Transducin ß-Like Protein 1 Controls Multiple Oncogenic Networks In

Transducin β-like protein 1 controls multiple oncogenic networks in diffuse large B-cell lymphoma by Youssef Youssef, Vrajesh Karkhanis, Wing Keung Chan, Frankie Jeney, Alessandro Canella, Xiaoli Zhang, Shelby Sloan, Alexander Prouty, JoBeth Helmig-Mason, Liudmyla Tsyba, Walter Hanel, Xuguang Zheng, Pu Zhang, Ji-Hyun Chung, David M. Lucas, Zachary Kauffman, Kari - lyn Larkin, Anne M. Strohecker, Hatice G. Ozer, Rosa Lapalombella, Hui Zhou, Zijun Y. Xu- Monette, Ken H. Young, Ruolan Han, Elmar Nurmemmedov, Gerard Nuovo, Kami Maddocks, John C. Byrd, Robert A. Baiocchi, and Lapo Alinari

Haematologica 2020 [Epub ahead of print]

Citation: Youssef Youssef, Vrajesh Karkhanis, Wing Keung Chan, Frankie Jeney, Alessandro Canella, Xiaoli Zhang, Shelby Sloan, Alexander Prouty, JoBeth Helmig-Mason, Liudmyla Tsyba, Walter Hanel, Xuguang Zheng, Pu Zhang, Ji-Hyun Chung, David M. Lucas, Zachary Kauffman, Karilyn Larkin, Anne M. Strohecker, Hatice G. Ozer, Rosa Lapalombella, Hui Zhou, Zijun Y. Xu-Monette, Ken H. Young, Ruolan Han, Elmar Nurmemmedov, Gerard Nuovo, Kami Maddocks, John C. Byrd, Robert A. Baiocchi, and Lapo Alinari. Transducin β-like protein 1 controls multiple oncogenic networks in diffuse large B-cell lymphoma. Haematologica. 2020; 105:xxx doi:10.3324/haematol.2020.268235

Publisher's Disclaimer. E-publishing ahead of print is increasingly important for the rapid dissemination of science. Haematologica is, therefore, E-publishing PDF files of an early version of manuscripts that have completed a regular peer review and have been accepted for publication. E-publishing of this PDF file has been approved by the authors. After having E-published Ahead of Print, manuscripts will then undergo technical and English editing, typesetting, proof correction and be presented for the authors' final approval; the final version of the manuscript will then appear in print on a regular issue of the journal. All legal disclaimers that apply to the journal also pertain to this production process. Transducin β-like protein 1 controls multiple oncogenic networks in diffuse large B-cell lymphoma

Youssef Youssef,1 Vrajesh Karkhanis,1 Wing Keung Chan,1 Frankie Jeney,1 Alessandro Canella,1 Xiaoli Zhang,2 Shelby Sloan,1 Alexander Prouty,1 JoBeth Helmig-Mason,1 Liudmyla Tsyba,1 Walter Hanel,1 Xuguang Zheng,1 Pu Zhang,3 Ji-Hyun Chung,1 David M. Lucas,1 Zachary Kauffman,1 Karilyn Larkin,1 Anne M. Strohecker,4,5 Hatice G. Ozer,6 Rosa Lapalombella,1 Hui Zhou,7 Zijun Y. Xu- Monette,7 Ken H. Young,7 Ruolan Han,8 Elmar Nurmemmedov,9 Gerard Nuovo,10 Kami Maddocks,1 John C. Byrd,1 Robert A. Baiocchi,1 and Lapo Alinari1

1. Department of Internal Medicine, Division of Hematology, The Ohio State University, Columbus, OH, USA. 2. Center for Biostatistics, Department of Biomedical Informatics, The Ohio State University, Columbus, OH, USA. 3. Division of Pharmaceutics and Pharmaceutical Chemistry, College of Pharmacy, The Ohio State University, Columbus, OH, USA 4. Department of Cancer Biology and Genetics, The Ohio State University Columbus, OH, USA. 5. Department of Surgery, Division of Surgical Oncology, The Ohio State University Columbus, OH, USA. 6. Department of Biomedical Informatics, The Ohio State University, Columbus, OH, USA. 7. Department of Pathology, Division of Hematopathology, Duke University, Durham, NC, USA. 8. Iterion Therapeutics, Huston, TX, USA 9. Department of Translational Neurosciences and Neurotherapeutics, John Wayne Cancer Institute, Providence Saint John's Health Center, Santa Monica, CA, USA. 10. Discovery Life Sciences, Powell, OH, USA.

Article type: Original Research Article Running title: TBL1 modulates prosurvival proteins stability Abstract: 237 words. Main text: 3901 words. Figures: 6 Supplemental figures: 15 Supplemental tables: 5 Number of references: 49

Corresponding Author: Lapo Alinari, MD, PhD Division of Hematology , The Ohio State University, 410 W. 12th Ave, 481A Wiseman Hall, Columbus, OH 43210. Phone: 614-293-5594. [email protected]

Abstract

Diffuse large B-cell lymphoma (DLBCL) is the most common Non-Hodgkin’s lymphoma and is characterized by a remarkable heterogeneity with diverse variants that can be identified histologically and molecularly. Large-scale gene expression profiling studies have identified the germinal center B- cell (GCB-) and activated B-cell (ABC-) subtypes. Standard chemo-immunotherapy remains standard front line therapy, curing approximately two thirds of patients. Patients with refractory disease or those who relapse after salvage treatment have an overall poor prognosis highlighting the need for novel therapeutic strategies. Transducin β-like protein 1 (TBL1) is an exchange adaptor protein encoded by the TBL1X gene and known to function as a master regulator of the Wnt signalling pathway by binding to β-CATENIN and promoting its downstream transcriptional program. Here, we show that, unlike normal B-cells, DLBCL cells express abundant levels of TBL1 and its overexpression correlates with poor clinical outcome regardless of DLBCL molecular subtype.

Genetic deletion of TBL1 and pharmacological approach using tegavivint, a first-in-class small molecule targeting TBL1 (Iterion Therapeutics), promotes DLBCL cell death in vitro and in vivo.

Through an integrated genomic, biochemical, and pharmacologic analyses, we characterized a novel,

β-CATENIN independent, post-transcriptional oncogenic function of TBL1 in DLBCL where TBL1 modulates the stability of key oncogenic proteins such as PLK1, MYC, and the autophagy regulatory protein BECLIN-1 through its interaction with a SKP1-CUL1-F-box (SCF) protein supercomplex.

Collectively, our data provide the rationale for targeting TBL1 as a novel therapeutic strategy in

DLBCL.

Introduction

DLBCL comprises the majority of adult Non-Hodgkin Lymphoma cases worldwide and is traditionally categorized, based on the molecular profile, into two major subtypes, germinal center B- cell (GCB) and activated B-cell (ABC), mirroring the cell of origin as well as reflecting the differences in clinical outcomes.1-3 DLBCL is curable in 40-60% of the cases following standard front-line immuno-chemotherapy.4 However, patients with refractory disease, those who relapse after salvage chemotherapy and autologous stem cell transplant or chimeric antigen receptor T-cell therapy have an overall poor prognosis highlighting the need for novel therapeutic approaches.4-6

Transducin β-like protein 1 (TBL1), encoded by the TBL1X gene, is an adaptor protein initially identified as a core component of the co-repressor silencing mediator for retinoid and thyroid hormone receptor (SMRT)–nuclear receptor co-repressor (N-COR) complex7-9 SMRT/NCOR form a complex with TBL1, BCL6, and other proteins leading to transcriptional repression of target genes through HDAC3-mediated H3K9 acetylation.7, 10 Subsequently, TBL1 was found to be a key player in enhancing the canonical Wnt signaling pathway by directly binding to β-CATENIN and recruiting it to the promoter of Wnt target genes (MYC, BIRC5, CCND1) to promote uncontrolled cell proliferation and survival.9, 11, 12 Previous work also showed that TBL1 binds to a S-phase kinase-associated protein 1 (SKP1)/Cullin-1(CUL1)/F-box protein complex (SCF complex) and protects β-CATENIN from proteasomal degradation.13, 14 Tegavivint (Iterion Therapeutics) is a first-in-class small molecule compound that binds directly to the N-terminal domain of TBL1 and disrupts TBL1/β-CATENIN interaction leading to β-CATENIN degradation and subsequent downstream inhibition of the Wnt transcriptional program.15 Tegavivint has shown significant activity in preclinical models of acute myeloid leukemia (AML) and multiple myeloma (MM).15-17

Here, we report that, unlike normal B-cells, DLBCL cells express abundant levels of TBL1 regardless of the molecular subtype and its overexpression correlates with poor clinical outcome. We demonstrate that both TBL1 genetic deletion and pharmacologic targeting with tegavivint induce

3 lymphoma cell death in vitro and in vivo. More importantly, we show that β-CATENIN is dispensable for DLBCL cell survival and plays a minimal role in the tegavivint-mediated attenuation of known Wnt targets. Through an integrated genomic, biochemical, and pharmacologic approach we detail the crucial role played by TBL1 in controlling the stability of critical regulatory proteins involved in cell proliferation and autophagy induction. Our findings provide the rational for targeting TBL1 in DLBCL.

Methods

Cell lines and primary samples

DLBCL cell lines were purchased from ATCC and identity confirmed via short tandem repeats profiling (University of Arizona).18 Peripheral blood and tonsils from healthy donors and peripheral blood, bone marrow samples, and lymph nodes from DLBCL patients were obtained following written informed consent under a protocol approved by the Institutional Review Board of OSU in accordance with the Declaration of Helsinki.

Tissue microarray (TMA) and IHC

Details on TMA preparation and IHC in the supplementary methods.

Gene knock-down using shRNA and CRISPR-Cas9 system

TBL1X, CTNNB1, and CAND1 were genetically silenced using pLKO.1 shRNA vectors (Mission® shRNA, Sigma). Details in supplementary methods.

Immunoblot and co-immunoprecipitation assay

Details in the supplementary methods (Supplemental Table 4).

RNA extraction and real time (RT)-PCR

Details in the supplementary methods (see Supplemental Table 5).

Chromatin-immunoprecipitation (ChIP)-Sequencing and ChIP-PCR

For ChIP-Seq assays, 5x106 Pfeiffer and Riva cells treated with either vehicle control or tegavivint for

12 hours were processed according to Active Motif protocols.19 All raw sequence data are available in the Gene Expression Omnibus (GSE148232). ChIP-PCR was performed according to standard procedures. Details in the supplementary methods.

Confocal microscopy, Proximity Ligation Assay (PLA), Cytotoxicity assay and Transmission

Electron Microscopy (TEM)

Details in the supplementary methods.

Animal studies

Animal studies were approved by the OSU Institutional Animal Care and Use Committee.

Experimental details for the animal studies with tegavivint are included in the result section. For the in vivo TBL1X-knock-down experiment, luciferase and Cas9 positive Riva cells were infected with either doxycycline-inducible sgRNA targeting TBL1X or vector control only using a lentiviral luciferase vector previously described.20 20×106 of engineered Riva cells were engrafted into NSG mice via tail vein injection. On day 23, the tumor burden was visualized after intraperitoneal administration of 200 µL of

D-luciferin potassium salt (Caliper Life Sciences) in PBS using in vivo live-imaging system (IVIS)

(Perkin Elmer Inc.). Additional details in supplementary methods.

Statistics

All the experiments include at least 3 independent replicates and results were expressed as the mean

± standard error of the mean (SEM). For correlated data such as patient cells treated with different

5 conditions, paired t-tests for two group comparisons or linear mixed effects models for two or more group comparisons were used for analysis. Otherwise, for independent data, two-sample t-tests and analysis of variance (ANOVA) were used for comparisons between two groups and among more than two groups respectively. Survival analysis was performed using log-rank test or Cox proportional hazard regression models. p-values <0.05 were considered significant after Holm’s procedure adjustment for multiple comparisons.

Results

TBL1 is overexpressed and correlates with clinical outcome in DLBCL regardless of its molecular subtype

Immunoblot analysis showed abundant levels of TBL1 protein expression in 7 DLBCL cell lines regardless of the molecular subtype (4 GCB and 3 ABC, Figure 1A) and in primary DLBCL patient samples (n=4) (Figure 1B) (patient characteristics in Supplemental Table 1) compared to normal peripheral blood B-cells and CD10+CD19+ sorted germinal center B-cells from reactive tonsils

(sorting strategy in Supplemental Figure 1). To validate our findings, we performed immunohistochemistry (IHC) staining of TBL1 in tissue samples obtained from 83 primary de novo

DLBCL patients uniformly treated with front-line RCHOP (46 GCB- and 37 ABC-subtype, Figure 1C,

Supplemental Figures 2-4, patient characteristics in Supplemental Tables 2 and 3). In our cohort, the percentage of DLBCL cells with strong (brown) TBL1 staining (similar to the intensity seen in the positive controls) ranged from 5% to 90% (GCB-DLBCL: mean percentage of positive cells 46%; range 5% to 85%; ABC-DLBCL: mean percentage 49%; range 5% to 90%) compared to 5% to 15% with a mean of 8% in tonsils with reactive lymphoid hyperplasia (n=5; p=0.0001). Normal germinal center B-cells were identified by co-staining with CD10 (brown) and TBL1 (red) and co-expression showed in yellow (Figure 1C) or with TBL1 single staining (Supplemental Figure 2).21 No significant difference in the TBL1 expression patterns was seen between GCB- and ABC-DLBCL (dot plot in

Figure 1C). Using the median percentage of TBL1 positive lymphoma cells (50% for GCB-DLBCL and

55% for ABC-DLBCL), patients were classified into high or low expressors. Importantly, a high percentage of TBL1 positive DLBCL cells was associated with significantly inferior progression free survival (PFS) and overall survival (OS) in both ABC- and GCB-DLBCL (Figure 1D). However, the negative prognostic impact of TBL1 positivity on PFS/OS did not persist after adjusting for

International Prognostic Index (IPI) and IPI factors in a Cox multivariate model.

TBL1 is critical for DLBCL cell survival

In vitro TBL1 knock-down using either a doxycycline-inducible CRISPR-Cas9 system or a TBL1X- specific shRNA was associated with significant DLBCL cell death regardless of the molecular subtype

(Figure 2A) (p<0.05). Next, luciferase positive Riva cells were stably transduced with Cas9 and a doxycycline-inducible sgRNA targeting TBL1X exon 2 (guide 1). Prior to engraftment, efficient TBL1 knock-down after doxycycline induction was verified in vitro (Supplemental Figure 5). NSG mice (n=8 per group) received doxycycline (100mg/kg) daily via oral gavage starting at day 24 after successful engraftment was verified by IVIS on day 23. TBL1 knock-down induced tumor regression after 5 days of doxycycline treatment as documented by IVIS and significantly prolonged the survival of these animals compared to control [median OS 42 days vs 32 days, respectively, (p=0.0004)] (Figure 2B).

Next, we tested the in vitro cytotoxicity of tegavivint. As shown in Figure 3A-B, 24 hours treatment with tegavivint (2-100nM) induced significant DLBCL cell death in 8 DLBCL cell lines (4 GCB and 4

ABC) and primary DLBCL samples (n=8; patient characteristics in Supplemental Table1) compared to the vehicle control (p<0.001). Importantly, cytotoxic studies on normal resting and activated immune cell subsets from the peripheral blood of healthy donors (n=3) incubated for 24 hours with the highest concentration of tegavivint (100nM) showed no significant toxicity on B- and NK-cells while marginal toxicity (<10% decrease in viability) was noted in activated (p=0.0287) but not resting T-cells, consistent with their low TBL1 expression (Supplemental Figure 6A-B). In support of a selective

TBL1-mediated DLBCL cell death, tegavivint lost the majority of its cytotoxic activity in DLBCL cells when TBL1 was efficiently knocked-down via shRNA (Supplementary Figure 7).

To evaluate the activity of tegavivint in vivo, we employed three murine models representative of systemic ABC- and GCB-DLBCL: a) Seven-week-old NSG mice were engrafted with 20×106 Riva cells (ABC-DLBCL) via tail vein injection. Mice in groups of 5 were then randomized to receive either vehicle control or tegavivint at 25mg/kg every 3 days via tail vein injection starting at day 7 post engraftment. Tegavivint significantly prolonged survival in the treated group (median OS of tegavivint vs vehicle control mice 33 vs 27 days, p=0.0031) (Figure 3C). Notably, 4 of 5 tegavivint-treated mice died of central nervous system (CNS) involvement with minimal systemic disease evaluated by the spleen size (Figure 3C) consistent with the known limited CNS penetration property of tegavivint. b)

Seven-week-old NSG mice were engrafted with 40×106 SU-DHL10 cells (GCB-DLBCL) via tail vein injection. Mice in groups of 10 were then randomized to receive either vehicle control or intravenous tegavivint at 25mg/kg starting at day 3 post engraftment on a twice weekly schedule (Monday-

Thursday). Tegavivint significantly prolonged survival in the treated group (median OS of tegavivint vs vehicle control mice 47 vs 34.5 days, p=0.0002) (Figure 3D). c) Seven-week-old NSG mice were tail vein engrafted with 1×106 of passage 3 DFBL-18689 cells from an ABC-DLBCL patient derived xenograft (PDX) mouse model obtained from the public repository of xenografts (Proxe).22 Mice in group of 5 were then randomized to receive either vehicle control or tegavivint at 25mg/kg via tail vein injection on a twice weekly schedule (Monday-Thursday) starting at day 3 post engraftment.

Tegavivint significantly prolonged the survival of the treated animals (median OS of tegavivint vs vehicle control mice 55 vs 37 days, p=0.0026) (Figure 3E and Supplemental Figure 8).

Collectively, these data demonstrate that TBL1 is a novel therapeutic target in DLBCL and that tegavivint induces the desired pharmacologic effect by exhibiting selective TBL1-mediated lymphoma cell death with very limited toxicity to normal immune cells.

TBL1 modulates Wnt targets in a novel, post-transcriptional, β-CATENIN independent manner

Co-immunoprecipitation (Co-IP) experiments in Riva and Pfeiffer cells showed that treatment with tegavivint for 12 hours led to the disruption of TBL1/β-CATENIN interaction (Figure 4A) which was further confirmed via PLA (Figure 4A and Supplemental Figure 9A-B). ChIP-seq analysis in the same cell lines revealed that genome-wide TBL1 promoter occupancy was not significantly affected by tegavivint treatment for 12 hours (Supplemental Figure 10A) while ChIP-PCR validation of selected

Wnt/β-CATENIN target genes showed selective loss of β-CATENIN recruitment to the promoters of

MYC and BIRC5 (Figure 4B) upon treatment (p<0.0005) while TBL1 recruitment was unaffected.

Treatment of 4 DLBCL cell lines with tegavivint for 24 hours resulted in decreased protein levels of β-

CATENIN, MYC and SURVIVIN with no effect on TBL1 expression when compared to the untreated control (Figure 4C). Similarly, shRNA-mediated TBL1 knock-down led to significant decrease in MYC, and SURVIVIN expression compared to controls (Figure 4D). Despite the loss of β-CATENIN recruitment to the promoter regions of MYC and BIRC5 following treatment with tegavivint, qRT-PCR showed significant increase in MYC transcript in Riva (p<0.005) and Pfeiffer (p<0.05) and no changes in OCI-Ly3 and WSU-NHL (Figure 4E). A similar pattern was observed with the transcript of two other well-known β-CATENIN target genes, WISP1 and AXIN2 (Supplemental Figure 10B) while treatment with tegavivint resulted in a minimal but significant decrease in BIRC5 transcript in 3 out of 4 DLBCL cell lines tested (Figure 4E). Lastly, efficient shRNA knock-down of β-CATENIN did not significantly affect MYC or SURVIVIN protein expression and had minimal effect on cell viability in all the tested

DLBCL cell lines (Figure 4F). In further support of a TBL1-mediated and β-CATENIN independent process, tegavivint maintained both its cytotoxic activity and its effect on MYC and SURVIVIN (protein and transcript) in the context of efficient β-CATENIN knock-down (Supplementary Figure 11).

Collectively our data provide evidence that β-CATENIN contributes minimally to the transcriptional regulation of critical Wnt targets and to DLBCL cell survival and that TBL1 regulates the stability of MYC in a β-CATENIN independent fashion.

TBL1 is a critical component of the SCF complex in DLBCL

The SCF complex controls the ubiquitination and degradation of a large number of proteins through the specificity-conferring F-box proteins such as βTRCP and FBW7.23, 24 Its activity is regulated by the interaction of CUL1 with two proteins: 1) the cullin associated and neddylation dissociated 1 (CAND1) and 2) NEDD8. CAND1 functions as an inhibitor of the SCF complex while

NEDD8 binding to CUL1 promotes the targeted substrate ubiquitination and degradation.23 We hypothesized that TBL1 modulates target protein stability through the SCF complex and that targeting

TBL1 with tegavivint enhances SCF complex activity leading to increased downstream degradation of key regulatory proteins. Co-IP and PLA in Riva and Pfeiffer cells showed TBL1-SKP1-CUL1 association which was unaffected by tegavivint (Figure 5A and Supplemental Figure 12). Importantly, tegavivint treatment of Riva and Pfeiffer induced significant decrease of CAND1-CUL1 association by

PLA (p<0.005) (Figure 5B and Supplemental Figure 13) with no effect on the expression of these two proteins (Supplemental Figure 14A).

These results suggest a model in which TBL1, as a critical component of an SCF complex, modulates the stability of critical regulatory proteins (Figure 5C).

TBL1 modulates MYC stability through the SCF complex

Polo-like kinase-1 (PLK1), an essential regulator of mitosis and a novel target in lymphoma, is degraded by SCFβTrCP.25 Importantly, PLK1 plays a critical protective role in MYC stability by preventing its SCFFbw7-mediated proteasomal degradation.26-28 We hypothesized that targeting TBL1 with tegavivint may enhance the proteasomal-mediated degradation of MYC directly by SCFFbw7 and indirectly by enhancing SCFβTrCP–mediated PLK1 degradation. Time course analysis showed that treatment of 4 DLBCL cell lines with tegavivint led to PLK1 downregulation while leaving its transcript unaffected (Figure 5D and Supplemental Figure 14B-C), decreased MYC levels with concomitant increase in its phosphorylation on Ser62 and Thr58 residues at 10 hours, consistent with enhanced

MYC proteasomal degradation (Supplemental Figure 14D). Adding the translation inhibitor cycloheximide (CHX) to tegavivint significantly shortened the half-life of PLK1 protein compared to

CHX alone (Figure 5E and Supplemental Figure 14E). Consistent with this, the proteasomal inhibitor

MG132 partially restored PLK1 and MYC protein levels as well as cell viability when combined with tegavivint versus tegavivint alone (p<0.05) (Figure 5F-G). In further support of an SCF/TBL1- modulated process, efficient shRNA knock-down of CAND1 (and TBL1) in Riva and Pfeiffer cells led to significant decrease in viability as well as downregulation of PLK1 and MYC protein expression

(Supplemental Figure 14F-G).

Altogether, our data support a novel model in which TBL1 regulates MYC stability directly

β through SCFFbw7 and indirectly by enhancing SCF TrCP-mediated PLK1 degradation.

TBL1 is involved in the modulation of cytoprotective autophagy

Autophagy is a catabolic pathway that sustains metabolism by recycling cytoplasmic contents to support macromolecular synthesis during periods of cellular stress with accumulating evidence supporting its pro-survival role in lymphoma.29,30

As SCF complexes regulate the stability of key autophagy inducers such as BECLIN-1 and

MYC,31-33 we hypothesized that tegavivint impairs the autophagy flux by affecting these key

autophagy inducers. Immune blot analysis showed downregulation of BECLIN-1 protein level in 4

DLBCL cell lines at 12 and 24 hours (Figure 6A) whereas BECN1 transcript level was not affected

by tegavivint (Supplemental Figure 15). Co-treatment with CHX and tegavivint significantly

shortened the half-life of BECLIN-1 compared to either CHX or tegavivint alone (Figure 6B) and

incubation with MG132 partially restored BECLIN-1 protein level when combined with tegavivint

compared to tegavivint alone (Figure 6C). To evaluate the effect of tegavivint on the autophagy flux,

four DLBCL cell lines were treated with either DMSO control, chloroquine (autophagy blocker), the

rapamycin (autophagy inducer), tegavivint or combinations. The increase in the amount of LC3-II

11 observed in chloroquine-treated cells represents the amount of LC3 delivered to the lysosome for degradation and is a well-established measure of autophagic flux.34 As shown in Figure 6D, treatment with tegavivint led to an increase in the levels of LC3-II which could be due to autophagy induction or inhibition. Co-treatment with chloroquine led to further accumulation of LC3-II, suggesting autophagy induction in 2 of the 4 cell lines tested. However, treatment with rapamycin in combination with either tegavivint or chloroquine resulted in an increase of LC3-II compared with rapamycin alone across the 4 cell lines tested, suggesting that tegavivint may impair autophagic flux. To clarify this discordance, we transfected Riva and Pfeiffer cells with a GFP–mCherry-LC3 construct. This fluorescence reporter monitors LC3 flux by relying on both GFP sensitivity and mCherry resistance to the acidic/proteolytic environment in the lysosomes. With autophagy blockade, both mCherry (red fluorescence protein) and GFP (green fluorescence protein) co-localize in a vesicular compartment that has not fused with lysosomes (autophagosomes) resulting in yellow puncta, whereas autophagy induction leads to fusion of these compartments (autolysosomes) and subsequent quenching of GFP signal resulting in singly mCherry red puncta.35 As shown in Figure

6E, treatment of these cells with either chloroquine or tegavivint increased yellow puncta assessed by confocal microscopy whereas rapamycin treatment was associated with increased red puncta solely. By calculating the ratio between yellow and red puncta, we found that the yellow signal was significantly increased in chloroquine and tegavivint-treated DLBCL cells (p<0.05) compared to controls, suggesting autophagosome accumulation (Figure 6E-F). Ultrastructurally, Riva and Pfeiffer treated with chloroquine or tegavivint for 12 hours showed accumulation of enlarged vesicles with characteristic enclosed cytoplasmic ultrastructures, supporting the notion that tegavivint treatment impairs the terminal stages of autophagy (Figure 6G).

Collectively, these results indicate that tegavivint impairs cytoprotective autophagy by promoting

SCF-mediated degradation of key autophagy inducers such as MYC and BECLIN-1.

Discussion

In the work presented here, we showed that TBL1 encoded by the TBL1X gene is abundantly expressed in DLBCL cells regardless of the molecular subtype. While further studies are required to validate TBL1 on DLBCL samples using IHC, and for this to predict relapse in a prospective manner, in this initial evaluation we show that TBL1 overexpression correlates with poor survival in de novo

DLBCL patients uniformly treated with standard front-line immuno-chemotherapy.

Genetic deletion of TBL1X or pharmacologic treatment with tegavivint significantly affected

DLBCL cell viability in vitro while minimally toxic against normal immune cells and significantly prolonged the survival in four animal models of disseminated DLBCL, supporting the therapeutic potential of targeting TBL1 in this disease.

While tegavivint has been shown to mediate AML and MM cell death by disrupting the TBL1/β-

CATENIN interaction followed by proteasomal-mediated degradation of β-CATENIN and inhibition of the Wnt/transcriptional program,16, 17 our findings are novel and provide insight into a previously undiscovered post-transcriptional, β-CATENIN-independent oncogenic network modulated by TBL1 in

DLBCL. Despite β-CATENIN expression in a subset of DLBCL cases,36, 37 our data minimize the transcriptional relevance of β-CATENIN in promoting Wnt signalling and DLBCL cell survival while suggesting a possible role as a direct or indirect transcriptional repressor for some of the Wnt downstream targets (MYC, WISP1 and AXIN2) as previously shown in melanoma.38 However, this scenario may differ in a setting such as tumor microenvironment where cytokine-induced Wnt signaling is active.39

The SCF-type of E3 ubiquitin ligase is a multi-protein complex composed of three static subunits

RBX1, CUL1, SKP1 and variable F-box subunits which confer substrate selectivity to the complex.23,

24, 40-42 Here we showed that TBL1 modulates the stability of MYC, PLK1 and BECLIN-1 through its direct interaction with the SCF complex. Mechanistically, our data support the hypothesis that tegavivint binding to TBL1 leads to conformational changes in the SCF complex resulting in CAND1-

CUL1 dissociation followed by enhanced proteasomal degradation of these critical regulatory/prosurvival proteins. In further support of these findings, we showed that efficient shRNA knock-down of

CAND1 led to significant decrease in DLBCL viability as well as downregulation of PLK1 and MYC protein expression. CAND1 functions as an inhibitor of the SCF complex by preventing the access of

SKP1 and F-box proteins to the CUL1 catalytic core.23 In agreement with our work, C60, a small molecule inhibitor of CAND1, has been shown to destabilize CAND1-CUL1 interaction while increasing global ubiquitination in EBV-positive lymphoma cell lines.43

Interestingly, TBL1 itself has an F-box motif and a WD40 domain for substrate recognition42 suggesting the possibility that TBL1 directly interacts with specific targets to modulate their stability. In support of this hypothesis which warrants further investigation, data produced in human embryonic kidney 293T cells showed that TBL1 directly interacts with and protects β-CATENIN from Siah- mediated ubiquitination and degradation as part of an SCFTBL1 complex.13, 14

MYC rearrangement is identified in a subset of DLBCL cases (5%-15%), MYC protein overexpression is seen more commonly (30%–50% of DLBCL).44 Regardless of the genetic mechanism,

MYC over-expression is associated with a more aggressive clinical behaviour and inferior outcome in

DLBCL.45 In addition, it has been recently showed that PLK1 signalling promotes MYC protein stability, and in turn, MYC directly induces PLK1 transcription, establishing PLK1 as a therapeutic vulnerability in high grade B-cell lymphomas.28 Our findings are of particular importance because they indicate that tegavivint enhances the proteasomal-mediated degradation of MYC directly by SCFFbw7 and indirectly by enhancing SCFβTrCP–mediated PLK1 degradation supporting the notion that TBL1 holds promise as a therapeutic target particularly in MYC overexpressing DLBCL.

Cytoprotective autophagy plays an important role in promoting lymphoma cell survival in adverse conditions and adequate levels of BECLIN-1 are necessary to maintain the autophagy flux as supported by the fact that BECLIN-1+/- mice have defective autophagy.29, 30, 33, 46 Consistent with this published work and with that on MYC as a potent inducer of autophagy31, our data indicate that

14 tegavivint impairs pro-survival autophagy by promoting the SCF-mediated proteasomal degradation of both BECLIN-1 and MYC.

Although we have identified TBL1 to play a central role in modulating the turnover of PLK1, MYC, and BECLIN-1 through the SCF complex, there are a number of other potential SCF targets that will need to be explored to further define the role of TBL1 in the context of the SCF complex and to fully characterize the mechanism of action of tegavivint. For example, in addition to β-CATENIN and

PLK1, the TBL1 interacting SCF complex (SCFβTrCP) targets multiple proteins regulating mTOR signalling including Ras homolog enriched in brain protein, an upstream direct mTORC1 activator.47

Constitutive activation of mTOR signaling is critical for cellular growth and metabolism in both ABC- and GCB-DLBCL48 and mTORC1 inhibition is known to lead to lymphoma cell cycle arrest and compensatory autophagy induction.49 While TBL1 targeting represents an attractive approach to simultaneously affect multiple prosurvival signalling pathways leading to a more profound lymphoma cell death than that achievable with agents targeting each of these pathways individually, this approach, given the complexity of the prosurvival pathways regulated by TBL1, may induce upregulation of compensatory mechanisms ultimately leading to drug resistance and providing the rational for combination strategies.

While ongoing efforts are focused on studying the functional role of nuclear TBL1, the findings presented here demonstrate that TBL1 serves as a master regulator of a post-transcriptional oncogenic network and provide a rationale for TBL1 targeting in DLBCL with tegavivint which is currently in a phase-I clinical trial for patients with primary or recurrent desmoid tumors

(NCT03459469).

Acknowledgments

We are grateful to the patients and healthy volunteers who provided tissue samples for these studies, to the OSU Comprehensive Cancer Center Leukemia Tissue Bank Shared Resource (supported by

NCIP30 CA016058) for sample procurement, to the OSU Comparative Pathology and Mouse

Phenotyping Shared Resource (CPMPSR,supported in part by NCIP30 CA016058) for excellent histotechnologic support (Ms. Brenda Wilson and Ms. Tessa VerStraete), to the OSU Campus

Microscopy and Imaging Facility (CMIF supported in part by NIH Grant P30 CA016058) for excellent support with electron and confocal microscopy, to Somersault18:24 for the professional illustration of the proposed model of TBL1 post-transcriptional role within the SCF complex. The authors would also like to acknowledge Dr. Aharon Freud (OSU Department of Pathology) for providing normal pediatric tonsil controls.

Author contributions

YY and LA conceptualized and designed research studies and wrote the manuscript. YY, VK, WC, FJ,

AC, SS, AP, JHM, LT, WH, XZ, PZ, JC, ZK, EN, KL performed the experiments and analysed data.

HGO provided bioinformatics support. AMS and RH provided technical and material support. KHY provided the DLBCL primary samples for IHC studies. KHY, HZ, ZYX, and GV performed and scored the IHCs. XZ performed the statistical analysis. RL, KM, DML, JCB, and RAB provided conceptual advice and edited the manuscript. LA supervised the study and provided funding support.

Disclosure of Conflicts of Interest

Ruolan Han is an employee of Iterion therapeutics.The other authors have declared that no conflict of interest exists.

References

1. Teras LR, DeSantis CE, Cerhan JR, Morton LM, Jemal A, Flowers CR. 2016 US lymphoid malignancy statistics by World Health Organization subtypes. CA Cancer J Clin. 2016;66(6):443-459. 2. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-511. 3. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(25):1937-1947. 4. Sehn LH, Donaldson J, Chhanabhai M, et al. Introduction of combined CHOP plus rituximab therapy dramatically improved outcome of diffuse large B-cell lymphoma in British Columbia. J Clin Oncol. 2005;23(22):5027-5033. 5. Chow VA, Shadman M, Gopal AK. Translating anti-CD19 CAR T-cell therapy into clinical practice for relapsed/refractory diffuse large B-cell lymphoma. Blood. 2018;132(8):777-781. 6. Sarkozy C, Sehn LH. Management of relapsed/refractory DLBCL. Best Pract Res Clin Haematol. 2018;31(3):209-216. 7. Yoon HG, Chan DW, Huang ZQ, et al. Purification and functional characterization of the human N-CoR complex: the roles of HDAC3, TBL1 and TBLR1. EMBO J. 2003;22(6):1336-1346. 8. Guenther MG, Lane WS, Fischle W, Verdin E, Lazar MA, Shiekhattar R. A core SMRT corepressor complex containing HDAC3 and TBL1, a WD40-repeat protein linked to deafness. Gene Dev. 2000;14(9):1048-1057. 9. Perissi V, Aggarwal A, Glass CK, Rose DW, Rosenfeld MG. A corepressor/coactivator exchange complex required for transcriptional activation by nuclear receptors and other regulated transcription factors. Cell. 2004;116(4):511-526. 10. Hatzi K, Jiang Y, Huang C, et al. A hybrid mechanism of action for BCL6 in B cells defined by formation of functionally distinct complexes at enhancers and promoters. Cell Rep. 2013;4(3):578-588. 11. Li J, Wang C-Y. TBL1–TBLR1 and β-catenin recruit each other to Wnt target-gene promoter for transcription activation and oncogenesis. Nat Cell Biol. 2008;10(2):160-169. 12. Perissi V, Scafoglio C, Zhang J, et al. TBL1 and TBLR1 phosphorylation on regulated gene promoters overcomes dual CtBP and NCoR/SMRT transcriptional repression checkpoints. Mol Cell. 2008;29(6):755-766. 13. Dimitrova YN, Li J, Lee Y-T, et al. Direct ubiquitination of β-catenin by Siah-1 and regulation by the exchange factor TBL1. J Biol Chem. 2010;285(18):13507-13516. 14. Liu J, Stevens J, Rote CA, et al. Siah-1 mediates a novel β-catenin degradation pathway linking p53 to the adenomatous polyposis coli protein. Mol Cell. 2001;7(5):927-936. 15. Soldi R, Horrigan SK, Cholody MW, et al. Design, synthesis, and biological evaluation of a series of anthracene-9, 10-dione dioxime β-catenin pathway inhibitors. J Med Chem. 2015;58(15):5854-5862. 16. Fiskus W, Sharma S, Saha S, et al. Pre-clinical efficacy of combined therapy with novel β-catenin antagonist BC2059 and histone deacetylase inhibitor against AML cells. Leukemia. 2015;29(6):1267-1278. 17. Savvidou I, Khong T, Cuddihy A, McLean C, Horrigan S, Spencer A. β-catenin inhibitor BC2059 is efficacious as monotherapy or in combination with proteasome inhibitor bortezomib in multiple myeloma. Mol Cancer Ther. 2017;16(9):1765-1778. 18. ATCC S. Authentication of human cell lines: standardization of STR profiling. ATCC SDO Document ASN-. 2011. 19. de Almeida Nagata DE, Chiang EY, Jhunjhunwala S, et al. Regulation of tumor-associated myeloid cell activity by CBP/EP300 bromodomain modulation of H3K27 acetylation. Cell Rep. 2019;27(1):269-281. 20. Aubrey BJ, Kelly GL, Kueh AJ, et al. An inducible lentiviral guide RNA platform enables the identification of tumor-essential genes and tumor-promoting mutations in vivo. Cell Rep. 2015;10(8):1422- 1432. 21. Fiore C, Bailey D, Conlon N, et al. Utility of multispectral imaging in automated quantitative scoring of immunohistochemistry. J Clin Pathol. 2012;65(6):496-502.

22. Townsend EC, Murakami MA, Christodoulou A, et al. The public repository of xenografts enables discovery and randomized phase II-like trials in mice. Cancer Cell. 2016;29(4):574-586. 23. Goldenberg SJ, Cascio TC, Shumway SD, et al. Structure of the Cand1-Cul1-Roc1 complex reveals regulatory mechanisms for the assembly of the multisubunit cullin-dependent ubiquitin ligases. Cell. 2004;119(4):517-528. 24. Cardozo T, Pagano M. The SCF ubiquitin ligase: insights into a molecular machine. Nat Rev Mol Cell Biol. 2004;5(9):739-751. 25. Giráldez S, Galindo-Moreno M, Limón-Mortés MC, et al. G1/S phase progression is regulated by PLK1 degradation through the CDK1/βTrCP axis. FASEB J. 2017;31(7):2925-2936. 26. Popov N, Schülein C, Jaenicke LA, Eilers M. Ubiquitylation of the amino terminus of Myc by SCF β- TrCP antagonizes SCF Fbw7-mediated turnover. Nat Cell Biol. 2010;12(10):973-981. 27. Xiao D, Yue M, Su H, et al. Polo-like kinase-1 regulates Myc stabilization and activates a feedforward circuit promoting tumor cell survival. Mol Cell. 2016;64(3):493-506. 28. Ren Y, Bi C, Zhao X, et al. PLK1 stabilizes a MYC-dependent kinase network in aggressive B cell lymphomas. J Clin Invest. 2018;128(12):5517-5530. 29. Alinari L. Toward autophagy-targeted therapy in lymphoma. Blood. 2017;129(13):1740-1742. 30. Mathew R, Karantza-Wadsworth V, White E. Role of autophagy in cancer. Nat Rev Cancer. 2007;7(12):961-967. 31. Hart LS, Cunningham JT, Datta T, et al. ER stress–mediated autophagy promotes Myc-dependent transformation and tumor growth. J Clin Invest. 2012;122(12):4621-4634. 32. Cui D, Xiong X, Zhao Y. Cullin-RING ligases in regulation of autophagy. Cell Div. 2016;11(1):1-14. 33. Kang R, Zeh H, Lotze M, Tang D. The Beclin 1 network regulates autophagy and apoptosis. Cell Death Differ. 2011;18(4):571-580. 34. Mizushima N, Yoshimori T, Levine B. Methods in mammalian autophagy research. Cell. 2010;140(3):313-326. 35. Kimura S, Noda T, Yoshimori T. Dissection of the autophagosome maturation process by a novel reporter protein, tandem fluorescent-tagged LC3. Autophagy. 2007;3(5):452-460. 36. Ge X, Lv X, Feng L, Liu X, Wang X. High expression and nuclear localization of β-catenin in diffuse large B-cell lymphoma. Mol Med Rep. 2012;5(6):1433-1437. 37. Bognar M, Vincendeau M, Erdmann T, et al. Oncogenic CARMA1 couples NF-κB and β-catenin signaling in diffuse large B-cell lymphomas. Oncogene. 2016;35(32):4269-4281. 38. Spranger S, Bao R, Gajewski TF. Melanoma-intrinsic β-catenin signalling prevents anti-tumour immunity. Nature. 2015;523(7559):231-235. 39. Zhan T, Rindtorff N, Boutros M. Wnt signaling in cancer. Oncogene. 2017;36(11):1461-1473. 40. Ciechanover A. The ubiquitin–proteasome pathway: on protein death and cell life. EMBO J. 1998;17(24):7151-7160. 41. Hussain M, Lu Y, Liu Y-Q, et al. Skp1: implications in cancer and SCF-oriented anti-cancer drug discovery. Pharmacol Res. 2016;111:34-42. 42. Wang Z, Liu P, Inuzuka H, Wei W. Roles of F-box proteins in cancer. Nat Rev Cancer. 2014;14(4):233- 247. 43. Tikhmyanova N, Tutton S, Martin KA, et al. Small molecule perturbation of the CAND1-Cullin1- ubiquitin cycle stabilizes p53 and triggers Epstein-Barr virus reactivation. PLoS Pathog. 2017;13(7):e1006517. 44. Savage KJ, Johnson NA, Ben-Neriah S, et al. MYC gene rearrangements are associated with a poor prognosis in diffuse large B-cell lymphoma patients treated with R-CHOP chemotherapy. Blood. 2009;114(17):3533-3537. 45. Valera A, López-Guillermo A, Cardesa-Salzmann T, et al. MYC protein expression and genetic alterations have prognostic impact in patients with diffuse large B-cell lymphoma treated with immunochemotherapy. Haematologica. 2013;98(10):1554-1562. 46. Yue Z, Jin S, Yang C, Levine AJ, Heintz N. Beclin 1, an autophagy gene essential for early embryonic development, is a haploinsufficient tumor suppressor. Proc Natl Acad Sci U S A. 2003;100(25):15077-15082. 18

47. Harraz MM, Tyagi R, Cortés P, Snyder SH. Antidepressant action of ketamine via mTOR is mediated by inhibition of nitrergic Rheb degradation. Mol Psychiatry. 2016;21(3):313-319. 48. Sehn LH, Gascoyne RD. Diffuse large B-cell lymphoma: optimizing outcome in the context of clinical and biologic heterogeneity. Blood. 2015;125(1):22-32. 49. Wanner K, Hipp S, Oelsner M, et al. Mammalian target of rapamycin inhibition induces cell cycle arrest in diffuse large B cell lymphoma (DLBCL) cells and sensitises DLBCL cells to rituximab. Br J Haematol. 2006;134(5):475-484.

Figure 1. TBL1 is abundantly expressed and correlates with clinical outcome in DLBCL. (A)

Immunoblot showing abundant TBL1 expression in whole cell lysates from the indicated ABC- and

GCB-DLBCL cell lines compared to normal resting and activated peripheral blood (PB) B-cells and germinal center (GC) B-cells (n=3). (B) Immunoblot showing abundant TBL1 expression in whole cell lysates from 4 primary DLBCL patient (Pt) samples compared to normal B-cells and peripheral blood mononuclear cells (PBMCs). ACTIN was used as loading control; mantle cell lymphoma cell line

(Jeko) was used as positive control (+Ctl) for (A)-(B). (C) Immunohistochemical (IHC) staining of lymph nodes involved by ABC- and GCB-DLBCL (at 40x and 100x) showing strong (dark) nuclear and cytoplasmic positivity for TBL1. As comparison, tonsillar germinal center were co-stained with

CD10 (brown) and TBL1 (red). TBL1-CD10 co-expression at 40x and 100x is shown in yellow (black arrows) using the Nuance Multispectral Imaging System (model 3.0.2) and co-localization tool which allows to select any color for co-localization evaluation. Positive controls [tonsillar epithelium (left) and breast ductal carcinoma (right)] are both shown at 40x and 100x. Dot plot showing percentage of

TBL1 positive lymphoma cells by immunohistochemical staining in DLBCL patient samples [n=83 total with 46 germinal center B-cell subtype (GCB-DLBCL) and 37 activated B-cell subtype (ABC-DLBCL) compared to reactive tonsillar germinal center controls (n=5). Data represent median percentage of

TBL1 positive cells ± SD. ***p<0.0005 by ANOVA. (D) Percentage of TBL1 positive lymphoma cells

(≥55% for ABC-DLBCL and ≥50% for GCB-DLBCL) inversely correlated with progression free survival

(PFS) and overall survival (OS) in de novo GCB- and ABC-DLBCL patients treated with front-line R-

CHOP chemotherapy. p values were determined using the log-rank tests.

Figure 2. TBL1 genetic knock-down results in significant DLBCL cell death in vitro and in vivo.

(A) TBL1 was knocked-down (KD) using either a doxycycline-inducible CRISPR-Cas9 system (2 different guides: 1 and 2 targeting exon 2 and 3, respectively) or a TBL1-specific shRNA construct

(n=3). 72 hours after transduction, efficient TBL1 knock-down was verified via immunoblot and viability determined by annexin/propidium iodide staining and flow cytometry. Data represent means ±

SEM. ns: p>0.05, *p<0.05, **p<0.005, by paired t-test. Empty vector and scramble were used as controls. +Dox/-Dox: with (red) or without (black) doxycycline 0.75μg/mL. Numbers below the immunoblots reflect normalized value of quantified protein bands relative to the controls. (B) Kaplan–

Meier curve showing overall survival (OS) of NSG mice engrafted with Cas9+ Riva cells transduced with either a doxycycline-inducible TBL1-targeting sgRNA (n=8) or empty vector control (n=8). Daily doxycycline induction was started on day 24, IVIS images were obtained on day 23 and day 29 post engraftment. Median OS was 32 days for the empty vector controls (Ctl) versus 42 days for the TBL1 knock-down (KD) group (p=0.0004 using the log-rank test).

Figure 3. Tegavivint shows significant activity in in vitro and in vivo models of DLBCL. (A-B)

Cytotoxicity assay in the indicated DLBCL cell lines (n=3) and DLBCL patient samples [n=8: peripheral blood (n=4), lymph node (n=3) and bone marrow (n=1)] treated with the indicated concentration of tegavivint (T) in nM for 24 hours. Viability was determined by annexin/propidium iodide staining and flow cytometry. Data represent means ± SEM. *p<0.05, **p<0.005, ***p<0.0005 by linear mixed effects models with adjustment for multiple dose comparisons in (A) and with two sample t-test in (B). (C) Kaplan–Meier curve showing OS of NSG mice engrafted with Riva (ABC-DLBCL) and randomized to receive either vehicle control (n=5) or tegavivint (n=5) given at 25mg/kg via tail vein injection every 3 days starting at day 7 post engraftment. Median OS was 27 days for the controls

(Ctl) versus 33 days for the treated group (p=0.0031 using the log-rank test). Representative picture of the spleen size (left image) and H&E stain showing brain involvement by DLBCL cells (right image). (D) Kaplan–Meier curve showing OS of NSG mice engrafted with SU-DHL10 (GCB-DLBCL) and randomized to receive either vehicle control (n=10) or tegavivint (n=10) given at 25mg/kg via tail vein injection twice weekly starting at day 3 post engraftment. Median OS was 34.5 days for the controls (Ctl) versus 47 days for the treated group (p=0.0002 using the log-rank test). (E) Using an adaptive transfer model of DFBL-18689, recipient NSG mice (n=5/group) were randomized to receive either vehicle control or tegavivint at 25mg/kg via tail vein injection on a twice weekly schedule

(Monday-Thursday) starting at day 3 post engraftment. Kaplan–Meier curve showing OS (p=0.0026 using the log-rank test). Median OS: 55 days (tegavivint) and 37 days (control, ctl). Two mice in the treated group were censored due to engraftment failure assessed by flow cytometry for circulating lymphoma cells and MRI for spleen volume.

Figure 4. TBL1 modulates MYC levels in a post-transcriptional/β-CATENIN independent manner. (A) DLBCL cell lines were treated with either DMSO control (NT) or tegavivint (T) for 12 hours. Lysates were immunoprecipitated with anti-TBL1 or anti-IgG and then immunoblotted with the indicated antibodies. Confocal images of PLA (60x) using anti-TBL1 and β-CATENIN antibodies after treatment of DLBCL cells with either DMSO control (NT) or tegavivint (T) for 12 hours. Red indicates

PLA signal, blue indicates cell nuclei (Dapi) (n=3). (B) ChIP assay on crosslinked chromatin from

DLBCL cells after treatment with either DMSO control (NT) or tegavivint (T) for 12 hours. Either preimmune (PI) or the indicated immune antibodies were used and the retained DNA was amplified using MYC or BIRC5 promoter-specific primers. Fold enrichment with each antibody was calculated relative to the PI. n=3; data represent means ± SEM. ns: p>0.05, *p<0.05, ***p<0.0005 by linear mixed effects models. (C) Immunoblot showing TBL1, β-CATENIN, MYCc-Myc and SURVIVIN expression after treatment of DLBCL cells with either DMSO control (NT) or tegavivint (T) for 24 hours (n= 3). (D) Immunoblot showing MYC and SURVIVIN protein levels after TBL1 knock-down

(KD) using two TBL1 specific shRNA constructs (n=3). Lysates are the same used in Figure 2A however probed for different proteins. (E) qRT-PCR showing the mRNA fold changes of the indicated oncogenes after treating DLBCL cells with tegavivint (T) for 12 hours relative to the untreated control

(NT). n=3, data represent means ± SEM. ns: p>0.05, *p<0.05, **p<0.005 by paired t-test. (F)

Immunoblot showing MYC and SURVIVIN expressions after efficient β-CATENIN KD using two β-

CATENIN specific shRNA constructs. Viability obtained by trypan blue exclusion 72 hours post- transduction. n=3, data represent means ± SEM. ns: p>0.05 by paired t-test (ns=not significant).

Tegavivint (T): Riva, Pfeiffer: 70nM; OCI-ly3: 50nM; WSU-NHL: 15nM (for all experiments).

Figure 5. TBL1 is a critical component of a SCF complex in DLBCL through which it modulates

MYC stability. (A) The indicated DLBCL cell lines were treated with either DMSO control (NT) or tegavivint (T) for 12 hours, immunoprecipitated with either an anti-TBL1 or anti-SKP1, and then immunoblotted with the indicated antibodies. β-CATENIN and Histone H3 were used as positive and negative controls, respectively. (B) Confocal images (60x) of PLA showing the CAND1-CUL1 interaction in the indicated DLBCL cells lines after treatment with either DMSO control (NT) or tegavivint (T) for 12 hours. Red indicates PLA signal, blue indicates cell nuclei (Dapi). Graph represents the mean value of PLA signal from no less than 300 cells (n=3). **p<0.005 by paired t-test.

(D) Immunoblot showing PLK1 and MYC expression levels after treatment with either DMSO control

(NT) or tegavivint (T) at the indicated time points. (E) Immunoblot showing PLK1 level after treatment with either the translation inhibitor cycloheximide (CHX), tegavivint (T) or the combination (CHX+T) for the indicated time points. (F) Immunoblot showing PLK1 and MYC expression after treatment with either DMSO control (NT), tegavivint (T), the proteasome inhibitor MG132 or the combination

(T+MG132) for 24 hours. (G) Cytotoxicity assay in DLBCL cell lines treated with either DMSO control

(NT) or tegavivint (T), MG132 or the combination (T+MG132). Viability was determined by annexin

/propidium iodide staining and flow cytometry at 24 hours. n= 3, data represent means ± SEM.

*p<0.05, **p<0.005 by linear mixed effects models including interaction test between T and MG132.

CHX: 70ug/mL added for 1 hour and then washed before adding tegavivint treatment. MG132 (Riva,

OCI-Ly3 and WSU-NHL: 0.5uM and Pfeiffer: 0.3uM) was added 6 hours after tegavivint treatment was started.

Figure 6. TBL1 modulates cytoprotective autophagy through the SCF complex. (A) Immunoblot showing BECLIN-1 expression in the indicated DLBCL cell lines after treatment with either DMSO control (NT) or tegavivint (T) for 12 hours and 24 hours (n=3). (B) Immunoblot showing BECLIN-1 expression in the indicated DLBCL cell lines after treatment with either cycloheximide (CHX), tegavivint (T) or the combination (CHX+T) (n=3). (C) Immunoblot showing BECLIN-1 expression in the indicated DLBCL cell lines after treatment with either DMSO control (NT), tegavivint (T), MG132 or the combination (T+MG132) at 24 hours (n=3). (D) Immunoblot showing BECLIN-1 and microtubule-associated protein light-chain 3 (LC3-I and LC3-II) expression in the indicated DLBCL cell lines after treatment with either DMSO control (NT), tegavivint (T), chloroquine (CQ), rapamycin

(Rapa) or combinations at 12 hours and 24 hours (n=3). (E) Top: schematic depicting tandem LC3 reporter. Bottom: Confocal microscopy images (120x) of the indicated DLBCL cell lines transduced with a lentivirus expressing a tandem LC3 plasmid (mCherry-eGFP-LC3) and subsequently treated with either DMSO control (NT), tegavivint (T), chloroquine (CQ) or rapamycin (Rapa) for 12 hours. (F)

Histograms represent quantification of yellow (GFP/mCherry) puncta/total red (mCherry) puncta (%) of no less than 100 cells per condition. n= 4, data represent means ± SEM. *p<0.05, **p<0.005 by a linear mixed effects model with Holm’s adjustment for multiple comparisons for each cell line. (G)

Representative transmission electron microscopy images showing ultrastructural changes in the indicated DLBCL cell lines after treatment with either DMSO control (NT), tegavivint, chloroquine

(CQ) or rapamycin (Rapa) at 12 hours. The arrows indicate accumulation of autophagic vacuoles containing cytoplasmic material after exposure to tegavivint or chloroquine (n=3). Chloroquine: 50uM and Rapamycin: 10uM for all cell lines.

Supplementary methods

Cell lines and Primary samples

DLBCL cell lines were grown in standard RPMI-1640 medium supplemented with 10% FBS, 1 mM sodium pyruvate and 1% Penicillin-Streptomycin and routinely confirmed to be Mycoplasma negative

(MycoAlert Detection Kit, Lonza).

Primary DLBCL cells were isolated using CD19 magnetic beads and cultured according to standard methods. Patient characteristics in Supplemental Tables 1 and 2. Normal B-, T- and NK-cells of peripheral blood obtained from healthy volunteers were isolated using positive selection magnetic beads (Miltenyi Biotech). Purity of the isolated immune cells was determined by FACS analysis using immune cell subset specific markers. For activating the lymphocytes: B-cell and T-cells were treated with 3.2uM CpG and anti-CD3/CD28 respectively and harvested after 24 hours (doxorubicin was used as positive control for B-cells); NK-cells were treated with IL-2/IL-12/18 and harvested after 15 hours following standard methods.

For germinal center B-cells, total mononuclear cells were isolated from pediatric tonsils as described in1 without performing any further selection. Cells were then stained with the following antibodies:

CD10-PE, CD19-FITC, CD56-APC, CD3-APC, and CD14-PerCP (BD Biosciences) according to the manufacture protocol. CD14-CD3-CD56-CD19+CD10+ cells were then sorted to ≥97% purity with a

FACSAriaII sorter (BD Biosciences).

Reagents

Tegavivint was provided by Iterion Therapeutics. MG132, cycloheximide and chloroquine were all purchased from Sigma. Rapamycin and doxorubin were purchased from Selleckchem. Doxycycline was purchased from Tocris.

Tissue microarray (TMA) and Immunohistochemistry (IHC) 1

For IHC on human samples, hematoxylin-eosin stained slides from each of the 83 cases of DLBCL were reviewed, and TMA was constructed from tumor cell-rich areas. The 83 DLBCL cases were previously reported by Visco C et al. and classified into ABC and GCB based on integrated gene expression profiling and immunohistochemical studies.2

IHC studies for a variety of markers using a streptavidin-biotin complex technique were performed on

4-µm TMA sections in all cases. The markers assessed included: CD30, CD10, Bcl-6, GCET1,

MUM1, FOXP1, BCL2, MYC, p53, p21 and Ki-67. A cut-off value for each marker was established from analysis of receiver-operating characteristic curves to achieve maximum specificity and sensitivity as described previously. Immunohistochemistry for TBL1 and CD10 (for germinal center identification within normal tonsils) was performed on formalin-fixed paraffin-embedded tissues using an anti-TBL1 antibody (abcam, ab24548) and anti-CD10 antibody (Ventana, 790-4506) according to the manufacturer’s instructions. Enzo Life Sciences HRP-conjugated secondary antibody was used for detection. Tissue underwent antigen retrieval in 10mM citrate buffer (pH 9.0) at 95˚C for 30 minutes. TBL1 and CD10 co-expression analysis and quantification was performed with the Nuance

Multispectral Imaging System and Fast Red (TBL1: red), and DAB (CD10: brown) with yellow indicating co-expression (Nuance software using the co-localization tool).3

TBL1 protein expression was recorded in 5% increments as the percentage of positive lymphoma cells. Tonsillar epithelium and a breast ductal carcinoma tissue biopsy were used as positive controls.

Immune cells in the background were used as negative control. After TBL1 staining, patients were dichotomized based on the median percentage of TBL1 positive cells: GCB-DLBCL cases with ≥50% and ABC-DLBCL cases with ≥55% of TBL1 positive cells were defined as high expressors, the remaining patients were defined as low expressors. For samples obtained from the human cell line xenograft model (Riva), all tissues were fixed by immersion in 10% neutral buffered formalin, processed by routine methods and embedded in paraffin. Six µM-thick sections were stained with hematoxylin and eosin on a Leica ST5020 Multistainer (Leica Biosystems, Buffalo Grove, IL) by the

2 histology laboratory of the Comparative Pathology and Mouse Phenotyping Shared Resource. A board-certified veterinary anatomic pathologist assessed the brain sections under bright-field microscopy (Nikon Ci-L microscope, Nikon Instruments, Melville, NY; representative photomicrographs taken using an Olympus SC180 digital camera, B & B Microscopes Limited,

Pittsburg, PA).

Gene knock-down using shRNA and CRISPR-Cas9 systems

Lentiviral particles were produced by transient transfection of 10 × 106 Lenti-X 293T cells (Takara Bio

USA Inc.) with 15 µg of vector DNA along with the packaging constructs psPAX (15 μg), and pVSG (3

μg) using Lipofectamine 2000 (Thermo Fisher Scientific). Virus containing supernatants were collected at 72 hr after transfection, and used to infect DLBCL cell lines using spinoculation. Briefly, the cells were mixed with the indicated lentiviral particles at the MOI of no more than 10 and centrifuged at 3900 rpm at 32°C for 90 min. The cells were recovered, and either sorted (GFP+) or incubated at 37°C for 48 hours in puromycin (1ug/mL) to select the stable clones. Three days after selection, the cells were withdrawn with the puromycin selection for 2 days before downstream experiments were performed. The successful knock-down were confirmed by RT-PCR with gene- specific probes/primers in each experiment. shRNA from Sigma catalogue number for each gene: β-CATENIN: TRCN0000314921,

TRCN0000314991; TBL1: TRCN0000299525. CAND1: TRCN000000346. TBL1X Human shRNA

Plasmid Kit, CAT#: TL316751 from OriGene Technologies (Rockville, MD). sgRNA sequences for TBL1 knock-down: TBL1X: Guide 1 targeting exon 2: 5’-

TCCCAAACGTGAAAGCCGAG-3’. Guide 2 targeting exon 3 Oligo1 5’-

AGACGCCGTGATGCCCGACG-3’. Guide 3 targeting exon 5 Oligo1 5’-

CTTTGTTACTCGGGACGTCA-3’

Immunoblot and co-immunoprecipitation assay

Cell were treated with either vehicle control or the indicated concentration of tegavivint at the indicated time points, then for whole cell extracts, cells were lysed in RIPA buffer [10 mM Tris-HCl

(pH 7.4), 150 mM NaCl, 1% Triton X-100, 0.1% SDS and 1% sodium deoxycholate] containing protease and phosphatase inhibitors (all from Sigma) and cell lysates were clarified by centrifugation.

Nuclear and cytosolic extracts were prepared using the NE-PER kit (Pierce) according to the manufacturer's recommendations. Proteins were analysed by immunoblot using standard procedures.

To show TBL1, β-CATENIN and SKP1 interaction, preimmune immune, anti-TBL1, anti-β-CATENIN and anti-SKP1 antibodies were crosslinked to the magnetic beads according to the manufacturer's protocol (abcam). Whole cell extracts were incubated with these crosslinked antibodies overnight at

4°C. Beads were collected by centrifugation, washed five times with 0.5 mL of washing buffer [40 mM

Tris-HCl (pH 8.0), 180 mM NaCl], and bound proteins were analyzed by western blotting. Western blots were captured by either using an enhanced chemiluminescence substrate for detection of scant

HRP (Thermo Fisher Scientific; Ref 32106) or by LI-COR Biosciences Odyssey Infrared Imaging

System using IRDye antibodies (Two-color 181 multiplex detection). Protein bands were normalized using LI-COR Image Studio software (Version 5.2). Western blot bands intensity, for each protein, were first calculated (normalized) considering control sample(s) as the reference. Then the calculated values were normalized based on the difference between the loading control (ACTIN or GAPDH).

Details on antibodies used are in Supplemental Table 4.

RNA extraction and RT-PCR

Total RNA from DLBCL cells was isolated using TRIzol reagent (Thermo Fisher Scientific) and RNA

Clean-Up and Concentration Kit (Cat. 23600) (Norgen Biotek Corp) according to the manufacture protocol. The mRNA levels of target genes were determined by RT-PCR using predesigned Taqman gene expression assay accordingly to the manufacturer's instruction (ABI Biosystems, Thermo

Scientific). The assays were run on Viia 7 real-time PCR system (ABI Biosystems) and the data was analyzed by ΔCT calculation relative to the reference gene GAPDH. Details on specific primers and probes are in Supplemental Table 5.

Chromatin-immunoprecipitation (ChIP)-Sequencing and ChIP-PCR

For Chip-Seq assays, cell pellets were snap frozen and shipped on dry ice to Active Motif where subsequent steps were performed according to previously published pipeline.4 TBL1 ChIP reactions were performed using 30 μg of DLBCL cell lines chromatin and 8 μl of TBL1 antibody (abcam, ab24548). Illumina sequencing libraries were prepared from the ChIP and Input DNAs by the standard consecutive enzymatic steps of end-polishing, dA-addition, and adaptor ligation. After a final

PCR amplification step, the resulting DNA libraries were quantified and sequenced on Illumina’s

NextSeq 500 (75 nt reads, single end). Raw sequence reads (75bp single end) were aligned to human reference genome (hg19) using bwa version 0.7.12.5 Normalized (reads per million) bigwig files were generated using bamCoverage function of deepTools.5 Heatmaps depicting genome-wide enrichment of the ChIP signal around transcription start site were generated using computeMatrix and plotHeatmap functions of deepTools.6

For ChIP-PCR, cross-linked chromatin was prepared from Pfeiffer, Riva and OCI-Ly3 cell lines treated with either vehicle control or biologically relevant concentration of tegavivint for 12 hours. After sonication, cross-linked chromatin was digested with micrococcal nuclease (1 unit/ml) at 37oC for 20 min before adding 200 μL of stop buffer (100 mM Tris-HCl [pH 8.6], 0.45% SDS, 2.5% Triton X-100, 5 mM EDTA [pH 8.0], and protease inhibitors). Chromatin was analyzed by agarose gel electrophoresis to ensure that DNA fragment sizes did not exceed 500 bp, and ChIP assays were carried out essentially as described previously.7 To amplify DNA sequences of MYC, pre-immune and immune antibodies raised against TBL1 (abcam, ab24548), and β-CATENIN (CST; mAb#8814), after extensive washing, nucleoprotein complexes were eluted in 200 μl of elution buffer (50 mM Tris-HCl [pH 8.0],

10 mM EDTA, 1% SDS), and cross-links were reversed by incubating samples at 65oC for 12 hr. Immunoprecipitated DNA was purified by phenol and chloroform extraction, and resuspended in

40 μl of TE buffer (10 mM Tris-Hcl [pH 8.0], 1 mM EDTA). mRNA promoter sequences were amplified using gene-specific primers and locked nucleic acid probes from the Human Universal ProbeLibrary

Set (Millipore Sigma). The following primers and probes were used in this study: MYC (forward: 5’-

GAAATTGGGAACTCCGTGTG-3’, reverse: 5’-CTAGGGCGAGAGGGAGGTT-3’; probe 53) and

BIRC5 (forward: 5’-AAGGAGGAGTTTGCCCTGAG-3’, reverse: 5’-GGGCCACTACCGTGATAAGA-3’; probe 24).

Confocal microscopy

LC3 tandem reporter (pCDH-CMV-mC-G-LC3B-P was a gift from Kazuhiro Oka (Addgene plasmid #

124974). Accumulation and co-localization of GFP and mCherry fluorescence signals from the GFP– mCherry–LC3 positive B-cell NHL cell lines were examined by laser scanning confocal microscopy.

After treatment with the indicated reagents, cells were fixed with 4% paraformaldehyde for 10 minutes at 37°C and then washed with PBS (pH 7.4), and then cytospun on the glass slides. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) containing mount medium (abcam, ab104139) for 5 minutes at room temperature and then coversliped for confocal microscopy observation.

Fluorescence of GFP (green), mCherry (red) and their colocalization (yellow) were analyzed by

Olympus Filter FV1000 Laser Scanning Confocal Imaging Systems and IMARIS microscopy image analysis software. At least 100 cells were counted in control and each treatment group.

Proximity Ligation Assay (PLA)

The tegavivint-treated Riva and Pfeiffer cell lines cells were fixed with 4% paraformaldehyde for 10 minutes at 37°C. They were permeabilized with 0.1% Triton-X-100 for 15 minutes at 37°C, washed in

PBS and blocked in Duolink Blocking buffer for 30 min at room temperature. Primary antibodies were

6 diluted in Duolink Antibody Diluent and incubated overnight at 4 °C. Where appropriate, cells were counterstained using Anti-IgG (CST#3900S and CST#5415S), mouse anti-TBL1 (Sigma;

SAB1411330), rabbit anti-β-CATENIN (Cell Signalling Technology; 2677), rabbit anti-CUL1 (Sigma:

SAB1411513), and mouse anti-CAND1 (Sigma: SAB1405164) primary antibodies overnight. The cells were washed for 20 min in a large volume of PBS with 1% BSA, followed by addition of the appropriate Duolink secondary antibodies. Cells were incubated for 1 h at 37 °C, after which cells were washed in TBST with 0.5% Tween-20 for 10 min. Ligation and amplification steps of the PLA were performed using the Duolink in situ Detection Reagents red kit (Sigma) according to the manufacturer’s instructions. Following the PLA, Nuclei were stained with 4′,6-diamidino-2- phenylindole (DAPI) containing mount medium (abcam, ab104139) for 5 minutes at room temperature and then coversliped for confocal microscopy observation. PLA signal was analyzed by using

Olympus Filter FV1000 Laser Scanning Confocal Imaging Systems. PLA spots were counted in cell lines using IMARIS software. PLA scores were determined by normalizing the number of PLA spots counted in each sample to the number of cell counted in the same sample. At least 300 cells were counted for each condition.

Cytotoxicity assay

DLBCL Cell lines and primary DLBCL samples were treated with either vehicle control or the indicated concentration of tegavivint 24 hours, then cells were stained with 5 μl of FITC-Annexin V antibody (BD Bioscience) and 5 μl of propidium iodide solution (BD Biosciences) for 15 min at room temperature in the provided buffer before they were analyzed on a BD LSRFortessa™ flow cytometer. Data were analyzed with Flowjo version 10 software (BD Bioscience).

Transmission Electron Microscopy (TEM)

DLBCL cells were treated as indicated and then fixed for 30 minutes in 2.5% glutaraldehyde in 0.1M phosphate buffer, pH 7.4. Further processing was performed by the OSU Campus Microscopy and

Imaging Facility. Images were obtained with a FEI Tecnai G2 Spirit TEM.

Animal studies

All animals were monitored daily for signs of tumor burden, including hind limb paralysis, respiratory distress, weight loss, ruffled coat, and distended abdomen. Body weight was measured weekly for the first 3 weeks and then daily. Animals were sacrificed when these symptoms were observed and subjected to histopathologic evaluation to confirm the presence of DLBCL. The primary endpoint of these studies was survival defined as the time from engraftment to the development of the defined clinical criteria leading to removal from the study. NSG mice were purchased from Jackson.

Supplemental Table 1: Patients’ characteristics of the DLBCL patient samples (n=8).

DLBCL Stage at Source of Previous Patient no. Age/Sex subtype diagnosis tumor cells therapy ABC, double Peripheral 1 62/M expressor IVA blood Yes GCB, with Peripheral 2 70/F t(8;14) IVB blood No GCB, with Peripheral 3 62/M t(8;14) IVB blood Yes 4 55/F ABC IIIA LN No 5 67/M GCB-DH IVA LN No DLBCL- NOS, DE, 6 48/M HIV IVB LN Yes DLBCL- 7 59/M NOS IVB BM No Peripheral 8 68/F GCB IVA blood Yes

Supplemental Table 2. Patients’ characteristics of the 83 DLBCL cases included in Figure 1, D and E.

Front-line Percentage OS PFS status Female B- chemo ABC of TBL1 status (1= Patients (Yes: symptoms response (Yes: positive OS (0=alive; PFS relapsed (n=83) 1) (Yes: 1) (R-CHOP) 1) DLBCL cells (months) 1=dead) (months) /refractory) 1 1 0 PR 0 70 13.25 1 13.25 1 2 0 0 CR 0 80 44.35 0 44.35 0 3 1 0 PD 1 90 3.58 1 3.58 1 4 0 1 PR 1 80 11.57 1 11.57 1 5 1 0 PD 0 85 38.6 0 38.6 0 6 0 0 SD 0 5 39.16 0 39.16 0 7 1 1 CR 0 50 36.1 0 36.1 0 8 1 1 CR 0 20 32.25 0 32.25 0 9 1 1 CR 1 5 27.95 0 27.95 0 10 1 0 CR 1 40 29.19 0 29.19 0 11 1 1 CR 0 10 25.58 0 25.58 0 12 0 0 CR 0 70 22.78 0 22.78 0 13 0 1 CR 0 55 16.83 0 16.83 0 14 1 0 CR 0 50 20.91 0 20.91 0 15 0 0 SD 0 65 46.68 0 46.68 0 16 1 1 PR 1 75 3.72 1 3.72 1 17 1 0 CR 0 65 50.63 0 43.89 1 18 0 1 CR 1 75 6.74 1 6.74 1 19 0 1 CR 1 55 57.04 0 57.04 0 20 1 1 CR 0 15 50.4 0 50.4 0 21 1 0 CR 1 35 36.89 0 36.89 0 22 0 0 CR 1 85 15.25 1 15.25 1 23 0 0 CR 0 35 53.95 0 53.95 0 24 0 0 CR 0 30 61.87 0 61.87 0 25 0 1 CR 0 50 53.69 0 53.69 0 26 0 0 CR 1 10 38.89 1 38.89 1 27 1 1 CR 1 65 74.5 0 74.5 0 28 0 1 CR 0 5 73.78 0 73.78 0 29 1 0 CR 1 75 75.06 0 75.06 0 30 0 0 CR 1 40 36.16 0 36.16 0 31 0 0 CR 1 45 66.31 1 66.31 1 32 0 0 CR 1 30 78.12 0 78.12 0 33 0 0 PD 0 65 1.94 1 1.22 1 34 1 1 CR 0 30 70.98 0 70.98 0 35 1 0 CR 1 10 79.76 0 41.06 1 36 0 0 PR 0 60 8.32 1 8.32 1 37 0 1 CR 1 65 84.3 0 11.47 1 38 1 0 CR 0 40 84.2 0 84.2 0 39 0 0 CR 1 60 68.75 0 68.75 0

40 0 1 PD 0 70 9.37 1 9.37 1 41 0 1 CR 0 10 88.54 0 88.54 0 42 0 1 CR 0 55 91.43 0 91.43 0 43 0 0 CR 0 80 91.86 0 91.86 0 44 0 0 CR 0 50 63.81 1 63.81 1 45 1 0 CR 0 10 92.84 0 92.84 0 46 0 0 CR 0 20 105.9 0 105.9 0 47 1 1 CR 0 55 36.53 1 35.7 1 48 1 . PR 1 30 113.03 0 23.18 1 49 1 1 CR 0 70 10.19 1 10.19 1 50 1 0 CR 0 40 102.67 0 102.67 0 51 0 0 CR 1 30 98.99 1 98.99 1 52 0 0 CR 0 40 25.35 1 25.35 1 53 0 0 CR 1 60 99.42 0 53.52 1 54 0 1 CR 0 80 56.45 1 56.45 1 55 0 0 CR 0 80 91.96 0 91.96 0 56 1 1 CR 1 30 91.53 0 91.53 0 57 1 0 CR 0 15 85.48 0 85.48 0 58 1 0 CR 1 35 88.14 0 88.14 0 59 0 0 CR 0 20 86.53 0 86.53 0 60 1 0 CR 0 80 91.53 0 7.36 1 61 0 0 CR 0 65 16.87 1 16.87 1 62 1 1 CR 1 35 77.42 0 77.42 0 63 1 0 CR 0 40 72.92 0 72.92 0 64 0 0 CR 1 15 72.59 0 72.59 0 65 0 0 CR 1 25 67.86 0 67.86 0 66 1 0 CR 0 60 66.97 0 66.97 0 67 0 0 CR 1 70 55.76 0 55.76 0 68 0 1 CR 0 25 53.95 0 53.95 0 69 0 0 CR 1 50 59.15 0 59.15 0 70 0 1 PD 0 75 5.85 1 4.87 1 71 0 0 CR 1 70 86.47 0 86.47 0 72 1 0 CR 1 60 125.16 0 125.16 0 73 0 0 CR 1 15 66.38 0 66.38 0 74 1 0 CR 1 70 21.27 1 8.42 1 75 0 0 CR 1 60 58.62 1 53.59 1 76 1 1 CR 0 20 98.53 0 98.53 0 77 0 1 PR 0 35 93.14 1 93.14 1 78 0 1 PR 1 55 10.82 1 10.65 1 79 0 1 PD 1 80 5.06 1 5.06 1 80 1 0 CR 1 35 25.84 1 25.84 1 81 0 0 CR 0 5 124.04 0 124.04 0 82 1 0 CR 0 65 43.96 1 26.17 1 83 1 1 PD 1 85 0.69 1 0.69 1

Supplemental Table 3. International Prognostic Index (IPI) of the 83 DLBCL cases included in Figure 1, D and E.

Patients stage III-IV Serum LDH >1 extranodal ECOG>1 (n=83) Age>60 (Yes=1) (High:1) sites (yes=1) (yes=1) IPI>2 final 1 0 0 1 0 0 0 2 0 1 1 0 0 0 3 1 1 1 0 0 1 4 0 1 1 1 0 1 5 0 1 1 0 0 0 6 0 0 1 0 1 0 7 0 1 1 0 0 0 8 0 0 . 0 . 0 9 1 0 1 0 0 0 10 1 0 1 0 0 0 11 0 1 1 0 0 0 12 0 0 1 0 0 0 13 1 0 1 0 0 0 14 0 1 1 0 1 1 15 0 1 1 0 0 0 16 1 1 1 0 1 1 17 0 0 0 0 0 0 18 0 1 1 1 1 1 19 0 1 0 1 0 0 20 1 0 0 0 0 0 21 0 1 1 1 0 1 22 1 0 0 0 0 0 23 0 0 0 0 0 0 24 0 0 0 0 0 0 25 0 1 1 0 0 0 26 1 0 1 0 0 0 27 1 0 1 0 0 0 28 1 1 0 0 0 0 29 1 0 0 0 0 0 30 1 0 1 1 0 1 31 1 0 1 0 0 0 32 1 1 1 0 0 1 33 1 1 1 0 0 1 34 0 1 1 0 0 0 35 1 0 0 0 0 0 36 1 1 1 0 0 1 37 0 0 1 0 0 0 38 1 0 0 0 0 0 39 0 0 1 0 0 0 40 1 0 1 0 0 0

41 0 1 1 0 0 0 42 1 0 0 0 0 0 43 0 0 1 0 0 0 44 1 0 0 0 0 0 45 0 0 0 0 0 0 46 1 0 0 0 0 0 47 1 0 1 0 0 0 48 0 . . . . . 49 1 1 1 1 0 1 50 1 0 0 0 0 0 51 1 0 0 0 0 0 52 1 0 0 0 0 0 53 1 1 1 0 1 1 54 1 1 1 0 1 1 55 1 1 0 1 0 1 56 1 1 0 0 1 1 57 0 0 0 0 0 0 58 0 1 1 0 0 0 59 1 1 0 0 0 0 60 0 1 1 1 0 1 61 1 0 0 0 0 0 62 0 0 0 0 0 0 63 1 0 0 0 0 0 64 1 1 0 1 0 1 65 0 1 1 0 0 0 66 1 0 0 0 1 0 67 1 0 0 0 0 0 68 0 1 1 1 1 1 69 1 0 0 0 0 0 70 0 1 1 1 1 1 71 1 0 0 0 0 0 72 0 1 1 0 0 0 73 1 1 0 0 1 1 74 1 0 1 0 0 0 75 0 1 1 0 0 0 76 0 0 1 0 0 0 77 0 1 0 0 1 0 78 1 1 1 1 0 1 79 0 1 1 1 1 1 80 1 1 1 0 0 1 81 0 0 1 0 0 0 82 1 1 0 0 0 0 83 1 1 1 1 0 1

Supplemental Table 4. Immunoblot Antibodies. Western blot antibodies Source TBL1X GeneTex (GTX107736), abcam (ab24548) and Invitrogen (MA5-24874). ACTIN Santa Cruz Biotechnology (sc-47778) Non-phospho active β-CATENIN CST#2745 MYC CST#18583 Phospho MYC (Ser62) CST#13748 GAPDH CST#5174 and Invitrogen (39-8600) β-CATENIN CST#8480 microtubule-associated protein 1 light CST#12741 chain 3 (LC3A/B) BECLIN-1 CST#4122 CUL1 abcam (ab11047) Phospho MYC (thr58) Abcam (ab28842) CAND1 CST#8759 and Sigma Aldrich (SAB1405164) PLK1 CST #4513 and abcam (ab17056) SKP1 CST #12248 Histon H3 CST #4499 SURVIVIN CST #2808

Supplemental Table 5. RT-PCR primers. Gene name PCR primer MYC assay ID Hs00153408_m1 BIRC5 assay ID Hs00153353_m1 PLK1 assay ID Hs00983227_m1 Endogenous control GAPDH assay ID Hs02758991_g1 WISP1 assay ID Hs00180245_m1 AXIN2 assay ID Hs00610344_m1 CTNNB1 assay ID Hs00355045_m1 TBL1X assay ID Hs00959540_m1 CAND1 assay ID Hs00218384_m1 BECN1 Assay ID Hs01007018_m1

Supplemental Figures

S. Figure 1. Sorting strategy for normal GC B-cells isolation from pediatric tonsils. Dot plots of flow cytometry analysis showing sorting strategy and post-sort purity check of the germinal center (GC) B-cells used in Figure 1A. Mononuclear cells were isolated from two different pediatric tonsils (n=2) and then stained before sorting with anti-CD10, anti-CD19, anti-CD14, anti-CD56 and anti-CD3. CD19+CD10+ gated germinal center B-cells were sorted after excluding Monocytes (CD14+) , T-cells (CD3+) and NK-cells (CD56+). Numbers above the black boxes represent post-sort purity of the sorted GC B-cells from each donor.

S. Figure 2. TBL1 is expressed at low level in reactive germinal center. Representative images of TBL1 immunohistochemical staining of reactive germinal center from pediatric tonsil (donor 5) (at 20x, 40x and 100x) showing weak positivity for TBL1 (brown). Tonsillar epithelium (black arrows) served as an internal positive control.

S. Figure 3. TBL1 expression in de novo ABC-DLBCL. Representative images of TBL1 immunohistochemical (IHC) staining of lymph nodes involved by ABC-DLBCL (at 40x and 100x) showing diffuse, strong nuclear and cytoplasmic positivity for TBL1.

S. Figure 4. TBL1 expression in de novo GCB-DLBCL. Representative images of TBL1 immunohistochemical (IHC) staining of lymph nodes involved by GCB-DLBCL (at 40x and 100x) showing diffuse, strong nuclear and cytoplasmic positivity for TBL1.

S. Figure 5. Immunoblot validating efficient TBL1 knock-down in the luciferase positive CRISPR- Cas9 engineered Riva cells prior to engraftment into NSG mice (in vivo experiment described in Figure 2B). Empty vector was used as control (n=1). ACTIN was used as a loading control. Numbers below the immunoblots reflect normalized value of quantified protein bands relative to the controls.

S. Figure 6. Tegavivint is minimally cytotoxic against normal immune cell subsets. (A) Cytotoxicity assay in normal resting (Rest) and activated (Act) lymphocytes (B-cells, T-cells and NK- cells) from 3 healthy donors treated with the indicated concentration of tegavivint (T) in nM. Viability was determined by annexin/propidium iodide (PI) staining and flow cytometry at 24 hours. The graph shows normalized percentage of live cells (ann-/PI- cells) relative to the untreated control (NT). n=3, data represent means ± SEM. *p<0.05 by linear mixed effects models with testing the interaction between treatment and activation status. (B) Immunoblot showing abundant TBL1 expression in the indicated DLBCL cell lines compared to normal resting and activated NK and T lymphocytes. ACTIN was used as a loading control (n=1).

S. Figure 7. Tegavivint selective cytotoxicity: TBL1 was knocked-down in the indicated DLBCL cell lines using a TBL1-specific shRNA construct. Scramble was used as control. Following verification of efficient TBL1 knock-down via immunoblot (72 hours after transduction), DLBCL cells were incubated with 70nM of tegavivint (T70) for 12 and 24 hours. Viability was assessed by annexin/propidium iodide (PI) staining and flow cytometry. Graphs show normalized percentage of live cells (ann-/PI- cells) relative to the untreated scramble control (NT) from three separate experiments. ns: p>0.05, *p<0.05, **p<0.005 by paired t-test. For immunoblot, GAPDH or actin were used as loading control. Numbers below the immunoblot reflect normalized value of quantified protein bands relative to control (1).

S. Figure 8. TBL1 is abundantly expressed in purified DFBL-18689 cells: Immunoblot showing abundant TBL1 expression in whole cell lysates from DFBL-18689 cells compared to normal human peripheral blood (PB) B-cells. Mantle cell lymphoma cell line (Jeko) was used as positive control (+ Ctl). GAPDH was used as a loading control.

S. Figure 9. TBL1/β-CATENIN loss of proximity upon treatment with tegavivint: (A) Confocal images of PLA (60x) showing the negative IgG control in the indicated DLBCL cells lines. Blue indicates cell nuclei (Dapi) (n= 3). (B) Confocal images of PLA (100x) using anti-TBL1 and β- CATENIN antibodies in the indicated DLBCL cells lines after treatment with either DMSO control (NT) 24 or tegavivint (T) (70nM) for 12 hours. Red indicates PLA signal, blue indicates cell nuclei (Dapi) (n= 3). Graph represents the mean value of PLA signal from no less than 300 cells. *p<0.05, **p<0.005 by paired t-test.

S. Figure 10. Genome-wide TBL1 promoter occupancy is unffected by tegavivint treatment. (A) Representative heat map of ChIP-seq read densities for TBL1 in the indicated DLBCL cell lines after treatment with either DMSO control (NT) or tegavivint (T, 70nM) ranked by decreasing combined occupancy signals centered ± 2.5 kb relative to transcriptional start site (TSS) of transcriptionally active peak regions (n= 3). (B) qRT-PCR showing the mRNA fold changes of the indicated genes after treating the indicated DLBCL cell lines with tegavivint (T) for 12 hours relative to the untreated control (NT). n=3, data represent means ± SEM. ns: p>0.05, **p<0.005 by paired t-test.

S. Figure 11. TBL-1 mediated/β-CATENIN independent effects on known Wnt/β-CATENIN targets: β-CATENIN was knocked-down (KD) in the indicated DLBCL cell lines using a β-CATENIN specific shRNA construct. Scramble was used as control. 72 hours after transduction, cells were treated with 70nM of tegavivint (T70) for 12 hours. β-CATENIN, MYC and SURVIVIN protein expression levels and mRNA fold changes were assessed by immunoblot (left) and qRT-PCR (right), respectively. Immunoblot represent one gel. GAPDH was used as loading control. Numbers below the immunoblot images reflect normalized value of quantified protein bands relative to the controls. n=3, data represent means ± SEM. ns: p>0.05, *p<0.05, **p<0.005, ***p<0.0005 by linear mixed effects models incdluing testing the interaction between shRNA and treatment condition.

S. Figure 12. Tegavivint does not affect CUL1/TBL1 proximity. Confocal images of PLA (100x) (red) using anti-CUL1 and TBL1 antibodies in the indicated DLBCL cells lines treated with either DMSO control (NT) or tegavivint (T) (70nM) for 12 hours. Red indicates PLA signal, blue indicates cell nuclei (Dapi) (n=3). Graph represents the mean value of PLA signal from no less than 300 cells. ns: p>0.05 by paired t-test.

S. Figure 13. Loss of CUL1/CAND1 proximity upon treatment with tegavivint. Confocal images of PLA (100x) (red) using anti-CUL1 and anti-CAND1 antibodies in the indicated DLBCL cells lines treated with either DMSO control (NT) or tegavivint (T) (70nM) for 12 hours. Red indicates PLA signal, blue indicates cell nuclei (Dapi) (n=3).

S. Figure 14. TBL1 is part of a SCF complex that modulates the stability of critical oncoproteins. (A) Immunoblot showing CAND1 and CUL1 expression in the indicated DLBCL cell lines after treatment with either DMSO control (NT) or tegavivint (T) for 12 and 24 hours (n=3). (B) Immunoblot showing time course expression of PLK1 in the indicate DLBCL cell lines after treatment with either DMSO control (NT) or tegavivint (T) (n=3). (C) qRT-PCR showing mRNA fold changes of PLK1 at 12 hours after treating the indicated DLBCL cell lines with tegavivint (T) relative to the

30 untreated control (NT). n=3, data represent means ± SEM. ns: p>0.05 by paired t-test for each cell line. (D) Immunoblot showing time course expression of PLK1, MYC, and phosphorylated MYC (Thr58 and Ser62) in the indicated DLBCL cell lines after treatment with either DMSO control (NT) or tegavivint (T) (from the same gel, n=3). (E) Immunoblot showing PLK1 expression in the indicated DLBCL cell lines after treatment with either DMSO control (NT), cycloheximide (CHX, 70ug/mL), tegavivint (T) or the combination (CHX+T) for 8 and 12 hours (n=3). (F) Immunoblot showing PLK1 expression after TBL1 knock-down (KD) in the indicated DLBCL cell lines using a TBL1 specific shRNA construct (n=3). (G) Immunoblot showing MYC and PLK1 expression after CAND1 knock- down (KD) in the indicated DLBCL cell lines using a CAND1 specific shRNA. Numbers in the blue box below the immunoblot indicate percentage of viable DLBCL cells by trypan blue exclusion from one experiment (n=1). Numbers below the immunoblot images reflect normalized value of quantified protein bands relative to the controls (1). Tegavivint (T): Riva and Pfeiffer: 70nM, OCI-ly3: 50nM and WSU-NHL: 15nM. MG132: Riva, OCI-Ly3 and WSU-NHL: 0.5uM and Pfeiffer: 0.3uM added 6 hours following tegavivint treatment.

S. Figure 15. Tegavivint does not significantly affect Becn-1 transcript. qRT-PCR showing the mRNA fold changes of Becn-1 after treating the indicated DLBCL cell lines with tegavivint (T) for 12 hours relative to the untreated control (NT). n=3, data represent means ± SEM. ns: p>0.05 by paired t-test for each cell line. Tegavivint (T): Riva and Pfeiffer: 70nM, OCI-ly3: 50nM and WSU-NHL: 15nM.

Supplementary References

1. Scoville SD, Keller KA, Cheng S, et al. Rapid column-free enrichment of mononuclear cells from solid tissues. Sci Rep. 2015;5(1):1-6. 2. Visco C, Li Y, Xu-Monette ZY, et al. Comprehensive gene expression profiling and immunohistochemical studies support application of immunophenotypic algorithm for molecular subtype classification in diffuse large B-cell lymphoma: a report from the International DLBCL Rituximab-CHOP Consortium Program Study. Leukemia. 2012;26(9):2103-2113. 3. Fiore C, Bailey D, Conlon N, et al. Utility of multispectral imaging in automated quantitative scoring of immunohistochemistry. J Clin Pathol2012;65(6):496-502. 4. de Almeida Nagata DE, Chiang EY, Jhunjhunwala S, et al. Regulation of tumor-associated myeloid cell activity by CBP/EP300 bromodomain modulation of H3K27 acetylation. Cell Rep 2019;27(1):269-281. e264. 5. Li H, Durbin R. Fast and accurate short read alignment with Burrows–Wheeler transform. bioinformatics. 2009;25(14):1754-1760. 6. Ramírez F, Ryan DP, Grüning B, et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 2016;44(W1):W160-W165. 7. Pal S, Baiocchi RA, Byrd JC, Grever MR, Jacob ST, Sif S. Low levels of miR‐92b/96 induce PRMT5 translation and H3R8/H4R3 methylation in mantle cell lymphoma. EMB J. 2007;26(15):3558-3569.