|Home | About | Journals | Submit | Contact Us | Français|
Diffuse large B-cell lymphomas (DLBCLs) can be classified into two subtypes: Germinal-center B-cell (GCB)-like and Activated B-cell (ABC)-like tumors, which are associated with longer or shorter patient overall survival, respectively. In our previous studies, we have shown that, although DLBCL tumors of GCB-like and ABC-like subtypes express similar levels of IL4 mRNA, they exhibit distinct patterns of IL-4-induced intracellular signaling and different expression of IL-4 target genes. We hypothesized that these differences may contribute to the different clinical behavior and outcome of DLBCL subtypes. Herein, we demonstrated that IL-4 increased the sensitivity of GCB-like DLBCL to doxorubicin-induced apoptosis and complement-dependent rituximab cell killing. In contrast, IL-4 protected ABC-like DLBCL from the cytotoxic effects of doxorubicin and rituximab. The distinct effects of IL-4 on doxorubicin sensitivity in GCB-like and ABC-like DLBCL cells may be partially attributed to the contrasting effects of the cytokine on Bcl-2 and Bad protein levels in the DLBCL subtypes. These findings suggest that the different effects of IL-4 on chemotherapy and immunotherapy-induced cytotoxicity of GCB- and ABC-like DLBCL could contribute to the different clinical outcomes exhibited by patients with these two subtypes of DLBCL.
Diffuse large B-cell lymphoma (DLBCL) represents a diverse group of neoplasms with heterogeneous genetic abnormalities, clinical features, prognoses, and treatment responses. Gene expression profiling subdivided DLBCL into two clinically distinct types: germinal center B-cell (GCB)-like and activated B-cell (ABC)-like(Alizadeh, et al 2000, Rosenwald, et al 2002). GCB-like DLBCLs express genes that are characteristic of germinal center lymphocytes(Alizadeh, et al 2000, Huang, et al 2002), exhibit immunoglobulin V (IGHV@) intraclonal heterogeneity(Lossos, et al 2000), frequently harbor the t(14:18)(q32:q21) translocation of the BCL2 gene(Huang, et al 2002), and commonly exhibit amplification of the REL locus(Rosenwald, et al 2002). ABC-like DLBCLs are characterized by high expression of genes that are induced by in vitro activation of peripheral blood B cells(Alizadeh, et al 2000), demonstrate no intraclonal heterogeneity of IGHV@ genes(Lossos, et al 2000), and are dependent on constitutive activation of the nuclear factor-κB (NF-κB) signaling pathway for their survival(Davis, et al 2001).
In addition to biological differences, these two subtypes of DLBCL demonstrate distinct clinical behavior: patients with GCB-like tumors exhibit significantly longer overall survival compared to patients with ABC-like DLBCL(Alizadeh, et al 2000, Rosenwald, et al 2002). However, the pathophysiological reasons for the distinct clinical outcome of patients with GCB-like and ABC-like DLBCL are presently unknown. Recently we demonstrated that, while GCB-like and ABC-like DLBCL tumors and cell lines exhibit similar interleukin-4 (IL4) mRNA and surface receptor levels, respectively, they differ in expression levels of IL-4 target genes(Lu, et al 2005). Furthermore, we have demonstrated qualitatively different effects of IL-4 on GCB-like and ABC-like DLBCL(Lu, et al 2005). In GCB-like DLBCL cells, IL-4 induced activation of signal transducer and activator of transcription (STAT) 6 and expression of IL-4 target genes(Lu, et al 2005). In contrast, in ABC-like DLBCL cells IL-4 activated Akt but did not induce the expression of IL-4 target genes nor did it induce a sustained increase in levels of nuclear phosphorylated STAT6(Lu, et al 2005). The reduced activation of the Janus-activated protein kinase (JAK)-STAT6 pathway in the ABC-like cell lines was attributed to increased cytoplasmic and nuclear STAT6 dephosphorylation(Lu, et al 2005). We have demonstrated that two phosphatases, protein tyrosine phosphatase 1B (PTP1B) and T-cell protein tyrosine phosphatase (TCPTP), dephosphorylate cytoplasmic and nuclear STAT6 in DLBCL(Lu, et al 2007, Lu, et al 2008). Protein expression levels of these phosphatases differ between the two subtypes with high expression in ABC-like DLBCL and low expression in GCB-like DLBCL, which may account for some of the differences in intracellular IL-4 signaling in the DLBCL subtypes. Whether the differences in IL-4 signaling in GCB-like and ABC-like DLBCL contribute to the differing response of these two DLBCL subtypes to chemotherapy is unknown.
This study evaluated whether IL-4 may alter the biological sensitivities of DLBCL tumors to commonly used therapeutic agents, such as doxorubicin and rituximab, which represent major components of the current standard therapy for DLBCL: R-CHOP (Rituximab, Cyclophosphamide, Doxorubicin, Vincristine, and Prednisone). We demonstrated that IL-4 distinctively modified responsiveness of the GCB-like and ABC-like DLBCL cells to these immuno-chemotherapeutic agents. IL-4 increased the sensitivity of GCB-like DLBCL to doxorubicin and rituximab chemotherapies yet protected ABC-like DLBCL from these insults. The effects of IL-4 on chemotherapy sensitivity in ABC-like DLBCL can be at least partially attributed to upregulation of Bcl-2 and the inactivation of pro-apoptotic Bad proteins. We have shown that the anti-apoptotic effect of IL-4 in ABC-like DLBCL is dependent on phosphatidylinositol 3-kinase (PI3K) activation and therefore probably mediated by Akt. Since IL-4 is expressed at similar levels in both GCB-like and ABC-like DLBCL primary tumors(Lu, et al 2005), our observations suggest that the different clinical outcomes of patients with DLBCL subtypes may be partially attributed to IL-4-induced modulation of tumor responsiveness to chemo-immunotherapy. Furthermore, these studies suggest that addition of IL-4 to standard therapy for GCB-like DLBCL and conversely, IL-4 ablation in ABC-like DLBCL may further improve responsiveness of these tumors and warrants further investigation.
Recombinant human IL-4 was purchased from R&D Systems (Minneapolis MN) and used in cell culture at a dose of 10ng/ml (100 U/ml) which is similar or lower than those used in previous studies(Carey, et al 2007, Gee, et al 2001, Jinquan, et al 2003). Doxorubicin was purchased from Sigma-Aldrich (St. Louis, MO) and rituximab was obtained from Genentech (San Francisco CA). Propidium Iodide (PI) was purchased from Invitrogen Corporation (Carlsbad CA) and 7-Amino-actinomycin D (7-AAD) was purchased from BD Pharmingen (San Diego CA). Antibodies to Bad (C-7), Mcl-1 (22), and pBad (Ser136) were purchased from Santa Cruz Biotechnology (Santa Cruz CA). Antibodies to CD46, -CD55, and -CD59 and mouse IgG2a κ isotype control were purchased from BD Pharmingen (San Jose CA). Bcl-2 (50-E3) antibody was purchased from Cell Signaling Technologies (Boston MA).
GCB-like (SUDHL-4, SUDHL-6, OCILY-19) and ABC-like (OCILY-3, OCILY-10) DLBCL cell lines (Alizadeh, et al 2000, Davis, et al 2001) were selected for this study. SUDHL-4 and SUDHL-6 cell lines were grown in RPMI 1640 medium (Fisher Scientific, Santa Clara CA) supplemented with 10% fetal calf serum, 2 nM glutamine (Gibco BRL, Grand Island NY), and penicillin/streptomycin (Gibco BRL). The OCILY-3, OCILY-10, and OCILY-19 cell lines were grown in Iscove’s modified Dulbecco medium (IMDM; Fisher Scientific) supplemented with 20% fresh human plasma, 2nM glutamine, penicillin/streptomycin, and 50 μM 2-β mercaptoethanol (Gibco BRL).
A fresh primary DLBCL tumor, obtained from a routine biopsy after the patient signed an informed consent approved by the Institutional Review Board, was used for preparation of a viable single cell suspension. The lymph node was cut sterilely and forced through a metal sieve. Mononuclear cells were obtained after centrifugation of the cell suspension over Ficoll/Hypaque gradient. B cell purification was performed by negative selection using a cocktail of biotinylated CD-2, CD-14, CD-16, CD-36, CD-43, and CD-235a (Glylcophorin A) antibodies (Miltenyi Biotec, Auburn CA). Magnetically labeled cells were separated using an autoMACS magnetic sorter (Miltenyi Biotec). Purity was assessed by anti-CD-19 (BD Biosciences) staining and analysis on a Becton-Dickinson LSR analyzer (BD Biosciences). The tumor was of the GCB-like subtype, as determined by gene expression analysis and immunohistochemical studies performed as previously reported(Alizadeh, et al 2000, Natkunam, et al 2008, Natkunam, et al 2005). The primary tumor sample was cultured in RPMI 1640 medium (Fisher Scientific) supplemented with 10% fetal bovine serum, 2 nM glutamine, and penicillin/streptomycin and used for subsequent experiments.
Cells were incubated at 105 cells/ml with or without IL-4 (100 U/ml), doxorubicin (25 or 100 ng/ml), and/or rituximab (50 μg/ ml ) for indicated periods of time, harvested, washed with 1x phosphate-buffered saline (PBS), and re-suspended in PBS. Cells were stained with 7-AAD as per the manufacturer’s instructions and analyzed on a Becton-Dickinson LSR analyzer (BD Biosciences, San Jose CA). To activate, BCR cells were cultured in the presence of 5μg/ml anti-IgM F(ab’)2 (R&D Systems, Minneapolis MN).
To assay caspase activation, cells were pre-treated with IL-4 (100 U/ml) for 48 h and then treated with doxorubicin (25 ng/ml) for 24 h. Caspase-Glo 3 reagent (Promega, Madison WI) was then added to cells and luminescence was measured on a Luminoskan Ascent (Thermo Labsystems, Waltham MA) luminometer following a 30-min incubation. Readings were normalized to total amount of protein.
Methocult HM-4100 2.6% methyl cellulose (StemCell Technologies, Vancouver BC) was diluted in 20% human serum and IMDM to a final concentration of 0.9% methyl cellulose. Cells were incubated with or without IL-4 (100 U/ml) for 48 h and doxorubicin (25 ng/ml) was added. After 36 h, aliquots of cells in medium were diluted 1/10 in 0.9% methyl cellulose media in a 48-well format. Two weeks later colonies were counted using a 10x magnification on a CE microscope (VWR, Batavia IL).
Whole human cell extracts were prepared by lysing 5×106 cells in RIPA buffer (1× PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate (SDS), and 10mM phenylmethylsulfonyl fluoride, 1μg/ml aprotinin, 100mM sodium orthovanadate) on ice for 30 min. Protein concentration in cell lysates was determined using Coomassie Protein Assay Reagent (Pierce, Rockford IL) and a Genesys 10UV Spectrophotometer (Thermo Labsystems). For Western blot analysis, 20 μg of whole cell lysates per experimental condition were separated by electrophoresis on 10% SDS-polyacrylamide gel, transferred to nitrocellulose membrane (BioRad Laboratories Inc., Hercules CA) and immunoblotted with specific antibodies.
Isolation of RNA, its quantification, and the RT reactions were performed as reported previously(Lossos, et al 2002). BCL2 mRNA expression levels were measured by real time PCR using the Applied Biosystems Assays-on-Demand™ Gene Expression Product on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City CA). Expression of the gene was normalized to RPS18 expression that was used as an endogenous RNA/cDNA quantity control, as we reported previously. (Lossos, et al 2002)
CDC assays were performed with the use of 20% human AB serum (Gemini Bio-Products, Calabasas CA) as the source of human complement. The cells were pre-incubated with IL-4 (100 U/ml) for 36 h and then treated with rituximab (50 μg/ml) for 15 min before finally adding human AB serum for one hour. Cells were then washed in PBS containing 1:20 7-AAD solution and cell viability was determined by flow cytometry on a Becton-Dickinson LSR analyzer (BD Biosciences).
ADCC was performed as previously described by Wilkinson et al(2001). Briefly, lymphoma cells were stained with PKH26 red fluorescent linker kit (Sigma-Aldrich) at 2 μM as described by the manufacturer. Mononuclear cells were obtained by Ficoll-Paque PLUS (GE Healthcare, Piscataway NJ) centrifugation of peripheral blood from the same donor for all experiments, and were used as effector cells. 1× 104 PKH26 stained target cells were incubated for 4 h at 37°C with 30× effector cells in 250 μl of RPMI medium with or without rituximab (50 μg/ml). 7-AAD dye was then added and the fraction of dead target cells was assessed by fluorescence-activated cell sorting analysis as described by Wilkinson et al. (2001) and normalized to the number of dead cells observed without addition of rituximab. The specific cell death was determined by subtracting the spontaneous cell death without antibody from that of the treated cells and then dividing the result by the maximum percentage of live cells at the beginning of the experiment.
Cells (1.0 × 106) were incubated with either isotype control or anti-human CD46, CD55, or CD59 mouse IgG2a antibodies labeled with fluorescein isothiocyanate in a total volume of 100 μl on ice for 30 mins. After staining, the cells were washed three times with cold PBS supplemented with 1% bovine serum albumin. Cells were analyzed by flow cytometry on a Becton-Dickinson LSR analyzer (BD Biosciences).
The 2-tailed Student t test was used test the differences in IL-4 responses . P-values less than 0.05 were considered statistically significant.
Doxorubicin is the most effective chemotherapeutic agent in the treatment of DLBCL. Treatment of GCB-like and ABC-like cell lines with doxorubicin resulted in a strong induction of apoptosis and cell death in both DLBCL subtypes, and the response was dose-dependent (Figure 1A). We examined whether IL-4 may distinctively modulate doxorubicin-induced apoptosis and cell death in GCB-like and ABC-like DLBCL. Pre-incubation of both the GCB- and ABC-like DLBCL cell lines with IL-4 differently modified the cells’ apoptotic response to doxorubicin, as assessed by staining with 7-AAD. Specifically, in the ABC-like cell lines pre-incubation with IL-4 protected cells from the doxorubicin-induced cell death while in the GCB-like cell lines, IL-4 pre-incubation augmented the doxorubicin cell killing and the difference was statistically significant (p<0.01) (Figure 1A and B).
Doxorubicin-induced apoptosis and cell death is associated with an activation of Caspases 3 and 7 in neoplastic lymphocytes(Gamen, et al 1997). We therefore examined the effect of IL-4 on doxorubicin-induced activation of Caspase 3/7. In the GCB-like cell lines, as shown for SUDHL-4, pretreatment with IL-4 alone did not activate Caspase 3/7 yet it enhanced doxorubicin-induced Caspase 3/7 activation (Figure 1C). In contrast, pretreatment of the ABC-like cell lines with IL-4 decreased doxorubicin-induced activation of Caspases 3/7, as shown for OCILY-10 cells (Figure 1C).
To demonstrate that the IL-4—induced modulation of the doxorubicin cytotoxicity was not restricted to established DLBCL cell lines, we evaluated the effects of IL-4 on a de novo untreated primary GCB-like DLBCL tumor. While stimulation of the DLBCL primary cells with IL-4 prior to doxorubicin treatment improved their viability, it also enhanced doxorubicin-mediated cell death similarly to our observations in the GCB-like cell lines.
We next utilized colony assays to determine colony formation potential in response to IL-4-doxorubicin treatment. Although previous experiments showed that IL-4 pre-treatment alone did not have a cytotoxic effect on either GCB-like and ABC-like DLBCL cell lines, it did reduce the ability of the ABC-like DLBCL cell lines to form viable colonies (Figure 2A and B) while increasing the number of colonies in the GCB-like DLBCL. These observations corroborated the observed distinct effects of IL-4 on the cell-cycle and proliferation rates of the two DLBCL subtypes previously reported (Lu, et al 2005). Furthermore, IL-4 differentially modulated the effects of doxorubicin on colony formation in the distinct DLBCL subtypes: it reduced the inhibitory effects of the doxorubicin on cell colony formation of ABC-like cell lines while it augmented the inhibitory effect of the doxorubicin on the colony formation of GCB-like cell lines. The observed differences were statistically significant (p< 0.05)
Addition of rituximab to the standard CHOP chemotherapy was recently reported to improve outcome of DLBCL patients(Coiffier, et al 2002, Habermann, et al 2006). Rituximab exerts its anti tumor effects via 3 mechanisms: direct cellular cytotoxicity; complement-dependent cytotoxicity (CDC); and antibody-dependent cellular cytotoxicity (ADCC). We therefore examined the effects of IL-4 on these 3 mechanisms in the different DLBCL subtypes. While IL-4 pre-treatment did not affect rituximab—mediated direct cellular cytotoxicity in either GCB-like or ABC-like cell lines (data not shown), differential effects of IL-4 were observed on rituximab-mediated CDC. In the GCB-like cell lines, pretreatment with IL-4 enhanced rituximab-mediated CDC, while in the ABC-like cell lines IL-4 protected the tumor cells (p=0.02, Figure 3A and B).
Previous studies have implicated complement inhibitors as potential regulators of CDC with CD55 exhibiting this role most clearly(Golay, et al 2000). In order to determine whether IL-4 modulated the expression of complement inhibitors that are expressed on the cell surface, we tested cell surface expression of CD46, CD55, and CD59 in unstimulated and IL-4-stimulated cells. No changes in surface expression of CD46 and CD59 were observed in both GCB-like and ABC-like cell lines (data not shown). However, in the OCILY-3 ABC-like cell line, IL-4 treatment did increase the cell surface expression of CD55 (Figure 3C), which could contribute to the protection from cell death by rituximab. Major changes in CD55 expression were not observed the remaining DLBCL cell lines tested.
Although ADCC is regarded as one of the major mechanisms by which Rituximab exerts its anti-tumor effects, no changes were observed in ADCC in the various DLBCL cell lines with IL-4 pretreatment (Figure 3D).
To assess the mechanisms by which IL-4 may modulate doxorubicin-induced apoptosis in the DLBCL subtypes, we studied the expression of several proteins that are known to regulate apoptosis. IL-4 potently induced expression of the anti-apoptotic protein Bcl-2 in the ABC-like cell lines, while its expression in the GCB-like DLBCL was unchanged (Figure 4A). Additionally, both ABC-like cell lines exhibited consistent levels of pro-apoptotic Bad but increasing levels of pBad, the inactive form of Bad, suggesting that Bad was inactivated upon IL-4 stimulation. In contrast, all analyzed GCB-like cell lines exhibited a decrease in pBad while levels of Bad remained constant or increased, leading to a net pro-apoptotic effect. No changes in levels of the Mcl-1 protein were observed in any analyzed cell lines (data not shown). IL-4-induced changes in the ratio of pBad to Bad may prime the cells for apoptosis or suppress the pro-apoptotic signaling upon treatment with doxorubicin and rituximab. Similar changes in expression of apoptosis-regulating proteins were observed in DLBCL cell lines in the presence of doxorubicin (Figure 4B).
To determine whether the changes in expression of Bcl-2 were due to the effects of IL-4 on mRNA transcription we measured BCL2 mRNA levels by real-time PCR. IL-4 treatment led to decreased BCL2 mRNA levels in one GCB-like DLBCL cell line (SUDHL-6) but significantly upregulated BCL2 mRNA transcripts in the ABC-like DLBCL cell lines (Figure 4C). These data suggest that in ABC-like cell lines, changes in Bcl-2 protein expression may be due to IL-4-induced changes in BCL2 mRNA levels.
We previously demonstrated that IL-4 induced PI3K and Akt activation in ABC-like cell lines, but not in the GCB-like cell lines(Lu, et al 2005). Other studies have reported that Akt can be involved in the development of chemoresistance in DLBCL tumors(Gagnon, et al 2004, Uddin, et al 2006). We therefore examined the effects of blocking the PI3K-Akt pathway with LY294002, a potent and specific inhibitor of PI3K, on IL-4 modulation of chemotoxicity. Pretreatment of the ABC-like cell line OCILY-10 with 5μM or 10μM LY294002 reversed the protective effects of the IL-4 on doxorubicin-induced cell killing and resulted in increased cytotoxicity (Figure 5A). LY294002 did not have any effects on doxorubicin cytotoxicity in the GCB-like DLBCL cell lines (data not shown).
To determine whether cell activation status could change the effects of IL-4 on doxorubicin-induced apoptosis the GCB-like cell line SUDHL-6 was treated with anti-human IgM and its effects examined on the IL-4 modification of doxorubicin cytotoxicity. B-cell receptor stimulation with anti-IgM prevented IL-4 enhancement of doxorubicin cytotoxicity in GCB-like SUDHL-6 (Figure 5B) and, in fact, IL-4 became protective for these cells. These observations suggest that B cell receptor stimulation may block and reverse IL-4-induced potentiation of doxorubicin cytotoxicity in GCB-like DLBCL cell lines.
IL-4 plays a major role in B-cell maturation and immunological responses. IL-4 has been shown to elicit potent antitumor activity against carcinoma in in vitro and animal models; however, the effect of IL-4 on in vivo and in vitro behavior of B cell lymphomas is not well understood. It has been demonstrated that IL-4 can act as a growth factor for B-lymphocytes and lymphoma cells, as follicular lymphoma cells can be grown in vitro in the CD40/stromal cell system only in the presence of IL-4. In contrast, Taylor et al (1990) reported that IL-4 inhibited growth in tissue culture of 60% of primary lymphoma specimens. Studies by Jones et al (2002) indicated that IL-4 provides growth-inhibitory signals to lymphoma cells activated through their surface Ig receptors. However, in each of these studies only one of the tumors was of DLBCL type.
We recently demonstrated that IL-4 is expressed in all analyzed DLBCL tumors (Lu, et al 2005), and is most likely secreted by non-malignant cells, because none of the examined DLBCL cell lines expressed IL4 mRNA; however, whether the presence of IL-4 in DLBCL tumors may affect neoplastic cell sensitivity to chemotherapies has not been explored. In the present study it was observed that, while IL-4 alone did not have a significant effect on cell viability in either subtype of DLBCL, it enhanced doxorubicin-induced apoptosis and rituximab-induced CDC of GCB-like DLBCL while protecting ABC-like DLBCL cell lines and primary tumors. Although the extent of the IL-4—mediated enhancement of cell death or protection of cells from immuno-chemotherapy ranged from 10-50%, these seemingly small changes may translate into significant differences in patient survival rates.
Cancer cells that survive chemotherapy and have the ability to proliferate are of major importance, because these cells most likely lead to cancer relapses. Notably, patients with GCB-like and ABC-like DLBCL do not differ in their response to initial chemotherapy(Alizadeh, et al 2000, Lossos, et al 2004), but do exhibit differences in overall and progression-free survival due to different relapse rates. In our studies, IL-4 blocked colony formation of GCB-like DLBCL cell lines that were treated with doxorubicin, but rescued colony formation of the ABC-like subtype after treatment with doxorubicin. In theory, IL-4-mediated preservation of malignant cells that retain the ability to proliferate following chemotherapy may predispose the patient to tumor relapse, as observed in the ABC-like DLBCL.
The expression levels of several apoptosis-regulating proteins were studied in order to determine the mechanism by which IL-4 modulates the effects of doxorubicin . IL-4 activated the pro-apoptotic Bad protein in the GCB-like DLBCL, most probably “priming” the cells for apoptosis upon treatment with the chemotherapeutic agent doxorubicin. No such activation was observed in ABC-like DLBCL in which IL-4 inactivated Bad and increased mRNA and protein expression of the anti-apoptotic protein Bcl-2. These changes might contribute to the IL-4—mediated blocking of the pro-apoptotic effects of doxorubicin in ABC-like DLBCL. Pre-treatment of cell lines with doxorubicin did not alter IL-4—induced changes in apoptosis-regulating proteins.
Bad phosphorylation can be mediated by Akt, which is activated in response to IL-4 in the ABC-like cell lines. We therefore examined whether prevention of Akt activation by inhibition of the upstream PI3K might reverse the protective effects of IL-4 in the ABC-like DLBCL tumors. Indeed, blocking of PI3K activity with specific inhibitor LY294002 eliminated the protective effect of the IL-4 on doxorubicin cytotoxicity in the ABC-like DLBCL, thus suggesting that the anti-apoptotic effects of IL-4 in ABC-like DLBCL cells are at least partially mediated by IL-4-induced activation of the PI3K-Akt signaling pathway in this DLBCL subtype.
We have previously shown that IL-4 may increase proliferation of GCB-like DLBCL cell lines, while reducing proliferation of ABC-like cell lines(Lu, et al 2005). Because doxorubicin is known to better target neoplastic cells that exhibit higher proliferation rates, it is possible that both the distinct effects of IL-4 on proliferation of the GCB-like and ABC-like cell lines and the observed differences in the expression of apoptosis-regulating proteins may lead to different modulation of doxorubicin cytotoxicity.
Since the gene expression profile of the ABC-like DLBCL is similar to that of normal peripheral blood lymphocytes activated via stimulation of their B-cell receptor, we examined whether IL-4 effects on the GCB-like DLBCL cells can be reversed by stimulation of B cell receptor. Indeed, stimulation of the GCB-like SUDHL-6 cells with anti-IgM altered their response to IL-4 and doxorubicin treatment, changing it to an ABC-like response manifested by reduced cytotoxicity and cell protection. The mechanism underlying this change is presently unclear. Moreover, it is unknown whether lymphoma B cell receptor is stimulated in vivo; however tonic stimulation of B cell receptor was shown to be important for survival of normal B cells(Lam, et al 1997). If such stimulations occur in the lymphoma microenvironment, they may affect tumor sensitivity to chemotherapy.
Our studies also showed that complement-mediated cytotoxicity of rituximab in DLBCL can also be modulated by IL-4. IL-4 pretreatment enhanced complement-dependent rituximab cytotoxicity in GCB-like DLBCL while protecting ABC-like DLBCL cells from this immunotherapy. Complement inhibitors CD55 and CD59 have been suggested to alter rituximab’s antitumor effects because blocking antibodies specific to CD55 or CD59 increased tumor susceptibility to rituximab-induced CDC, as reported by Golay et al (2000). To determine whether the effects of IL-4 are mediated by these proteins, we studied the cell surface expression of complement inhibitors CD46, CD55, and CD59. Although levels of CD46 and CD59 did not change following IL-4 treatment, an upregulation of cell surface expression of CD55 was observed in the OCILY-3 ABC-like DLBCL cell line. Up-regulation of CD55 following IL-4 treatment may at least partially contribute to the observed rescue of these cells from rituximab-mediated CDC. However, the mechanism of IL-4-mediated alteration of complement-mediated cytotoxicity of rituximab in other cell lines of GCB-like and ABC-like subtypes is unknown. Overall, the role of complement in rituximab therapy remains controversial. Complement is consumed during rituximab therapy(van der Kolk 2001) and on some occasions, cells remaining or emerging after treatment appear to have been selected to express increased levels of complement defense molecules(Bannerji, et al 2003, Treon, et al 2001), however it has also been shown that the expression of complement defence molecules on tumor cells does not predict patients’ clinical outcome(Weng and Levy 2001). ADCC is considered to be the major mechanism of rituximab-induced cytotoxicity. Although ADCC has been reported to act in a complementary manner to CDC in rituximab-mediated cell death (Godal, et al 2006, van Meerten, et al 2006) we did not observe any change in rituximab ADCC with IL-4 pretreatment of cells. Although IL-4-induced upregulation of Bcl-2 protein expression was observed in the ABC-like DLBCL, it did not affect rituximab-mediated ADCC, which is concordant with the previously reported lack of Bcl-2 effects on ADCC(Godal, et al 2006).
Overall, our observations suggest that IL-4 that is present in the tumor milieu may rescue ABC-like cells from chemotherapy- and immunotherapy-induced cell death, promoting their survival and possibly contributing to the relapses and inferior outcomes observed in these patients. Furthermore, our study demonstrates that in vitro testing of tumor sensitivity to chemotherapy may not necessarily reflect the in vivo response of the tumor, since the presence of cytokines or other substances secreted by the tumor microenvironment may alter tumor responsiveness. These interactions need to be considered when evaluating the efficacy of new therapeutic agents; equal consideration may apply for biological agents. Previously, IL-4 was used in two phase II clinical studies of relapsed non-Hodgkin lymphomas. Taylor et al (1990) reported that one of 15 patients with aggressive lymphoma demonstrated a partial response lasting longer than 15 months. In an additional study, Kurtz et al. (2007) reported complete and partial responses in 1 and 4 patients, respectively, for an overall response rate of 13%. Both studies observed anti-tumor activity with IL-4, and none of them achieved the predetermined targeted response rate. However, the observed low response rates might simply reflect the non-selective enrollment of DLBCL patients whose specific lymphoma subtypes might exhibit distinct responses to this cytokine, as demonstrated in our study. In addition, in these studies IL-4 was used alone without chemotherapy and required high doses that caused side effects. It is possible that combining IL-4 at therapeutically tolerated dose levels with chemotherapy may allow potentiation of doxorubicin-induced cell killing of the GCB-like tumor cells. Selective targeting of this therapeutic approach to GCB-like DLBCL may increase the observed response rate, while opposite and non-beneficial effects may be seen for ABC-like DLBCL. Consequently, it is possible that a patient-tailored use of IL-4 or other biological or therapeutic agents based on specific tumor biology would result in higher therapeutic efficiency while eliminating the unnecessary toxicity in patients that are unlikely to respond. Further clinical studies evaluating the manipulation of IL-4 levels in DLBCL subtypes are warranted.
Supported by RO1 CA109335 and RO1 CA122105 from the United States Public Health Service—National Institutes of Health and the Dwoskin Family Foundation (ISL)