin, and ranged from (Table 2). WB was also performed on lysates of B and T cell + (weak signal) to +++ (strong signal). fractions purified from total lymphoma populations. FasL pro- A strong expression of FasL mRNA was observed in acti- tein expression was detected in purified B cells from B-NHL vated Jurkat cells, whereas non-activated cells were negative cases as well as in the reactive T cell populations from the (Figure 2). Positive FasL bands were clearly detectable in 38 same samples (Figure 4 and Table 2). In one HD case, FasL out of the 63 lymphoma specimens. Positive cases were dis- expression was observed in both the CD30+ and CD3+ cell tributed as follows: 7/15 HD cases, 22/34 B cell lymphoma subpopulations (Figure 4). cases and 8/14 T cell NHL cases. The levels of expression, as judged by band intensity, were not correlated with a particular lymphoma subtype (Figure 1 and Table 1). A faint FasL band The neoplastic and reactive cell populations from was observed in one out of three benign reactive lymph nodes lymphoma tissues display different sensitivities to Fas- by RT-PCR. mediated apoptosis Fas mRNA expression was observed in 40 out of the 63 cases analyzed. None of the Fas-positive tumors expressed Apoptosis was evaluated by the immunodetection of the anti- sFas mRNA. Fas and FasL were often co-expressed (Table 1), gen 7A6, which defines a novel epitope on the mitochondrial except for 10 cases which only expressed FasL, and five cases membrane, that appears to be exposed on cells undergoing which only expressed Fas. apoptosis.26 The anti-7A6 positivity on permeabilized cells is strictly correlated with the apoptosis induced in PBL and in lymphoid cell lines.26 Fresh neoplastic B cells from the four FACS analysis detects FasL and Fas proteins in both B cell tumors tested were significantly less sensitive to Fas- neoplastic and reactive cell populations mediated apoptosis than the reactive T cells isolated from the corresponding lymphoma tissues (Figure 5 and Table 3). The 16 lymphoma cases analyzed for surface FasL expression Higher concentrations of 7C11 and/or longer incubation per- showed variable percentages of positive populations, as iods were indeed not sufficient to obtain similar rates of cell detailed in Tables 2 and 3. FACS detection of surface FasL death in the neoplastic B cells and in the reactive T cells correlated well with the results of RT-PCR analysis of lym- (Table 3). The relative resistance of the different cell popu- phoma samples and control Jurkat cells (Figure 2). Two cases lations was not strictly related to Fas expression levels of lymphocytic NHL and one case of follicular NHL were con- (Table 3). The susceptibility of malignant T cells to 7C11- sidered as negative. In the other cases, bright FasL staining induced apoptosis appeared heterogeneous in the two T cell was observed in 31% to up to 75% of the total cell population lymphomas tested. Malignant T cells from one case were (Table 2). Dual-color flow cytometric analysis showed that almost as sensitive as control Jurkat cells. In the other case, FasL was expressed by T cells in B cell NHLs, HD and T cell the neoplastic T cell population displayed slightly higher NHLs (Figure 3). A significant percentage of B cells in various apoptotic rates than the reactive/residual B cell population histologic subtypes of B cell NHLs also stained positively with (Table 3).

Discussion

The Fas/FasL apoptotic pathway was shown to play a key role in the immune system by triggering the activation-induced suicide of T cells (for review see Ref. 27). Resting T cells nor- mally express low levels of Fas; this expression, however, increases significantly within hours of engagement of their antigen receptor.2,3 Several days after the initial activation, this pathway becomes effective, and T cells undergo apoptosis via Figure 1 RT-PCR detection of FasL transcripts in lymphoma cross-linking of the FasL concomitantly upregulated in acti- tissues. cDNAs from biopsy samples of various NHL and HD subtypes vated T cells.28,29 This key regulating function is underlined (lanes 1–25) were independently amplified using FasL and actin pri- by the prevention of graft rejection and other dangerous mers. Aliquots of amplified samples were submitted to electrophoresis inflammatory responses due to the elimination of Fas-positive in a 1.5% agarose gel containing ethidium bromide. The intensity of infiltrating T cells by FasL.30,31 It appears from recent reports the positive FasL bands (845 base pairs) varies from +, as in lanes 1 and 5 (corresponding to Burkitt’s NHL and mantle zone NHL samples, that FasL may also play a role in the homeostasis with B respectively) to +++, as in lanes 10, 12 and 14 (corresponding to one cell population.32,33 HD and two PTL samples, respectively). In the present report we show that FasL is expressed by both Fas ligand in human lymphomas L Xerri et al 1872

Figure 2 Correlation between FACS and RT-PCR analysis of FasL expression in Jurkat cells. Jurkat cells were successively incubated with anti-FasL polyclonal antibody, and FITC-conjugated anti-rabbit antibody and analyzed by flow cytometry. FasL immunodetection performed on non-activated Jurkat cells was negative (upper left). The positive control corresponds to labeling of the same population (upper right) with the anti-␤2 microglobulin mAb B1G6 (Immunotech). In comparison, significant FasL expression was observed after PHA activation of Jurkat cells (bottom left). The corresponding RT-PCR analysis is shown on the bottom right: FasL mRNA was detected in activated Jurkat cells (second left lane). Non-activated Jurkat cells were negative for FasL (third left lane) and positive for the actin control (first right lane). First left lane contains the molecular weight marker.

neoplastic and reactive cell components in the majority of appears unlikely; we could not detect sFas encoding tran- human lymphoma tissues, and that FasL can be coexpressed scripts in any of the lymphoma samples analyzed. We determ- with Fas on malignant cells. Thus, intratumoral triggering of ined the levels of Bcl-2 expression by RT-PCR in these NHL Fas via FasL probably occurs in lymphoma cells in vivo.We cases, and found no correlation with the profiles of Fas/FasL also report here the application of a new method for monitor- expression, ie coexpression was not associated with high lev- ing apoptosis in fresh cells isolated from human tissues, and els of Bcl-2 (data not shown). However, this does not rule out show that malignant and reactive cells from B cell NHLs the possibility of an upregulation of the Bcl-2 protein. Simi- display a different sensitivity to Fas-mediated apoptosis. larly, we cannot exclude the influence of another anti-apop- It has been previously demonstrated that the level of Fas totic factor downstream of Fas, like Bcl-xL, which we pre- expression in lymphoid malignancies is not always correlated viously found to be hyperexpressed in human lymphomas.36 with anti-Fas-mediated cell death.34,35 We extend these earlier Another possibility could be that lymphoma cells are not results by showing that FasL expressing B cells from lym- sensitive to Fas triggering because of interfering activation sig- phoma tissues are much more resistant to Fas triggering than nals. The activation state of B cells is indeed critical for their the corresponding intratumoral reactive T cells. This differ- sensitivity to Fas.37 Under certain circumstances, Fas was ence in apoptosis susceptibility is likely to explain why the shown to deliver a positive activation signal, resulting in cell Fas/FasL pathway does not allow for an effective autocrine proliferation.38 One can also speculate that Fas/FasL interac- and/or paracrine regulation of tumor growth. The tion is not an impediment for the development of malignant mechanism(s) responsible for insensitivity to Fas-mediated lymphoma because the primary structure of Fas is altered, thus killing remain unclear. The possibility that the effects of FasL rendering it inactive. Fas structural defects were identified in might be neutralized by the production of soluble Fas (sFas) some Fas-resistant cell lines which harbor a truncated Fas Fas ligand in human lymphomas L Xerri et al 1873 Table 2 Correlations between FACS and Western blot analysis of FasL expression in NHL and HD

Case No.a/Diagnosisb

5203 5305 6627 4670 4434 4451 4355 5196 6603 6300 PTL CLL FL DLCL HD DLCL DLCL DLCL FL FL

Phenotypec CD3 58% 4% 27% 39% 75% 5% 21% 33% 34% 14% CD19 21% 94% 68% 53% ND 94% 79% 60% 84% 74% CD30 ND ND ND ND 21% ND ND ND ND ND

FasL expression (FACS analysis) Total population 75% 2% 73% ND ND 37% ND 65% 3% 31% CD3 cells 59% ND 12% 12% ND ND 67% 40% 1% 27% CD19 cells 22% ND 69% 21% ND ND 45% 23% 0% 13%

FasL expression (Western blot analysis) Total population + − +++ + +++ + +++ ++ − +++ CD3+ cells ND ND ND ND ++ ND ++ + ND + CD19+ cells ND ND ND ND ND ++ND ++ CD30+ cells +

Fas expression (immunohistochemistry) Malignant cells +−+++++−+− Reactive cells + − + + ++ ++ + ++ + + aRefers to Table 1. bPTL, peripheral T cell lymphoma; CLL, lymphocytic lymphoma/chronic lymphocytic leukemia; FL, follicular non-Hodgkin’s lymphoma; DLCL, diffuse large B cell lymphoma; HD, Hodgkin’s disease; ND, not done. c% of positive cells.

Table 3 Fas-induced apoptosis in isolated B and T cell populations from fresh NHL samples, as determined by the anti-7A6 immunodetection

Casea/Diagnosisb Fas-induced apoptosis (% of 7A6-positive cells) Fas expression (%) FasL expression (%)

7C11 24 h 48 h B cells T cells B cells T cells (ng/106 cells) B cells T cells B cells T cells

6631/FL 0 3 9 1 14 37 56 33 40 10 4 37 ND 60 100 4 52 15 71 6634/FL 0 6 5 17 8 39 81 19 69 10 ND ND 23 50 100 14 13 20 71 4842/DLCL 0 11 1 11 7 18 79 22 59 10 ND 40 10 39 100 16 71 13 82 4098/CLL 0 13 no T cells 12 no T cells 6 no T cells 2 no T cells 10 17 no T cells 11 no T cells 100 15 no T cells 13 no T cells 4725/PTL 0 no B cells 23 no B cells 8 no B cells 83 no B cells 41 10 no B cells ND no B cells ND 100 no B cells 64 no B cells 75 4566/PTL 0 1 2 7 10 19 76 8 81 10 ND ND ND ND 100 22 9 49 35 Jurkat cells 0 5 2 Ϸ100 Activation 10 63 95 +− 100 78 97 Ϸ100 Ϸ0 aRefers to Table 1. bRefers to footnotes for Table 2. molecule.39 None the less, our previous search for rearrange- unlikely, as a recent description of the human auto-immune ments and/or deletions of the Fas gene in a large series of fresh lymphoproliferative syndrome showed that such mutations lymphoma samples did not reveal a frequent occurrence of either result in lymphoid hyperplasia in childhood, or such abnormalities.40 The presence of point mutations of the remain silent.15,16 Fas gene in lymphomas cannot be excluded, but appears What might be the relevance of in vivo expression for FasL Fas ligand in human lymphomas L Xerri et al 1874

Figure 3 FACS analysis of FasL expression in fresh lymphoma cell populations. The upper panel (T-CELL LYMPHOMA) shows the surface expression of FasL in a case of PTL, detected by single-color ﬂow cytometry on the total cell population (top). Double-color ﬂow cytometry (bottom) revealed that FasL was present on both malignant T cells and residual/reactive B cells. The lower panel (B-CELL LYMPHOMA) is a double-color FACS analysis performed on separated cell populations from a follicular NHL case, showing FasL expression on both malignant and reactive cell compartments, CD19+ B cells and CD3+ T cells, respectively.

by neoplastic cells of the T and B lineages? The expression of recently suggested in extra-lymphoid tumors, like colon can- FasL in malignant B cells which we report here may simply cer cell lines,41 metastatic melanoma42 and hepatocellular reflect the occurrence of FasL expression in some normal acti- carcinoma.43 Our observation that double positive FasL+/Fas+ vated B cells.33 Alternatively, FasL expressing lymphoma B reactive T cells, but not FasL+/Fas+ malignant B cells, are cells could eliminate Fas expressing reactive T cells and by efficiently killed by Fas triggering is evocative of an in vivo this means contribute to tumoral development. Existence of role of FasL in the immune escape of lymphoid tumors of B local immunosuppression due to tumor-derived FasL was cell origin. Apoptosis-resistant FasL+/Fas+ lymphoma cells Fas ligand in human lymphomas L Xerri et al 1875 could kill surrounding Fas+ reactive T cells, while being pro- tected from their own suicide. Such a functional scheme is reminiscent of the role established for FasL in the mechanism of immune previlege. Immune privileged tissues are particular anatomic sites, like the testis and eye, which express FasL to prevent dangerous inflammatory responses by killing infiltrating T cells.31 This hypothesis seems attractive but the function of FasL is probably more complex in the pathogenesis of + Figure 4 WB analysis of FasL expression in separated cell popu- NHLs, as the results reported here indicate that Fas malignant lations from NHLs and HD. Western blot was performed as described T cells display a heterogeneous sensitivity to apoptosis. in Materials and methods. B-NHL (1) and B-NHL (2) represent follicu- Although this observation needs to be confirmed in a larger lar B-NHL samples which displayed a positive FasL signal (40 kDa series, it may result from different states of T cell activation, band) in the total cell population, as well as in sorted CD3+ and + as activation is known to interfere with T cell resistance to Fas- CD19 cell fractions. FasL was also immunodetected in the different 44 cell populations isolated from a HD sample (HD), including a weak mediated death. The fact that inhibition of FasL expression 20 signal in CD30+ cells. Human placenta was chosen as a positive con- inhibits chemotherapy-induced apoptosis, supports the view trol as this tissue exhibits the highest levels of FasL expression among that high intra-tumoral levels of FasL may be unfavorable the human organs tested.45 to lymphoma growth under certain circumstances. FasL expression is induced upon treatment with anticancer drugs,

Figure 5 Monitoring of Fas-induced apoptosis using the anti-7A6 mAb. Jurkat cells, and freshly separated cells from a sample of FasL+/Fas+ follicular B cell NHL were incubated with variable amounts of the apoptotic inducer anti-Fas mAb 7C11, for increasing lengths of time. Treatment with increasing doses of 7C11 mAb induced a rapid and dose-dependent 7A6 positivity on Jurkat cells (upper panel), indicative of high sensitivity to Fas-mediated death. Fresh reactive T cells (middle panel) displayed significant, although slightly lower apoptotic rates following 7C11 triggering. In contrast, neoplastic B cells (lower panel) were resistant to apoptosis under the same conditions, as suggested by the lack of significant 7A6 positivity following 7C11 triggering. Fas ligand in human lymphomas L Xerri et al 1876 and it mediates apoptosis by either fratricide (or autocrine) 17 Mo¨ller P, Henne C, Leithaüser F, Eichelman A, Schmidt A, Bru¨d- and/or paracrine death.20 FasL therefore appears as a double- erlein S, Dhein J, Krammer PH. Coregulation of the APO-1 antigen faced agent, being capable of both protecting and destroying with intercellular adhesion molecule-1 (CD54) in tonsillar B cells and coordinate expression in follicular center B cells and in fol- lymphoma cells. However, the parameters modulating its licle center and mediastinal B-cell lymphomas. Blood 1993; 81: effects have to be further understood before a therapeutic 2067–2075. strategy can be developed. 18 Xerri L, Carbuccia N, Parc P, Hassoun J, Birg F. Frequent expression of FAS/APO-1 in Hodgkin’s disease and anaplastic large cell lymphomas. Histopathology 1995; 27: 235–241. Acknowledgements 19 Coney LR, Daniel PT, Sanborn D, Dhein J, Debatin KM, Krammer PH, Zurawski VR. Apoptotic cell death induced by a mouse- human anti-APO-1 chimeric antibody leads to tumor regression. This work was supported by INSERM, Institut Paoli Calmettes, Int J Cancer 1994; 58: 562–567. and grants from The ‘Fe´de´ration Nationale des Centres de 20 Friesen C, Herr I, Krammer PH, Debatin KM. Involvement of CD95 Lutte Contre le Cancer’, and the ‘Comite´ des Bouches du (Apo-1/Fas) receptor/ligand system in drug induced apoptosis in Rhoˆne de la Ligue Contre le Cancer’. We thank C Mawas and leukemia cells. Nature Med 1996; 5: 574–577. M Hirn for critical comments on the manuscript. 21 Tanaka M, Suda T, Haze K, Nakamura N, Sato K, Kimura F, Motoyoshi K, Mizuki M, Tagawa S, Ohga S, Hatake K, Drummond H, Nagata S. Fas Ligand in human serum. Nature Med 1996; 2: 317–322. References 22 Lee Harris N, Jaffe ES, Stein H, Banks PM, Chan JKC, Cleary ML, Delsol G, De Wolf-Peeters C, Falini B, Gatter KC, Grogan TM, 1 Nagata S, Golstein P. The Fas Death Factor. Science 1995; 267: Isaacson PG, Knowles D, Mason DY, Muller-Hermelink H-K, 1449–1456. Pileri SA, Piris MA, Ralfkiaer E, Warnke RA. A revised European– 2 Oehm A, Behrmann I, Falk W, Pawlita M, Maier G, Klas C, Li- American classification of lymphoid neoplasms: a proposal from Weber M, Richards S, Dhein J, Trauth BC, Ponsting H, Krammer the International Lymphoma Study Group. Blood 1994; 84: P. Purification and molecular cloning of the APO-1 cell surface 1361–1392. antigen, a member of the tumor necrosis factor/nerve growth factor 23 Xerri L, Mathoulin MP, Birg F, Bouabdallah R, Stoppa AM, Has- receptor superfamily. J Biol Chem 1992; 267: 10709–10715. soun J. Heterogeneity of rearranged T-cell receptor V-␣ and V-␤ 3 Itoh N, Yonehara S, Ishii A, Yonehara M, Mizushima S, Sameshima transcripts in tumor-infiltrating lymphocytes from Hodgkin’s dis- M, Hase A, Seto Y, Nagata S. The polypeptide encoded by the ease and non-Hodgkin’s lymphoma. Am J Clin Pathol 1994; 101: cDNA for human cell surface antigen Fas can mediate apoptosis. 76–80. Cell 1991; 66: 233–243. 24 Mori S, Murakami-Mori K, Jewett A, Nakamura S, Bonavida B. 4 Armitage RJ. Tumor necrosis factor receptor superfamily members Resistance of AIDS-associated Kaposi’s sarcoma cells to Fas- and their ligands. Curr Biol 1994; 6: 407–413. mediated apoptosis. Cancer Res 1996; 56: 1874–1879. 5 Debatin KM, Goldman CK, Bamford R, Waldmann TA, Krammer 25 Ponte P, Ng SY, Engel J, Gunning P, Kedes L. Evolutionary conser- PH. Monoclonal antibody-mediated apotosis in adult T cell leukemia. Lancet 1990; 335: 497–500. vation in the untranslated regions of actin mRNAs: DNA sequence 6 Debatin KM, Goldman CK, Waldmann TA, Krammer PH. APO-1 of a human beta-actin cDNA. Nucleic Acids Res 1984; 12: induced apoptosis of leukemia cells from patients with adult T- 1687–1696. cell leukemia. Blood 1993; 81: 2972–2977. 26 Zhang C, Ao Z, Seth A, Schlossman SF. A mitochondrial mem- 7 Suda T, Takahashi T, Golstein P, Nagata S. Molecular cloning and brane protein defined by a novel monoclonal antibody is preferen- expression of the Fas ligand, a novel member of the tumor necrosis tially detected in apoptotic cells. J Immunol 1996; 157: 3980– factor family. Cell 1993; 75: 1169–1178. 3987. 8 Suda T, Nagata S. Purification and characterization of the Fas 27 Steller H. Mechanisms and genes of cellular suicide. Science ligand that induces apoptosis. J Exp Med 1994; 179: 873–878. 1995; 267: 1445–1449. 9 Tanaka M, Suda T, Takahashi T, Nagata S. Expression of the func- 28 Alderson MR, Tough TW, Davis-Smith T, Braddy S, Falk B, tional soluble form of human Fas ligand in activated lymphocytes. Schooley KA, Goodwin RG, Smith CA, Ramsdell F, Lynch DH. EMBO J 1995, 14: 1129–1135. Fas ligand mediates activation-induced cell death in human T lym- 10 Takahashi T, Tanaka M, Inazawa J, Abe T, Suda T, Nagata S. phocytes. J Exp Med 1995; 181: 71–77. Human Fas ligand: gene structure, chromosomal localization and 29 Suda T, Okazaki T, Naito Y, Yokota T, Arai N, Ozaki S, Nakao species specificity. Int Immunol 1994; 6: 1567–1574. K, Nagata S. Expression of the Fas Ligand in cells of T cell lineage. 11 Watanabe-Fukunaga R, Brannan CI, Copeland NG, Jenkins NA, J Immunol 1995; 154: 3806–3813. Nagata S. Lymphoproliferation disorder in mice explained by 30 Bellgrau D, Gold D, Selawry H, Moore J, Franzusolf A, Duke RC. defects in Fas antigen that mediates apoptosis. Nature 1992; 56: A role for CD95 ligand in preventing graft rejection. Nature 1995; 314–317. 377: 630–632. 12 Lynch DH, Watson ML, Alderson MR, Baum PR, Miller RE, Tough 31 Griffith TS, Brunner T, Fletcher SM, Green DR, Ferguson TA. Fas T, Gibson M, Davis-Smith T, Smith CA, Hunter K, Bhat D, Din W, ligand-induced apoptosis as a mechanism of immune previlege. Goodwin RG, Seldin MF. The mouse Fas-ligand gene is mutated in Science 1995; 270: 1189–1192. gld mice and is part of a TNF family gene cluster. Immunity 1994; 32 Rathmell JC, Cooke MP, Ho WY, Grein J, Townsend SE, Davis 1: 131–136. MM, Goodnow CC. CD95 (Fas)-dependent elimination of self- + 13 Ramsdell F, Seaman MS, Miller E, Tough TW, Alderson MR, Lynch reactive B cells upon interaction with CD4 T cells. Nature 1995; DH. gld/gld mice are unable to express a functional ligand for Fas. 376: 181–184. Eur J Immunol 1994; 24: 928–933. 33 Hahne M, Enno T, Schroeter M, Irmler M, French L, Bornand T, 14 Takahashi T, Tanaka M, Brannan CI, Jenkins NA, Copeland NG, MacDonald HR, Tschopp J. Activated B cells express functional Suda T, Nagata S. Generalized lymphoproliferative disease in FasL ligand. Eur J Immunol 1996; 26: 721–724. mice, caused by a point mutation in the Fas ligand. Cell 1994; 34 Owen-Schaub LB, Meterissian S, Ford RJ. Fas-Apo-1 expression 76: 969–976. and function on malignant cells of hematologic and non hematol- 15 Fisher GH, Rosenberg FJ, Straus SE, Dale JK, Middelton LA, Lin ogic origin. J Immunother 1993; 14: 234–241. AY, Strober W, Lenardo MJ, Puck JM. Dominant interfering Fas 35 Westendorf JJ, Lammert JM, Jelinek DF. Expression and function gene mutations impair apoptosis in a human autoimmune of Fas (APO-1/CD95) in patient myeloma cells and myeloma cell lymphoproliferative syndrome. Cell 1995; 81: 935–946. lines. Blood 1995; 85: 3566–3576. 16 Rieux-Laucat F, Le Deist F, Hivroz C, Roberts IAG, Debatin KM, 36 Xerri L, Parc P, Brousset P, Schlaifer D, Hassoun J, Reed JC, Kra- Fischer A, de Villartay JP. Mutations in Fas associated with human jewski S, Birnbaum D. Predominant expression of the long isoform lymphoproliferative syndrome and autoimmunity. Science 1995; of BCL-X (BCL-XL) in human lymphomas. Br J Haematol 1996; 268: 1347–1349. 92: 900–906. Fas ligand in human lymphomas L Xerri et al 1877 37 Krammer PH, Behrmann I, Daniel P, Dhein J, Debatin KM. Regu- implications for tumor immune escape. Science 1996; 274: lation of apoptosis in the immune system. Curr Opin Immunol 1363–1366. 1994; 6: 279–284. 43 Strand S, Hofmann WJ, Hug H, Muller M, Otto G, Strand D, Mari- 38 Alderson MR, Armitage RJ, Marakovsky E, Tough TW, Roux E, ani SM, Stremmel W, Krammer PH, Galle PR. Lymphocyte Schooley K, Ramsdell F, Lynch DH. Fas transduces activation sig- apoptosis induced by CD95 (Apo-1/Fas) ligand expressing tumor nals in normal human T lymphocytes. J Exp Med 1993; 178: cells-A mechanism of immune evasion? Nature Med 1996; 2: 2231–2236. 1361–1366. 39 Cascino I, Papoff G, De Maria R, Testi R, Ruberti G. Fas/Apo-1 44 Suda T, Tanaka M, Miwa K, Nagata S. Apoptosis of mouse naive (CD95) receptor lacking the intracytoplasmic domain protects T-cells induced by recombinant soluble Fas ligand and activation tumor cells from Fas-mediated apoptosis. J Immunol 1996; 156: induced resistance to fas ligand. J Immunol 1996; 157: 3918– 13–17. 3924. 40 Xerri L, Carbuccia N, Parc P, Birg F. Search for rearrangements and/or allelic loss of the Fas-APO1 gene in 101 human lymphomas. Am J Clin Pathol 1995; 104: 424–430. Reference added in proof 41 O’Connel J, O’Sullivan GC, Collins JK, Shanahan F. The Fas coun- ter attack: Fas-mediated T-cell killing by colon cancer cells expressing Fas ligand. J Exp Med 1996; 184: 1075–1082. 45 Xerri L, Devilard E, Hassoun J, Mawas C, Birg F. Fas ligand is not 42 Hahne M, Rimoldi D, Scroter M, Romero P, Schreier M, French only expressed in immune privileged human organs but is also LE, Schneider P, Bornand D, Fontana A, Lienard D, Cerottini JC, coexpressed with Fas in various epithelial tissues. J Clin Pathol Tschopp J. Melanoma cell expression of Fas (Apo1/CD95) Ligand: Mol Pathol 1997; 50: 87–91.