Search

In spite of newly emerging therapies and the improved survival of patients with non-Hodgkin lymphoma (NHL), relapses or primary refractory disease are commonly observed and associated with dismal prognosis. Although discovery of the anti-CD20 antibody rituximab has markedly improved outcomes in B-cell NHL, rituximab resistance remains an important obstacle to successful treatment of these tumors. To improve the efficacy of CD20-targeted therapy, we fused interleukin 21 (IL-21), which induces direct lymphoma cytotoxicity and activates immune effector cells, to the anti-CD20 antibody (αCD20-IL-21 fusokine). We observed substantially enhanced IL-21R-mediated signaling by the fusokine compared with native IL-21 at equimolar concentrations. Fusokine treatment led to direct apoptosis of lymphoma cell lines and primary tumors that otherwise were resistant to native IL-21 treatment. In addition to direct cytotoxicity, the fusokine enhanced NK cell activation, effector functions, and interferon γ production, resulting in greater antibody-dependent cell-mediated cytotoxicity compared with IL-21 and/or anti-CD20 antibody treatments. Further, the αCD20-IL-21 fusokine stabilizes IL-21 and prolongs its half-life. In vivo αCD20-IL-21 therapy resulted in a significant tumor control in the rituximab-resistant A20-huCD20 tumors. Collectively, the dual functional ability of the αCD20-IL-21 fusokine to induce direct apoptosis and activate immune effector cells may provide benefit over existing treatments for NHL.

Regulating B-cell receptor (BCR) signaling after antigenic stimulation is essential to properly control immune responses. Currently known mechanisms of inhibiting BCR signaling are via co-receptor stimulation and downstream immunoreceptor tyrosine-based inhibition motif (ITIM) phosphorylation. Herein we demonstrate that BCR stimulation induces rapid and reversible palmitoylation of the SCF-FBXO10 ubiquitin E3 ligase. This results in FBXO10 relocation to the cell membrane, where it targets the human germinal center-associated lymphoma (HGAL) protein for ubiquitylation and degradation, leading to decreases in both BCR-induced calcium influx and phosphorylation of proximal BCR effectors. Importantly, FBXO10 recognition and degradation of HGAL is phosphorylation independent and instead relies on a single evolutionarily conserved HGAL amino acid residue (H91) and FBXO10 relocalization to the cytoplasmic membrane. Together our findings demonstrate the first evidence of negative BCR signaling regulation from direct BCR stimulation and define the temporospatial functions of the FBXO10-HGAL axis. FBXO10 is infrequently mutated in DLBCL but some of these mutations deregulate BCR signaling. These observations may have important implications on lymphomagenesis and other immune processes.

Key Points CRISPR/Cas9 genomic editing of wild-type hematopoietic stem cells generates Npm1-Alk, leading to ALK+ large-cell lymphomas in recipients. CD30+ postthymic T-cell lymphomas are polyclonal but transplantable to secondary recipients with long latency.

GLI1 oncogene has been implicated in the pathobiology of several neoplasms including diffuse large B-cell lymphoma (DLBCL). However, mechanisms underlying GLI1-increased activity in DLBCL are poorly characterized. Herein, we demonstrate that IKKβ phosphorylates GLI1 in DLBCL. IKKβ activation increased GLI1 protein levels and transcriptional activity, whereas IKKβ silencing decreased GLI1 levels and transcriptional activity. Tumor necrosis factor-α (TNFα) mediated IKKβ activation-impaired GLI1 binding with the E3 ubiquitin ligase-ITCH, leading to decreased K48-linked ubiquitination/degradation of GLI1. We found 8 IKKβ-dependent phosphorylation sites that mediate GLI1 stability. Mutating or deleting these residues facilitated GLI1-ITCH interaction and decreased the protective effect of TNFα on GLI1 stability. IKKβ-GLI1 crosstalk is significant because combined inhibition of both molecules resulted in synergistic suppression of DLBCL viability in vivo and in vitro. By linking IKKβ-mediated nuclear factor-κB activity with GLI1, we identified a crosstalk between these 2 pathways that can inform the design of novel therapeutic strategies in DLBCL.

To mount an effective immune response against infectious pathogens or tissue injury, hematopoietic stem cells (HSCs) increase their proliferation and production of myeloid cells, including macrophages, which destroy the pathogens and repair the damaged tissue. Proper resolution of inflammation is essential to restore hematopoietic homeostasis, as unrestrained inflammation can result in life-threatening pathologies such as sepsis, autoimmune disorders and cancer. The molecular mechanisms that control the resolution of inflammation, and how these contribute to disease phenotypes, are poorly understood. BAF (SWI/SNF) complexes are ATPase dependent chromatin-remodeling complexes that play fundamental roles in transcription. BAF complexes use the energy of ATP to modulate the accessibility of transcription factors to DNA and thus, orchestrate the gene expression programs that control proliferation and cellular identity. Genes encoding for BAF subunits are frequently mutated in cancer and developmental disorders. In hematopoietic malignancies, loss-of-function mutations and low expression of specific BAF subunits have been reported in patients with anemia and bone marrow failure. Work from our lab previously demonstrated that the hematopoietic-specific BAF complex subunit Dpf2 cooperates with the transcription factor Runx1 to regulate myeloid differentiation. Based on these studies, we generated a hematopoietic-specific Dpf2 knock-out mouse model and found that mice lacking Dpf2 develop pancytopenia, anemia and an uncontrolled inflammatory response that results in early death. Dpf2-/- peripheral blood samples showed dysplastic features including increased number of polychromatophilic blood cells and Howell-Jolly bodies in erythrocytes. Histopathological analyses revealed the presence of fibrosis and prominent infiltration of histiocytes in multiple organs, including lungs, liver and spleen. Detailed chemical profiling of plasma showed increased levels of multiple pro-inflammatory cytokines, indicative of systemic inflammation. Flow cytometry analyses and Mass cytometry profiling further revealed an expansion of myeloid lineages, specifically monocytes and macrophages, concomitant with severe defects in lymphoid and erythroid differentiation. We also found that Dpf2-/-HSCs had increased serial re-plating capacity and a marked myeloid differentiation bias. To identify the transcription factor networks underlying these phenotypes, we performed RNAseq and ATACseq on control and Dpf2-/- HSCs. Gene Set Enrichment Analyses indicated that Dpf2-/- HSCs have extensive gene expression alterations in immune signaling and interferon response pathways, as well as leukocyte and erythroid differentiation. We also found that Dpf2 loss results in pronounced changes in expression and genomic accessibility of specific transcription factors that control macrophage differentiation and proliferation. Together, our mechanistic studies support a model whereby the absence of Dpf2 results in misregulation of the transcription factor networks that establish macrophage cell identity, leading to a marked increase in macrophage infiltrations and shortened survival of the mice. Treatment of the Dpf2-/-mice with clodronate-containing liposomes, which deplete macrophages from bone marrow and spleen, prolonged survival of the mice. Our work uncovers a novel role of Dpf2 in restraining inflammatory responses by controlling macrophage proliferation and function. Moreover, we propose that, in addition to their tumor suppressive roles in cancer, BAF complexes may have a central role in the prevention of immunopathologies. Disclosures Kadoch: Foghorn Therapeutics: Consultancy, Current equity holder in private company, Membership on an entity's Board of Directors or advisory committees, Other: Scientific founder, fiduciary board of directors member, scientific advisory board member, shareholder, and consultant for Foghorn Therapeutics (Cambridge, MA). . Vega:NCI: Research Funding.

Diffuse large B-cell lymphoma (DLBCL) is an aggressive disease with a high proliferation rate. However, the molecular and genetic features that drive the aggressive clinical behavior of DLBCL are not fully defined. Here, we have demonstrated that activated Jun signaling is a frequent event in DLBCL that promotes dissemination of malignant cells. Downregulation of Jun dramatically reduces lymphoma cell adhesion to extracellular matrix proteins, subcutaneous tumor size in nude mice, and invasive behavior, including bone marrow infiltration and interaction with bone marrow stromal cells. Furthermore, using a combination of RNA interference and gene expression profiling, we identified Jun target genes that are associated with disseminated lymphoma. Among them, ITGAV, FoxC1, and CX3CR1 are significantly enriched in patients with 2 or more extranodal sites. Our results point to activated Jun signaling as a major driver of the aggressive phenotype of DLBCL.

Constitutive PI3K/AKT activation is relevant to multiple aspects of tumor growth and survival in numerous cancers including diffuse large B cell lymphoma (DLBCL). For example, PTEN loss is one of the mechanisms leading to constitutive PI3K/AKT activation in a subset of DLBCL. Smoothened (SMO) is a seven transmembrane spanning and Frizzled-class G-protein coupled receptor that functions as a Hedgehog (Hh) signal transducer. SMO is overexpressed in DLBCL cell lines and tumors. While canonical Hh signaling culminates in the activation of GLI transcription factors and is best understood in the context of cilia, "noncanonical" Hh signaling does not involve GLI transcriptional activity and remains less well characterized. Here, we found that SMO is not only an integral component of lipid rafts but also plays an unexpected central role in the organization of raft microdomains (specialized glycolipid-enriched microdomains known to serve as a highly dynamic signaling platform for cell surface receptors and signaling proteins) and in the sorting of lipid raft-associated proteins. To address whether SMO co-localizes to lipid rafts in the context of DLBCL, HBL1 cells were engineered to stably overexpress a C-terminal SMO-mCherry-fusion protein and were incubated with FITC-Cholera toxin. Within 15 min, a large fraction of SMO-mCherry was co-localized with FITC-Cholera toxin in lipid raft clusters. We also performed immunostaining of endogenous SMO and CD59 in HBL1 cells. CD59 is a glycol-phosphatidyl inositol-anchored lipid raft protein. Immunostaining of live HBL1 cells revealed the colocalization of SMO and CD59 on the extracellular surface of HBL1 cells. No colocalization was found between SMO and the transferrin receptor (TF-R), a plasma membrane protein not associated with lipid rafts. Finally, immunoblotting analysis of detergent-free fractionation further corroborated SMO as a bona fide component of lipid rafts. We then tried to explore the functional relevance of SMO association to the lipid compartment. First, we examined the effects of SMO stable knockdown in DLBCL. We observed a marked decrease in the expression of raft-associated receptors and signaling proteins (e.g. IGFR1, EGFR, IRS1) while caveolin and flotillin, two functional components of lipid rafts, remained unaltered in their levels and distribution. Although only a portion of the overall pool of AKT and pAKT were localized to lipid rafts, SMO loss significantly reduced raft-localized total AKT and pAKT (T308/S473). Consistent with the well-established role of AKT in cell survival, SMO silencing also resulted in reduced DLBCL cell viability. To evaluate whether SMO regulated IGF1R expression at the transcriptional level, we analyzed the mRNA levels of GLI1, the immediate transcriptional target of canonical SMO signaling, and IGF1R in SMO-/- MEFs. Even if GLI1 transcript levels were reduced, consistent with the established mode of GLI1 regulation by SMO through the GLI2 transcription factor, IGF1R mRNA levels remained unchanged. Using GLI1-/- MEFs, we could confirm that the effect of SMO on surface receptors was independent of canonical Hh signaling. We also determined that the rate of IGFR degradation was comparable in the presence and absence of SMO. Total IGF1R receptor levels at steady state represent the balance of total protein synthesis and the fraction of existing receptors that is directed towards lysosomal degradation. To test whether the absence of SMO results in increased degradation, we inhibited lysosomal function with chloroquine (CHLQ). CHLQ had a far more pronounced impact on restoring IGF1R protein levels in SMO deficient cells than it did in control cells. These last results suggested that in the absence of SMO, a larger fraction of this receptor is directed towards lysosomal degradation, and thus resulting in lower steady state levels. In summary, our data confirm that SMO is localized to raft microdomains in lymphoma cells and play a novel role in the sorting of surface proteins for degradation or recycling. In particular, SMO increases the levels of raft resident receptors and facilitates the assembly of an AKT activating machinery to enhance lymphoma cell survival. This novel role of SMO in signal regulation at the level of lipid rafts has broad implications for cancer biology. Disclosures Vega: National Cancer Institute, national Institutes of Health: Other: Grant Funding-R01CA222918.

Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoid malignancy in adults. Aberrant activation of Hedgehog (Hh) and nuclear factor (NF)-κB pathways is ubiquitously observed and known to mediate tumor growth, survival, and chemoresistance in DLBCL. Here, we find that activation of Hh signaling is positively correlated with NF-κB pathway in DLBCL tumors, and that smoothened (SMO), the signal transducer subunit of Hh pathway, contributes to NF-κB activation through recruiting G protein subunits Gαi and Gα12 to activate PKCβ/CARMA1/TRAF6/NEMO signaling axis followed by assembling of the CARMA1/BCL10/MALT1/TRAF6 complex to SMO. Moreover, functional inhibition of SMO enhances the cytotoxic effects of NF-κB inhibitor. Altogether, our study reveals a noncanonical Hh signaling pathway in which SMO activates trimeric G proteins and CARMA1-associated signaling complex, leading to NF-κB activation. This signaling cascade contributes to the survival of DLBCL and may serve as a potential target for combination therapies in DLBCL.

Hedgehog (Hh) signaling pathway is activated in diffuse large B-cell lymphoma (DLBCL). Genetic abnormalities that explain activation of Hh signaling in DLBCL are unknown. We investigate the presence of amplifications of Hh genes that might result in activation of this pathway in DLBCL. Our data showed few extra copies of GLI1 and SMO due to chromosomal aneuploidies in a subset of DLBCL cell lines. We also showed that pharmacologic inhibition of PI3K/AKT and NF-κB pathways resulted in decreased expression of GLI1 and Hh ligands. In conclusion, our data support the hypothesis that aberrant activation of Hh signaling in DLBCL mainly results from integration of deregulated oncogenic signaling inputs converging into Hh signaling.

Chromosomal rearrangements resulting in generation of novel fusion oncogenes are common in hematologic malignancies. These disease drivers are key therapeutic targets and form the basis of animal model development for preclinical studies. For example, retroviral introduction of the chronic myeloid leukemia (CML) fusion BCR-ABL to hematopoietic stem cells (HSCs) results in a myeloproliferative disease similar to accelerated-phase human CML when transplanted to recipient mice. This model and many others are established traditionally through retroviral or germline introduction of human fusion oncogenes to the murine genome. Recent advances in CRISPR/Cas9 technology now permit direct editing of the murine genome to create endogenous genotypes that more accurately reflect configurations found in human diseases. To date, these techniques have not been successfully applied to the modeling of fusion oncogene-driven hematologic malignancies. Anaplastic Large Cell lymphoma (ALCL) is a T-cell non-Hodgkin lymphoma common in adolescents and young adults and driven in ~70% of cases by chromosomal rearrangements involving Anaplastic Lymphoma Kinase (ALK). Most commonly, t(2:5; p23:q35) fuses the 3' ALK kinase domain to the 5' oligomerization domain of the constitutively expressed Nucleophosmin1 (NPM1) gene. An existing immunocompetent model of ALK+ lymphoma employs a CD4 promoter-driven NPM1-ALK transgene but results in immature T-lymphomas in about two-thirds and B-lineage plasma-cell lymphomas in the rest of animals. We employed CRISPR/Cas9 vectors containing guide strands designed to generate double-stranded DNA cleavage in mouse chromosomes 11 and 17 at breakpoints predicted to generate an in-frame Npm1-Alk fusion oncogene. Wild-type HSCs derived from fetal livers were divided and subjected to either transient CRISPR or mock transfection during a brief ex vivo passage followed by immediate transplantation to sub-lethally irradiated wild-type recipients. Mice initially transplanted with CRISPR-modified HSCs developed an ALK+ large cell lymphoma with a latency of ~9 months, while mock transfected controls sacrificed in parallel were phenotypically normal. qPCR analysis of lymphoid organs from mice that developed disease showed extremely high expression of Alk and Tnfrsf8, which encodes CD30. Pathologically, tumors contained large malignancy T cells with anaplastic morphology. Immunohistochemistry confirmed ALK protein expression, including nuclear localization classic for NPM1-ALK, and intense CD30 cell-surface staining. One recipient instead developed a CD30-negative ALK+ diffuse large B-cell lymphoma. Transplantation of primary T-lymphoma cells to secondary recipients resulted also in ALK+ T-cell lymphomas in recipients with more rapid onset infiltrating lymph nodes, spleen, liver, and other organs. T-cell receptor clonality analyses through β-chain deep sequencing show an oligoclonal T-cell disease in both primary and secondary recipients. We therefore demonstrate successful genomic editing of transplantable murine hematopoietic stem cells to generate a novel model of a fusion oncogene-driven hematologic malignancy. These methods are widely applicable to additional lymphomas and leukemias and could fuel development of improved in vivo preclinical model systems. Disclosures No relevant conflicts of interest to declare.

Background: Ocular adnexa lymphomas (OAL) consist of a heterogeneous group of tumors that can arise in the lacrimal gland, orbit, conjunctiva and eyelid. More than 95% of these lymphomas are of B-cell origin and extranodal marginal zone lymphoma (EMZL) of mucosa-associated lymphoid tissue is the most common subtype of OAL, accounting for 55-85% of the cases. Ocular adnexa extranodal marginal zone lymphoma (OA-EMZL) patients have a persistent risk of relapses and in 4% of the cases, the disease transforms into an aggressive lymphoma. A comprehensive list of all somatic nonsilent mutations remains currently unknown thus hampering the complete understanding of the disease and the development of new diagnostic and therapeutic approaches. Previous studies analyzed only a small number of OA-EMZL patients using gene targeted sequencing. In this study we utilized whole-exome sequencing to define the genetic landscape of 69 cases of OA-EMZL. Our data provide an unbiased identification of genetically altered genes and pathways that may play a role in the molecular pathogenesis of OA-EMZL. Methods: GATK HaplotypeCaller was used for joint variant calling on all samples, while MuTect was used to detect somatic variants in seven tumor samples with available paired normal germline DNA. We focused on variants that were exonic, not synonymous, and deleterious (CADD score > 10), while excluding any variants found with a frequency > 1% in control populations. For copy number (CN) analysis we used EXCAVATOR on each sequenced exome in pooled mode against the seven paired normal samples. The output was provided to GISTIC 2.0 to determine recurrent arm-level and gene-level CN changes. Results: We identified 125 candidate cancer driver genes with the following characteristics: 1) have genomic alteration in more than 5% of the cases, 2) are present in the list of COSMIC and DLBCL driver genes, 3) are considered essential genes or tumor suppressor genes as previously defined in a functional CRISPR screen (Reddy et al., Cell 2017). Among the most frequent alterations there were mutations and CN losses of CABIN1 (32%), inactivation of TNFAIP3 (26%) and KMT2D (22%), mutations and CN changes in CARD11 (25%), RHOA (25%), CREBBP (20%) and TBL1XR1 (23%). Gene set enrichment analysis of candidate driver genes showed statistically significant enrichment for genes involved in cell junction, adhesion, cytoskeleton regulation, chromatin organization and for genes harboring at least one highly conserved NFAT binding motif TGGAAA in their promoter. In addition, we identified recurrent lesions affecting several genes belonging to the NFkB, NOTCH and B cell receptor (BCR) signaling pathways. The most frequent alterations in the NFkB pathway included inactivation of TNFAIP3 (26%), mutations and CN alterations of CARD11 (25%), MYD88 (12%), BCL10 (10%) and CRD10 (9%). Mutations affecting NOTCH signaling were found in NOTCH3 (22%), DTX1 (12%), MAML1 (12%), MAML2 (12%), DLK1 (10%) and SPEN (9%). BCR signaling genes alterations were observed in PIK3R2 (12%), VAV1 (10%), PTPN6 (7%) and BTK (4%). CABIN1 was one of the most commonly altered genes. In T cells, CABIN1 is implicated in inhibiting calcineurin-mediated signal transduction, including NFAT activation. In reactive lymph node and tonsil tissue, CABIN1 was expressed in reactive B cells (CD20+ and CD3-) localized mainly in the mantel zone area, as demonstrated by IHC and IF. We hypothesized that it may act as a negative regulator of BCR signaling in MZL. To verify this possibility we used shRNAs to knock down CABIN1 in MZL cell lines and measured NFAT activity in response to IgM stimulation using a luciferase reporter system. IgM stimulation of CABIN1-depleted MZL cells induced a more pronounced activation of NFAT signaling compared to wild-type cells. Conclusions: The genetics of OA-EMZL is characterized by somatic alterations in NFkB, NOTCH, and BCR signaling pathways. Candidate cancer driver genes are implicated in chromatin organization, cytoskeleton regulation and in adhesion, and they are also enriched for NFAT target genes. One of the most altered candidate cancer driver genes was CABIN1, which is a calcineurin inhibitor that acts as a negative regulator of NFAT activity. CABIN1 depletion in MZL cells boosts BCR-stimulated NFAT activity. Our data suggest a widespread dysregulation of NFAT signaling cascade downstream of BCR stimulation as a possible mechanism of OA-EMZL pathogenesis. Disclosures Lossos: Affimed: Research Funding.

The pathogenesis of Chlamydophila psittaci-negative ocular adnexal extranodal marginal zone lymphomas (OAEMZLs) is poorly understood. OAEMZLs are monoclonal tumors expressing a biased repertoire of mutated surface immunoglobulins. Antigenic activation of the B-cell receptor (BCR) may have a role in the pathogenesis of these lymphomas. We have analyzed the reactivity of recombinant OAEMZL immunoglobulins. OAEMZL antibodies reacted with self-human antigens, as demonstrated by enzyme-linked immunosorbent assays, HEp-2 immunofluorescence and human protein microarrays. All the analyzed recombinant antibodies (rAbs) exhibited polyreactivity by comprehensive protein array antibody reactivity and some rAbs also demonstrated rheumatoid factor activity. The identity of several reactive antigens was confirmed by microcapillary reverse-phase high-performance liquid chromatography nano-electrospray tandem mass spectrometry. The tested rAbs frequently reacted with shared intracellular and extracellular self-antigens (for example, galectin-3). Furthermore, these self-antigens induced BCR signaling in B cells expressing cognate surface immunoglobulins derived from OAEMZLs. These findings indicate that interactions between self-antigens and cognate OAEMZL tumor-derived BCRs are functional, inducing intracellular signaling. Overall, our findings suggest that self-antigen-induced BCR stimulation may be implicated in the pathogenesis of C. psittaci-negative OAEMZLs.

Tumor necrosis factor receptor-associated factor 6 (TRAF6), an (K63) E3-ligase, plays a crucial role in many biological processes and its activity is relevant in the biology of multiple cancers including diffuse large B cell lymphoma (DLBCL). Although molecules that trigger TRAF6 activation have been defined, those that stabilize TRAF6 levels and/or enhance TRAF6 function remain largely unclear. Previously, we found that activation of smoothened (SMO) with recombinant Hedgehog (Hh) ligand increased the binding between SMO with TRAF6, as well as TRAF6 protein levels (Blood 2013; 121:4718-28). In addition, transient overexpression of SMO resulted in increased K63-Ub of both TRAF6 and NEMO indicating stabilization of these proteins resulting in NF-kB activation. This is relevant, as more recently we found that TRAF6 amplifies pAKT signaling in DLBCL and that TRAF6 is the dominant E3 ligase for the K63-Ub of AKT in DLBCL. Moreover, TRAF6 recruitment to the cell membrane, and stabilization of its ubiquitination profile are facilitated by SMO. SMO is a member of the Frizzled-class G-protein-coupled receptor (GPCRs) and is traditionally known for its role as signal transducer in canonical Hedgehog (Hh) signaling. These observations prompted us to investigate whether the ability of SMO to increase TRAF6 levels is limited to ligand induced signaling, whether it contributes to chemoresistance in DLBCL cells, and whether SMO directly participates in controlling TRAF6 levels. To confirm the regulatory role of SMO in the TRAF6/AKT axis in DLBCL cells (HBL1 and HT) and further outline the nature of the underlying regulation, we measured the impact of activation of the Hh pathway with recombinant Shh ligand on TRAF6 levels, with and without SMO knockdown or recombinant SMO overexpression. Canonical Hh signaling results in the activation of the GLI1 transcription factor and the subsequent elevation of GLI1 mRNA levels is an established indicator of activation of the Hh pathway. However, neither SMO activation nor the knockdown of GLI1 had a significant impact on TRAF6 mRNA levels. These findings indicate that TRAF6 is not transcriptionally regulated by SMO signaling through GLI1 (canonical Hh signaling). In contrast, overexpression of SMO or siRNA knockdown of SMO resulted in an increase or decrease of TRAF6 protein levels, respectively. Consistent with the decrease of AKT activation (pAKT T308 and S473) after TRAF6 knockdown, the increase in TRAF6 levels that follows SMO overexpression resulted in an increase in the levels of AKT phosphorylation. Altogether, these observations suggest a post-translational regulation of TRAF6 by SMO. Indeed, stable knockdown of SMO dramatically reduces the half-life of TRAF6 in both HBL1 and HT cells in the presence of cyclohexamide. Furthermore, overexpression of SMO increases K63-Ub of both TRAF6 and AKT. In contrast, the SMO induced decrease in K48-Ub occurred only for TRAF6 but not for AKT. These data link the SMO-stimulated activation of TRAF6 to the enhancement of AKT signaling and protection of TRAF6 from proteasomal degradation. Mechanistically, we found that SMO, through its C-terminal tail, stabilizes TRAF6 and protects TRAF6 from proteosomal degradation, an effect mediated by ubiquitin-specific protease-8 (USP8). Importantly, this functional link between SMO and TRAF6 is reflected in DLBCL patient samples where high expression of both molecules correlates with poor prognosis. Resistance to DXR is a serious challenge in the treatment of DLBCL, and activated AKT is known to contribute to DXR resistance in multiple cancers including DLBCL. We evaluated whether SMO and TRAF6 support resistance to DXR in DLBCL cell lines. We exposed HT and HBL1 cells as well as their counterparts with stable knockdown of TRAF6 or SMO to DXR for 96hrs. Cell viability after exposure to DXR was determined by an Annexin V and PI staining assay. Silencing SMO or TRAF6 dramatically decreased cell survival after treatment with DXR. In summary, we report that SMO is needed to facilitate and maintain TRAF6-dependent elevated pAKT levels in DLBCL cell lines of germinal (GC) and non-GC subtypes, and that the SMO/TRAF6 axis contributes to DXR resistance in DLBCL. Our study reveals a novel and potential central cell survival signaling mechanism in which SMO stabilizes and protects TRAF6 from proteosomal degradation. Disclosures Lossos: Affimed: Research Funding.

Previously, we have demonstrated that inhibition of Hedgehog pathway induces predominantly apoptosis in diffuse large B-cell lymphoma (DLBCL) cell lines of activated B-cell (ABC) type but predominantly cell cycle arrest in those of germinal center (GC). Here, we explored the possibility of overcoming the resistance to apoptosis to SMO inhibitors in five DLBCL cells of GC type using the combination of the SMO inhibitor HhAntag (Genentech Inc) with the BH3 mimetic ABT-737 (Abbott Laboratories). As controls we have used two DLBCL of ABC type (OCI-LY10 and OCI-LY3). Combinatorial treatments were performed with increasing concentrations of the HhAntag with low doses (equal or less than the IC20) of ABT-737. MTS assays were used to detect changes in cell viability and Annexin-V and PARP1 cleavage assays were used to detect apoptosis. Combining low doses of ABT-737 with increasing concentrations of HhAntag in GC DLBCL cell lines resulted in significantly increase of apoptosis in comparison to treatments with the SMO inhibitor alone. We concluded that in GC DLBCL cell lines, in contrast to those of ABC type, functional inhibition of BCL2 family members is usually needed to overcome the resistance to apoptosis to SMO inhibitors. These findings provide a rationale to explore the use of SMO and BCL2 inhibitors as adjuvant therapy for treatment of DLBCL of GC type.

Background The mechanisms responsible for chemoresistance in patients with refractory classical Hodgkin lymphoma (CHL) are unknown. ATP-binding cassette (ABC) transporters confer multidrug resistance in various cancers and ABCC1 overexpression has been shown to contribute to drug resistance in the CHL cell line, KMH2. Findings We analyzed for expression of five ABC transporters ABCB1, ABCC1, ABCC2, ABCC3 and ABCG2 using immunohistochemistry in 103 pre-treatment tumor specimens obtained from patients with CHL. All patients received first-line standard chemotherapy with doxorubicin (Adriamycin®), bleomycin, vinblastine, and dacarbazine (ABVD) or equivalent regimens. ABCC1 was expressed in Hodgkin and Reed-Sternberg (HRS) cells in 16 of 82 cases (19.5%) and ABCG2 was expressed by HRS cells in 25 of 77 cases (32.5%). All tumors were negative for ABCB1, ABCC2 and ABCC3. ABCC1 expression was associated with refractory disease (p = 0.01) and was marginally associated with poorer failure-free survival (p = 0.06). Multivariate analysis after adjusting for hemoglobin and albumin levels and age showed that patients with CHL with HRS cells positive for ABCC1 had a higher risk of not responding to treatment (HR = 2.84, 95%, CI: 1.12-7.19 p = 0.028). Conclusions Expression of ABCC1 by HRS cells in CHL patients predicts a higher risk of treatment failure and is marginally associated with poorer failure-free survival using standard frontline chemotherapy regimens.

Background: Constitutive activation of NF-κB signaling is a hallmark of DLBCL. Activation of NF-κB is a multifactorial process resulting from oncogenic mutations (CARD11, MYD88…), chromosomal abnormalities, chronic activation of B-cell receptor signaling (BCR) as well as stimuli from the microenvironment. Chronic activation of BCR is not only the result of gene mutations (e.g.CD79B) but also is the result of stimuli generated from the lymphoma microenvironment. We previously found that smoothened (SMO), transducer of hedgehog (Hh) signaling, enhanced NF-κB activation in lymphoma cells, independently of the presence of oncogenic mutations (Changju et al., Blood 2013). Hh ligands are provided by the lymphoma microenvironment (e.g. stromal cells) to activate SMO in the lymphoma cells. Activated SMO, recruits and activates trimeric-G-coupled proteins to activate PKCβ/CARMA1/TRAF6/NEMO axis, followed by assembling of the CARMA1/BCL10/MALT1/TRAF6 complex to SMO resulting in NF-κB activation. Now, we reveal an additional mechanism by which SMO further contributes to the activation of NF-κB in DLBCL. Summary of results: We explored if recombinant Hh ligand can activate NF-κB signaling in the presence of the BTK inhibitor (PCI-32765). Hh stimulation resulted in phosphorylation of PLCg2 (downstream of BTK) and partially rescued the inhibitory effect of the BTK inhibitor on phosphorylation of PLCg2 suggesting a crosstalk between SMO and BCR receptor signaling. We identified that SMO forms a complex with TRAF-6 and PLCg2 using immunoprecipitation and Duolink in situ proximity ligation assays. We found that SMO stabilizes and protects TRAF-6 from proteosomal degradation. SMO-dependent TRAF6 stabilization is mediated by the ubiquitin-specific protease-8 (USP-8) by removing ubiquitin from K48-linked lysine of TRAF6. Conclusions: Collectively, our data reveals multilayer crosstalk between Hh and BCR/NF-κB pathways and provide significant insights into how SMO contributes to chemotolerance and progression of DLBCL. We expect that our results will delineate novel molecular mechanisms involved in the pathobiology of DLBCL that may serve as therapeutic targets. Disclosures Vega: Seatle Genetics: Honoraria; NIH: Research Funding.

GLI1 is a Hedgehog (Hh) related transcription factor originally discovered as an amplified product in gliomas. Inappropriate activation of the GLI1 has been shown in many cancers including diffuse large B cell lymphoma (DLBCL). We previously showed that GLI1 mediated canonical Hh signaling is constitutive active in DLBCL and contributes to cell survival, proliferation and enhances chemotolerance. Although the importance of GLI1 in tumor development is well recognized, the molecular mechanisms controlling the transcriptional activity of GLI1 are poorly characterized. To identify regulatory components that participate in the transcriptional activity of GLI1, we explored GLI1 putative interacting proteins by liquid chromatography tandem mass spectrometry following immunoprecipitation of endogenous GLI1. We detected that the inhibitor of NF-KB kinase, IKKβ, is one of the proteins associated with GLI1 transcription factor. Here we investigate the regulatory role of IKKβ in the transcriptional activity of GLI1. We show that IKKβ regulates the transcriptional activity of GLI1 by phosphorylating GLI1 in C-terminal region and modulating its protein stability. Short stimulation of SUDHL4 and DOHH2 cells with TNF-α (20ng/mL) resulted in increased GLI1 protein levels. Similar results were observed in 293T cells transiently transfected with GLI1 and IKKβ kinase constructs. Moreover, silencing of IKKβ using siRNA and shRNAs led to decreased GLI1 protein levels and its transcriptional activity in DLBCL cell lines with constitutive activation of the NF-KB. Next, we characterized nine probable IKKβ dependent GLI1 phosphorylation sites (S543-S548, S1070, S1071 and S1074 identified by nanospray ion trap mass spectrometry) using mutational and deletions studies. We show that IKKβ phosphorylates GLI1 at Thr1074 and decreases binding between GLI1 and HECT-type E3 ubiquitin ligase (ITCH) resulting in reduced GLI1 polyubiquitination and degradation. Point mutation of Threonine 1074 to Alanine prevents IKKβ-mediated GLI1 phosphorylation and facilitates GLI1-ITCH interaction, polyubiquitination and degradation of GLI1 in the proteasome. Collectively, our data links IKKβ-mediated NF-kB signaling to the transcriptional activity of GLI1 and illustrates a novel cross talk between these two pathways. This is of clinical interest because activation of the NF-kB pathway is frequent in DLBCL and the connection between Hh and NF-kB pathways may open novel therapeutic avenues for DLBCL. Disclosures No relevant conflicts of interest to declare.

Hedgehog (Hh) signaling plays an important role in the pathobiology of B-cell malignancies including diffuse large B cell lymphoma (DLBCL). The GLI family members that include GLI1, GLI2 and GLI3 are Hh signaling transcription factors. While most research has focused on smoothened (SMO), Hh transducer receptor subunit, many lines of evidence indicate that other signaling pathways can activate GLI transcription factors. GLI1 is a zinc finger transcription factor and established oncogene. GLI1 expression increases in response to Hh signaling activation and potentiates the transcriptional output of Hh signaling. We previously demonstrated that the canonical Hh ligand-PTCH1-SMO-GLI axis is functional and plays an important role in cell proliferation, survival and chemotolerance in DLBCL (Leukemia 2010 and Oncogene 2011). Moreover, using stable GLI1 knockdown DLBCL cell lines we found that GLI1 contributes to cell survival and proliferation of DLBCL (J Biol Chem 2013). However, our understanding of the mechanisms controlling the transcriptional activity of GLI1 is limited. To identify regulatory components that participate in the transcriptional regulation of GLI1, we explored GLI1 putative interacting proteins by liquid chromatography tandem mass spectrometry following immuno-precipitation (IP) of endogenous GLI1. We found that the NF-κB kinase, IKKβ, is one of the proteins associated with GLI1. This finding was of interest as we found a positive correlation between NF-kB activity and Hh signaling in DLBCL (Blood 2013). To validate the functional interaction between endogenous GLI1 with IKKβ in-vivo, lysates of 293T cells were IP with GLI1 or control antibody and subjected to immunoblot analysis with antibodies recognizing IKKβ subunit and SuFu (known partner of GLI1) respectively. IKKβ and SuFu were co-immunoprecipitated with GLI1. To confirm these findings, 293T cells were transiently co-transfected with plasmids carrying GLI1 and FLAG-tagged IKKα and IKKβ wild-type or kinase dead (K44A) forms. Both forms (IKKα and IKKβ) were detected in the immunoprecipitated GLI1 protein complex, however only transfection of IKKβ resulted in abundance of GLI1 protein suggesting a functional role of IKKβ kinase in the regulation of GLI1 protein levels. The association (and nuclear co-localization) between IKKβ and GLI1 was confirmed in DLBCL cell lines and 293T cells by immunofluorescence studies. To investigate the role of IKKβ in the transcriptional activity of GLI1, we knocked down IKKβ gene in HBL1 (cells with CD79B mutation) cells using two independent IKKβ shRNAs. Silencing IKKβ resulted in a decrease of IKKβ expression associated with decreased levels of GLI1 and GLI1 downstrean targets (PTCH1, CCND1 and BCL2). To better understand the functional role of IKKβ kinase activity in the transcriptional regulation of GLI1 target genes, we explore the IKKβ dependent GLI1 phosphorylation sites. We co-transfected full length GLI1 and IKKβ plasmids in 293T cells and IKKβ-GLI1 complex were purified using anti-GLI1 antibody. Peptides resulting from digestion of GLI1 with several different proteases were analyzed by nanospray ion trap mass spectrometry. Seven phosphorylation sites in GLI1 C-terminal residues were identified during analysis of trypsin-digested GLI1 by mass spectrometry. These data suggest that IKKβ phosphorylates GLI1 and plays an important role in the regulation of GLI1 transcriptional activity. Because activation of NF-κB is a frequent event in DLBCL, our study will help to understand the cross talk between NF-κB and Hh signaling pathways and will open potential new avenues for targeted therapy in DLBCL. Disclosures: Vega: NIH: Research Funding; Leukemia & Lymphoma Society: Research Funding.

Abstract 2399 Activation of the Hedgehog (Hh)/glioma-associated oncogene (GLI) pathway has been found in a growing number of malignancies. We have provided evidence that canonical Hh signaling is required for cell survival and proliferation of DLBCL cell lines. To confirm the pathogenic role of GLI1 in DLBCL, we established GLI1 knock down DLBCL cell lines (OCI-Ly19, HBL-1 and BJAB) using a lentiviral shRNA system and performed cell viability and apoptosis assays. Cell viability assays demonstrated that GLI1 knockdown DLBCL cells experienced a statistically significantly decrease in the number of viable cells in comparison with control cells harboring scramble shRNA. To examine whether decreases number of cell viability in GLI1 knock down cells were due to apoptosis, we performed annexin V and PI assays. We observed marked increase of apoptosis in GLI1 knock down DLBCL cells versus controls (2.5 fold increase for OCI-Ly10, and 5 fold for HBL1 and BJAB). To investigate the mechanism by which GLI1 regulates tumorigenesis and cell survival, we searched for whole genome GLI1-target genes in DLBCL cells using CHIP sequencing technique and identified AKT genes as potential targets of GLI1. Using pharmacological and silencing approaches, we observed that Hh signaling modulates the expression of AKT genes in DLBCL cells. We further identified two putative binding sites for GLI1 in the AKT1 promoter region and confirmed their functionality using chromatin immunoprecipitation, luciferase reporter and site-directed mutagenesis assays. To investigate whether there is any correlation between AKT1 and GLI1 mRNA expression in human DLBCL tumors, we performed quantitative real-time PCR analyses in 17 frozen DLBCL specimens including apharesis samples from pleural effusions. The real time PCR analysis revealed a strong Spearmen correlation coefficient (R2=0.9) between GLI1 and AKT1 mRNA expression. In summary, we provide evidence of the role of GLI1 in the pathobiology of DLBCL and demonstrated a cross talk, at the transcriptional level, between Hh signaling and AKT in DLBCL. A link between these 2 pathways at the trasncriptional level was not previoulsy documented. This finding is of clinical interest as AKT has a key role in lymphoma cell survival and constitutive activation of AKT has been described in DLBCL. Disclosures: No relevant conflicts of interest to declare.

Abstract 1298 Aberrant activation of hedgehog (Hh) and NF-kB pathways contribute to tumor cell growth, survival and chemotolerance in diffuse large B-cell lymphoma (DLBCL). Previously, we documented a functional crosstalk between hedgehog (Hh) and NF-kB pathways that contribute to tumor cell growth and survival in diffuse large B-cell lymphoma (DLBCL). However, the molecular mechanisms that link Hh with NF-kB pathway have not been defined. Based on that smoothened (SMO) has been associated with heterotrimeric G protein members of the Gα i family and established as a GPCRs-like protein, we hypothesize that GPCRs-related mechanisms such as the protein kinase C (PKC)-CARMA1-BCL10/MALT1/TRAF6 axis may contribute to SMO-dependent activation of NF-kB. First, we confirmed that SMO contributes to the activity of total PKC in 293T and DLBCL cells. Transiently or stably silencing of SMO resulted in decreased total PKC activity in comparison with the controls, while overexpression of SMO resulted in increased activity of PKC. Cyclopamine-KAAD or recombinant Shh N-terminal peptide resulted in decrease or increase of the total activity of PKC, respectively. As PKC isoforms β-1 and β-2 are the major isoforms expressed in B-lymphocytes that mediates NF-kB activation induced by activation of the B-cell receptor, we assessed the activation status of these two isoforms in response to changes of SMO activity. Inhibiting SMO with cyclopamine-KAAD or silencing SMO by siRNA decreased the phosphorylation status of PKCβ-1 and −2. In contrast, activating SMO with Shh N-terminal peptide increased the phosphorylation of PKCβ-1 and −2. Next, we assessed if SMO can modulated the activity of CARMA1. CARMA1 is a scaffold protein that serves to integrate the upstream signal of PKCs with downstream effectors in hematopoietic cells. We found that cyclopamine-KAAD or silencing SMO by siRNA decreased the phosphorylation status (inactivation) of CARMA1 and that activation of SMO with Shh N-terminal peptide increased the phosphorylation (activation) of CARMA 1. Because the polyubiquitination of TRAF6 and NEMO (IKKg) at lysine 63 (K63) are important events in propagating NF-kB signaling we examined the effect of overexpressing SMO on K63 polyubiquitination of TRAF6 and NEMO. Overexpression of SMO resulted in increased polyubiquitination of TRAF6 and NEMO at K63 supporting NF-kB pathway activation. TRAF6 is involved in activation of TAK1 and the IKK complex, resulting in translocation of NF-kB to the nucleus and activation of NF-kB. Silencing SMO by shRNA decreased total levels of TRAF6 that was associated with faster proteosomal degradation of TRAF6 in 293T cells. Inhibition of SMO with cyclopamine-KAAD also decreased the total levels of TRAF6 in DLBCL cells. In addition, overexpression of SMO (including a constitutively active mutated SMO) in 293T cells caused increase of TRAF6 expression by decreased polyubiquitination of TRAF6 at lysine 48 (K48) that targets TRAF6 for proteosomal degradation. In summary, altogether, these findings support that at least one mechanism by which SMO contributes to modulate activation of NF-kB is through the activation of the axis PKCβ/CARMA1/TRAF6/NEMO and stabilization and blocking degradation of TRAF6. Disclosures: No relevant conflicts of interest to declare.

Abstract 3483 Activated Hedgehog (HH) signaling was identified in our laboratory to contribute to cell survival, proliferation and chemotolerance of diffuse large B-cell lymphoma (DLBCL) (Leukemia 2010;24:1025 & Oncogene 2011, in press). The HH receptor complex is integrated by two main proteins, smoothened (SMO) and patched-1 (PTCH1). SMO is a seven-transmembrane G protein-coupled receptor that transduces HH signal to the cytoplasm and has glioma-associated oncogene homologue (GLI) proteins as major signaling transcription effectors. PTCH1 is a 12 transmembrane protein that inhibits SMO in the absence of HH ligands. Here, we investigated potential cross talk between SMO with the activation status of PI3K/AKT and NF-kB, two relevant oncogenic pathways in DLBCL. Using a small interfering (si)RNA approach and DLBCL cell lines of germinal center (GC) and activated B-cell (ABC) type we found that the expression levels of SMO modulate the activation status of AKT and canonical NF-KB pathways in DLBCL cells of GC type and mainly AKT in those of ABC type. In DLBCL cells of GC cell type, silencing SMO resulted in decrease of the phosphorylation levels of ser473p-AKT and ser536p-P65 and silencing PTCH1 resulted in increase of the phosphorylation levels of both proteins. The same silencing experimental approach in DLBCL cells of ABC type resulted in similar modulation in the activation status of the AKT but not, or to less extent, in the activation of NF-kB pathway. The modulation of the activation status of the NF-KB pathway was also confirmed using protein nuclear extracts and DNA binding ELISA assays. In cells of both DLBCL subtypes, silencing of the SMO transcriptional effector GLI1 showed no changes in the activation status of both pathways. The modulation in the activation status of AKT and NF-KB was also detected using SMO inhibitors, cyclopamine-KAAD and HhAntag (Genentech Inc) or activators, purmorphamine and recombinant HH protein. Combinatorial treatments with increasing concentrations of SMO inhibitors (cyclopamine and HhAntag [1.6, 3.2 and 4.8 μM]) with minimal lethal doses of Ly294002 (PI3K inhibitor) or BAY-11 (NF-KB inhibitor) were also performed. Using cell viability, and apoptosis (Annexin-V) assays, we found that combined treatments of PI3K or NF-KB inhibitors with a SMO inhibitor resulted in an additive/synergistic decrease of cell viability and increase of apoptosis in comparison to the treatments with SMO inhibitors alone. Taken together, our data shed novel light on the contribution of SMO on the activation of PI3K/AKT and NF-kB pathways in DLBCL. Our data also provide a rationale to use SMO inhibitors in combination with inhibitors of other oncogenic pathways such as PI3K/AKT and/or NF-KB for the treatment of patients with DLBCL. Disclosures: No relevant conflicts of interest to declare.

Abstract 2479 Generally in cancers, high expression of anti-apoptotic proteins induced by oncogenic signaling pathways is a major cause for treatment failure. Previously, we have demonstrated that Hedgehog (Hh) signaling is activated and functional in diffuse large B-cell lymphoma (DLBCL) and that its inhibition, using the Smoothened (SMO) inhibitor cyclopamine-KAAD, induces apoptosis in cell lines of DLBCL of activated B-cell (ABC) subtype and cell cycle arrest (G1-phase) in those of germinal center (GC) subtype. Induction of apoptosis in ABC-DLBCL was due to decrease of BCL2 expression (protein and mRNA levels), a direct target of Hh signaling (Leukemia 2010;24:1025–36). We also found that BCL2 is upregulated in the tumor cells in the presence of stroma and that Hh inhibition abrogates this upregulation (Oncogene 2011, in press). Here, we explored the possibility of overcoming resistance to apoptosis-induced by Hh inhibition in GC-DLBCL using a combination of the SMO inhibitor “Hh antagonist” (Genentech Inc) with the BH3 mimetic ABT-737 (Abbot laboratories) in a panel of 5 GC type DLBCL cell lines, 4 of them with the t(14:18)(q32;q21) (DOHH2, SUDHL4, Toledo and OCI-LY19) and one without the translocation (HT). ABT-737 is a functional inhibitor of three major anti-apoptotic proteins of the BCL2 family (BCL2, BCL-xL and BCL-W). Combinatorial treatments were performed with increasing concentrations of the Hh antagonist (2.5, 5.0 and 7.5 μM) and with variable but minimal lethal doses of ABT-737 (inducing cell death less than 20%). Using cell viability, apoptosis (annexin-V) and cell cycle based assays, we found that the combined treatments resulted in a additive/synergistic decrease of cell viability and increase of apoptosis in comparison to the treatment with the Hh antagonist alone. In contrast to ABC type DLBCL cell lines, in DLBCL cells with t(14;18), Hh antagonist did not result in decrease expression of BCL2, BCL-xL and BCL-W indicating that the additive effect of ABT-737 was due to their functional inhibition and not by suppressing expression levels of these proteins. On the other hand, treatments with the Hh antagonist resulted in a concentration dependent decrease of BCL2 protein levels in HT cells. In summary, our data confirm that in the presence the t(14;18), inhibition of SMO does not result in decrease of BCL2 expression. In these cases, concomitant inhibition of BCL2 function and SMO is needed to induce apoptosis. Combined inhibition of Hh signaling and BCL2 function may have a therapeutic value for lymphomas with the t(14;18)(q32;q21). Disclosures: No relevant conflicts of interest to declare.

Abstract 972 Successful treatment of diffuse large B-cell lymphoma (DLBCL) is frequently hindered by development of resistance to conventional chemotherapy resulting in disease relapse and high mortality. High expression of anti-apoptotic and/or drug transporter proteins induced by oncogenic signaling pathways has been implicated in the development of chemoresistance in cancer. However, the cellular mechanisms responsible for drug resistance and treatment failure in DLBCL are poorly understood and their elucidation could lead to development of new therapeutic approaches. Activated Hedgehog (HH) signaling was identified in our laboratory to contribute to DLBCL cell survival and proliferation. We also found that inhibiting HH signaling resulted in decrease of BCL-2 protein and mRNA levels in ABC DLBCL cells but not in GC DLBCL cells, and that the resistance to apoptosis to HH inhibition observed in GC DLBCL cells can be overcome using the BCL2 inhibitor YC137. (Leukemia 2010;24:1025). We also found that the drug transporter protein ABCG2 is a potential prognostic factor in DLBCL given that its expression (as detected by immunohistochemistry) inversely correlated with overall survival and failure-free survival (Mod Pathol 2009;22:1312). Here, we investigate potential mechanistic connections between HH signaling and chemoresistance in DLBCL. Using chromatin immunoprecipitation (ChIP), site directed mutagenesis and luciferase reporter based assays, we confirmed the presence of a GLI-1 transcription factor-binding site in the ABCG2 gene promoter and we established ABCG2 gene as a direct downstream target of HH signaling. We also characterized the baseline expression and functionality of ABCG2 in DLBCL cell lines and explored the role of stroma in modulating its expression levels as well as the expression of other drug transporters (MDR1 and ABCC1) and the levels of the anti-apoptotic proteins, BCL-2, BCL2A1 and BCL-xL. Co-culturing DLBCL cell lines (BJAB and DOHH2) with human bone marrow stromal cells (HS5) resulted in decreased chemosensitivity to doxorubicin and methotrexate that was associated to marked increased expression levels of the drug transporters ABCG2, MDR1 and ABCC1 as well as of the expression levels of BCL2, BCL2A1 and BCL-xL. Our results also showed that pharmacologic inhibition of the activity of ABCG2, using fumitremorgin C and glefanine, significantly increased the chemosensitivity of BJAB and DOHH2 cells in the presence of stromal cells. However, this effect was not seen in the absence of stromal cells. Collectively, our data confirm that the stromal cells plays a role in inducing chemoresistance in DLBCL and that this effect is in part mediated by inducing high expression of drug protein transporters and antiapoptotic proteins. Inhibition of HH signaling as well as inhibition of drug transporters abrogates the stroma-induced chemoresistance suggesting that targeting these molecules may have a therapeutic value for the treatment of DLBCL. Disclosures: No relevant conflicts of interest to declare.