Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 https://doi.org/10.1186/s13046-019-1458-7 RESEARCH Open Access Notch1 signaling in NOTCH1-mutated mantle cell lymphoma depends on Delta- Like ligand 4 and is a potential target for specific antibody therapy Elisabeth Silkenstedt1,2,3, Fabian Arenas1,4, Berta Colom-Sanmartí1, Sílvia Xargay-Torrent1, Morihiro Higashi1, Ariadna Giró1,4, Vanina Rodriguez1, Patricia Fuentes5, Walter E. Aulitzky6, Heiko van der Kuip3ˆ, Sílvia Beà4,7, Maria L. Toribio5, Elias Campo4,7,8,9, Mònica López-Guerra1,4,8† and Dolors Colomer1,4,8,9*† Abstract Background: NOTCH1 gene mutations in mantle cell lymphoma (MCL) have been described in about 5–10% of cases and are associated with significantly shorter survival rates. The present study aimed to investigate the biological impact of this mutation in MCL and its potential as a therapeutic target. Methods: Activation of Notch1 signaling upon ligand-stimulation and inhibitory effects of the monoclonal anti-Notch1 antibody OMP-52M51 in NOTCH1-mutated and -unmutated MCL cells were assessed by Western Blot and gene expression profiling. Effects of OMP-52M51 treatment on tumor cell migration and tumor angiogenesis were evaluated with chemotaxis and HUVEC tube formation assays. The expression of Delta-like ligand 4 (DLL4) in MCL lymph nodes was analyzed by immunofluorescence staining and confocal microscopy. A MCL mouse model was used to assess the activity of OMP-52M51 in vivo. Results: Notch1 expression can be effectively stimulated in NOTCH1-mutated Mino cells by DLL4, whereas in the NOTCH1-unmutated cell line JeKo-1, less effect was observed upon any ligand-stimulation. DLL4 was expressed by histiocytes in both, NOTCH1-mutated and –unmutated MCL lymph nodes. Treatment of NOTCH1-mutated MCL cells with the monoclonal anti-Notch1 antibody OMP-52M51 effectively prevented DLL4-dependent activation of Notch1 and suppressed the induction of numerous direct Notch target genes involved in lymphoid biology, lymphomagenesis and disease progression. Importantly, in lymph nodes from primary MCL cases with NOTCH1/2 mutations, we detected an upregulation of the same gene sets as observed in DLL4-stimulated Mino cells. Furthermore, DLL4 stimulation of NOTCH1-mutated Mino cells enhanced tumor cell migration and angiogenesis, which could be abolished by treatment with OMP-52M51. Importantly, the effects observed were specific for NOTCH1-mutated cells as they did not occur in the NOTCH1-wt cell line JeKo-1. Finally, we confirmed the potential activity of OMP-52M51 to inhibit DLL4-induced Notch1- Signaling in vivo in a xenograft mouse model of MCL. (Continued on next page) * Correspondence: dcolomer@clinic.cat Monica López-Guerra and Dolors Colomer share the seniorship of this manuscript 1 Experimental Therapeutics in Lymphoid Malignancies Group, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain 4 Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), Madrid, Spain Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 2 of 15 (Continued from previous page) Conclusion: DLL4 effectively stimulates Notch1 signaling in NOTCH1-mutated MCL and is expressed by the microenvironment in MCL lymph nodes. Our results indicate that specific inhibition of the Notch1-ligand-receptor interaction might provide a therapeutic alternative for a subset of MCL patients. Keywords: Mantle cell lymphoma, Notch1, Delta-like ligand, Angiogenesis, OMP-52 M51 Background that finally release NICD from its membrane receptor Although with current standard therapy high initial re- and lead to its nuclear translocation [7]. sponse rates can be achieved, early relapses and rapid Targeting Notch signaling has been studied in various disease progression determine the clinical course of cancer types and particularly using gamma-secretase in- most mantle cell lymphoma (MCL) patients [1]. In the hibitors (GSI) in hematological malignancies [6, 10, 11]. last years, new therapies were approved targeting the However, the clinical applicability of GSI is limited as it proteasome, the PI3K-Akt-mTOR pathway, the B-cell can cause severe diarrhea resulting from simultaneous receptor (BCR) signaling pathway and the anti-apoptotic inhibition of Notch1 and Notch2 signaling in gut epithe- Bcl-2 protein family. Yet, MCL remains an incurable dis- lial stem cells [12, 13]. Thus, alternative strategies for ease [2]. The t (11;14)(q13;q32) translocation leading to therapeutic targeting of Notch1 are highly warranted. Cyclin D1 overexpression is the primary oncogenic event Recently, antibodies that inhibit signaling of both, wild- in MCL pathogenesis [3]. Furthermore, constitutive acti- type and mutated Notch1 receptors have been character- vation of BCR signaling plays an important role in dis- ized [14]. OMP-52M51 (brontictuzumab) is a full length ease development [4]. Additionally, genomic profiling IgG2 humanized monoclonal antibody that selectively revealed a high number of secondary genetic alterations binds the negative regulatory region of the Notch1 re- and recurrent mutations affecting cell cycle, DNA dam- ceptor leading to inhibition of Notch1 signaling [15]. A age response and apoptosis pathways that contribute to phase I study has been conducted in subjects with solid pathogenesis and aggressiveness of MCL [3]. Among tumors showing efficacy in cases with Notch1 pathway them, NOTCH1 gene mutations have been described activation [16] . In this study, we investigated the role of with a frequency of 5–10% and were shown to be associ- the Notch ligands in activating Notch1 signaling in ated with shorter survival rates [5, 6]. Therefore, further NOTCH1-mutated and - unmutated MCL cell lines and investigation of the biological effect of this mutation in evaluated the effects of OMP-52M51 in these cell lines. MCL and its potential as a therapeutic target is of great interest. The majority of the previously described NOTCH1- Methods mutations in MCL consist of either small frameshift- Cell lines and primary MCL cells causing indels or nonsense mutations in exon 34 [6]. MCL cell lines Mino (CRL-3000), JeKo-1 (CRL-3006) These mutations lead to truncation of the C-terminal and REC-1 (CRL-3004), were obtained from American PEST-domain, thereby removing the recognition site Type Culture Collection (ATCC). To avoid Mycoplasma from the ubiquitin ligase degradation complex, resulting contamination, cell lines were routinely tested for Myco- in a more stable and transcriptionally active form of plasma infection by PCR. The identity of all cell lines Notch1-intracellular domain (NICD). The Notch signal- was verified by using GenePrint® kit (Promega, Madison, ing pathway operates in a context- and tissue-dependent WI, USA). MCL cell lines were cultured in RPMI 1640 way participating in diverse cellular processes, such as complemented with 10–20% fetal bovine serum (FBS), 2 cancer stem cell biology, angiogenesis, cell proliferation mM L-glutamine and 50 μg/mL penicillin/streptomycin and survival [7]. In addition to the well-known Notch (Life Technologies, Carlsbad, CA, USA) and grown in a target genes HES1 and HEY1, two basic-helix-loop-helix humidified atmosphere at 37 °C with 5% CO2. The mur- (Bhlh)-class of transcription factors, numerous genes ine bone-marrow derived stromal cell line OP9 (CRL- have been identified as directly regulated by activated 2749; ATCC) overexpressing DLL4 (OP9-DLL4) was Notch1 [8]. Some cancer-related target genes of Notch generated and grown as described [17, 18]. Primary cells include NF-κB family members, CYCLIN D1, p21, from MCL patients were isolated and cultured as de- GATA3, MYC and DTX1 [9]. In mammals, Notch signal- scribed [19] and conserved within the Hematopathology ing is usually activated upon interactions with Delta-like collection of our institution registered at the Biobank ligands (DLL1, DLL3, DLL4) and Jagged ligands (JAG1, from Hospital Clínic-IDIBAPS (R121004–094). The eth- JAG2), resulting in a series of proteolytic cleavage events ical approval for this project including informed patient Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 3 of 15 consent was granted following guidelines of the Hospital Cruz) and anti-human CD68 (Dako, Glostrup, Germany). Clínic Ethics Committee. Background staining was determined by incubating with ir- relevant antibodies. Unspecific fluorescence was quenched Ligand stimulation and cell treatment by incubating with avidin/biotin blocking solutions (Vector Four micrograms per milliliter of the recombinant Lab, Burlingame, CA, USA). For DLL4 detection, tissue slides Notch-receptor-ligands DLL1, DLL4, JAG1 and JAG2 were incubated with anti-rabbit IgG-HRP antibody (Dako). (R&D Systems, Minneapolis, MN, USA) were resus- Signal was amplified using the Cyanine-3 Tyramide Signal pended in phosphate buffered saline (PBS) and stored in Amplification Kit (TSA; NEL 744, Perkin Elmer, Waltham, culture plates for 4 h at 4 °C to let them attach to the MA). For CD68 detection, a biotinylated anti-mouse IgG plates. MCL cells were treated with 25 μg/mL of OMP- (Vector Lab) was added following incubation with avidin/ 52M51 or human IgG2 (both kindly provided by biotin complex (Elite Vectastain ABC Complex kit, Vector Oncomed Pharmaceuticals, Redwood City, CA, USA) and Lab). Signal was developed by adding Alexa-488-conjugated incubated at 37 °C for 2 h prior to adding them to the streptavidin. Nuclei were stained with Topro-3 (Invitrogen, ligand-coated plates for stimulation. For coculture experi- Carlsbad, CA, USA) and mounted with Fluoromount-G ments, OP9 cells were plated overnight, and then medium (Southern Biotech., Birmingham, AL, USA). Images were ac- was replaced by MCL cells (0.5 × 106 cells/mL) previously quired using a LSM510 laser scan confocal microscope treated for 2 h with OMP-52M51. After 24 h of coculture, (Zeiss, Oberkochen, Germany) coupled to an Axiovert200 MCL cells were collected by carefully rinsing the wells (Zeiss) microscope, using 63x Plan-Neofluar magnification. without disturbing the stroma monolayer and processed as required. Cell cycle assay A total of 3 × 105 MCL cells were collected after stimula- Protein analysis tion with DLL4 and incubated with OMP-52M51 for 48 Whole-cell protein extracts were obtained using Triton h. Then cells were fixed in ice-cold 70% ethanol while containing lysis buffer (Tris-HCl pH 7.6 20 mmol/L, NaCl being gently vortexed, incubated at − 20 °C for 24 h, 150 mmol/L, EDTA 1 mmol/L, 1% Triton X-100) supple- washed in PBS and resuspended in 500 μL of staining so- mented with protease and phosphatase inhibitors (10 μg/ lution containing 20 μg/mL propidium iodide (Invitro- mL leupeptin, 10 μg/mL aprotinin, 1 mmol/L phenylmetha- gen) and 100 μg/mL RNase A (Thermofisher) in PBS. nesulfonyl fluoride, 5 mmol/L NaF, 2 mmol/L Na3VO4). Cells were incubated during 30 min at 37 °C and ana- Solubilized proteins were quantified by Bradford protein lysed using BD LSRFortessa 4 L cytometer (Becton Dick- assay and 50 μg of protein was analyzed by Western Blot- inson, Franklin Lakes, NJ, USA). Cell cycle analysis was ting. The following primary antibodies were used: cleaved- performed using FlowJo software (Becton Dickinson). NOTCH1 (Val1744)(D3B8), MEK1/2, phospho-MEK1/2 (Ser217/221)(41G9) (Cell Signaling, Boston, MA, USA), Chemotaxis assay ERK1 (K-23), phospho-ERK (Thy202/204) (E-4) (Santa MCL cells were stimulated with DLL4 and incubated Cruz, Dallas, TX, USA), α-tubulin and β-actin (Sigma Al- with OMP-52M51 for 48 h. Transwell culture polycar- drich, Saint Louis, MI, USA). Chemiluminescence was de- bonate inserts (6.5 mm diameter, 8 μm of pore size) tected with ECL substrate (Pierce Biotechnology Waltham, (Corning, NY, USA) were transferred to wells containing MA, USA) on a mini-LAS4000 Fujifilm device (Fujifilm, 600 μL of RPMI supplemented with 0.5% BSA with 200 Valhalla, NY, USA). Protein expression was densitometric- ng/mL of human recombinant CXCL12 (Peprotech). One ally quantified using the Image Gauge software (Fujifilm). hundred microliter of cell suspension (0.5 x 106cells) was then added into the transwell inserts. Input cell count was Immunohistochemistry and confocal microscopy obtained from adding 100 μL of cell suspension to wells Lymph node (LN) biopsies from MCL cases were obtained containing 600 μL of 0.5% BSA in RPMI 1640. After 3 h, from the Hematopathology collection of our institution 100 μL were collected in triplicates from each lower cham- registered at the Biobank from Hospital Clínic-IDIBAPS ber and input well, viable cells were counted on an Attune (R121004–094). Formalin-fixed paraffin-embedded (FFPE) cytometer (Thermofisher, Waltham, MA, USA) under tissue slides (serial 8 μm sections) were dewaxed and tissue constant flow rate. Migration is represented as the ratio antigens were retrieved by boiling for 10–15 min in sodium between migrated cells and total viable input cells. citrate (10 mM, pH 6.0). Slides were allowed to cool down to room temperature (RT) and then washed in distilled water Human umbilical vein endothelial cells (HUVEC) tube and PBS. Quenching was carried out using 1% H2O2 in formation assay 100% methanol (40 min, RT) and permeabilization with 0,3% HUVEC, kindly provided by Dr. MC Cid (IDIBAPS), Triton-X-100 (20 min, Sigma-Aldrich) in PBS. Sections were were cultured as described [20]. Supernatants from MCL incubated overnight with anti-human DLL4 (H-70) (Santa cells (1 × 106 cells/mL) were collected after 48 h of Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 4 of 15 OMP-52M51 treatment of DLL4-stimulated cells. 24- described in the GO canonical pathways analysis and well plates were coated with 300 μL of Matrigel (Becton GSEA analysis of transcripts increased by Notch in Mino Dickinson) before 500 μL of HUVEC (0.5 × 105 cells/mL] cells (except MYC), and the significant genes were and 500 μL of the supernatants were added. After 24 h, grouped into specific pathways (Additional file 1: Table number of branch points was quantified as a measure of S1). The “NOTCH1 direct targets” gene set was also in vitro angiogenesis as the mean of five randomly constructed based on the 79 genes described by Ryan chosen fields from each well. Images were taken with a et al. [26]. Similarly, the “NOTCH1 custom” gene set DM IL LED microscope coupled to a DFC295 camera was designed by manually selecting significant Notch1 (magnification 100x) with Leica Application Suite v 3.7 related genes found in the literature [6, 11, 15, 27–30]. software (Leica, Wetzlar, Germany). Two-class analysis with 1000 permutations of gene sets and a weighted metric was used for all cases. The resulting Gene expression profiling gene sets with a false discovery rate (FDR) < 0.05 were Total RNA was extracted using the TRIzol method (Life considered to be significant, except for the “NOTCH1” technologies) according to manufacturer’s instructions. gene set, where a FDR < 0.12 was used. GSEA analysis was RNA integrity was examined with the Agilent 2100 Bioa- performed of MCL cell lines stimulated or not with DLL4 nalyzer (Agilent Technologies, Santa Clara, CA, USA) and treated or not with OMP-52M51 and of MCL lymph and only high quality RNA samples were further proc- node tissues with or without mutations in a NOTCH gene. essed. RNA from MCL cell lines was hybridized to a HeatMaps were created using the Morpheus software Gene Chip HT HG-U219 array, according to Affymetrix (https://software.broadinstitute.org/morpheus/) followed (Santa Clara, CA, USA) standard protocols. RNA from by hierarchical clustering using one minus Pearson correl- MCL lymph nodes was hybridized to a Gene Chip Human ation of the average of gene expression in order to illus- Genome U133 plus 2.0 arrays, according to Affymetrix trate the differential expression of those genes significantly standard protocols. Determination of the detection call for modulated by DLL4 stimulation and OMP-52M51 treat- each probe set of the array was obtained with GeneChip® ment in the MCL cell lines and by NOTCH gene mutation Command Console® Software (AGCC) (Affymetrix). Raw in MCL lymph node tissues for the all custom gene sets data was normalized using the Robust Multichip Analysis analysis performed. (RMA) algorithm of the BioConductor Affy Package. Differ- ential expression data analysis was carried out using the In vivo mouse model Multiexperiment Viewer Platform (TM4-MEV) [21]. The NSG (NOD-scid-gamma) mice were injected intraven- number of statistically significant up- and down-regulated ously (i.v.) with 10 × 106 Mino cells. MCL cell engraftment genes was determined using Rank Products methodology was periodically monitored over a 3 months period. After [22] for cell lines, and Volcano plot [23] for MCL lymph 3 months, mice presented infiltration in several organs. nodes, both setting up a false discovery rate (FDR) < 0.2 and Tumor cells from lymph nodes were collected, cultured in an absolute fold change (FC) > 1.75. PANTHER (http:// RPMI 1640 + 10% FBS as described above and cryopre- pantherdb.org/) was used to perform gene ontology (GO) served after several passages. We next confirmed that pathways analysis [24] to visualize the relationships between these Mino cells engraft faster in a secondary transplant the significantly modulated genes of NOTCH-mutated MCL (45–60 days). Again, these cells obtained from lymph lymph nodes and proteins in known pathways of biological nodes were collected and cryopreserved. These “fast processes [25]. Primary microarray data of MCL cell lines engrafting” tumor cells were then thawed and expanded and primary MCL lymph nodes are available at the Gene to get enough cells for in vivo studies. 225 × 106 Mino Expression Omnibus (GEO) of the National Center for Bio- cells were then stimulated ex vivo by coculturing them technology Information under accession Nos. GSE125349, with OP9-DLL4 cells (7.5 × 106 Mino cells/plate 100 × 20 GSE36000 and GSE46969. mm2 (Corning). After 24 h of incubation, 15 × 106 stimu- lated Mino cells were injected into the intraperitoneal cav- Gene set enrichment analysis (GSEA) ity (IC) of 12 female NSG mice at the age of 10 weeks. For GSEA, the desktop application version 2.0 (GSEA, Mice were treated intraperitoneally 1 day prior to injec- Broad Institute at MIT, Cambridge, MA; http://www. tion of cells and then every 4 days with 20 mg/kg of OMP- broadinstitute.org/gsea/) was applied using experimen- 52M51 or control antibody human IgG2 (6 mice/group). tally derived custom genes. The “Custom MCL” gene set After 10 days, mice were sacrificed and a peritoneal lavage was designed by manually grouping gene sets involved (PL) was done by injecting the cavity with 5 mL of cold in pathways of Notch-activated genes in MCL according PBS. Human B-cells were purified by using human CD19 to the results described by Ryan et al. [26]. Briefly, we beads. Protein extracts were obtained and expression of searched molecular signatures (http://software.broadin- cleaved Notch1 was analyzed by Western Blot. Procedures stitute.org/gsea/msigdb/) corresponding to the pathways involving animals and their care are conforming to Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 5 of 15 institutional guidelines that comply with national and Studio v1.5.20 software (Benooist-Mathis Lab, Harvard international laws and policies (EEC Council Directive 86/ Medical School, MA, USA). Comparisons between 2 groups 609, OJ L 358, 12 December, 1987) and were authorized of samples were analyzed with Kruskall-Wallis nonparamet- by the local ethical committee. ric test followed by Mann-Whitney U test. Results were considered statistically significant when P < 0.05 (*P < 0.05; **P < 0.01; ***P < 0.001). Statistical analysis Data is represented as the mean ± SD of 3 independent ex- Results periments. All statistical analyses were done by using Activation of Notch1 signaling can effectively be achieved GraphPad Prism 6.01 software (GraphPad Software, La by stimulation with DLL4 in NOTCH1-mutated MCL cells Jolla, CA, USA). Volcano plot of P values as a function of A NOTCH1-mutation affecting the PEST-domain in weighted FC for mRNA was performed by using Multiplot exon 34 (p.Q2487*) was described in the established Fig. 1 Activation of Notch1 signaling can effectively be achieved by stimulation with DLL4 in NOTCH1-mutated MCL cells. a Mino, REC-1 and JeKo-1 cells were stimulated with the different soluble ligands JAG1, JAG2, DLL1 and DLL4 (10 μg/mL). Protein expression levels of cleaved Notch1 were determined by Western Blot analysis after 48 h (n = 3, one representative experiment is shown). b Mino cells were cocultured with mesenchymal stem cells expressing DLL4 (OP9-DLL4) or GFP-transfected control cells for 48 h and protein expression of cleaved Notch1 was analyzed by Western Blot. [(n=3, one representative experiment is shown). c Double staining of MCL lymph node (LN) with anti-CD68 (marker of histiocytes, green) and anti-DLL4 (red). Nuclei were stained with Topro-3 (blue). Images were acquired using a LSM510 laser scan confocal microscope coupled to an Axiovert 200 microscope, using 63x Plan-Neofluar magnification. Images from a representative case per group are shown Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 6 of 15 MCL cell line Mino [5, 6]. The effect of this mutation is also in the coculture system of Mino cells with OP9- considered to be ligand-dependent. In contrast, the DLL4 (Fig. 2b). Again, no effect was detected in the MCL cell line REC-1 presented an intragenic deletion in NOTCH1-wt JeKo-1 cell line. exon 28 encoding for a truncated Notch1 protein with in- To evaluate the effect of OMP-52M51 on downstream creased activity in a ligand-independent fashion [26, 31]. signaling of Notch1, we performed a gene expression First, we analyzed which Notch ligand was the most ef- profile analysis in JeKo-1 and Mino cells stimulated with fective to stimulate Notch1 signaling in MCL. NOTCH1- DLL4 and treated with OMP-52M51 for 48 h. Using the mutated cell lines Mino and REC-1 and the NOTCH1-wt GSEA software analyzing a customized NOTCH1 set of cell line JeKo-1 were stimulated with the recombinant li- genes (NOTCH1 custom) [6, 11, 15, 27–30], we detected gands DLL1, DLL4, JAG1 and JAG2 and the Notch activa- 44 significantly upregulated leading edge genes (Fig. 2c, tion status in these samples was determined by Western Additional file 2: Table S2) upon DLL4-treatment in Blot analysis of cleaved Notch1. As shown in Fig. 1a, Mino cells (FDR < 0,123; p < 0,005, NES > 1.48). As dis- DLL4 and DLL1 activated the expression of cleaved played in the Z score heatmap, the significantly modu- Notch1 in Mino cells, DLL4 being the most potent ligand. lated genes upon DLL4 stimulation in Mino cells were In contrast, these ligands induced only a minor Notch1 effectively downregulated by treatment with OMP- activation in unmutated JeKo-1 cells. We confirmed that 52M51 (Fig. 2c). REC-1 cells overexpressed cleaved Notch1 independently When we applied a signature of NOTCH1 target genes of the ligands. The effect of DLL4 was then confirmed by described in MCL cells (NOTCH1 direct targets) [26], a coculturing of Mino cells with the bone-marrow derived strong upregulation of these genes was detected in Mino mesenchymal stem cell line OP9 overexpressing DLL4 cells stimulated with DLL4 (FDR < 0.001; p < 0.001, (OP9-DLL4), which represents a cell culture model better NES = 3.00). A good correlation was also observed in reflecting the situation in vivo, where Notch1 ligands are JeKo-1 stimulated with DLL4 (FDR = 0.025; p = 0.025, usually presented by microenvironmental cells [26, 31]. NES = 1.40) (Additional file 3: Table S3 and Figure S1), Increased expression of cleaved Notch1 in Mino cells confirming again that in NOTCH1-unmutated cells the upon coculture with OP9-DLL4 cells was detected by activation of Notch signaling is less potent than in mu- Western Blot (Fig. 1b). In view of the remarkable in vitro tated cells. This may be due to the fact that the signaling effect of DLL4 in NOTCH1-mutated MCL, we sought to would not be sustained enough as the wild type Notch1 characterize which cells could express this ligand and trig- protein has a rapid turnover [7]. Again, OMP-52M51 ger Notch activation in vivo. We therefore analyzed the was able to revert the expression of these direct Notch- expression of DLL4 in MCL lymph nodes (LN) using im- target genes in Mino cells (Fig. 2d). munofluorescence staining and confocal microscopy. We found that DLL4 was widely expressed in the vascular OMP-52M51 significantly impedes DLL4-induced endothelium (data not shown) and, importantly, it was upregulation of genes involved in lymphoid biology, expressed by some histiocytic (CD68+) cells in both lymphomagenesis and disease progression NOTCH1-mutated and -unmutated cases (Fig. 1c). Our gene expression results were analyzed by GSEA using a custom gene set (Custom MCL) generated with data ob- Treatment of NOTCH1-mutated MCL cells with OMP- tained from an integrative analysis of Notch-regulated 52M51 effectively prevents DLL4-dependent activation of transcripts, genomic binding of Notch transcription com- Notch1 plexes and genome conformation data in MCL cell lines We next investigated the effect of the humanized mono- [26]. We observed that DLL4 upregulated genes related to clonal Notch1 antibody OMP-52M51 in NOTCH1-mu- angiogenesis, apoptosis, migration and adhesion, cell cycle, tated (Mino) and unmutated (JeKo-1) MCL cell lines. cytokine signaling, DNA damage and repair, MTOR and Since the recognition-binding site of OMP-52M51 in the MAPK signaling, leukocyte proliferation and defense re- EGF-like domain is missing in the REC-1 cell line due a sponse (Fig. 3a). OMP-52M51 was able to modulate all deletion in this site [26], this cell line was not suitable these gene sets only in the NOTCH1-mutated Mino cell for further investigation. line (Additional file 4: Table S4). We observed that treatment of DLL4-stimulated Mino Next, we analyzed the gene expression profile of cells with OMP-52M51 for 24 and 48 h resulted in in- lymph nodes from primary MCL cases (n = 21), 3 of hibition of DLL4-mediated cleaved-Notch1 overexpres- them harboring NOTCH1/2 mutations. By using the sion. This effect could also be observed in primary cells same GSEA analysis, in lymph nodes from primary MCL from a NOTCH1-mutated MCL case [MCL#1] carrying cases with NOTCH1/2 mutations, we detected an upreg- the typical 2-bp deletion in exon 34 (c.7541_7542delCT) ulation of the same gene sets as observed in DLL4- (Fig. 2a). We confirmed the potential of OMP-52M51 to stimulated Mino cells (Fig. 3b, Additional file 5: Table inhibit DLL-stimulated induction of Notch1 signaling S5). In these cases, to maintain Notch signaling, a Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 7 of 15 Fig. 2 Treatment of NOTCH1-mutated MCL cells with OMP-52M51 effectively prevents DLL4-dependent activation of Notch1. a Mino and JeKo-1 MCL cells and NOTCH1-mutated primary MCL cells (patient sample MCL) were treated with OMP-52M51 or IgG2 control antibody and stimulated with human recombinant DLL4 (4 μg/mL). Protein expression of cleaved Notch1 was determined by Western Blot analysis after 24 and 48 h. b Mino and JeKo-1 MCL cells were treated with OMP-52M51 and cocultured with DLL4-overexpressing OP9-stromal cells or GFP-transfected control cells. After 48 h, Western Blot analysis of cleaved Notch1 protein expression was performed. c Significantly modulated genes (FDR < 0.123; p < 0.005; NES > 1.48) related to the Notch1 signaling pathway (NOTCH custom) upon stimulation of Mino and JeKo-1 cells with DLL4 after 48 h obtained with the Gene Set Enrichment Analysis (GSEA) software. d Significantly upregulated genes (FDR < 0.001; p < 0.001, NES = 3.00) upon stimulation of Mino and JeKo-1 cells with DLL4 after 48 h obtained with GSEA applying a signature of Notch target genes described in MCL cells (NOTCH direct targets). Heatmaps were hierarchical clustered by one minus Pearson correlation of the average of gene expression Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 8 of 15 Fig. 3 OMP-52M51 significantly impedes DLL4-induced upregulation of numerous genes involved in lymphoid biology, lymphomagenesis and disease progression mimicking the signature of NOTCH mutated MCL lymph nodes. Heatmaps displaying the significantly upregulated genes and grouped by pathway signaling obtained after GSEA analysis (Custom MCL set of genes) of (a) Mino cells treated with OMP-52M51 after DLL4- stimulation and (b) MCL lymph nodes with or without mutation in NOTCH genes. c Volcano plot filtering of 55 genes differentially expressed in NOTCH-mutated MCL lymph nodes. d Enrichment of biological processes found by PANTHER gene ontology analysis of the volcano-filtered genes differentially expressed in NOTCH-mutated MCL lymph nodes Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 9 of 15 crosstalk between tumor MCL cells and accessory cells, branch points as a measure of in vitro angiogenesis com- probably histiocytes, is required. In addition, the PAN- pared to those of unstimulated cells. Importantly, the THER GO analysis of the 55 genes differentially proangiogenic effect of these supernatants could be ef- expressed in MCL lymph nodes carrying NOTCH gene fectively abolished by treatment of cells with OMP- mutations compared to NOTCH-unmutated MCL lymph 52M51. Again, no differences were detected in unmu- nodes, filtered by Volcano plot (Fig. 3c, Additional file 6: tated JeKo-1 cells (Fig. 4d). Table S6), revealed 818 signaling pathways associated with leukocyte biology primarily enriched for biological OMP-52 M51 effectively inhibits DLL4 induced activation processes associated with disease progression such as of Notch1 in an in vivo model regulation of tethering/rolling, adhesion and migration Finally, we aimed to confirm the activity of OMP-52M51 of leukocytes (Fig. 3d). in an in vivo MCL model. As NOTCH1-mutated Mino Given that most of these gene sets were related to B cells are dependent on ligand activation, these cells were cell activation, regulation of leukocyte tethering or roll- stimulated ex vivo with OP9-DLL4 cells to ensure hu- ing and leukocyte adhesion and migration, we selected man Notch1 activation. After stimulation of Mino cells these signatures for further functional validation. As the for 24 h, cells were injected into the IC of NSG mice and activation of ERK is an important integration point for B treated every 4 days with 20 mg/kg of OMP-52M51. cell activation [32], we analyzed phosphorylation of After 10 days, mice were sacrificed and cells were recol- ERK1/2 and MEK by Western Blot after stimulation of lected from the IC (Fig. 5a). Total recovery of the periton- Mino and JeKo-1 cells with DLL4 and treatment with eal cells was evaluated by flow cytometry after staining OMP-52M51. As displayed in Fig. 4a, DLL4 stimulation with huCD45/CD19/antibodies. This short-term treat- increased phosphorylation of both, MEK and ERK in ment with the anti-Notch1 antibody did not affect cell via- Mino cells, indicating that aberrant Notch1 signaling bility or tumor cell counts (data not shown). Protein stimulates B cell activation in MCL. This effect could be extracts were obtained and Western Blot analysis con- reduced by treatment with OMP-52M51, a process that firmed that OMP-52M51 was able to inhibit cleaved was not observed in NOTCH1-unmutated JeKo-1 cells Notch1 in vivo (Fig. 5b), although this short-term OMP- (Fig. 4a). To corroborate whether the upregulation of 52M51 treatment was not enough to cause a significant genes related with cell cycle is effectively dependent of efficacy in tumor growth. aberrant expression of NOTCH1 genes, we analyzed the modulation of cell cycle in NOTCH1-mutated Mino cells Discussion and NOTCH1-unmutated JeKo-1 cells treated or not Activating NOTCH1 mutations are among the most fre- with DLL4 and/or OMP-52M51. We observed that quent secondary genetic alterations in MCL detected in DLL4-stimulated Mino cells showed a significant in- 5–10% of cases [5, 6, 35]. Similar to CLL, the majority of crease of cell proportion in G2-GM phase that was par- NOTCH1 alterations described in MCL are nonsense tially reverted (p = 0.057) by incubation with OMP- truncating mutations and small frame-shift indels lo- 52M51 for 48 h (Fig. 4b). cated in exon 34 which encodes the PEST domain, We next studied the effect of aberrant Notch1 signal- resulting in enhanced Notch1 signaling [7, 30]. The clin- ing on cell migration by means of chemotaxis assays. As ical significance of this mutation is highlighted by the shown in Fig. 4c, DLL4 stimulation of Notch1 signifi- fact that patients harboring a NOTCH1-mutation had cantly increased CXCL12-induced migratory capacity of significantly shorter overall survival rates [5, 6]. Yet, the Mino cells that could be abrogated by treatment with molecular impact of NOTCH1 mutations in MCL is not OMP-52M51. In contrast, no effect could be observed in well understood. Recently, it has been postulated that unmutated JeKo-1 cells (Fig. 4c). Notch signaling regulates genes involved in BCR and The DLL4-Notch1 signaling pathway plays an import- cytokine signaling as well as the oncogene MYC, which ant role in regulating blood vessel formation during sustains proliferation of Notch-dependent MCL cell lines physiological and pathological angiogenesis [33, 34]. Fur- [26]. Targeting Notch signaling has been studied in thermore, data from our gene expression arrays revealed various cancer types [6, 10, 36], and particularly in that DLL4 stimulation induces several genes related to hematological malignancies carrying NOTCH1 muta- angiogenesis. This prompted us to investigate the impact tions [10, 37]. It has been reported that GSI-mediated of activating NOTCH1 mutations on tumor angiogen- inhibition of Notch signaling in NOTCH1-mutated MCL esis. We therefore used supernatants of Mino and JeKo- cells resulted in reduction of cell proliferation and apop- 1 cells stimulated or not with DLL4 and treated or not tosis induction [6]. However, the clinical applicability of with OMP-52M51 in a tube formation assay with GSI is limited due to severe gastrointestinal toxicities HUVEC. Supernatants of DLL4 stimulated NOTCH1- [12, 13]. Thus, alternative strategies for targeting Notch mutated Mino cells significantly increased number of signaling are highly warranted. In this way, we analyzed Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 10 of 15 Fig. 4 (See legend on next page.) Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 11 of 15 (See figure on previous page.) Fig. 4 Activating NOTCH1 mutations promote BCR signaling, proliferation, angiogenesis and enhance tumor cell migration upon DLL4-stimulation that can be abolished by treatment with OMP-52M51. a Effect of DLL4-stimulation and treatment with OMP-52M51 on ERK and MEK phosphorylation. Mino and JeKo-1 MCL cells were stimulated with DLL4 (4 μg/mL) and treated with OMP-52M51 or IgG2. After 48 h, protein expression of p-ERK, total ERK1, p-MEK1/2 and total MEK1/2 was assessed by Western Blot [n = 3; one representative experiment is shown]. b Relative cell proportion in G2-GM phase of cell cycle in Mino and JeKo-1 cells stimulated or not with DLL4 and treated with OMP-52M51 or IgG2 for 48 h. Cell cycle phases were measured by flow cytometry using propidium iodide and analysed using FlowJo software (n = 4, *p = 0.0286, bars represent the mean ± SD). c Mino and JeKo-1 cells were stimulated with DLL4 and treated with OMP-52M51 or IgG2 for 48 h. Migration of cells was assessed in a transwell system with inserts of 8 μm pore size. Migration was defined by counting the cells that migrated to the lower chambers containing medium with the chemoattractant CXCL12 (200 ng/mL) (n = 5, **p = 0.0079, bars represent the mean ± SD). d Cells were stimulated with DLL4 and treated with OMP-52M51 or IgG2 for 48 h. Supernatants were then harvested and added to HUVEC. After 24 h, the number of branch points was quantified as the mean of five randomly chosen fields from each well. Pictures were taken with a DM IL LED microscope coupled to a DFC295 camera (magnification 100x) (n = 5, bars represent the mean ± SD, *p = 0.05; ***p = 0.001). Microscope images from one representative experiment are shown the effect of a humanized IgG2 monoclonal antibody heterogeneous, suggesting that leukemia cells can adopt OMP-52M51 that showed encouraging antitumor effi- several strategies to evade Notch inhibition according to cacy in T-acute lymphoblastic leukemia (ALL) xenograft the therapeutic drug used [38]. models [15]. Recently, phenotypic and molecular fea- In Non-Hodgkin lymphomas, only little is known tures of resistance after long-term treatment with OMP- about the dependency of Notch1 signaling activation 52M51 have been reported, that are highly upon a certain ligand. Candidate cellular sources include Fig. 5 OMP-52M51 effectively inhibits DLL4 induced activation of Notch1 in an in vivo model. a Mino cells were stimulated ex vivo by coculturing them with DLL4 expressing stromal cells. After stimulation for 24 h, cells were injected into the intraperitoneal cavity of NSG mice. Mice were treated IP 1 day prior to injection of cells and then every 4 days with 20 mg/kg of OMP-52M51. After 10 days, mice were sacrificed and cells were recollected from the intraperitoneal cavity. b Expression of cleaved Notch1 was analyzed by Western Blot Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 12 of 15 other hematopoietic cells; endothelial cells, which are [5] associated with NOTCH mutations in MCL. Our next capable of inducing Notch signaling in B lymphoma cells aim was to analyze the potential effect of the anti-Notch1 [39] and DLL1/4-expressing fibroblastic cells in second- antibody OMP-52M51 on blocking the Delta-like ligand- ary lymphoid organs [40]. In this study, we show for the induced signal in MCL. Treatment of Mino cells with first time that DLL4 is a potent ligand to activate Notch1 OMP-52M51 could effectively inhibit DLL4-dependent signaling in MCL cells harboring a NOTCH1 mutation Notch1-activation and suppress transcriptional expression in the PEST domain, whereas JAG1 and JAG2 could not of several direct Notch target genes described in MCL [26]. sufficiently stimulate Notch1 activity. DLL4 was shown to These findings are in line with results observed in T-ALL, be constitutively expressed in some lymphoid organs, where OMP-52M51 treatment of T-ALL patient samples where it influences regulation of Notch signaling during harboring mutations in the NOTCH1-PEST-domain hematopoiesis [41]. In T-ALL cells, where mutations in caused strong inhibitory effects on the expression of NOTCH1 are frequent and well characterized, DLL4 plays Notch-target genes [15]. Furthermore, we confirmed that an important role as part of the tumor microenvironment in our model anti-Notch1 therapy attenuated the expres- contributing to early steps of T-ALL cell growth [42]. Al- sion levels of the four well known Notch-target genes that though immobilized DLL1 was also used to model ligand- have been described as a marker of the effect of OMP- dependent Notch1 activation in MCL [26], the effect of 52M51 in T-ALL (CR2, DTX1, HES1 and HES4) [15]. DLL1 in MCL cells is lower than of DLL4. Only a minor Moreover, we elucidated a functional relationship between effect upon any ligand stimulation was observed in Notch1 signaling and microenvironment processes related NOTCH1-wt cells corroborating the fact that NOTCH1 to MCL aggressiveness such as cell proliferation, cell mi- mutations in the PEST domain lead to a more stabilized gration and angiogenesis. Notch1 signaling has been shown protein due to loss of the recognition site from the ubiqui- to play a role in CCL19-driven homing of CLL cells [49] tin ligase degradation complex [7]. Furthermore, we and Notch1 signaling inhibition in multiple myeloma was provide evidence that DLL4 is expressed in the lymph described to prevent tumor cell migration [50]. Accord- node MCL compartment, where it could promote Notch ingly, we showed that OMP-52M51 reverts the strong in- activation. In particular, we detected DLL4 to be highly duction of gene signatures related to tumor cell migration expressed in the vascular endothelium and by some and adhesion upon Notch1 activation and prevents DLL4- CD68-positive cells, indicating that DLL4 expressed by stimulated migratory capacity of MCL cells. the LN microenvironment might provide a specific niche The DLL4-Notch1 axis is known to play an important for Notch activation. role in regulating angiogenesis. Previous studies have Furthermore, our results confirm that stimulation of shown that productive tumor angiogenesis requires Notch signaling by Delta-like ligands could have a critical cooperation between VEGF-A, which induces prolifera- role in MCL pathogenesis as we observed upregulation of tion of endothelial ‘tip’ cells and expression of DLL4 in several direct Notch target genes [26] involved in angio- ‘stalk’ cells [51]. In this context, DLL4 inhibits endothe- genesis, apoptosis, migration and adhesion, cell cycle, lial proliferation and promotes branching morphogen- cytokine signaling, DNA damage and repair, MTOR and esis, and the balance between proliferation and MAPK signaling, leukocyte proliferation and defense re- branching is key to the formation of a functional capil- sponse of B cell activation, cell cycle progression and onco- lary network. As such, treatment with anti-DLL4 anti- genesis both in DLL4-stimulated Mino cells and in lymph bodies resulted in disorganized angiogenesis, nodes from primary MCL cases carrying NOTCH muta- characterized by endothelial proliferation without for- tions. In this way, we detected the induction of transcrip- mation of functional capillaries [52]. In line with this, tion factors genes implicated in B-cell differentiation and we showed that supernatants of NOTCH1-mutated activation, PAX5 and IRF8 [43–46] as well as of MYBL2, a MCL cells stimulated with DLL4 increased HUVEC transcription factor participating in cell cycle progression tube formation, whereas OMP-52M51 blocked this [47, 48] and recently described as a direct Notch1 target in proangiogenic effect. We therefore postulate that by pro- B cell lymphomas [26]. We confirmed a possible inter- moting pronounced vasculature development, DLL4- action between Notch and BCR signaling in B cell malig- stimulated Notch1 signaling might contribute to the clin- nancies [26] as we observed the upregulation of several ically observed aggressive behavior of NOTCH1-mutated genes related to B cell activation (FYN, FGR, NEDD9 and MCL. Interestingly, NRARP, one of the genes with the SH2B2) and an increase of cell proliferation after DLL4 strongest induction upon Notch1 stimulation in our gene stimulation in NOTCH1-mutated MCL cells. Moreover, expression array, has been described to be related to the results obtained for NOTCH -mutated MCL lymph Notch signaling [53] and shown to act as a molecular link nodes could be considered a proof-of-concept for those between Notch and Wnt signaling in endothelial cells to observed in DLL4-stimulated Mino cells, and are in con- control stability of new vessel connections [54]. In this cordance with the poor overall survival [6] and prognosis sense, our results suggest that in MCL, the link between Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 13 of 15 Notch1 and NRARP might promote angiogenesis and JAG: Jagged ligands; LN: Lymph nodes; MCL: Mantle cell lymphoma; needs to be further explored. NICD: NOTCH1 Intracellular domain; NSG: NOD Scid Gamma; PBS: Phosphate buffered saline; RMA: Robust Multichip Analysis; RT: Room temperature Importantly, we showed that even if DLL4 could po- tentially activate Notch1 signaling irrespective of the mutational status of NOTCH1, its functional effects are Acknowledgements The authors thank Jocabed Roldan and the Genomics Unit of IDIBAPS for specific for NOTCH1-mutated MCL cells. This might be their technical support. due to the fact that in NOTCH1-unmutated MCL, the weak expression of intracellular cleaved Notch1 upon Authors’ contributions ligand-activation seems to be very unstable due to rapid Conception and design: ES, MLG, DC. Development of methodology and proteasomal degradation and might not be potent performance of research: ES, FA, BCS, SXT, MH, AG, VR, PF. Acquisition and enough to cause functionally relevant transcriptional ef- management of samples: ES, FA, BCS, SXT, MH, SB, AG, VR, PF. Analysis and interpretation of data: ES, FA, PV, MLT, MLG, DC; Writing of manuscript: ES, fects. This observation is of clinical relevance as a spe- MLG, DC. Review of manuscript: SXT, WRA, SB, MLT, EC, FA. Study cific Notch1-antibody therapy might be a promising supervision: MLG, DC. All authors read and approved the final manuscript. therapeutic alternative for the subgroup of patients with NOTCH1-PEST-mutations. Funding The study was supported by research funding from the Spanish Ministry of Conclusion Economy and Competitiveness through the Plan Estatal de Investigación Científica y Técnica y de Innovación [MINECO], SAF15–67633-R and RTI2018– We show for the first time that DLL4 is a potent stimu- 094584-B-I00 [to DC], SAF2016–75442-R [to MLT] and was cofunded by the lator of Notch1 signaling in NOTCH1-mutated MCL European Regional Development Fund [ERDF] and the CERCA program from and that expression of this ligand observed in histiocytic Generalitat de Catalunya, Centro de Investigación Biomédica en Cáncer [CIBERONC] [CB16/12/00334 and CB16/12/00225] and Generalitat de cells from MCL lymph nodes might provide a specific Catalunya Suport Grups de Recerca [2017 SGR 1009] and the Robert-Bosch- niche for Notch activation. We propose a link between Foundation. Notch1-induced expression of tumor-promoting genes and activation of processes contributing to a more Availability of data and materials aggressive MCL phenotype. Furthermore, our findings The datasets used and/or analyzed during the current study are available indicate that specific inhibition of the Notch1-ligand- from the corresponding author on reasonable request. Primary microarray receptor interaction provides an efficient and specific data are available at the Gene Expression Omnibus (GEO) of the National Center for Biotechnology Information Nos. GSE125349, GSE36000 and therapeutic alternative for a subset of MCL patients. GSE46969. Supplementary information Supplementary information accompanies this paper at https://doi.org/10. Ethics approval and consent to participate 1186/s13046-019-1458-7. Primary cells from MCL patients conserved within the Hematopathology collection of our institution registered at the Biobank from Hospital Clínic- IDIBAPS (R121004–094). The ethical approval for this project including Additional file 1: Table S1. Gene sets included in the “Custom MCL” informed patient consent was granted following guidelines of the Hospital set of genes. Clínic Ethics Committee. Additional file 2: Table S2. Significantly regulated NOTCH1 target Procedures involving animals and their care are conforming to institutional genes upon DLL4 stimulation using a customized NOTCH1 set of genes guidelines that comply with national and international laws and policies (EEC (NOTCH1 custom) [6, 11, 15, 27–30] Council Directive 86/609, OJ L 358, 12 December, 1987) and were authorized by the local ethical committee. Additional file 3: Table S3. Significantly regulated direct NOTCH1 target genes upon DLL4 stimulation using a customized set of genes according to Ryan et al. (NOTCH1 direct targets) [26]. Figure S1. GSEA Enrichment plots upon DLL4 stimulation in Mino and JeKo-1 cells using a customized set of Consent for publication Not applicable. genes according to Ryan et al. (NOTCH1 direct targets) [26] Additional file 4: Table S4. Modulated gene sets upon treatment with OMP-52M51 of DLL4-stimulated Mino cells using a customized set of Competing interests genes (Custom MCL) The authors declare that they have no competing interests. Additional file 5: Table S5. Modulated gene sets comparing NOTCH- mutated and wild type lymph nodes from MCL patients using a Author details 1 customized set of genes (Custom MCL) Experimental Therapeutics in Lymphoid Malignancies Group, Institut Additional file 6: Table S6. Significantly modulated genes (FC=1.75) in d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain. 2 lymph nodes from NOTCH-mutated MCL patients Department of Internal Medicine III, University Hospital, Ludwig Maximilian University, Munich, Germany. 3Dr. Margarete Fischer-Bosch-Institute of Clinical Pharmacology and University of Tuebingen, Stuttgart, Germany. 4 Abbreviations Centro de Investigación Biomédica en Red de Cáncer (CIBERONC), Madrid, ALL: Acute lymphoblastic leukemia; ATCC: American Type Culture Collection; Spain. 5Centro de Biología Molecular Severo Ochoa, Consejo Superior de BCR: B-cell receptor; Bhlh: Basic-helix-loop-helix; Ct: Cycle threshold; Investigaciones Científicas, Universidad Autónoma de Madrid, Madrid, Spain. 6 DLL: Delta like ligands; FBS: Fetal bovine serum; FC: Fold change; FDR: False Department of Hematology and Oncology, Robert-Bosch-Hospital, Stuttgart, discovery rate; FFPE: Formalin-fixed paraffin-embedded; GEO: Gene Germany. 7Lymphoid Neoplasm Program, IDIBAPS, Barcelona, Spain. 8 Expression Omnibus; GSEA: Gene Set Enrichment analysis; GSI: Gamma- Hematopathology Section, Hospital Clínic, Barcelona, Spain. 9University of secretase inhibitors; IC: Intraperitoneal cavity; IgG2: Immunoglobulin G2; Barcelona, Barcelona, Spain. Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 14 of 15 Received: 22 May 2019 Accepted: 17 October 2019 24. Thomas PD, Campbell MJ, Kejariwal A, et al. PANTHER: a library of protein families and subfamilies indexed by function. Genome Res. 2003;13(9):2129–41. 25. Thomas PD, Kejariwal A, Guo N, et al. Applications for protein sequence- function evolution data: mRNA/protein expression analysis and coding SNP References scoring tools. Nucleic Acids Res. 2006;34(Web Server issue):W645–50. 1. Dreyling M, Geisler C, Hermine O, et al. Newly diagnosed and relapsed 26. Ryan RJH, Petrovic J, Rausch DM, et al. A B cell Regulome links notch to mantle cell lymphoma: ESMO Clinical Practice Guidelines for diagnosis, downstream oncogenic pathways in small B cell lymphomas. Cell Rep. treatment and follow-up. Ann Oncol. 2014;25(Suppl 3):iii83–92. 2017;21(3):784–97. 2. Fisher RI, Bernstein SH, Kahl BS, et al. Multicenter phase II study of 27. Shih Ie M, Wang TL. Notch signaling, gamma-secretase inhibitors, and bortezomib in patients with relapsed or refractory mantle cell lymphoma. J cancer therapy. Cancer Res. 2007;67(5):1879–82. Clin Oncol. 2006;24(30):4867–74. 28. Palomero T, Lim WK, Odom DT, et al. NOTCH1 directly regulates c-MYC and 3. Jares P, Colomer D, Campo E. Molecular pathogenesis of mantle cell activates a feed-forward-loop transcriptional network promoting leukemic lymphoma. J Clin Invest. 2012;122(10):3416–23. cell growth. Proc Natl Acad Sci U S A. 2006;103(48):18261–6. 4. Puente XS, Jares P, Campo E. Chronic lymphocytic leukemia and mantle cell 29. Real PJ, Tosello V, Palomero T, et al. Gamma-secretase inhibitors reverse lymphoma: crossroads of genetic and microenvironment interactions. glucocorticoid resistance in T cell acute lymphoblastic leukemia. Nat Med. Blood. 2018;131(21):2283–96. 2009;15(1):50–8. 5. Bea S, Valdes-Mas R, Navarro A, et al. Landscape of somatic mutations and 30. Puente XS, Pinyol M, Quesada V, et al. Whole-genome sequencing identifies clonal evolution in mantle cell lymphoma. Proc Natl Acad Sci U S A. 2013; recurrent mutations in chronic lymphocytic leukaemia. Nature. 2011; 110(45):18250–5. 475(7354):101–5. 6. Kridel R, Meissner B, Rogic S, et al. Whole transcriptome sequencing 31. Kluk MJ, Ashworth T, Wang H, et al. Gauging NOTCH1 activation in Cancer reveals recurrent NOTCH1 mutations in mantle cell lymphoma. Blood. using immunohistochemistry. PLoS One. 2013;8(6):e67306. 2012;119(9):1963–71. 32. Jacob A, Cooney D, Pradhan M, et al. Convergence of signaling pathways 7. Arruga F, Vaisitti T, Deaglio S. The NOTCH pathway and its mutations in on the activation of ERK in B cells. J Biol Chem. 2002;277(26):23420–6. mature B cell malignancies. Front Oncol. 2018;8:550. 33. Pitulescu ME, Schmidt I, Giaimo BD, et al. Dll4 and notch signalling couples 8. Meier-Stiegen F, Schwanbeck R, Bernoth K, et al. Activated Notch1 sprouting angiogenesis and artery formation. Nat Cell Biol. 2017;19(8):915–27. target genes during embryonic cell differentiation depend on the 34. Zhao D, Xue C, Lin S, et al. Notch signaling pathway regulates angiogenesis via cellular context and include lineage determinants and inhibitors. PLoS endothelial cell in 3D co-culture model. J Cell Physiol. 2017;232(6):1548–58. One. 2010;5(7):e11481. 35. Zhang J, Jima D, Moffitt AB, et al. The genomic landscape of mantle cell 9. Gu Y, Masiero M, Banham AH. Notch signaling: its roles and therapeutic lymphoma is related to the epigenetically determined chromatin state of potential in hematological malignancies. Oncotarget. 2016;7(20):29804–23. normal B cells. Blood. 2014;123(19):2988–96. 10. Lopez-Guerra M, Xargay-Torrent S, Rosich L, et al. The gamma-secretase 36. Wu CX, Xu A, Zhang CC, et al. Notch inhibitor PF-03084014 inhibits inhibitor PF-03084014 combined with fludarabine antagonizes migration, hepatocellular carcinoma growth and metastasis via suppression of Cancer invasion and angiogenesis in NOTCH1-mutated CLL cells. Leukemia. 2015; Stemness due to reduced activation of Notch1-Stat3. Mol Cancer Ther. 29(1):96–106. 2017;16(8):1531–43. 11. Samon JB, Castillo-Martin M, Hadler M, et al. Preclinical analysis of the 37. Chiang MY, Radojcic V, Maillard I. Oncogenic notch signaling in T-cell and gamma-secretase inhibitor PF-03084014 in combination with B-cell lymphoproliferative disorders. Curr Opin Hematol. 2016;23(4):362–70. glucocorticoids in T-cell acute lymphoblastic leukemia. Mol Cancer Ther. 38. Agnusdei V, Minuzzo S, Pinazza M, et al. Dissecting molecular mechanisms 2012;11(7):1565–75. of resistance to Notch1-targeted therapy in T-cell acute lymphoblastic 12. Krop I, Demuth T, Guthrie T, et al. Phase I pharmacologic and leukemia xenografts. Haematologica. 2019. https://doi.org/10.3324/ pharmacodynamic study of the gamma secretase (notch) inhibitor MK-0752 in haematol.2019.217687 [Epub ahead of print]. adult patients with advanced solid tumors. J Clin Oncol. 2012;30(19):2307–13. 39. Cao Z, Ding BS, Guo P, et al. Angiocrine factors deployed by tumor vascular 13. Messersmith WA, Shapiro GI, Cleary JM, et al. A phase I, dose-finding study niche induce B cell lymphoma invasiveness and chemoresistance. Cancer in patients with advanced solid malignancies of the oral gamma-secretase Cell. 2014;25(3):350–65. inhibitor PF-03084014. Clin Cancer Res. 2015;21(1):60–7. 40. Fasnacht N, Huang HY, Koch U, et al. Specific fibroblastic niches in 14. Wu Y, Cain-Hom C, Choy L, et al. Therapeutic antibody targeting of secondary lymphoid organs orchestrate distinct notch-regulated immune individual notch receptors. Nature. 2010;464(7291):1052–7. responses. J Exp Med. 2014;211(11):2265–79. 15. Agnusdei V, Minuzzo S, Frasson C, et al. Therapeutic antibody targeting 41. Yan XQ, Sarmiento U, Sun Y, et al. A novel notch ligand, Dll4, induces T-cell of Notch1 in T-acute lymphoblastic leukemia xenografts. Leukemia. leukemia/lymphoma when overexpressed in mice by retroviral-mediated 2014;28(2):278–88. gene transfer. Blood. 2001;98(13):3793–9. 16. Ferrarotto R, Eckhardt G, Patnaik A, et al. A phase I dose-escalation and 42. Minuzzo S, Agnusdei V, Pusceddu I, et al. DLL4 regulates NOTCH signaling dose-expansion study of brontictuzumab in subjects with selected solid and growth of T acute lymphoblastic leukemia cells in NOD/SCID mice. tumors. Ann Oncol. 2018;29(7):1561–8. Carcinogenesis. 2015;36(1):115–21. 17. Ayllon V, Bueno C, Ramos-Mejia V, et al. The notch ligand DLL4 specifically 43. Usui T, Wakatsuki Y, Matsunaga Y, et al. Overexpression of B cell-specific marks human hematoendothelial progenitors and regulates their activator protein (BSAP/Pax-5) in a late B cell is sufficient to suppress hematopoietic fate. Leukemia. 2015;29(8):1741–53. differentiation to an Ig high producer cell with plasma cell phenotype. J 18. Garcia-Peydro M, de Yebenes VG, Toribio ML. Notch1 and IL-7 receptor Immunol. 1997;158(7):3197–204. interplay maintains proliferation of human thymic progenitors while 44. Carotta S, Willis SN, Hasbold J, et al. The transcription factors IRF8 and suppressing non-T cell fates. J Immunol. 2006;177(6):3711–20. PU.1 negatively regulate plasma cell differentiation. J Exp Med. 2014; 19. Horing E, Montraveta A, Heine S, et al. Dual targeting of MCL1 and NOXA as 211(11):2169–81. effective strategy for treatment of mantle cell lymphoma. Br J Haematol. 45. Xu Y, Jiang L, Fang J, et al. Loss of IRF8 inhibits the growth of diffuse large 2017;177(4):557–61. B-cell lymphoma. J Cancer. 2015;6(10):953–61. 20. Rosich L, Montraveta A, Xargay-Torrent S, et al. Dual PI3K/mTOR inhibition is 46. Xu H, Chaudhri VK, Wu Z, et al. Regulation of bifurcating B cell trajectories required to effectively impair microenvironment survival signals in mantle by mutual antagonism between transcription factors IRF4 and IRF8. Nat cell lymphoma. Oncotarget. 2014;5(16):6788–800. Immunol. 2015;16(12):1274–81. 21. Saeed AI, Sharov V, White J, et al. TM4: a free, open-source system for 47. Gonda TJ, Ramsay RG. Adenoid cystic carcinoma can be driven by MYB or microarray data management and analysis. Biotechniques. 2003;34(2):374–8. MYBL1 rearrangements: new insights into MYB and tumor biology. Cancer 22. Breitling R, Armengaud P, Amtmann A, et al. Rank products: a simple, yet Discov. 2016;6(2):125–7. powerful, new method to detect differentially regulated genes in replicated 48. Martinez I, Dimaio D. B-Myb, cancer, senescence, and microRNAs. Cancer microarray experiments. FEBS Lett. 2004;573(1–3):83–92. Res. 2011;71(16):5370–3. 23. Li W. Volcano plots in analyzing differential expressions with mRNA 49. Arruga F, Gizdic B, Bologna C, et al. Mutations in NOTCH1 PEST domain microarrays. J Bioinform Comput Biol. 2012;10(6):1231003. orchestrate CCL19-driven homing of chronic lymphocytic leukemia cells Silkenstedt et al. Journal of Experimental & Clinical Cancer Research (2019) 38:446 Page 15 of 15 by modulating the tumor suppressor gene DUSP22. Leukemia. 2017; 31(9):1882–93. 50. Mirandola L, Apicella L, Colombo M, et al. Anti-notch treatment prevents multiple myeloma cells localization to the bone marrow via the chemokine system CXCR4/SDF-1. Leukemia. 2013;27(7):1558–66. 51. Hellstrom M, Phng LK, Hofmann JJ, et al. Dll4 signalling through Notch1 regulates formation of tip cells during angiogenesis. Nature. 2007;445(7129):776–80. 52. Gu JW, Rizzo P, Pannuti A, et al. Notch signals in the endothelium and cancer "stem-like" cells: opportunities for cancer therapy. Vasc Cell. 2012;4:7. 53. Krebs LT, Deftos ML, Bevan MJ, et al. The Nrarp gene encodes an ankyrin- repeat protein that is transcriptionally regulated by the notch signaling pathway. Dev Biol. 2001;238(1):110–9. 54. Phng LK, Potente M, Leslie JD, et al. Nrarp coordinates endothelial notch and Wnt signaling to control vessel density in angiogenesis. Dev Cell. 2009; 16(1):70–82. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.