Article https://doi.org/10.1038/s41467-024-54361-4 Exploitation of the fibrinolytic system by B-cell acute lymphoblastic leukemia and its therapeutic targeting Valentina R. Minciacchi 1, Jimena Bravo2, Christina Karantanou3, Raquel S. Pereira 4, Costanza Zanetti5, Rahul Kumar2, Nathalie Thomasberger6, Pablo Llavona7, Theresa Krack2, Katrin Bankov8, Melanie Meister9, Sylvia Hartmann 10, Véronique Maguer-Satta 11, Sylvain Lefort 11, Mateusz Putyrski12, Andreas Ernst13, Brian J. P. Huntly 14, Eshwar Meduri14, Wolfram Ruf 1,15 & Daniela S. Krause 2,16,17,18 Fibrinolysis influences themobilization of hematopoietic stem cells from their bone marrow microenvironment (BMM). Here we show that activation of plasmin, a key fibrinolytic agent, by annexin A2 (ANXA2) distinctly impacts progression of BCR-ABL1+ B-cell acute lymphoblastic leukemia (B-ALL) via modulation of the extracellular matrix (ECM) in the BMM. The dense ECM in a BMMwith decreased plasmin activity entraps insulin-like growth factor (IGF) 1 and reduces mTORC2-dependent signaling and proliferation of B-ALL cells. Conversely, B-ALL conditions the BMM to induce hepatic generation of plas- minogen, the plasmin precursor. Treatment with ε-aminocaproic acid (EACA), which inhibits plasmin activation, reduces tumor burden and prolongs survi- val, including in xenogeneic models via increased fibronectin in the BMM. Human data confirm that IGF1 and fibronectin staining in trephine biopsies are correlated. Our studies suggest that fibrinolysis-mediated ECM remodeling and subsequent growth factor release influence B-ALL progression and inhi- bition of this process by EACA may be beneficial as adjunct therapy. The type of leukemia and even the type of oncogene1 determine the reciprocal interactions of leukemia cells with cellular and acellular components of the bone marrow (BM) microenviron- ment (BMM) in a highly specific way. Such interactions lead to establishment of a distinct, protective and supportive environ- ment for leukemic stem cells (LSC) and are major contributors to inefficient disease eradication by many treatments2, raising the need for novel therapies. Received: 5 October 2023 Accepted: 6 November 2024 Check for updates 1Center for Thrombosis and Hemostasis (CTH), Johannes Gutenberg University Medical Center, 55131 Mainz, Germany. 2Institute of Transfusion Medicine – TransfusionCenter, JohannesGutenberg UniversityMedical Center, 55131Mainz, Germany. 3Department of Vascular Dysfunction -Medical FacultyMannheim of Heidelberg University, Mannheim, Germany. 4Institute for Experimental Pediatric Hematology and Oncology, Goethe-University Frankfurt, Frankfurt am Main, Germany. 5Division of mRNA Cancer Immunotherapy, Helmholtz Institute for Translational Oncology Mainz, Mainz, Germany. 6BioNTech SE, Mainz, Germany. 7The Institute of Molecular Biology, Mainz, Germany. 8Department of Pediatrics (Hematology/Oncology), Charité-Universitätsmedizin, Berlin, Germany. 9AbbVie, Wiesbaden, Germany. 10Department of Pathology, Goethe University, Frankfurt am Main, Germany. 11CRCL, Inserm U1052-CNRS UMR5286, Centre Léon Bérard, Lyon, France. 12Fraunhofer Institute for Molecular Biology and Applied Ecology IME, Project Group Translational Medicine & Pharmacology TMP, Frankfurt am Main, Germany. 13Pharmazentrum/ZAFES Frankfurt, Faculty of Medicine, Goethe-University Frankfurt, Frankfurt am Main, Germany. 14Department of Haematology, University of Cambridge, Cambridge, UK. 15Department of Immunology and Microbiology, Scripps Research, La Jolla, CA, USA. 16German Cancer Research Center (DKFZ), Heidelberg, Germany. 17German Cancer Consortium (DKTK), Heidelberg, Germany. 18Research Center for Immunotherapy (FZI), University Medical Center, University of Mainz, Mainz, Germany. e-mail: krauseda@uni-mainz.de Nature Communications | (2024) 15:10059 1 12 34 56 78 9 0 () :,; 12 34 56 78 9 0 () :,; http://orcid.org/0000-0002-9296-9403 http://orcid.org/0000-0002-9296-9403 http://orcid.org/0000-0002-9296-9403 http://orcid.org/0000-0002-9296-9403 http://orcid.org/0000-0002-9296-9403 http://orcid.org/0000-0002-9863-6538 http://orcid.org/0000-0002-9863-6538 http://orcid.org/0000-0002-9863-6538 http://orcid.org/0000-0002-9863-6538 http://orcid.org/0000-0002-9863-6538 http://orcid.org/0000-0003-3424-1091 http://orcid.org/0000-0003-3424-1091 http://orcid.org/0000-0003-3424-1091 http://orcid.org/0000-0003-3424-1091 http://orcid.org/0000-0003-3424-1091 http://orcid.org/0000-0002-1556-068X http://orcid.org/0000-0002-1556-068X http://orcid.org/0000-0002-1556-068X http://orcid.org/0000-0002-1556-068X http://orcid.org/0000-0002-1556-068X http://orcid.org/0000-0001-7320-4256 http://orcid.org/0000-0001-7320-4256 http://orcid.org/0000-0001-7320-4256 http://orcid.org/0000-0001-7320-4256 http://orcid.org/0000-0001-7320-4256 http://orcid.org/0000-0003-0312-161X http://orcid.org/0000-0003-0312-161X http://orcid.org/0000-0003-0312-161X http://orcid.org/0000-0003-0312-161X http://orcid.org/0000-0003-0312-161X http://orcid.org/0000-0002-6064-2166 http://orcid.org/0000-0002-6064-2166 http://orcid.org/0000-0002-6064-2166 http://orcid.org/0000-0002-6064-2166 http://orcid.org/0000-0002-6064-2166 http://orcid.org/0000-0003-3603-1119 http://orcid.org/0000-0003-3603-1119 http://orcid.org/0000-0003-3603-1119 http://orcid.org/0000-0003-3603-1119 http://orcid.org/0000-0003-3603-1119 http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-54361-4&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-54361-4&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-54361-4&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-54361-4&domain=pdf mailto:krauseda@uni-mainz.de www.nature.com/naturecommunications Accordingly, we found that Col2.3 kb GFP+ mesenchymal (osteo- blastic) cells3 from mice with BCR-ABL1+ chronic myeloid leukemia (CML) versus MLL-AF9+ acute myeloid leukemia (AML) showed higher expression of the annexins (Anxa) A2 and A54, which belong to the family of Ca2+-dependent membrane proteins5. This suggested possi- ble regulation of leukemia progression by these BMM factors. Indeed, we recently showed a deficiency of ANXA5 to be involved in the creation of an inflammatory BMM, accelerating AML progression4. ANXA2, on the other hand, facilitates the homing of multiple myeloma cells to the BM6. In leukemia, a role for ANXA2 has only been reported on tumor cells7. Cell surface ANXA2 forms a heterotetramer with S100A10, amember of the S100 family of proteins, facilitating the conversion of plasminogen (PLG) to plasmin by tissue plasminogen activator (tPA) and, thereby, acting as activator of fibrinolysis8. In solid cancers, activation of plasmin, a protease essential for fibrinolysis, promotes disease progression and metastasis via degradation of extracellular matrix (ECM) proteins9, activation of matrix metallopro- teinases (MMPs)10 and release of growth factors (GFs) from the ECM11. Therefore, we hypothesized that the fibrinolytic system, previously implicated in the mobilization of hematopoietic stem and progenitor cells12, including ANXA2, have an impact on leukemia progression, possibly via involvement of the natural anticoagulant pathway and ECM remodeling. Focusing on another BCR-ABL1+ disease, B-cell acute lympho- blastic leukemia (B-ALL), a target of several novel therapies13, in which the BMM is known to play an important role14,15, we unravel here how the fibrinolytic system and its degradation of the ECM in the BMM specifically promotes B-ALL progression. In turn, conditioning of hepatocytes by a leukemic environment leads to the secretion of plasminogen, thereby, perpetuating this leukemia-propagating circuit. We demonstrate the beneficial effect of combination treatment with cytarabine plus the inhibitor of plasmin activation, ε-aminocaproic acid (EACA), in B-ALL, in preclinical models and provide human data supporting the prognostic and therapeutic relevance of interfering with fibrinolysis in human B-ALL. Results ANXA2-deficiency in the BMM leads to survival extension in BCR-ABL1+ leukemia Based on our evidence that Anxa2 was differentially regulated in Col2.3+ cells in AML versus CML (Supplementary Fig. 1A, B)4, we showed thatAnxa2 expressionwas higher in Sca1+ CD73+mesenchymal stromal cells (MSC) from an AML versus a CML BMM (Supplementary Fig. 1C, D, Tables 1, 2). In healthy mice levels of ANXA2 were variable (Supplementary Fig. 1E). These data suggest that gene expression of mesenchymal cells in the murine BMM is differentially impacted by MLL-AF9+ AML versus CML cells. To identify the contribution of BMM-associated ANXA2 to leuke- miaprogression, as suggested by ourmicroarray data,we transplanted WT leukemia-initiating cells intoWTorANXA2-deficient recipientmice using the retroviral transduction/transplantationmodels of BCR-ABL1- induced CML or B-ALL or MLL-AF9-induced AML1. Hereby, the bone marrow of WT donors (pretreated with 5-fluorouracil in the case of CML and AML, but not B-ALL) is transduced with retroviri expressing BCR-ABL1 (for CMLand B-ALL) orMLL-AF9 (for AML) and transplanted into WT or ANXA2-KO recipients. Healthy ANXA2 KO mice, which are viable and fertile16, had no abnormalities in relevant basic hematolo- gical parameters (Supplementary Fig. 2) and hematopoietic stem and progenitor cells (Supplementary Figs. 3, 4) compared towildtype (WT) mice. In follow-up of our microarray, we first initiated CML experi- ments. This revealed that induction of CML led to a significant survival extension in ANXA2-deficient compared to WT recipient mice (Sup- plementary Fig. 5A, B), but homing of CML-initiating cells to WT or ANXA2-deficient environments did not differ (Supplementary Fig. 5C). Myeloid colony formationbyBMcells fromANXA2-deficient recipients with CML was reduced (Supplementary Fig. 5D), and survival of sec- ondaryWT recipients of BM, but not spleen cells fromANXA2-deficient donors with CML was significantly prolonged (Supplementary Fig. 5E). In contrast, no significant differences were observed in survival between WT or ANXA2-deficient recipients with MLL-AF9-induced AML (Supplementary Fig. 5F, G). Induction of BCR-ABL1+ B-ALL (henceforth termed B-ALL), how- ever, led to a significant reduction of GFP+ (BCR-ABL1)+ BP1+ pre-B cells in peripheral blood (PB) (Fig. 1A), as well as survival extension in ANXA2-deficient compared to WT recipient mice (Fig. 1B). BM cells fromANXA2-deficientmicewith B-ALL showed reduced ability to form colonies (Fig. 1C) and mostly failed to induce disease in secondaryWT recipients (Supplementary Fig. 5H). GFP (BCR-ABL1)+ BP1+ B-ALL- initiating cells homed less to the BM, but not the spleen of ANXA2 KO mice (Supplementary Fig. 5I). Intrafemoral transplantation of B-ALL- initiating cells partially rescued disease induction in ANXA2 KO reci- pients (Supplementary Fig. 5J, K). Non-irradiation of recipients prior to transplantation14 led tomore pronounced survival extension in ANXA2 KOmice (Fig. 1D, E). Further, infiltration of the majority of organs with B-ALL cells was reduced in irradiated (Supplementary Fig. 6A–G) ANXA2 KO compared to WT recipients. In contrast, ANXA2 deficiency in B-ALL-initiating cells (LIC) did not lead to changes in survival com- pared to WT LIC (Supplementary Fig. 6H). Lastly, transplantation of empty vector-transduced BM cells into WT versus ANXA2 KO reci- pients only led to slightly increased percentages of GFP+ CD11b+ mye- loid cells in PB and BM of ANXA2 KO mice, but otherwise showed no differences (Supplementary Fig. 7A–J). In summary, with a focus on B-ALL due to the greater need for improved therapies compared to CML, which in most patients is well controlled by treatment with tyr- osine kinase inhibitors17—these results suggest that microenviron- mental ANXA2 deficiency attenuates induction of BCR-ABL1+ B-ALL, possibly due to inhospitality of the BMM. Survival extension in BCR- ABL1+ B-ALL was partially due to reduced LIC homing to an ANXA2 KO BMM, but incomplete rescue of B-ALL in intrafemorally transplanted ANXA2 KOmice andmore pronounced prolongation of B-ALL survival in non-irradiated recipients suggest major contributory roles of an ANXA2-deficient BMM for B-ALL progression. Plasminogen activation contributes to ECM remodeling in the leukemic BMM As ANXA2 supports plasminogen activation, we hypothesized that an ANXA2-deficient BMM impairs homing, engraftment and progression of BCR-ABL1+ B-ALL via an accumulation of ECM proteins. Indeed, levels of the ECM protein fibronectin were increased in the BMM of ANXA2-deficient compared to WT mice with B-ALL (Fig. 2A, Supple- mentary Fig. 8A). Laminin, another ECM protein in the BM18, in con- trast, did not differ significantly in ANXA2 KO versus WT mice (Supplementary Fig. 8B, C). Fibronectin staining in other tissues,which was largely localized around vessels, was only increased in spleens of mice with BCR-ABL1+ B-ALL (Supplementary Fig. 8D). Fibronectin levels were also increased in the BMM of ANXA2-deficient mice transplanted with empty vector-transduced BM and of non-irradiated ANXA2-deficient recipient mice with B-ALL compared to WT mice (Supplementary Fig. 8E, F). Irradiation did not alter preexisting dif- ferences in fibronectin levels between WT and ANXA2 KO mice (Sup- plementary Fig. 8G). Next, we focusedonMSCas important producersof ECMproteins and components of the B-ALL-niche19 for in vitro experiments asmodel system, as primary osteoblastic cells, the object of our earlier micro- array studies (Supplementary Fig. 1A), which are derived from MSC, cannot be cultured easily. BMM-derived primarymurineMSC20 indeed express high levels of ANXA2 (Supplementary Fig. 9A–C). Immuno- phenotyping (Supplementary Fig. 9D) and RNA-sequencing (Supple- mentary Fig. 10A, B) showed no relevant differences between WT and ANXA2-deficient MSC, including with regards to the percentage of Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 2 www.nature.com/naturecommunications MSC of total live stromal cells (Supplementary Fig. 10C). Anxa2 expression was also similar between primary murine MSC at baseline and after differentiation into osteoblasts (Supplementary Fig. 10D). A trend towards reduced generation of fibroblast colony forming units (Supplementary Fig. 11A) and a significant decrease of osteoblastic differentiation (Supplementary Fig. 11B) were found in ANXA2 KO compared to WT MSC, whereby adipocyte differentiation did not dif- fer (Supplementary Fig. 11C). The percentage of F4/80+ CD169+ mac- rophages (Supplementary Fig. 11D), recently shown to play a role in the B-ALL BMM14,15, and fibroblast morphology (Supplementary Fig. 11E) also did not differ between WT and ANXA2 KO mice. ANXA2 expression was similar between primary murine MSC, macrophages and fibroblasts (Supplementary Fig. 11F), as well as in murine stroma (MS5), endothelial (H5V) andfibroblastic (3T3) cell lines (Supplementary Fig. 11G), underlining the likely contributory roles of other cell types of the BMM in addition to our chosen model system, the MSCs. Testing a possible ANXA2-mediated contribution to neoangiogenesis21 to our observed leukemiaphenotypes, we found the percentage of CD45- CD31+ EMCN (endomucin)+ endothelial cells to be increased in the BM of ANXA2 KO mice (Supplementary Fig. 11H, I). We confirmed by co-immunoprecipitation that ANXA2 stabilizes and interacts with S100A10 (Supplementary Fig. 12A). Plasminogen activation assays with WT versus ANXA2-deficient cells confirmed the lack of a functional ANXA2/S100A10 complex and plasminogen acti- vation on ANXA2-deficient MSC, macrophages, fibroblasts (Fig. 2B) and ANXA2-deficient MS5, H5V and 3T3 cells (Supplementary Fig. 12B, C, Table 3). This was accompanied by reduced extracellular laminin and fibronectin associated with WT, but not ANXA2-deficient MSC, whereby fibronectin has previously been described as plasmin substrate22,23 (Fig. 2C, Supplementary Fig. 12D). Lack of ANXA2 onMSC prevented the invasion of BCR-ABL1+ BA/F3 cells, a frequently used model for BCR-ABL1+ B-ALL cells19, through a layer of MSC, embedded in matrigel (Fig. 2D; Supplementary Fig. 12E). However, transient overexpression of ANXA2 in ANXA2-deficient MSC (Supplementary Fig. 12F-G) restored BCR-ABL1+ BA/F3 invasion through MSC embed- ded in matrigel (Fig. 2D). ANXA2-dependent activation of plasmin also resulted in increased activation of MMP924 in WT, but not ANXA2- deficientMSC, specifically after stimulation with tumor necrosis factor (TNF)α, which is known to be secreted byB-ALL cells19 (Supplementary Fig. 12H). To confirm the importance of plasminogen activation for B-ALL progression, we transplanted BCR-ABL1-transduced BM cells into WT or tPA-deficient mice. PB leukocytes and GFP+ (BCR-ABL1)+ BP1+ pre-B cells (Supplementary Fig. 12I) were reduced, and survival was sig- nificantly prolonged in tPA-deficient compared toWTmice (Fig. 2E). A similar result was obtained in secondary recipients transplanted with BM cells from WT or tPA-deficient mice with B-ALL (Supplementary Fig. 12J, K). Furthermore, BM cells isolated from leukemic tPA-deficient mice failed to form colonies in vitro (Fig. 2F). In line with data from ANXA2 KO mice, immunofluorescence (IF) analysis of bone sections revealed higher fibronectin levels in the BM of tPA-deficient compared toWTmice with BCR-ABL1+ B-ALL (Fig. 2G). Consistently, intrafemoral administration of tPA to WT or ANXA2-deficient mice, in order to bypass ANXA2-dependent plasminogen activation, significantly decreased fibronectin levels in ANXA2 KO, but not in WT mice, in which plasmin is normally activated (Fig. 2H, Supplementary Fig. 12L). Systemic plasmin administration to mice with B-ALL, which accumulated in the BMM (Supplementary Fig. 13A), however, led to acceleration and a ‘rescue’ of the disease (Fig. 2I), as well as significant reduction of fibronectin in the BM of ANXA2-deficient mice (Fig. 2J, Supplementary Fig. 13B). In support of specific roles of plasmin for cleaving protein targets25, recipient mice deficient for protease- activated receptor (PAR)-1, which is activated by plasmin26, trans- planted with B-ALL LIC phenocopied the survival extension observed B C 0 5 20 30 40 50 60 70 0 50 100 B- AL L ov er al l s ur vi va l ( % ) Days after transplant WT ANXA2 KO P = 0.009 Recipients D E 0 200 400 600 Co lo ni es p er p la te WT ANXA2 KO WT ANXA2 KO BM Spleen P = 0.038 P = 0.0008 BM Spleen A 0 20 40 60 80 % G FP + B P1 + P = 0.013 0 5 10 15 20 25 W BC /μ l ( x1 03 ) WT ANXA2 KO Day 25 W BC /μ l ( x1 03 ) WT ANXA2 KO Day 17 0 10 20 30 40 50 % G FP + B P1 + P < 0.0001 0 5 10 15 0 20 20 40 60 80 100 120 0 50 100 Days after transplant B- AL L ov er al l s ur vi va l ( % ) (n on -ir ra di at ed ) WT ANXA2 KO P = 0.01 Recipients Fig. 1 | ANXA2-deficiency in the BMM leads to survival extension in BCR-ABL1+ leukemia. AWBC count per µl (left) and percentage of GFP (BCR-ABL1)+ BP1+ cells of all cells (right) in the peripheral blood of WT (black) or ANXA2 KO (gray) reci- pient mice with BCR-ABL1+ B-ALL (P =0.013, two-tailed t test, WT n = 5, ANXA2 KO n = 10, mean± SD).BKaplan–Meier-style survival curve ofWT (black) or ANXA2KO (gray) recipient mice with BCR-ABL1+ B-ALL (P =0.009, Log-rank test, WT n = 14, ANXA2 KO n = 17). C Number of colonies per plate derived from total BM or spleen cells from WT (black) or ANXA2 KO (gray) recipient mice with BCR-ABL1+ B-ALL (P =0.038, P =0.0008, two-way ANOVA, Tukey test, n = 3 (3 replicates of 3 indivi- dual mice per group), mean± SD).DWBC count per µl (left) and percentage of GFP (BCR-ABL1)+ BP1+ cells (right) of all cells in the peripheral blood of non-irradiated WT (black) or ANXA2 KO (gray) recipient mice with BCR-ABL1+ B-ALL (P <0.0001, two-tailed t test, WT n = 9, ANXA2 n = 9, mean± SD). E Kaplan–Meier-style survival curve of non-irradiated WT (black) or ANXA2 KO (gray) recipient mice with BCR- ABL1+ B-ALL (P = 0.01, Log-rank test,WT= 5, ANXA2n = 4). Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 3 www.nature.com/naturecommunications in ANXA2-deficient mice (Supplementary Fig. 13C, D). Consistently, urokinase-type plasminogen activator receptor (uPAR), the receptor for urokinase plasminogen activator (uPA), known to be involved in degradation of the ECM, as well as invasion and metastasis of malig- nant tumors27, was higher in the plasma of WT mice with B-ALL than healthy mice (Supplementary Fig. 13E). Thus, ANXA2-mediated plas- minogen activation is central to BCR-ABL1+ B-ALL development. B-ALL promotes hepatic generation of plasminogen Next, we tested levels of plasminogen, plasmin and tPA in WT or ANXA2 KO mice with B-ALL. Consistent with ANXA2’s role in con- verting plasminogen to plasmin, plasmin, but not plasminogen or tPA, was reduced in the BMor plasma of healthy ANXA2 KOmice or ANXA2 KO mice with B-ALL compared to the respective WT controls (Fig. 3A–C, Supplementary Fig. 14A–C). However, levels of plasmin, A Ctrl tPA and PLG WT ANXA2 KO FibronectinDAPI B 0 10 20 30 40 % A re a of fi br on ec tin s ta in in g WT ANXA2 KO C WT ANXA2 KO 0 1 2 3 Fo ld c ha ng e in va si on BC R- AB L1 + B A/ F3 P = 0.0004 P = 0.0010 P = 0.0005 P = 0.0013 WT WT + PLG ANXA2 KO ANXA2 KO + PLG ANXA2 KO + ANXA2 OE ANXA2 KO + ANXA2 OE + PLG D 0 40 40 60 80 100 120 0 50 100 Days after transplant B- AL L ov er al l s ur vi va l ( % ) WT tPA KO P = 0.0017 RecipientsE Nu m be r o f c ol on ie s pe r p la te WT tPA KO WT tPA KO BM Spleen P = 0.023 G I J Fi br on ec tin n or m al iz ed m ea n in te ns ity WT WT + tPA and PLG ANXA2 KO ANXA2 KO + tPA and PLG P = 0.0047 Fo ld c ha ng e pl as m in a ct iv at io n (u ni ts /m l p la sm in ) P = 0.010 WT ANXA2 KO 0 Plasmin 2 Days 9 16 23 ANXA2 KO vehicle Merge DAPI Fibronectin ANXA2 KO plasmin 0 15 20 30 40 50 0 50 100 Days after transplant B- AL L ov er al l s ur vi va l ( % ) ANXA2 KO vehicle ANXA2 KO plasmin P = 0.037 Recipients tPA PLG I ANXA2/S100A10 II tPA PLG ANXA2/S100A10 Plasmin III tPA PLG ANXA2/S100A10 tPA D -Val -Leu -LysPlasmin IV ANXA2/S100A10 WT cells + + tPA PLG I II III D -Val -Leu -Lys IV ANXA2 KO cells + + tPA PLG tPA PLG tPA PLG 100 μm WT tPA KO Merge DAPI Fibronectin 40 μm 40 μm 100 μm n. s. MSC MΦ Fibroblasts 0 0.5 1.0 1.5 2.0 P = 0.0035 P = 0.009 P = 0.04 WT B-ALL i.f. vehicle WT B-ALL i.f. tPA Merge DAPI Fibronectin ANXA2 B-ALL i.f. vehicle ANXA2 B-ALL i.f. tPA F H 50 μm 0 20 40 50 100 150 0 1 2 3 4 Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 4 www.nature.com/naturecommunications plasminogen and tPA were higher in the BM of WT mice with B-ALL compared to healthy mice (Fig. 3D–F), although we cannot exclude that the employed antibody to plasminogen may also be detecting plasmin. D-dimers, representing fibrinogen degradation products as a result of degradation by plasmin, were similar in the plasma of healthy WT versus ANXA2 KOmice (Supplementary Fig. 14D), but increased in the plasma of WT mice with B-ALL compared to healthy controls (Supplementary Fig. 14E). Plasminogen activator inhibitor (PAI)-1, an inhibitor of tPA, was significantly reduced inWTmicewith B-ALL compared to healthymice (Fig. 3G). The anti-protease α2-macroglobulin was significantly decreased in healthy ANXA2 KO compared to healthy WT mice (Sup- plementary Fig. 14F), while there was a trend towards decreased levels ofα2-macroglobulin in the plasmaofWTmicewith B-ALL compared to healthymice (Supplementary Fig. 14G).α2-antiplasminwasunchanged (Supplementary Fig. 14H, I). Next, we hypothesized, that leukemia cells influence the genera- tion of plasminogen by hepatic cells, either directly or by conditioning of the BMM, which may secrete a mediator of hepatic plasminogen production. Given that interleukin-6 (IL-6) is known to regulate gene expression of plasminogen in hepatocytes28, andmay be secreted by B cells29 or cells of the BMM, we measured IL-6 levels. We observed that IL-6 was higher in the plasma and liver of WT mice with B-ALL com- pared to healthy mice (Fig. 3H, I). Consistently, incubation of Huh7 hepatic cells with recombinant IL-6 increased the expression of plas- minogen (Fig. 3J). In a chromatin immunoprecipitation (ChIP) assayusingHuh7 cells treatedwith vehicle or IL-6, we showed that IL-6 significantly increased the binding of cAMP Responsive Element Binding Protein 1 (CREB1), oneof the transcription factors downstreamof the IL-6 receptor, to the PLG regulatory element (Fig. 3K, Supplementary Table 4). In light of our published studies that B-ALL cells produce high amounts of TNFα19,30 and that TNFα increases IL-6 expression in MSC4, we specu- lated that MSC and/or other cell types in the BMM conditioned by B-ALL cells are the likely source of IL-6 leading to hepatic generation of plasminogen. Indeed, treatment of human MSC with TNFα or cocul- ture of human MSC with the BCR-ABL1+ B-ALL cell line SUPB15, which produces TNFα (Supplementary Fig. 14J), and subsequent culture of Huh7 cells in the conditioned medium of these MSC led to increased levels of plasminogen in the conditioned medium of Huh7 cells (Sup- plementary Fig. 14K). There was a strong trend towards inhibition of this increase in plasminogen in the conditionedmedium of Huh7 cells, if the Huh7 cells, cultured in the conditioned medium from the TNFα- exposed MSC, were simultaneously treated with an antibody to IL-6 (Supplementary Fig. 14L). Taken together, these data suggest that B- ALL-cell derived TNFα conditions MSC and possibly other cell types in the BMM to secrete IL-6, which promotes hepatic generation of plas- minogen. In the presence of B-ALL levels of some inhibitors of the plasmin system are reduced. Restricted access to growth factors in the BMM leads to reduced mTORC2 activation and proliferation of BCR-ABL1+ cells Using ANXA2-deficientMSC as a BMMmodel of impaired plasminogen activation and ECM accumulation, we evaluated the proliferation of BCR-ABL1+ BM cells plated in limiting dilution31 on MSC and showed that leukemia cells were slower at reaching confluence on ANXA2- deficient versus WT MSC (Supplementary Fig. 15A). Immunoblot ana- lysis of sorted primary murine B-ALL cells for downstream targets of the BCR-ABL1 oncoprotein revealed that AKT phosphorylation was significantly decreased in B-ALL cells from the BM (Fig. 4A), but not the spleen (Supplementary Fig. 15B), of ANXA2-deficient compared to WT mice. BCR-ABL1 mediates AKT phosphorylation on the T308 residue (pAKTT308)32, while phosphorylation of the S473 residue (pAktS473) is mainly regulated by mTOR complex 2 (C2)33,34. mTORC2 lies down- stream of receptors for growth factors such as insulin-like growth factor 1 (IGF1)35,36. Indeed, IF analysis showed increased colocalization of fibronectin and IGF1 in the BM of mice lacking ANXA2 compared to WT mice, both in the healthy condition (Supplementary Fig. 15C–F) and in B-ALL (Fig. 4B, Supplementary Fig. 15G, H). A similar increased colocalizationwas observed in the BMof tPA-deficientmicewith B-ALL (Fig. 4C; Supplementary Fig. 15I, J). Treatment with IGF1 modestly increased the numbers of BCR- ABL1+ BA/F3, but not MLL-AF9+ BA/F3 cells (Fig. 4D). This effect was blocked in BCR-ABL1+ BA/F3 cells using the mTOR inhibitor KU- 0063794 (Supplementary Fig. 15K). These data suggest that growth factors such as IGF1 colocalize with ECM proteins in the BM of ANXA2 KO or tPA KO compared to WT mice and may be trapped there. Decreased mTORC2 levels within the IGF1 signaling pathway may lead to decreased numbers and/or survival of BCR-ABL1+ cells. Differential sensitivity of leukemias to signaling events downstream of IGF1R We then hypothesized that there is differential oncogene- and/or lineage-dependent sensitivity of leukemias to mTORC2 signaling downstream of the IGF1R, which may explain the observed pheno- types of CML, B-ALL and AML in ANXA2 KO mice. Indeed, IGF1 stimulated an increase in protein levels of rapamycin- insensitive companion of mammalian target of rapamycin (Ric- tor), one of the main components of mTORC236, and an increase of pAKTS473 in BCR-ABL1+ BA/F3 (Fig. 5A) and primary murine B-ALL cells (Fig. 5B), but not in empty vector+, MLL-AF9+ (BA/F3) or pri- mary murine MLL-AF9+ AML counterparts. In contrast, the regulatory-associated protein of mTOR (Raptor), a component of mTOR complex 1 (C1), and pS6K1, downstream of mTORC1, were not affected by IGF1 treatment (Supplementary Fig. 16A, B). Similarly, highest levels of mTORC2, particularly Rictor and pAKTS473, were seen in K562 (BCR-ABL1+ CML cell line) and NALM6 cells (B-ALL cell line), compared with the AML cell lines THP1 and Kasumi (Supplementary Fig. 16C). However, IGF1 treatment resulted Fig. 2 | Plasminogen activation contributes to ECM remodeling in the leukemic BMM. A Immunohistochemistry of fibronectin in bones of WT (black) or ANXA2 KO (gray) recipient mice with BCR-ABL1+ B-ALL (P =0.04, two-tailed t test, n = 5 mice per group, mean± SD). B Active plasmin in the medium of primary WT (black) or ANXA2 KO (gray) mesenchymal stromal cells (MSC), macrophages (MΦ) or fibroblasts (P =0.01, P =0.0035, P =0.009, two-way ANOVA, Sidak test, n = 5 biological replicates, mean ± SD), normalized to WT. C Immunofluorescence of fibronectin (green) and 4′,6-diamidino-2-phenylindole (DAPI, blue) with or without activation of plasmin in cultures of WT (black) or ANXA2 KO (gray) MSC (P =0.0047, two-way ANOVA, Sidak test, n = 6 biological replicates, mean ± SD), normalized tocell number.D InvasionofBCR-ABL1+ BA/F3 cells platedon topofWT (black) or ANXA2 KO MSC (gray) or ANXA2 KO MSC overexpressing (OE) ANXA2 (light gray) (P =0.0004, P =0.0005, P =0.0013, P =0.001, two-way ANOVA, Tukey test n = 3 biological replicates,mean± SD), normalized toWT.EKaplan–Meier-style survival curve ofWT (black) versus tPA-knockout (tPAKO; gray) recipientmicewith BCR-ABL1+ B-ALL (P =0.0017, Log-rank test, WT n = 5, tPA KO n = 5). F Number of colonies per plate derived from total bonemarrow or spleen cells fromWT (black) or tPA KO (gray) recipient mice with BCR-ABL1+ B-ALL (P =0.023, two-way ANOVA, Tukey test, n = 3 (3 replicates of 3 individual mice per group), mean ± SD). G Immunofluorescence of bones, stained for fibronectin (green) and DAPI (blue), from WT or tPA KO recipient mice with BCR-ABL1+ B-ALL (n = 3 mice per group). H Immunofluorescence of bones, stained for fibronectin (green) and DAPI (blue), from WT or ANXA2 KO recipient mice with BCR-ABL1+ B-ALL. Vehicle or tPA was administered by intrafemoral (i.f.) injection (n = 5mice per group). I Kaplan–Meier- style survival curve of ANXA2 KO recipient mice with B-ALL treated with vehicle (solid line) or plasmin (dashed line) (P =0.037, Log-rank test, vehiclen = 10, plasmin n = 12). J Immunofluorescence of bones, stained for fibronectin (green) and DAPI (blue), fromANXA2KOrecipientmicewith BCR-ABL1+ B-ALL treatedwith vehicle or plasmin (n = 4 mice per group). Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 5 www.nature.com/naturecommunications in increased levels of both Rictor and pAKTS473 only in NALM6 cells (Fig. 5C). Consistently, highest levels of IGF1R were found on leu- kemia cells from mice with BCR-ABL1+ B-ALL (Fig. 5D) and BCR- ABL1+ BA/F3 cells (Supplementary Fig. 16D) compared to MLL-AF9+ cells. Testing differences between the BCR-ABL1+ cell lines K562 and SUPB15, whereby the former is of myeloid and the latter of lymphoid lineage, we found higher expression of IGF1R on SUPB15 cells (Fig. 5E). After stimulation with IGF1 Rictor and—as a trend—pAKTS473 were higher in SUPB15 compared to K562 cells (Fig. 5F). Taken together, these data suggest that IGF1 differentially affects signaling in BCR-ABL1+ versus MLL-AF9+ cells, whereby the effect is greatest on BCR-ABL1+ lymphoid compared to BCR-ABL1+ myeloid cells, possibly due to higher expression of IGF1R on B-ALL cells. G A Pl as m in (u ni ts /m l) (x 10 -3 ) P = 0.0317 Bone marrow Liver WT ANXA2 KO B-ALL B 0 0.2 0.4 0.6 0.8 Bone marrow P = 0.0608 0 0.05 0.10 0.15 0.20 Liver WT healthy WT B-ALL 0 100 100 125 150 Pl as m in og en (n g/ m l) Pl as m in og en (n g/ m l) (x 10 3 ) Bone marrow Plasma B-ALL 0 100 100 125 150 WT ANXA2 KO Liver C 0 200 400 600 800 1000 tP A (p g/ m l) Bone marrow Plasma Liver B-ALL 0 1000 2000 3000 WT ANXA2 KO E 0 50 100 150 200 0 500 1000 1500 2000 2500 0 100 200 300 400 500 P < 0.0001 D 0 200 250 300 350 400 450 500 P = 0.0061 Bone marrow Liver WT healthy WT B-ALL P = 0.0188 P = 0.0018 F WT healthy WT B-ALL Bone marrow Liver 0 100 200 300 PA I-1 (p g/ m l) P = 0.0079 Bone marrow WT healthy WT B-ALL J K IL -6 (p g/ m l) P = 0.0309 Plasma WT healthy WT B-ALL H 0 30 30 40 50 60 Huh7 vehicle Huh7 + IL-6 PL G m RN A re la tiv e ex pr es si on 0 0.8 1.0 1.5 2.0 2.5 P = 0.0023 IgG IP CREB IP % In pu t P = 0.0191 P = 0.0115 Set 1, PLG promoter Pl as m in (u ni ts /m l) (x 10 -3 ) Pl as m in og en (n g/ m l) tP A (p g/ m l) tP A (p g/ m l) Pl as m in (u ni ts /m l) (x 10 -3 ) Pl as m in (u ni ts /m l) (x 10 -3 ) Pl as m in og en (n g/ m l) Pl as m in og en (n g/ m l) tP A (p g/ m l) tP A (p g/ m l) 0 0.05 0.10 0.15 0.20 0.25 Pl as m in (u ni ts /m l) (x 10 -3 ) P = 0.0048 Plasma 0 5000 10000 15000 0 200 400 600 800 Plasma 0 0.5 1.0 1.5 2.0 2.5 Pl as m in (u ni ts /m l) (x 10 -3 ) 0 1000 2000 3000 4000 5000 tP A (p g/ m l) Plasma Pl as m in og en (n g/ m l) (x 10 3 ) Plasma 0 200 400 600 IgG IP CREB IP % In pu t Huh7 vehicle Huh7 + IL-6 P = 0.0510 P = 0.0114 Set 2, PLG promoter 0 0.2 0.4 0.6 I 0 50 100 150 200 250 IL -6 (p g/ m l) P = 0.0126 Liver WT Healthy WT B-ALL 0 0.2 0.4 0.6 0 0.01 0.02 0.03 0.04 0 0.05 0.10 0.15 Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 6 www.nature.com/naturecommunications Plasmin regulates IGF1 availability by contributing to breakdown of the ECM in an ANXA2-dependent manner Hypothesizing that MSC-mediated plasmin activation may regulate IGF1 availability, we found that primary BCR-ABL1+ B-ALL cell numbers significantly increased when IGF1 was released from matrigel containingWTMSC after plasmin activation by plasminogen (Fig. 6A). This increase of BCR-ABL1+ B-ALL cell numbers was due to a decrease of late apoptosis, when the BCR-ABL1+ B-ALL cells were cultured on WT, but not ANXA2 KO MSC embedded in matrigel containing IGF1 and PLG (Supplementary Fig. 17). However, in absence of ANXA2 in Fig. 3 | B-ALL promotes hepatic generation of plasminogen. A Active plasmin in supernatants of BM, plasma or liver fromWT (black) or ANXA2 KO (gray) recipient mice with B-ALL (P =0.0317, two-tailed Mann–Whitney test, n = 5 mice per group, mean ± SD).B Plasminogen in supernatants of BM, plasma or liver fromWT (black) or ANXA2 KO (gray) recipient mice with B-ALL (n = 5 mice per group, mean ± SD). C Tissue plasminogen activator (tPA) in supernatants of BM, plasma or liver from WT (black) or ANXA2 KO (gray) recipient mice with B-ALL (WT n = 4, ANXA2 KO n = 5, mean ± SD). D Active plasmin in supernatants of BM, plasma or liver from healthyWT (black) orWT recipientmicewithB-ALL (purple) (P =0.0608, two-tailed t test, n = 5 mice per group, mean ± SD). E Plasminogen in supernatants of BM, plasma or liver from healthy WT (black) or WT recipient mice with B-ALL (purple) (P <0.0001, P =0.0061, two-tailed t test n = 5 mice per group, mean ± SD). F tPA in supernatants of BM, plasma or liver from healthy WT (black) or WT recipient mice with B-ALL (purple) (P =0.0188, P =0.0018, two-tailed t test n = 5 mice per group, mean ± SD).G Plasminogen-activator inhibitor 1 (PAI-1) in the BM supernatant from healthyWTmice (black) orWT recipientmicewith B-ALL (purple) (P =0.0079, two- tailed Mann–Whitney test, n = 5 mice per group, mean± SD). Mice in (A)–(G) were unirradiated. H, I Interleukin 6 (IL-6) in plasma (H) and liver (I) from healthy WT (black) orWT recipientmicewith B-ALL (purple) (P =0.0309, P =0.0126, two-tailed t test, n = 5mice per group,mean ± SD). J Relative expression of plasminogen (PLG) in Huh7 cells treated with vehicle (circles) or IL-6 (squares) (P =0.0023, two-tailed t test,n = 4 biological replicates,mean ± SD).KCREBbinding to the promoter of PLG in Huh7 cells treated with vehicle (circles) or IL-6 (squares) in a chromatin immu- noprecipitation (ChIP) assay (P =0.0191, P =0.0510, two-way ANOVA, Sidak test, n = 4 biological replicates, mean ± SD). We cannot exclude that the employed antibody to plasminogen may also be detecting plasmin. Source data are provided as a Source Data file. A C B D WT ANXA2 KO Merge DAPI Fibronectin IGF1 Fibronectin + IGF1 40 μm WT tPA KO Merge DAPI Fibronectin IGF1 Fibronectin + IGF1 Empty vector BCR-ABL1 MLL-AF9 Nu m be r o f c el ls (x 10 6 ) BA/F3 P = 0.002 Ctrl IGF1 0 1 2 3 4 40 μm 188 mTOR 188 Rictor 64 pAKTS473 64 pAKTT308 62 AKT 38 pCRKL 38 1 2 3 4 WT 1 2 3 4 ANXA2 KO kDa CRKL 38 GAPDH mTOR Rictor pAKTS473 pAKTT308 AKT 0 0.5 1.0 1.5 Si gn al in te ns ity (n or m al iz ed to G AP DH ) P = 0.040 P = 0.014 P = 0.003 P = 0.043 WT ANXA2 KO Fig. 4 | Restricted access to growth factors in the BMM leads to reduced mTORC2 activation and proliferation of BCR-ABL1+ cells. A Immunoblot for the indicated proteins in lysates of sorted GFP (BCR-ABL1)+ BP1+ BM cells from WT (black) or ANXA2 KO (gray) recipient mice with BCR-ABL1+ B-ALL. Each column represents a single mouse (P =0.04, P =0.014, P =0.003, P =0.043, two-way ANOVA, Sidak test, n = 4 mice per group, mean ± SD). B Immunofluorescence of bones, stained for fibronectin (green), DAPI (blue) and insulin-like growth factor (IGF)1 (purple), from WT or ANXA2 KO recipient mice with BCR-ABL1+ B-ALL (n = 3 mice per group). C Immunofluorescence of bones, stained for fibronectin (green), DAPI (blue) and IGF1 (purple), from WT or tPA KO recipient mice with BCR-ABL1+ B-ALL (n = 3 mice per group). D Number of empty vector+, BCR-ABL1+ or MLL-AF9+ BA/F3 cells after 48h of culture in medium containing vehicle (circles) or murine recombinant IGF1 (squares) (12 ng/ml) (P =0.002, two-way ANOVA, Sidak test, n = 5 biological replicates, mean ± SD). Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 7 www.nature.com/naturecommunications A Empty vector BCR-ABL1 MLL-AF9 0 1 2 3 R ic to r s ig na l i nt en si ty (n or m al iz ed to G A PD H ) BA/F3 P = 0.049 Ctrl IGF1 pA K TS4 73 s ig na l i nt en si ty (n or m al iz ed to G A PD H ) Empty vector BCR-ABL1 MLL-AF9 0 1 2 3 P = 0.004 BA/F3 Ctrl IGF1 BCR-ABL1 Ctrl IGF1 Empty vector Ctrl IGF1 MLL-AF9 39 GAPDH 191 mTOR kDa 191 Rictor 64 pAKTS473 64 AKT BA/F3 B Rictor pAKTS473 B-ALL AML Si gn al in te ns ity (n or m al iz ed β -a ct in ) P = 0.0001 % G FP + I G F1 R+ c el ls P < 0.0001 B-ALL AML D 0 10 20 30 40 50 0 0.5 1.0 1.5 P = 0.016 188 188 62 62 β-actin pAKTS473 AKT mTOR 2 3 4 B-ALL 1 2 3 4 1kDa Rictor AML 38 C 188 188 64 pAKTS473 64 AKT mTOR Rictor 50 β-actin kDa Ctrl 5´ 10´ IGF1 K562 Ctrl 5´ 10´ IGF1 NALM6 Ctrl 5´ 10´ IGF1 THP1 0 0.5 1.0 1.5 pA KT S4 73 s ig na l i nt en si ty (n or m al iz ed to β -a ct in ) P = 0.011 Ctrl 5‘ IGF1 10‘ IGF1 K562 NALM6 THP1 K562 NALM6 THP1 0 0.5 1.0 1.5 2.0 2.5 Ri ct or s ig na l i nt en si ty (n or m al iz ed to β -a ct in ) P = 0.018 Ctrl 5‘ IGF1 10‘ IGF1 Raptor pS6K1 S6K1 188 64 64 P < 0.0001 K562 SUPB15 Ctrl IGF1 K562 39 GAPDH 191 mTOR kDa 191 Rictor 64 pAKTS473 64 AKT Ctrl IGF1 SUPB15 Rictor pAKTS473 0 0.5 1.0 1.5 2.0 2.5 K652 Pr ot ei n in te ns ity si gn al (n or m al iz ed to G A PD H ) Vehicle IGF1 Rictor pAKTS473 0 0.5 1.0 1.5 SUPB15 Pr ot ei n in te ns ity si gn al (n or m al iz ed to G A PD H ) P = 0.0260 P = 0.0868 Vehicle IGF1 E F 0 20 40 60 % IG F1 R+ c el ls Fig. 5 | Differential sensitivity of leukemias to signaling events downstream of IGF1R.A Immunoblot for the indicated proteins in lysates of emptyvector+, BCR- ABL1+ andMLL-AF9+ BA/F3 cells treated with vehicle (circles) or IGF1 (squares) and quantification (right) of the band intensity for Rictor and pAKTS473 normalized to GAPDH (P =0.049, P =0.004, two-way ANOVA, Sidak test, n = 6 biological repli- cates,mean± SD).B Immunoblot for the indicated proteins in lysates of sorted GFP (BCR-ABL1 or MLL-AF9)+ BM cells from individual WT recipient mice with BCR- ABL1+ B-ALL (circles) or MLL-AF9+ AML (squares) and quantification of the band intensity for Rictor and pAKTS473 normalized to β-actin (P =0.0001, P =0.016, two- way ANOVA, Sidak test, n = 4 biological replicates, mean± SD). C Representative immunoblot for the indicated proteins in lysates of K562, NALM6 or THP1 cells treated with vehicle (circles) or for 5 (squares) or 10 (triangles) minutes human recombinant IGF1 (12 ng/ml) and its quantification (right) of the band intensities for Rictor and pAKTS473 normalized to β-actin (P =0.018, P =0.011, two-way ANOVA, Dunnett test, n = 4 biological replicates, mean± SD). The samples are derived from the same experiment, but one gel was used for mTOR, Rictor, pAktS473, Akt and β-actin. Another gel was used for Raptor, pS6K1 and S6K1. D Percentage of IGF1- receptor (IGF1R)+ cells of all GFP (BCR-ABL1)+ B-ALL (circles) versus all GFP (MLL-AF9)+ AML cells (squares) in the murine system (P <0.0001, two-tailed t test, B-ALL n = 4, AML n = 7 biological replicates, mean± SD). E Percentage of IGF1- receptor (IGF1R)+ of all K562 (circles) versus SUPB15 (squares) cells (P <0.0001, two-tailed t test, n = 4 biological replicates, mean ± SD). F Representative immu- noblot for the indicated proteins in lysates of K562 and SUPB15 cells treated with vehicle (circles) or human recombinant IGF1 (squares) and quantification (right) of the band intensities for Rictor and pAKTS473 normalized to GAPDH (P =0.026, P =0.0868, two-way ANOVA, Sidak test, n = 6 biological replicates, mean ± SD). Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 8 www.nature.com/naturecommunications MSC or in presence of the plasmin activation inhibitor ε-aminocaproic acid (EACA) no increase of B-ALL cell numbers was observed (Fig. 6A). In order to test the relevanceof plasmin for breakdownof the ECMand subsequent release of IGF1 from the ECM, we, firstly, showed that plasminogen activation was reduced in ANXA2 KO MSC, in WT MSC treated with an inhibitor of the ANXA2/S100 A10 heterotetramer (A2ti)37 or treatedwith an antibody that blocks the tPA-binding domain of ANXA2 (ANXA2 Ab)38 (Fig. 6B). Secondly, addition of A2ti or ANXA2 Ab to matrigel containing WT MSC, IGF1 and PLG prevented an increase of primary BCR-ABL1+ B-ALL cells via IGF1 release from the matrix (Fig. 6C). And thirdly, plasmin inhibition by aprotinin sig- nificantly decreased BCR-ABL1+ B-ALL cell numbers cultured on top of an ECM containing IGF1 and PLG (Fig. 6D). Lastly, treatment of WT mice with B-ALL with an inhibitor of IGF1R significantly extended sur- vival (Fig. 6E). In sum, these data suggest that ANXA2 is involved in degradation of the ECM via its mediation of plasminogen activation. Breakdown of the ECM, in turn, releases growth factors such as IGF1, which promote the proliferation of B-ALL cells. Targeting fibrinolysis-associated pathways may extend survival of leukemic mice We next analyzed therapeutic implications by treating mice with EACA, an approved anti-hemorrhagic drug. Indeed, treatment of non-leukemic mice with EACA led to a significant increase in mIGF1 + mPLG mIGF1 + mPLG + EACA P = 0.01 P = 0.02 A Ctrl sIGF1 mIGF1 Ce ll nu m be r/m l ( x1 03 ) P = 0.002 P = 0.0002 P = 0.006 P = 0.0007 P = 0.0002 P = 0.006 WT ANXA2 KO P = 0.059 Primary BCR-ABL1+ BP1+ MSC WT or ANXA2 KO sIGF1 mIGF1 mIGF1 mPLG EACA mIGF1 mPLG mPLGCtrl 0 40 80 120 160 PLG + tPA PLG + tPA + A2ti PLG + tPA + ANXA2 Ab 0 0.4 0.4 0.8 1.2 Fo ld c ha ng e pl as m in a ct iv at io n (u ni ts /m l p la sm in ) P = 0.0013 P = 0.0010 P < 0.0001 0 50 100 150 200 250 WT ANXA2 KO Ce ll nu m be r/m l ( x1 03 ) P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 Ctrl sIGF1 mIGF1 mIGF1 + mPLG mIGF1 + mPLG + ANXA2 Ab mIGF1 + mPLG + A2ti B C WT ANXA2 KO Ctrl sIGF1 mIGF1 mIGF1 + mPLG mIGF1 + mPLG + aprotinin 0 100 200 300 400 Ce ll nu m be r/m l ( x1 03 ) P = 0.0038 P = 0.0002 P < 0.0001 P = 0.0038P < 0.0001 WT ANXA2 KO D E Days after transplant B- AL L ov er al l s ur vi va l ( % ) Vehicle Linsitinib (IGF1R inhibitor) P = 0.0028 Recipients0 Linsitinib 9 Days 0 10 10 15 20 25 30 35 40 45 50 55 0 50 100 Fig. 6 | Plasmin regulates IGF1 availability by contributing to breakdown of the ECM in an ANXA2-dependent manner. A Number of sorted primary GFP (BCR- ABL1)+ BP1+ cells from the BMofWTmice with BCR-ABL1-induced B-ALL. Leukemia cells were plated on either WT (black) or ANXA2 KO (gray) MSC embedded in matrigel in different conditions (schematic on left). IGF1 was added in solution (sIGF1) or embedded in matrigel (mIGF1). mPLG = plasminogen (1 ng/ml) and EACA= ε-aminocaproic acid were embedded in matrigel (two-way ANOVA, Tukey test, n = 5 biological replicates, mean± SD). B Active plasmin in the medium of primary WT (black) or ANXA2 KO (gray) mesenchymal stromal cells (MSC) in pre- sence or absence of the ANXA2/S100A10 heterotetramer inhibitor (A2ti) (50 μm)or an antibody against ANXA2 (ANXA2 Ab) (60 μg/ml) (P =0.0013, P =0.001, P <0.0001, two-way ANOVA, Tukey test, n = 9 biological replicates, mean± SD), normalized toWT.CNumber of sortedprimaryGFP (BCR-ABL1)+ BP1+ cells from the BM of WT mice with BCR-ABL1-induced B-ALL. Leukemia cells were plated on WT (black) or ANXA2KO (gray)MSCembedded inmatrigel indifferent conditions. IGF1 was added in solution (sIGF1) or embedded in matrigel (mIGF1). mPLG = plasmi- nogen, ANXA2 Ab (50μm) and A2ti (60 μg/ml) were embedded in matrigel (two- wayANOVA, Tukey test,n = 8biological replicates,mean ± SD).DNumberof sorted primary GFP (BCR-ABL1)+ BP1+ cells from the BM of WT mice with BCR-ABL1- induced B-ALL. Leukemia cells were plated onWT (black) or ANXA2 KO (gray) MSC embedded inmatrigel in different conditions. IGF1 was added in solution (sIGF1) or embedded in matrigel (mIGF1). mPLG = plasminogen and aprotinin were embed- ded in matrigel (two-way ANOVA, Tukey test, n = 6 biological replicates, mean± SD). E Kaplan–Meier-style survival curve of WT recipient mice with BCR-ABL1+ B-ALL treated with vehicle (solid line) or the insulin-like growth factor receptor 1 inhibitor linsitinib (dotted line) (P =0.0028, Log-rank test, vehicle n = 6, linsitinib n = 7). Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 9 www.nature.com/naturecommunications fibronectin levels in the BMM (Fig. 7A). Treatment of WT mice with BCR-ABL1+ B-ALLwith EACA after presumably completed homing1,39, led to significant reduction of the GFP (BCR-ABL1)+ BP1+ tumor load (Supplementary Fig. 18A), leukemic infiltration into various organs (Supplementary Fig. 18B–H) and survival extension (Fig. 7B). ANXA2-deficient mice were resistant to these effects (Supplemen- tary Fig. 18I). EACA treatment of mice with BCR-ABL1+ B-ALL resul- ted in significantly higher fibronectin levels as seen in ANXA2- deficientmice (Fig. 7C; Supplementary Fig. 18J). B-ALL cells from the BM of EACA-treated mice also showed a significant reduction in the levels of Rictor and pAKTS473 (Fig. 7D). Combination treatment with EACA and low dose cytarabine (ara-C), considered standard therapy in B-ALL and initiated when disease was detectable (Supplementary Fig. 18K), resulted in a significant survival extension compared to mice receiving vehicle (Fig. 7E, Supplementary Fig. 18L). Lastly, intravenous treatment of mice with B-ALL with EACA significantly reduced levels of plasmin/-ogen, plasminogen and tPA in the BM compared to vehicle treatment (Supplementary Fig. 19A–C). Local administration of EACA to the BMM by intrafemoral injection led to increased fibronectin levels in the BMM of WT, but not ANXA2 KO mice (Supplementary Fig. 19D). These data support the concept that targeting plasminogen activation via administration of EACA may represent an adjunct therapy in B-ALL. While the beneficial effect of EACA in the BMM was also achieved by systemic administration, intrafemoral administration of EACA directly increased fibronectin levels in the BMM. Days after transplant B- AL L ov er al l s ur vi va l ( % ) Vehicle EACA P = 0.005 Recipients0 EACA 2 Days C D Merge DAPI Fibronectin IGF1 Fibronectin + IGF1 WT Vehicle WT EACA B 0 25 30 40 50 60 70 0 50 100 E 0 EACA 10 Days 12 19 Ara-C Ara-C Days after transplant B- AL L ov er al l s ur vi va l ( % ) Vehicle Ara-C P = 0.055 Recipients EACA P = 0.031 0 22 30 40 50 60 0 50 100 Ara-C + EACA 40 μm 191 mTOR 191 Rictor 64 pAKTS473 64 pAKTT308 64 AKT 38 pCRKL 1 2 3 Vehicle 1 2 3 EACA kDa Vehicle EACA Rictor pAKTS473 pAKTT308 0 0.5 1.0 1.5 Si gn al in te ns ity (n or m al iz ed to v in cu lin ) P = 0.0006 P = 0.009 P = 0.051 38 CRKL 97 Vinculin Vehicle 3 weeks EACA Merge DAPI Fibronectin 100 μm 0 60 60 80 100 120 FI br on ec tin m ea n in te ns ity Vehicle EACA P = 0.0144 A Fig. 7 | Targeting fibrinolysis-associated pathways may extend survival of leukemicmice. A Immunofluorescence of bones of normal WTmice treated with vehicle (circles) or ε-aminocaproic acid (EACA) (squares) (1.2mg/kg daily for 3 weeks), stained for fibronectin (green) and DAPI (blue) (P = 0.0144, two-tailed t test, n = 3 mice (with 2–3 images per mouse), mean ± SD). B Kaplan–Meier-style survival curve of WT recipient mice with BCR-ABL1+ B-ALL treated with vehicle (solid line) or ε-aminocaproic acid (EACA; dotted line) (1.2mg/kg daily) starting from day 2 after transplant (P = 0.005, Log-rank test, vehicle n = 5, EACA n = 6). C Immunofluorescence of bones from WT recipient mice with BCR-ABL1+ B-ALL treated with vehicle or EACA starting from day 2 after transplant, as in (B), stained for fibronectin (green), IGF1 (purple) and DAPI (blue) (n = 3 mice pre group). D Immunoblot analysis of sorted GFP (BCR-ABL1)+ BP1+ BM cells from WT reci- pient mice with BCR-ABL1+ B-ALL treated with vehicle (circles) or EACA (squares) starting from day 2 after transplantation as in (B). The quantification on the right shows the band intensity of Rictor, pAKTS473 and pAKTT308 normalized over the loading control vinculin (two-way ANOVA, Sidak test, n = 3 mice per group, mean ± SD). E Kaplan–Meier-style survival curve of WT recipient mice with BCR- ABL1+ B-ALL treated with vehicle (black), cytarabine (ara-C; blue), EACA (purple) or a combination of ara-C and EACA (green) as from day 10 after transplantation (P = 0.031, P = 0.055, Log-rank test, vehicle n = 14, ara-C n = 16, EACA n = 15 and ara-C and EACA n = 14). The analyses in (C, D) were performed on day 20 after transplantation. Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 10 www.nature.com/naturecommunications Fibrinolysis-associated pathways may influence leukemia progression in human patients To test whether our findings in mice may be transferrable to human leukemia cell lines, we transplanted the BCR-ABL1+ B-ALL cell line SUP- B15 into NOD SCID interleukin-2 receptor γ knockout mice (NSG), whichwere treatedwith vehicle, ara-Cor the combination of ara-C plus EACA. This revealed a reduction of the human CD45+ CD19+ tumor burden in peripheral blood (Supplementary Fig. 20A), a significant increase of fibronectin levels in the BMM of ara-C plus EACA-treated mice (Supplementary Fig. 20B) and survival extension in mice treated with the combination of ara-C plus EACA compared to treatment with ara-C alone (Fig. 8A). Transplantation of primary human B-ALL cells into at least three NSG mice per donor, whereby each mouse was randomly assigned to the treatment groups vehicle, ara-C alone or ara- C plus EACA, led to significant reduction of the percentage of human CD45+ CD19+ cells in PB (Supplementary Fig. 20C, D), significant extension of survival (Fig. 8B) and significantly increased levels of fibronectin in the BMM in most xenotransplanted mice (Supplemen- tary Fig. 20E) in the double-treated cohort compared to mice treated with ara-C alone. In contrast, treatment of NSGmice transplanted with THP1 (AML) cells with ara-C plus EACA did not impact survival (Sup- plementary Fig. 20F). Hypothesizing that higher levels of the main natural inhibitor of plasmin, α2-antiplasmin (encoded by the SERPINF2 gene), in B-ALL C E Months B- AL L ov er al l s ur vi va l ( % ) low SERPINF2 expression high SERPINF2 expression P = 0.0182 D Myeloma B-ALL AML ANXA2 Plasmin/-ogen Healthy Fibronectin IGF1 25 μm 0 EACA 15/33 Days 20/37 Ara-C 21/38 R IC TO R lo g ex pr es si on Healthy bone marrow c-/Pre-B-ALL t(9;22) MLL-rearranged B-ALL MLL-rearranged AML P < 0.0001 P = 0.035 P = 0.004 P = 0.0003 A B 0 54 60 70 80 90 100 0 50 100 Days after transplant B- AL L ov er al l s ur vi va l ( % ) Vehicle Ara-C P = 0.0002 Recipients Ara-C + EACA P = 0.083 P = 0.007 0 EACA 13 Days 17 Ara-C 18 0 50 100 150 0 50 100 Days after transplant B- AL L ov er al l s ur vi va l ( % ) Vehicle Ara-C P = 0.020 Recipients Ara-C + EACA P = 0.049 SUPB15 Patient samples 0 50 100 150 0 50 100 0 5 10 15 Fig. 8 | Fibrinolysis-associated pathways may influence leukemia progression in humanpatients. AKaplan–Meier-style survival curve of NODSCID interleukin-2 receptor γ knockout (NSG) mice transplanted with SUP-B15 cells (BCR-ABL1+), treated with vehicle (black), ara-C (blue) or with a combination of ara-C and EACA (green) (P =0.0002, P =0.007, Log-rank test, vehicle n = 7, ara-C n = 7, ara-C and EACA n = 8). BKaplan–Meier-style survival of NSGmice transplanted with leukemia cells from 6 patients with B-ALL. Each sample was transplanted into at least 3 mice, whereby onemouse was treated with vehicle (black), one with ara-C (blue) and one with a combination of ara-C and EACA (green). Treatment started on day 15 or day 33 after transplantation depending on the successful detection of hCD45+ cells in theperipheral bloodof transplantedmiceafter transplantation (P =0.02,P =0.049, Log-rank test, vehicle n = 8, ara-C n = 7, ara-C + EACA n = 10). If enough B-ALL cells were available, more mice were transplanted. C Kaplan–Meier-style survival of patients with B-ALL with high (n = 32) or low (n = 39) expression of SERPINF2 (α2- antiplasmin). The curves were generated using data from the publicly available dataset TARGET (phase II), which was accessed via the cBioPortal (P =0.0182, Log- rank test). D Immunohistochemistry of bones from healthy individuals, patients withmultiplemyeloma, B-ALL and AML stained for the indicated proteins (n = 5 per condition). E Log2 expression of RICTOR in BM cells of healthy individuals (n = 73; black), patients with c-/pre-B-ALL positive for the BCR-ABL1 oncoprotein (t(9;22)) (n = 122; gray), MLL-rearranged pro-B-ALL (n = 73; yellow) or MLL-rearranged AML (n = 38; green), taken from the BloodSpot portal (MILE study) (two-way ANOVA). Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 11 www.nature.com/naturecommunications cells may influence survival in human patients, we analyzed publicly available datasets40–44. In line with the mouse experiments, high SER- PINF2 expression in leukemia cells was associated with a significant survival extension in patients with B-ALL (Fig. 8C; Supplementary Fig. 20G). In addition, SERPINF2 expression in leukemia cells was sig- nificantly lower in patients with B-ALL compared to AML (Supple- mentary Fig. 20H), but significantly higher in BCR-ABL1+ B-ALL cells versus healthy BM (Supplementary Fig. 20I). Although patient num- bers are too small to draw definitive conclusions, immunohistochem- istry of bone sections ofpatientswith B-ALL, AMLormultiplemyeloma (the latter acting as control) may tentatively support our murine data with regards to expressionofANXA2 (Fig. 8D; Supplementary Fig. 21A), plasmin/-ogen (Fig. 8D; Supplementary Fig. 21B), fibronectin and IGF1 (Fig. 8D; Supplementary Fig. 21C, D) in some B-ALL samples compared to multiple myeloma. However, fibronectin staining in B-ALL, but not AML, seemed to significantly correlate with IGF1 in the same patients (Fig. 8D; Supplementary Fig. 21E, F). Further analysis of publicly available datasets40,43,44 revealed higher expression of RICTOR in patients with (BCR-ABL+) B-ALL compared to healthy controls and AML, including AML patients with MLL rearrangements (Fig. 8E; Sup- plementary Fig. 21G). Taken together, these data suggest the concept that BMM-associated ANXA2/plasmin/IGF1-signaling may also be important for human B-ALL, but larger studies are needed to address this. Discussion Here we show that the fibrinolytic pathway including ANXA2 plays a differential role for progression of BCR-ABL1+ B-ALL versus MLL-AF9+ AML. We implicate plasmin-mediated degradation of the ECM in the BMM and IGF1 release from this ECM in the regulation of B-ALL expansion. In turn, via TNFα19,30 B-ALL cells conditionMSCandpossibly other cells of the BMM to secrete IL-6, which, in turn, activates hepa- tocytes to produce plasminogen. Whether B-ALL progression was impaired in an ANXA2-deficient BMM, characterized by a denser ECM than in WT mice, due to entrapment ofgrowth factors alone or, additionally, due tomechanical hindrance of B-ALL progression could not be completely clarified. However, we did observe decreased homing of B-ALL LIC in an ANXA2 KO BMM and more pronounced survival extension in unirradiated ANXA2 KO recipients, suggesting that mechanical or irradiation- associated factors on the ECM may be contributory, as described45. In support of this and ourwork, irradiation has been shown to reduce the stiffness of an in vitro collagen matrix46, which may lead to reduced cancer spread and prolonged survival. However, irradiation effects may be dependent on the dominant ECM protein in the respective tissue47. Albeit, in our model quantitative irradiation-induced ECM changes were ruled out. Additionally, endothelial cells were increased in healthy, unirradiated ANXA2 KO mice, suggesting that decreased angiogenesis in unirrradiated ANXA2 KO mice, as it has been described48, is less likely to be the reason for prolonged B-ALL survival in this model. While our studies were largely focused on MSC, we show that at least monocytes/macrophages14,15, endothelial cells, major producers of tPA, for instance in vascular niches of the BMMor hepatic sinusoids, and fibroblasts participate in plasminogen activation, but contribu- tions of other cell types cannot be excluded. Furthermore, consistent with the increase of plasminogen in B-ALL, uPAR, an important com- ponent of the plasminogen activation system and contributor to ECM lysis49, was also increased. Future studies are needed to address the role of this and the PAR pathways to B-ALL. In various solid cancers50,51, expression of ANXA2 has been cor- related with invasion, metastasis, resistance to treatments and decreased patient survival via regulation of plasmin-dependent degradation of the ECM52, consistent with our study. Our work extends knowledge of ECM remodeling and release of cytokines and GFs in bone metastasis of solid tumors53 to leukemia, where ECM proteins are known to influence disease progression19,54,55. In line with a previously published report on the leukemia-specific influence of the BMMon leukemic course1, we showhere that plasmin- mediated release of IGF1 specifically mediates pro-survival signaling in B-ALL cells via stimulation of themTORC2/AKT pathway, as previously shown56. In CML, this effect was less pronounced, possibly reflecting differences between BCR-ABL1+ myeloid and lymphoid leukemias or differences between CML in chronic versus more advanced stages57,58 with regards to IGF1/mTORC2 signaling. Other cytokines entrapped in the ECM18 may be contributing to the observed survival extension in CML in ANXA2 KO mice. In contrast, in MLL-AF9+ AML IGF1 activates mTORC1, and mTORC1 inhibition impairs AML progression59. Little is knownon IGF1-mediated activation ofmTORC2 signaling inAML60, but we show that IGF1R expression is diminished on MLL-AF9+ AML cells, suggesting that AML cells may be less sensitive to variation in IGF1 levels andmTORC2 signaling. Butwhether these findings are restricted toMLL-AF9+ AML or are applicable to other AML subtypes needs to be studied in the future. In PML-RARα-associatedAPMLexpression of the ANXA2/S100A10 complex causes hyperfibrinolysis due to the accumulation of plasmin on the surface of leukemia cells7. The contribution of ANXA2/S100A10 from the BMM, however, had not been investigated previously, although the in vitro migration of T-ALL cell lines or T-cell lymphoma growth25 are dependent on PLG. Lastly, our results suggest that manipulation of the levels of ECM proteins and of GF availability by EACA in addition to standard che- motherapymay efficiently reduce tumorburden or extend the survival of mice with B-ALL. Given the approved use of EACA in the hemor- rhagic setting, administration of EACA to B-ALL patients who are fre- quently thrombocytopenic and at riskof bleeding,would likely be safe. In conclusion, our data demonstrate that ANXA2 and the asso- ciated fibrinolytic pathway have highly differential roles for progres- sion of BCR-ABL1+ B-ALL and CML versus MLL-AF9+ AML. Adjunct treatments targeting plasmin activation, to be tested in clinical trials, raise hopes for consistently successful treatment of B-ALL in future. Methods All research complied with relevant ethical regulations and approved by the local government (Regierungspräsidium Darmstadt and the Landesuntersuchungsamt Rheinland-Pfalz, Germany) for murine stu- dies and the ethics committees of the respective universities for human studies. All authors compliedwith the guidelines on inclusion and ethics in global research. Mice 5–6-week- or 8–10-week-old C57/BL6 mice were purchased from Charles River Laboratories (Sulzfeld, Germany) and were used as wildtype (WT) donors orWT recipients, respectively, in all transplants. 8–10-week-old C57/BL6 mice were also used for the isolation of pri- mary mesenchymal stromal cells (MSC), fibroblasts andmacrophages. ANXA2 knockout (KO) (C57/BL6 background) mice16 were a kind gift fromProf. KatherineHajjar and bred in the animal facility of theGeorg- Speyer-Haus, Institute for Tumor Biology and Experimental Therapy. They were used at an age of 8–10 weeks. 5–6-week-old tissue plasmi- nogen activator (tPA) KO recipients (C57/BL6 and SJL backgroundmix) were purchased from Molecular Innovation (Novi, MI, USA). The respective control mice (B6SJLF1/J) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and used at an age of 5–6 weeks as donors and at anageof 8–10weeks as recipients. NODSCID interleukin (IL)-2 receptor γ knockout (NSG)mice were bred in the inhouse animal facility. Animals of different sexes and ages were randomly assigned to experimental groups. All mouse strains had been backcrossed for at least 7 generations. All animal studies were approved by the local Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 12 www.nature.com/naturecommunications government (Regierungspräsidium Darmstadt and the Land- esuntersuchungsamt Rheinland-Pfalz, Germany). The maximally allowed tumor burden of 80% (oncogene+ Gr1+, oncogene+ CD11b+ or oncogene+ BP1+ cells) was not exceeded. Mice were sacrificed according to the criteria stated in the animal studies approved by the local government. The mice were housed in individually ventilated cages at an ambient temperaturebetween 20–26degreesCelsius. A 14- h light/10 h dark cycle was used. Human samples Studies on cryopreserved unsorted cells from bone marrow or per- ipheral blood from patients with B-ALL were approved by the local ethics committee bylaws in the hematology departments of the Hos- pital Lyon Sud, Pierre Bénite and Centre Leon Bérard, Lyon, France. The use of bone sections from patients with B-ALL, AML and multiple myeloma was approved by the Ethics Committee of the University Clinic of theGoetheUniversity Frankfurt (Approval number 274/18 and SHN-5-2020). The age or gender of the patients, as well as the genetic subtype of the leukemia, were unknown, consistent with the ethics committee’s approval. Each patient had signed an informed consent. No compensation was provided to the patients for the inclusion of their samples in this study. Cell lines The human cell line 293T (ACC635) and the mouse cell line NIH/3T3 (ACC59) were purchased from the German Collection of Micro- organisms and Cell Cultures (DSMZ) and cultured in DMEM, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine. The medium for 293T cells was further supplemented with 1% non-essential amino acids. The murine cell line H5V (a kind gift from Prof. Stefanie Dimmeler) was cultured in DMEM, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine. The mouse cell line MS5 (ACC441) was purchased from the DSMZ andwas cultured inα-MEMwith 10% FBS, 1% penicillin/streptomycin and 1% L-glutamine. The murine pro-B cell line BA/F3 (ACC300) was purchased from the DSMZ and grown in RPMI, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine sup- plemented with 5% (v/v) WEHI medium as a source of interleukin 3 (IL-3). Cells were transduced on two consecutive days with cryo- preserved MSCV-IRES-GFP (empty vector), MSCV-IRES GFP BCR- ABL1- (as in vitro model for murine B-ALL) or MSCV-IRES GFP MLL- AF9- (as in vitro model for murine AML) expressing retroviri. The human leukemic cell lines K562 (BCR-ABL1+; myeloid, model of CML) (ACC10), NALM6 (lymphoid, model of BCR-ABL1-negative B- ALL) (ACC128), THP1 (MLL-AF9+; myeloid, model of AML) (ACC16), SUPB15 (lymphoid, model of BCR-ABL1-positive B-ALL) (ACC389) and Kasumi (AML1-ETO+; myeloid, model of AML) (ACC220) were purchased from the DSMZ and cultured in RPMI, 10% FBS, 1% penicillin/streptomycin and 1% L-glutamine. These cell lines were used to test the activation of mTORC2 in response to stimulation with IGF1, dependent on oncogene or lymphoid versus myeloid lineage. The human hepatocellular carcinoma cell line, Huh7, was a kind gift from Prof. Ivan Dikic and was cultured in DMEM, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine. The Huh7 cell line was used as in vitro model to test the production of PLG by hepatic cells in response to different conditions. Primary mouse BCR-ABL1+ B-ALL cells were cultured in RPMI, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine supplemented with IL-7 (10 ng/ml). Primary murine MSC were grown in MEM, 20% FBS, 1% penicillin/streptomycin and 1% L-glutamine. 100 μg/ml primocin (InvivoGen, Toulouse, France) was added to the culture medium until passage two. All cell lines were maintained in a 37 °C, 5% CO2 incubator. All cell lines were routinely tested for mycoplasma contamination and tested negative. Microarray Acute myeloid leukemia (AML) and chronic myeloid leukemia (CML) were induced in Col2.3 GFP reporter mice. On day 45 for AML and day 19 for CML, mice were sacrificed, and long bones were collected and crushed. Bones were further digested in 1.8mg/ml collagenase type I (Worthington, Biochemical Corporation, Lakewood,NJ) solution. Total RNA from 40,000–50,000 sorted GFP+ Ter119-PE- CD45-PE- osteo- blastic cells was extracted using RNeasy Plus Micro Kit (Qiagen, Ger- mantown, MD) according to the manufacturer´s protocol. RNA was amplified using Ovation Pico WTA reagent (NuGEN, Redwood City, CA). Amplified cRNAwas labeledwith the GeneChipwild-type terminal labeling kit (Affymetrix, Santa Clara, CA), hybridized toMouseGene ST 2.2 SST microarrays (Affymetrix, Santa Clara, CA), and scanned by GeneChip Scanner 3000 7G system (Affymetrix, Santa Clara, CA) according to standard protocols. The primary microarray data was normalized and analyzed using R with limma and lumi packages61,62. False discovery rate (FDR, cut off analysis <0.1) was used for multiple test correction. Retrovirus production To generate MSCV IRES GFP- (empty vector), MSCV IRES GFP BCR- ABL1- orMSCV IRESGFPMLL-AF9-expressing retroviri, 293 T cells were co-transfected with the respective plasmids (10 μg/3 × 106 cells) and ecopak plasmid (5 μg/3 × 106 cells). 48 h after transfection the condi- tioned medium containing the retroviri was harvested, and viral titers were tested via the transduction of NIH/3T3 cells. Bone marrow transduction/transplantation To induce CML, AML or the MSCV empty vector control condition, donor BM cells were harvested from 5-fluorouracil (5-FU)-pretreated mice (200mg/kg), which had received 5-FU 4 days prior to harvest. Subsequently, BM cells were prestimulated overnight with stem cell factor (Peprotech, Hamburg, Germany) (50ng/ml), interleukin (IL)-3 (Peprotech, Hamburg, Germany) (6 ng/ml) and IL-6 (Peprotech, Ham- burg, Germany) (10 ng/ml)1. Cells were subsequently transduced on two consecutive days with cryopreserved MSCV IRES GFP BCR-ABL1 (CML)- or MLL-AF9 (AML)-expressing retrovirus or MSCV IRES GFP empty vector and intravenously (i.v.) transplanted (2.5 × 105 cells/mouse for CML and 5 × 105 cells/mouse for AML) into sublethally irradiated (2 × 450 cGy) mice. For induction of B-cell acute lymphoblastic leukemia (B- ALL) BM cells were harvested from non-5-FU-treated mice, transduced once with MSCV IRES GFP BCR-ABL1-expressing retrovirus and trans- planted i.v. (1 × 106 cells/mouse) into sublethally irradiated (2 × 450cGy) or into non-irradiated (2 × 106 cells/mouse) mice. For the secondary transplants cells were harvested from the bones or spleens of mice with CML (sacrificed on day 14) or B-ALL (sacrificed on day 20). PB analysis prior to sacrifice confirmed full establishment of the disease. In the CML and B-ALL secondary trans- plantation models 3 × 106 BM or spleen cells were intravenously transplanted into each sublethally irradiated (2 × 450 cGy) WT sec- ondary recipient. For the xenotransplantation experiments, 1 × 106 SUPB15 cells (human BCR-ABL1+ B-ALL cell line) were injected into non-irradiated NSG mice, or (1.5–2) × 106 cells from the BM or PB of six different patients with B-ALL were intravenously transplanted into at least 3 sublethally irradiated (200 cGy) NSGmice. Individualmice fromeach group were randomly assigned to treatment with vehicle, cytarabine (ara-C) (Sigma-Aldrich, Darmstadt, Germany) or the combination of ara-C and ε-aminocaproic acid (EACA) (Sigma-Aldrich, Darmstadt, Germany) (for treatment details see below). In vivo drug treatment In the B-ALL rescue experiment ANXA2-deficientmicewere treated once a week with intraperitoneal (i.p.) injection of vehicle (PBS) or plasmin (Sigma-Aldrich, Darmstadt, Germany) (1mg/kg) starting 48 hrs after Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 13 www.nature.com/naturecommunications transplantation. For the B-ALL treatment experiments vehicle (PBS) or EACA (Sigma-Aldrich, Darmstadt, Germany) (1.2mg/kg) were injected subcutaneously (s.c.) on a daily basis, starting either on day 2 or 10 after transplantation (as indicated in the legends). Ara-C (0.5mg/kg) was administered i.p. on days 12 and 19 after transplantation. Vehicle (oil) or linsitinib (inhibitor of insulin-like growth factor receptor 1) (MedChem- Express, Sollentuna, Sweden) were administered by oral gavage as a daily regimen (75mg/kg starting from day 9-11 and 25mg/kg starting from day 12). In intrafemoral (i.f.) experiments mice with B-ALL were treated with EACA (Sigma-Aldrich, Darmstadt, Germany) (1.2mg/kg) by intrafemoral (i.f.) administration on day 15 after transplantation or tPA (Sigma-Aldrich, Darmstadt, Germany) (10mg/kg) by i.f. administration on day 19. 24h after treatment mice were sacrificed. In the xeno- transplantation experiments NSG mice were subcutaneously (s.c.) trea- ted with EACA (1.2mg/kg) or vehicle (PBS) daily starting on day 13 after transplantation for experiments involving SUPB15 cells and on day 33 after transplantation when human primary B-ALL cells were used. Ara-C was administered i.p. (75mg/kg) on days 17 and 18 after transplantation (SUPB15 experiment) and on days 20/21 or 37/38 after transplantation (human primary B-ALL cells), respectively. Homing experiment To assess homing in the B-ALL model, BM cells were harvested from non-5-FU-treatedmice and transduced oncewithMSCV IRES GFP BCR- ABL1-expressing retrovirus. After overnight culture in RPMI, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine supplemented with IL-7 (Peprotech, Hamburg, Germany) (10 ng/ml), 5 × 106 cells/mouse were intravenously transplanted into sublethally irradiated (2 × 450cGy)mice. 18 h after transplantation BM and spleen cells from recipient mice were analyzed for the percentage of GFP (BCR-ABL1)+ BP1+ cells. A list of all antibodies can be found in Supple- mentary Table 1. Analysis of diseased mice and tumor burden Disease progression in leukemic mice was assessed by analysis of peripheral blood using a Scil Vet animal blood counter (Scil Animal Care Company, Viernheim, Germany). In parallel, flow cytometry (BD Fortessa, Heidelberg, Germany) was used to determine the tumor burden represented by the percentage of GFP (BCR-ABL1)+ leukocytes positive for allophycocyanin (APC)-conjugated CD11b antibody (CML) or phycoerythrin (PE)-conjugated BP-1 antibody (B-ALL) and the per- centage of GFP (MLL-AF9)+ leukocytes positive for allophycocyanin- Cy7 (APC-Cy7)-conjugated Gr1 antibody (AML). PE-conjugated human CD45 and APC-conjugated CD19 antibodies were used to assess the tumor burden in xenotransplanted NSG mice. A list of all antibodies can be found in Supplementary Table 1. Flow cytometric analyses were performed using the FlowJo software (version 10). Gating strategies can be found in Supplementary Fig. 22. Progenitor analysis in normal and leukemic mice Bones, spleen andperipheral bloodwereharvested fromnormalWTor ANXA2-deficient mice or frommice transplanted with BM transduced with BCR-ABL1- (to induce CML) or empty vector-expressing retro- virus. Bones and spleens were crushed in PBS containing 2% FBS. Subsequently, cells were stained for the lineagemarkers CD11b, Ter119, CD5, B220 and F4/80 to identify Lin− cells. Multipotent progenitors (MPP) were identified as Lin− Thy1.1− Sca1+ c-Kit+ Flk2+ and common lymphoid progenitors (CLP) were identified as Lin− IL7Rα+ Thy1.1− Sca- 1lo c-Kitlo. A list of all antibodies can be found in Supplementary Table 1. Colony-formation assay 2 × 104 total BM or 2 × 105 spleen cells from individual mice with CML (sacrificed on day 15 after transplantation) were plated in triplicate in cytokine-free methylcellulose medium (M3234, Stemcell Technolo- gies, Cologne, Germany) supplemented with 100pg/ml IL-3 (Peprotech, Hamburg, Germany). For the B-ALL model, 2 × 105 total BM or 2 × 105 spleen cells from individual mice with B-ALL (sacrificed on day 20 after transplantation) were plated in triplicate in methyl- cellulose medium for pre-B lymphoid progenitor cells (M3630, STEMCELL Technologies, Cologne, Germany). In all conditions colonies were scored after 7 days. Immunohistochemistry Bones were collected from diseased mice and fixed in 10% formalin overnight at room temperature. After decalcification for 1–2 weeks in 0.5M EDTA bones were mounted in paraffin and sectioned. Immuno- histochemistry of bone sections frommice or humans was performed using antibodies for fibronectin, ANXA2, plasmin/-ogen and insulin- like growth factor (IGF1) according to standard procedures. A list of all antibodies can be found in Supplementary Table 1. Staining was detected by immunoperoxidase using a yellow-brown horseradish- peroxidase chromogen. Isolation of GFP+ stroma cells from the BMM To isolate stroma cells from the bone marrow microenvironment (BMM), femora, tibiae, pelvis, spine and humeri were harvested from nestin-GFP (GFP+ mesenchymal stromal cells), Col2.3-GFP (GFP+ osteoblastic cells) or Tie2-GFP (GFP+ endothelial cells) mice. Bones were crushed in PBS containing 2% FBS and stained with lineage mar- kers (see above). Lin+ cells were depleted with streptavidin beads (Miltenyi Biotec, Bergisch Gladbach, Germany). The Lin− fraction was further stained with antibodies to CD31, except when sorting for endothelial cells, as well as CD45 and Ter119. A list of all antibodies can be found in Supplementary Table 1. Cells positive for GFP, but negative for the markers above were sorted directly into RLT plus buffer (Qia- gen, Düsseldorf, Germany) or lysed with RIPA buffer (50mM Tris HCl pH 7.4, 150mM NaCl, 1% Triton X-100, 1% Na DOC, 0.1% SDS, 1mM EDTA). Gating strategies can be found in Supplementary Fig. 22. Isolation of primary stromal cells Longbones fromWTorANXA2-deficientmicewereharvested, cut into small pieces and digested for 1 h at 37 °C in collagenase. Subsequently, CD45− Ter119- PDGFRα+ Sca1+ MSC were sorted (BD FACSAria Fusion Cell Sorter, Franklin Lakes, NJ, USA) and cultured until passage six63. A list of all antibodies can be found in Supplementary Table 1. Gating strategies can be found in Supplementary Fig. 22. For the isolation of primary macrophages (MΦ), long bones from WT or ANXA2-deficient mice were harvested and crushed in phosphate-buffered saline (PBS) using mortar and pestle. After 1week non-adherent cells were removed and adherent cells were cultured in αMEM medium supplemented with 20% fetal bovine serum, 1% peni- cillin/streptomycin and 1% L-glutamine. The immunophenotype was confirmed using F4/80+ CD169+ to identify macrophages. For the isolation of fibroblasts, long bones from WT or ANXA2- deficient mice were harvested. After flushing, bones were cut into pieces, distributedon theplate and cultured in enoughDMEMmedium supplemented with 10% FBS, 1% penicillin/streptomycin and 1% L-glutamine to cover the pieces. After 1–2 weeks, when enough adherent cells were spreading out of the pieces of bones, the pieces were removed. The immunophenotypewas confirmedbypositivity for αSMA and vimentin to identify fibroblasts. Primarymacrophages andprimaryfibroblastswere used to repeat someof the experiments performedwithprimaryMSC to test, whether the role of ANXA2 in the BMM is specific to MSC. For the isolation of osteoblasts long bones from WT mice were harvested, cut into small pieces and digested for 1 h at 37 °C in collagenase. Subsequently, CD4− CD8− Ter119− CD11b− Gr1− CD31− Sca1− CD51+ osteoblastic cells were sorted (BD FACSAria III Cell Sorter, Franklin Lakes, NJ, USA), and RNA isolation was performed. A list of all antibodies can be found in Sup- plementary Table 1. Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 14 www.nature.com/naturecommunications For the isolation of endothelial cells long bones from WT mice were harvested, cut into small pieces and digested for 1 h at 37 °C in collagenase, as above. Subsequently, depletion of CD45+ cells was performed using CD45 MicroBeads (Miltenyi Biotec, Bergisch Glad- bach, Germany). Endothelial cells were then enriched by positive selection of CD31+ cells using CD31 MicroBeads (Miltenyi Biotec, Ber- gisch Gladbach, Germany). Cells were resuspended in TRIzol reagent (ThermoFisher, Darmstadt, Germany) to isolate RNA. These cells were used to test the levels of ANXA2 in different stromal cell populations. Differentiation and colony-forming unit fibroblast (CFU-F) assays of MSC 1 × 104 WT or ANXA2-deficient MSC were plated in 48-well plates. Cells were differentiated into adipocytes using the adipocyte differentiation and maintenance medium (LONZA, Cologne, Germany). After 10-14 days cells were stained with oil red O to identify adipocytes. To assess differentiation into osteoblasts, cells were cultured in αMEM con- taining 20% FBS, 1% penicillin/streptomycin and 1% L-glutamine sup- plemented with 50μg/ml ascorbic acid, 10mM β-glycerophosphate and 10mM dexamethasone. After 10–14 days cells were stained with alizarin red S solution to assess osteoblast differentiation. For the CFU-F assay 2 × 103 WT or ANXA2-deficient MSC were plated in 10 cm plates. Cells were cultured for 2 weeks in αMEM, 20% FBS, 1% penicillin/streptomycin and 1% L-glutamine. Twice a week half the volume of themediumwas carefully removed and substitutedwith fresh medium. After 2 weeks cells were fixed with 4% paraformalde- hyde (PFA) for 15min at room temperature (RT) and stained with 1% crystal violet solution (in 20% methanol) at room temperature for 30min. The colonies per plate were counted. RNAseq Primary MSC was sorted from WT or ANXA2-deficient mice and col- lected in TRIzol reagent (ThermoFisher, Darmstadt, Germany). Chloroform (1/5th of TRIzol volume) was added to allow RNA extrac- tion, and the RNA was further purified using an mRNA purification kit (Qiagen, Düsseldorf, Germany) following themanufacturer’s protocol. cDNA was generated and amplified using 8.7 ng of total RNA and the SMART-Seq® v4 Ultra® Low Input RNA Kit for Sequencing (Takara) following the manufacturer’s protocol. Then cDNA was sheared with the Covaris ultrasonicator followed by preparation of the sequencing library using NEBNext End Repair module, dA-Tailing module and Quick Ligation module (New England BioLabs). The final libraries were quality controlled by Agilent 4200 TapeStation System (Agilent Technologies) and Qubit ds DNA HS Assay kit (Life Technologies-Invitrogen). Based onQubit quantification and sizing analysis, sequencing libraries were normalized, pooled and clustered on the cBot (Illumina) with a final concentration of 250 pM (spiked with 1% PhiX control v3). 100bp paired-end sequencing was performed on the Illumina HiSeq 4000 instrument using standard Illumina protocols. Adapter sequences were trimmed from paired end RNA-seq reads using TrimGalore (https://github.com/FelixKrueger/TrimGalore). Quality-filtered reads were mapped using STAR64 against the mouse genome (GRCm39). Read counts were quantified with HTSeq65, and differential expression analysiswas carried outwith these counts using Bioconductor package DESeq266. Volcano plots were plotted using an R package EnhancedVolcano (https://github.com/kevinblighe/ EnhancedVolcano). Co-immunoprecipitation MSCwas lysed in RIPA buffer (50mMTrisHCl pH 7.4, 150mMNaCl, 1% Triton X-100, 1% Na DOC, 0.1% SDS, 1mM EDTA). Equal amounts of protein per sample were incubated with 2 μg of antibody against S100A10 (Invitrogen, Darmstadt, Germany) for 3 h at 4 °C followed by overnight incubation with magnetic beads (Dynabeads Protein G, ThermoFisher Scientific, Darmstadt, Germany) at 4 °C. After several washes with ice cold RIPA buffer the protein-bead complex was resuspended in Laemmli elution buffer and heated to 95 °C for 5min. The eluted proteins were run on an SDS page gel, as described below. Immunoblotting Culturedmouse or human cells were stimulated withmouse or human recombinant insulin-like growth factor (IGF) 1 (12 ng/ml), respectively, for 5min. Cultured cells or primary BMor spleen cells were lysed using RIPA buffer (50mMTris HCl pH 7.4, 150mMNaCl, 1% Triton X-100, 1% Na DOC, 0.1% SDS, 1mM EDTA), freshly supplemented with protease and phosphatase inhibitor cocktails (Sigma-Aldrich, Darmstadt, Ger- many). Lysates were kept on ice for 1 h and centrifuged at 15,000g at 4 °C for 20min. Protein concentrations were measured using the Bradford protein assay (Bio-Rad, Hercules, CA, USA). Equal amounts of protein were mixed with 4x Laemmli buffer, supplemented with β- mercaptoethanol, denatured and run on either 4-12% Bis-Tris poly- acrylamide gels (ThermoFisher Scientific, Darmstadt, Germany) or custom-made gels containing 10% acrylamide. Proteins were blotted on PVDF membranes (ThermoFisher Scientific, Darmstadt, Germany). Subsequently, the membranes were probed overnight at 4 °C in Tris- buffered saline, 0.1% Tween 20 (TBST) with antibodies to pAKTS473, pAKTT308, pCRKL, and pS6K1 (1:1000), in 2.5% BSA with antibodies to mTOR, Rictor and Raptor (1:1000) or in 5% milk with antibodies to ANXA2, S100A10, CRKL, AKT and S6K1 (1:1000), orGAPDH,β-actin and vinculin (1:2000). A list of all antibodies can be found in Supplemen- tary Table 1. After incubation with secondary horseradish-peroxidase (HRP)-conjugated antibodies, membranes were developed using X-ray films (Fujifilm, Düsseldorf, Germany). Band intensities were quantified using Image J software. Plasmin and MMP-9 activation assay To test plasmin activation, cells, which had been plated at 5 × 103 cells/well (96-well plate), were pre-incubated with plasminogen (PLG) (Sigma-Aldrich, Darmstadt, Germany) (170 nM) for 1 h. Then tissue plasminogen activator (tPA) (Sigma-Aldrich, Darmstadt, Ger- many) (12 nM) was added to the medium covering WT or ANXA2 KO MSC. After 5min of incubation at 37 °C plasmin substrate (D-Val-Leu- Lys-4-nitroanilid-dihydrochlorid) (Sigma-Aldrich, Darmstadt, Ger- many) was added to the culture medium, and after 10min the pro- duction of the cleaved colorogenic substrate, indicative of active plasmin, was measured at 405 nm. To test plasmin-dependent acti- vation of matrix metalloproteinase (MMP)-9, a 24-h-pretreatment of MSC with tumor necrosis factor (TNF)α (Peprotech, Hamburg, Ger- many) (15 ng/ml) in 1% FBS medium was followed by addition of tis- sue plasminogen activator (tPA) (Sigma-Aldrich, Darmstadt, Germany) and plasminogen (PLG) (Sigma-Aldrich, Darmstadt, Ger- many). To assess MMP-9 activity, 100 μl of conditioned medium was mixed with 100 μl of 2x assay buffer (400 nMNaCl, 100mMTris-HCL pH7.6, 10mM CaCl2, 40 μM ZnSO4, 0.1% Brij 35) and incubated for 1 h at 37 °C with MMP-9 fluorogenic substrate (10 μM) (Calbiochem, Darmstadt, Germany). Cleavage of the substrate was measured at 365 nm19. Only the MMP-9-activation assay was performed in 1% FBS. All the other assays were performed in the absence of serum. CRISPR The sgRNAs were designed to target the sequence of Anxa2 using the Benchling design tool (https://benchling.com/crispr) (Supplementary Table 3). The primers were cloned into lentiCRISPR.v2 (Addgene plas- mid #52961; http://n2t.net/addgene:52961; RRID:Addgene_52961)67 using BsmBI sites. A non‑targeting (NTC) sgRNAwas used as control and cloned into both vectors. Integration was confirmed by sequencing using the LKO.1 primer shown in Supplementary Table 3. The ANXA2KO cells were generated by transducingMS5,H5V and NIH/3T3 cellswith lentivirus expressing the sgRNA. After transduction, Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 15 https://github.com/FelixKrueger/TrimGalore https://github.com/kevinblighe/EnhancedVolcano https://github.com/kevinblighe/EnhancedVolcano https://benchling.com/crispr http://n2t.net/addgene:52961 www.nature.com/naturecommunications cells were cultured in 10 μg/ml puromycin to select for transduced cells. The KO efficiency was tested by immunoblot analysis. Immunofluorescence To test fibronectin or laminin degradation, 1 × 104 WT or ANXA2- deficient MSC were cultured in a 48-well plate for one week to allow deposition of fibronectin or laminin. After addition of tPA (Sigma- Aldrich, Darmstadt, Germany) (0.7 ng/ml) and PLG (Sigma-Aldrich, Darmstadt, Germany) (1 ng/ml) to the medium and culture for 6 h at 37 °C, cells were fixed in 4% PFA (Santa Cruz, Heidelberg, Germany) for 10min at roomtemperature (RT).After 30minofblockingwith 2%BSA at RT samples were incubated overnight with a primary antibody to fibronectin or laminin (Abcam, Cambridge, UK) (1:200 in 2% BSA). The following day samples were incubated with fluorophore-labeled sec- ondary antibody (Alexa Fluor 647, Invitrogen, Darmstadt, Germany) for 1 h at RT, and nuclei were counterstained with 5 μg/ml 4′,6-diami- din-2-phenylindol (DAPI). Images were acquired using the confocal quantitative image cytometer CQ1 (Yokogawa, Germany), and fibro- nectin staining was quantified using Image J software. Immunofluorescence of bone sections Bones were harvested from normal or diseasedmice and fixed with 4% PFA overnight at 4 °C. Following decalcification in 0.5M EDTA for 1–2- week bones were rehydrated in 2% polyvinylpyrrolidone (PVP) (Sigma- Aldrich, Darmstadt, Germany) overnight at 4 °C, frozen in Tissue-Tek optimum cutting temperature O.C.T. (Sakura, Alphen aan den Rijin, The Netherlands) and stored at −80 °C. Sections were 10μm thick. After thawing slides were blocked for 30min with 2% BSA and incu- bated overnight at 4 °C with primary antibodies to fibronectin and IGF1 (1:200 in 2%BSA) (Abcam,Cambridge,UK) or laminin (1:200 in 2%BSA) (Abcam,Cambridge, UK). The followingday slideswere incubatedwith the secondary antibodies Alexa Fluor 647 or 488 (Invitrogen, Darm- stadt, Germany) at a 1:400 dilution for 1 h at room temperature. Nuclei were counterstained with DAPI (5 μg/ml). Images were acquired using the confocal quantitative image cytometer CQ1 (Yokogawa, Germany) and staining was quantified using Image J software. Invasion assay 1 × 104 WT or ANXA2-deficient MSC, respectively, were mixed in a solution of 0.5mg/ml matrigel (Corning, Wiesbaden, Germany) ± PLG (Sigma-Aldrich, Darmstadt, Germany) (1 ng/ml) and plated on top of a transwell (3μm pore size, Corning, Wiesbaden, Germany). Upon matrigel polymerization 2.5 × 105 BCR-ABL1+ BA/F3 cells were plated on top of the matrigel in RPMI containing 2% FBS. RPMI containing 20% FBS was added to the lower chamber. After 18 h cells, which had invaded into the lower chamber, were counted. Plasmid for ANXA2 overexpression The Anxa2 cDNA template was amplified from NIH/3T3 cells by PCR and cloned as GST-fusion into the bacterial expression vector pGEX- 6P1. The insert was then recloned into amammalian expression vector similar to the commercial vector pEGFP-C1 (Clontech) with the sole difference that the EGFP sequence was replaced by the HA-tag sequence. The resulting vector was named ‘pHA-Annexin A2, mouse’. Transient ANXA2 overexpression ANXA2-deficient MSC was plated in a 6-well plate and transduced with the ANXA2-expressing plasmid (1.5 μg) using lipofectamine 2000 (ThermoFisher, Darmstadt, Germany) according to themanufacturer’s protocol. 48 h later the transduction efficiency was tested by qRT-PCR and immunoblot, and cells were used in the invasion assay. Quantitative real-time PCR Samples were collected in TRIzol reagent (ThermoFisher, Darmstadt, Germany). Chloroform (1/5th of TRIzol volume) was added to allow RNA extraction, and the RNA was further purified using an mRNA purification kit (Qiagen, Düsseldorf, Germany) following the manu- facturer’s protocol. Equal amounts of RNA,measuredwith aNanoDrop (ThermoFisher, Darmstadt, Germany), were reverse transcribed into cDNA using the ProtoScript First Strand cDNA Synthesis kit (New England Biolabs, Ipswich, MA). Quantitative RT-PCR was performed using SYBR green (ThermoFisher, Darmstadt, Germany). A list of pri- mers can be found in Supplementary Table 2. Enzyme-linked immunosorbent assay (ELISA) BM supernatant was collected by flushing a femur from each mouse with 300 μl of PBS. Liver and spleen supernatants were collected from eachmouse by crushing the same amount (in g) of liver or spleen with 300 μl of PBS. The concentration of plasminogen, tissue plasminogen activator (tPA), plasminogen activator inhibitor- 1 (PAI-1), urokinase plasminogen activator surface receptor (uPAR), D-dimers, α2-anti- plasmin, α2-macroglobulin and IL-6 in each sample was determined using themouseplasminogen (Abcam,Cambridge, UK),mouse/rat tPA (Abcam, Cambridge, UK), the mouse PAI-1 (Invitrogen, Darmstadt, Germany), mouse uPAR (Abcam,Cambridge, UK), D-dimer (LSBio,MA, USA), α2-antiplasmin (LSBio, MA, USA), mouse α2-macroglobulin (Novus Biologicals, MN, USA) and mouse IL-6 (Invitrogen, Darmstadt, Germany) ELISA kits, respectively, following the manufacturer’s pro- tocol. The supernatant of Huh7 cells cultured in the conditioned medium of human MSC exposed to vehicle, TNFα (Peprotech, Ham- burg, Germany) (15 ng/ml) or SUPB15 cellswere collected after 6 h. The concentration of plasminogen in each sample was determined using the human plasminogen (Abcam, Cambridge, UK) ELISA kit following the manufacturer’s protocol. We cannot exclude that the employed antibody to plasminogen may also be detecting plasmin. Chromatin immunoprecipitation (ChIP) assay Huh7 cells treated with vehicle (water) or human recombinant IL-6 (Peprotech, Hamburg, Germany) (20 ng/ml) for 3 h, were treated with formaldehyde (0.75%). 1M glycine was used to stop the reaction. Cells were collected in ChIP lysis buffer (50mMHEPES KOH, 140mM NaCl, 1% TritonX-100, 0.5mM EDTA pH8, 0.1% sodiumdeoxycholate, 0.1% SDS) and incubated for 1 h at 4°C followed by sonication to fragment the DNA. After removal of cellular debris (by centrifugation) the supernatant was snap-frozen in liquid nitrogen and stored at −80 °C. Antibodies to cAMP Responsive Element Binding Protein 1 (CREB1) or IgG control antibodies, together with protein A beads, previously blocked with RIPA buffer containing salmon sperm DNA (Sigma Aldrich, Darmstadt, Germany) (100 ng/μl) to prevent non-specific binding, were used for overnight immunoprecipitation at 4 °C. Several washes were then followed by DNA elution and incubation with pro- teinase K. The DNA was purified using the ChIP DNA Clean and Con- centrator kit (ZymoResearch, Irvine, CA, USA) and eluted in 60 μl of elution buffer. qPCR analysis was performed using 2 μl of chromatin (DNA)with twodifferent primer sets specific for thedistal regionof the plasminogen (PLG) promoter containing the CREB1 binding site. Pri- mers were designed to recognize the CREB1 binding site on the pro- moter region of plasminogen using the Integrative genomics viewer (IGV) (https://software.broadinstitute.org/software/igv/) and are listed in Supplementary Table 4. The DNA recovery was normalized over the input (as percent of the input)19. Proliferation assay 2.5 × 105 empty vector+, BCR-ABL1+ or MLL-AF9+ BA/F3 cells were cul- tured in RPMI containing 5% FBS with or without murine recombinant IGF1 (BioLegend, San Diego, CA, USA) (12 ng/ml), and after 48 h cell numbers were counted. Where indicated, KU-0063794 (Selleckchem, Munich, Germany) (5 μM) was added to the culture medium. In the IGF1 release experiment 5 × 103WTor ANXA2-deficientMSC were embedded in a matrigel solution (1.5mg/ml) with or without PLG Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 16 https://software.broadinstitute.org/software/igv/ www.nature.com/naturecommunications (Sigma-Aldrich, Darmstadt, Germany) (1 ng/ml), EACA (Sigma-Aldrich, Darmstadt, Germany) (12.5mM), ANXA2 antibody (Santa Cruz, Hei- delberg, Germany) (60 μg/ml), ANXA2/S100A10 heterotetramer inhi- bitor (A2ti) (MedChemExpress, Sollentuna, Sweden) (50 μM) or aprotinin (Sigma-Aldrich, Darmstadt, Germany) (60 nM) and plated in a 48-well plate. IGF1 (12 ng/ml) was either added to the solution or the matrigel. After polymerization of the matrigel (4–8) × 104 primary GFP (BCR-ABL1)+ BP1+ cells sorted from the BM of mice with BCR-ABL1- induced B-ALL were added to each condition. Cell were counted 48 h after plating. To analyze apoptosis, primary BM cells were transduced with MSCV IRES GFP BCR-ABL1-expressing retrovirus, cultured overnight in RPMI, 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and 1% L-glutamine supplemented with IL-7 (Peprotech, Hamburg, Germany) (10 ng/ml) and, finally, added to WT or ANXA2-deficient MSC embedded in a matrigel solution with PLG (Sigma-Aldrich, Darmstadt, Germany) (1 ng/ml) and IGF1 (12 ng/ml). Apoptosis was assessed by FACS after labelling the cells with Annexin 5 (BioLegend, San Diego, CA, USA) and 7-aminoactinomycin D (7-AAD) (Thermo- Fisher, Darmstadt, Germany). Annexin5+ 7-AAD+ cells were con- sidered as late apoptotic cells. Gating strategies can be found in Supplementary Fig. 22. Whitlock-Witte limiting dilution assay Primary BM cells from non-5-FU treated mice were transduced once with BCR-ABL1- expressing retrovirus and plated in triplicate in 24-well plates (on top of 1 × 104 WT or ANXA2-deficient MSC plated the night before) at the following dilutions: 1 × 102, 3 × 102, 1 × 103, 3 × 103, 1 × 104, 3 × 104, 1 × 105 and 3 × 105/well. Cells were cultured in RPMI containing 10% FBS, 200 μM L-glutamine, 50 μM β-mercaptoethanol and 1% penicillin/streptomycin (1ml/well). Twice a week half of the medium was replaced with freshmedium. Cells were counted starting from day 5. Confluence was considered at a leukemia cell count of 106/well31. Statistics and reproducibility All statistical analyseswereperformedusingGraphPadsoftware (Prism 8.0). Between three to six independent biological replicates were included in each experiment (except in Supplementary Fig. 10D). No statistical method was used to predetermine sample size. Data were excluded from the analysis only when they were defined as outlier by GraphPad software (Prism 8.0). The experiments were not rando- mized. The investigators were not blinded to allocation during experiments and outcome assessment. Shapiro–Wilks normality test was applied to all experiments to test, if the data followed a normal distribution. If the data were nor- mally distributed, a two-tailed t-test was used, when the means of two groups were being compared. Whenmultiple hypotheses were tested, one-way or two-way ANOVA and Tukey or Sidak tests (the latter for multiple comparisons) were used as post-hoc tests. When the data did not follow a normal distribution, nonparametric tests were used. Two variableswere comparedusing theunpaired, two-tailedMann-Whitney test and more than two variables were compared by applying the Kruskal-Wallis test. Survival curves were analyzed by Kaplan-Meier-style curves and Log-rank (Mantel-Cox) or Gehan-Breslow-Wilcoxon tests. In the Kaplan-Meier-style survival curve of patients with B-ALL, patients were stratified into high or low groups depending on whether their expression levels of SERPINF2 in B-ALL cells were above or below the average of all patients in the dataset. The data were presented as mean ± s.d (standard deviation). P values ≤0.05 were considered significant. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Data availability RNA sequencing and microarray data are available at Gene Expression Omnibus (GEO) under accession numbersGSE205762 andGSE205872, respectively. Data on the expression of genes in human samples and survival of patients with B-ALL were obtained from the publicly avail- able dataset TCGA (TARGET phase II), accessed through the cBioPortal portal (https://www.cbioportal.org) and through the BloodSpot portal (https://www.fobinf.com/). Supplementary information is available for this paper. Correspondence and requests for materials should be addressed to the corresponding author. Source data are providedwith this paper. References 1. Krause, D. S. et al. Differential regulation of myeloid leukemias by the bone marrow microenvironment. Nat. Med. 19, 1513–1517 (2013). 2. Méndez-Ferrer, S. et al. Bone marrow niches in haematological malignancies. Nat. Rev. Cancer 20, 285–298 (2020). 3. Kalajzic, I. et al. Use of type I collagen green fluorescent protein transgenes to identify subpopulations of cells at different stages of the osteoblast lineage. J. Bone Min. Res. 17, 15–25 (2002). 4. Minciacchi, V. R. et al. Differential inflammatory conditioning of the bone marrow by acute myeloid leukemia and its impact on pro- gression. Blood Adv. 8, 4983–4996 (2024). 5. Gerke, V., Creutz, C. E. & Moss, S. E. Annexins: linking Ca2+ sig- nalling to membrane dynamics. Nat. Rev. Mol. Cell Biol. 6, 449–461 (2005). 6. D’Souza, S. et al. Annexin II interactions with the annexin II receptor enhance multiple myeloma cell adhesion and growth in the bone marrow microenvironment. Blood 119, 1888–1896 (2012). 7. Huang,D. et al. Annexin A2-S100A10 heterotetramer is upregulated by PML/RARalpha fusion protein and promotes plasminogen- dependent fibrinolysis and matrix invasion in acute promyelocytic leukemia. Front Med. 11, 410–422 (2017). 8. Kwon, M., MacLeod, T. J., Zhang, Y. & Waisman, D. M. S100A10, annexin A2, and annexin a2 heterotetramer as candidate plasmi- nogen receptors. Front. Biosci. 10, 300–325 (2005). 9. Bazzi, Z. A. et al. Activated thrombin-activatable fibrinolysis inhi- bitor (TAFIa) attenuates breast cancer cell metastatic behaviors through inhibition of plasminogen activation and extracellular proteolysis. BMC Cancer 16, 328 (2016). 10. He, Y. et al. Interaction between cancer cells and stromalfibroblasts is required for activation of the uPAR-uPA-MMP-2 cascade in pan- creatic cancer metastasis. Clin. Cancer Res. 13, 3115–3124 (2007). 11. Pendurthi, U. R., Tran, T. T., Post, M. & Rao, L. V. Proteolysis of CCN1 by plasmin: functional implications. Cancer Res. 65, 9705–9711 (2005). 12. Gong, Y. & Hoover-Plow, J. The plasminogen system in regulating stem cell mobilization. J. Biomed. Biotechnol. 2012, 437920 (2012). 13. Jabbour, E. et al. The evolution of acute lymphoblastic leukemia research and therapy at MD Anderson over four decades. J. Hematol. Oncol. 16, 22 (2023). 14. Zanetti, C. et al. The age of the bone marrow microenvironment influences B-cell acute lymphoblastic leukemia progression via CXCR5-CXCL13. Blood 138, 1870–1884 (2021). 15. Witkowski, M. T. et al. Extensive Remodeling of the Immune Microenvironment in BCell Acute Lymphoblastic Leukemia.Cancer Cell 37, 867–882.e812 (2020). 16. Ling, Q. et al. Annexin II regulates fibrin homeostasis and neoan- giogenesis in vivo. J. Clin. Invest. 113, 38–48 (2004). 17. Leak, S., Horne, G. A. & Copland, M. Targeting BCR-ABL1-positive leukaemias: a review article.Camb. Prism Precis. Med. 1, e21 (2023). 18. Zanetti, C. & Krause, D. S. Caught in the net”: the extracellular matrix of the bone marrow in normal hematopoiesis and leukemia. Exp. Hematol. 89, 13–25 (2020). Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 17 https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE205762 https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE205872 https://www.cbioportal.org https://www.fobinf.com/ www.nature.com/naturecommunications 19. Verma, D. et al. Bone marrow niche-derived extracellular matrix- degrading enzymes influence the progression of B-cell acute lym- phoblastic leukemia. Leukemia 34, 1540–1552 (2020). 20. Houlihan, D. D. et al. Isolation ofmousemesenchymal stemcells on the basis of expression of Sca-1 and PDGFR-alpha. Nat. Protoc. 7, 2103–2111 (2012). 21. Sveeggen, T. M. et al. Annexin A2 modulates phospholipid mem- brane composition upstream of Arp2 to control angiogenic sprout initiation. FASEB J. 37, e22715 (2023). 22. Horowitz, J. C., Rogers, D. S., Simon, R. H., Sisson, T. H. & Than- nickal, V. J. Plasminogen activation induced pericellular fibronectin proteolysis promotes fibroblast apoptosis. Am. J. Respir. Cell Mol. Biol. 38, 78–87 (2008). 23. Quigley, J. P., Gold, L. I., Schwimmer, R. & Sullivan, L. M. Limited cleavage of cellular fibronectin by plasminogen activator purified from transformed cells. Proc. Natl Acad. Sci. USA 84, 2776–2780 (1987). 24. Ishihara, M. et al. Plasmin inhibitor reduces T-cell lymphoid tumor growth by suppressing matrix metalloproteinase-9-dependent CD11b(+)/F4/80(+) myeloid cell recruitment. Leukemia 26, 332–339 (2012). 25. Kuliopulos, A. et al. Plasmin desensitization of the PAR1 thrombin receptor: kinetics, sites of truncation, and implications for throm- bolytic therapy. Biochemistry 38, 4572–4585 (1999). 26. Majumdar, M. et al. Plasmin-induced migration requires signaling through protease-activated receptor 1 and integrin alpha(9)beta(1). J. Biol. Chem. 279, 37528–37534 (2004). 27. Zhai, B. T. et al. Urokinase-type plasminogen activator receptor (uPAR) as a therapeutic target incancer. J. Transl.Med.20, 135 (2022). 28. Jenkins, G. R., Seiffert, D., Parmer, R. J. & Miles, L. A. Regulation of plasminogen gene expression by interleukin-6. Blood 89, 2394–2403 (1997). 29. Vazquez,M. I., Catalan-Dibene, J. & Zlotnik, A. B cells responses and cytokine production are regulated by their immune microenviron- ment. Cytokine 74, 318–326 (2015). 30. Karantanou, C. et al. Impact of mesenchymal stromal cell-derived vesicular cargo on B-cell acute lymphoblastic leukemia progres- sion. Blood Adv. 7, 1190–1203 (2023). 31. Million, R. P., Harakawa, N., Roumiantsev, S., Varticovski, L. & Van Etten, R. A. A direct binding site for Grb2 contributes to transfor- mation and leukemogenesis by the Tel-Abl (ETV6-Abl) tyrosine kinase. Mol. Cell Biol. 24, 4685–4695 (2004). 32. Steelman, L. S. et al. JAK/STAT, Raf/MEK/ERK, PI3K/Akt and BCR- ABL in cell cycle progression and leukemogenesis. Leukemia 18, 189–218 (2004). 33. Carayol, N. et al. Critical roles for mTORC2- and rapamycin- insensitive mTORC1-complexes in growth and survival of BCR-ABL- expressing leukemic cells. Proc. Natl Acad. Sci. USA 107, 12469–12474 (2010). 34. Limon, J. J. & Fruman, D. A. Akt and mTOR in B cell activation and differentiation. Front Immunol. 3, 228 (2012). 35. Beauvais, D. M., Jung, O., Yang, Y., Sanderson, R. D. & Rapraeger, A. C. Syndecan-1 (CD138) suppresses apoptosis in multiple myeloma by activating IGF1 receptor: prevention by SynstatinIGF1R inhibits tumor growth. Cancer Res. 76, 4981–4993 (2016). 36. Vishwamitra, D., George, S. K., Shi, P., Kaseb, A. O. & Amin, H. M. Type I insulin-like growth factor receptor signaling in hematological malignancies. Oncotarget 8, 1814–1844 (2017). 37. Woodham, A. W. et al. Small molecule inhibitors of the annexin A2 heterotetramer prevent human papillomavirus type 16 infection. J. Antimicrob. Chemother. 70, 1686–1690 (2015). 38. Huang,D. et al. Annexin A2-S100A10 heterotetramer is upregulated by PML/RARα fusion protein and promotes plasminogen- dependent fibrinolysis and matrix invasion in acute promyelocytic leukemia. Front. Med. 11, 410–422 (2017). 39. Krause, D. S., Lazarides, K., von Andrian, U. H. & Van Etten, R. A. Requirement for CD44 in homing and engraftment of BCR-ABL- expressing leukemic stem cells. Nat. Med. 12, 1175–1180 (2006). 40. Bagger, F. O. et al. BloodSpot: a database of gene expression pro- files and transcriptional programs for healthy and malignant hae- matopoiesis. Nucleic Acids Res. 44, D917–D924 (2016). 41. Cerami, E. et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Dis- cov. 2, 401–404 (2012). 42. Gao, J. et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci. Signal 6, pl1 (2013). 43. Haferlach, T. et al. Clinical utility of microarray-based gene expression profiling in the diagnosis and subclassification of leu- kemia: report from the International Microarray Innovations in Leukemia Study Group. J. Clin. Oncol. 28, 2529–2537 (2010). 44. Verhaak, R. G. et al. Prediction of molecular subtypes in acute myeloid leukemia based on gene expression profiling. Haematolo- gica 94, 131–134 (2009). 45. Cai, H. et al. Critical role of Lama4 for hematopoiesis regeneration and acute myeloid leukemia progression. Blood 139, 3040–3057 (2022). 46. Miller, J. P. et al. Clinical doses of radiation reduce collagen matrix stiffness. APL Bioeng. 2, 031901 (2018). 47. Mohamed, F., Bradley, D. A. & Winlove, C. P. Effects of ionizing radiation on extracellular matrix. Nucl. Instrum. Methods Phys. Res. Sect. A: Accelerators Spectrometers Detect. Associated Equip. 580, 566–569 (2007). 48. Liu,W. &Hajjar, K. A. The annexin A2 system and angiogenesis. Biol. Chem. 397, 1005–1016 (2016). 49. Hamada, M. et al. Urokinase-Type Plasminogen Activator Receptor (uPAR) in inflammation and disease: a unique inflammatory path- way activator. Biomedicines 12, 1167 (2024). 50. Wang, T. et al. Anxa2 binds to STAT3 and promotes epithelial to mesenchymal transition in breast cancer cells. Oncotarget 6, 30975–30992 (2015). 51. Yang, X., Popescu, N. C. & Zimonjic, D. B. DLC1 interaction with S100A10 mediates inhibition of in vitro cell invasion and tumor- igenicity of lung cancer cells through a RhoGAP-independent mechanism. Cancer Res. 71, 2916–2925 (2011). 52. Sharma, M. C. Annexin A2 (ANX A2): An emerging biomarker and potential therapeutic target for aggressive cancers. Int. J. Cancer 144, 2074–2081 (2019). 53. Lu, X. et al. ADAMTS1 and MMP1 proteolytically engage EGF-like ligands in an osteolytic signaling cascade for bone metastasis. Genes Dev. 23, 1882–1894 (2009). 54. Wlodek, J. & Pituch-Noworolska, A. The influence of fibronectin on proliferation and apoptosis of acute lymphoblastic leukaemia cells in vitro. Pol. J. Pathol. 69, 62–66 (2018). 55. Kumar, R. et al. Specific, targetable interactions with the micro- environment influence imatinib-resistant chronic myeloid leuke- mia. Leukemia 34, 2087–2101 (2020). 56. Yamada, H. et al. Effects of insulin-like growth factor-1 on B-cell precursor acute lymphoblastic leukemia. Int. J. Hematol. 97, 73–82 (2013). 57. Lakshmikuttyamma, A. et al. Bcr-Abl induces autocrine IGF-1 sig- naling. Oncogene 27, 3831–3844 (2008). 58. Shi, P. et al. Inhibition of IGF-IR tyrosine kinase induces apoptosis and cell cycle arrest in imatinib-resistant chronic myeloid leukae- mia cells. J. Cell Mol. Med. 14, 1777–1792 (2010). 59. Tamburini, J. et al. Mammalian target of rapamycin (mTOR) inhibi- tion activates phosphatidylinositol 3-kinase/Akt by up-regulating insulin-like growth factor-1 receptor signaling in acute myeloid leukemia: rationale for therapeutic inhibition of both pathways. Blood 111, 379–382 (2008). Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 18 www.nature.com/naturecommunications 60. Park, S. et al. Role of the PI3K/AKT andmTOR signaling pathways in acute myeloid leukemia. Haematologica 95, 819–828 (2010). 61. Du, P., Kibbe, W. A. & Lin, S. M. lumi: a pipeline for processing Illumina microarray. Bioinformatics 24, 1547–1548 (2008). 62. Eyanga, E. et al. Research on antibodies against BHV-1, BHV-2, BHV- 4, BVD-MD virus, bovine adenovirus A and B, rotavirus and cor- onavirus in cattle in western Zaire: complementary results. Rev. Elev. Med. Vet. Pays Trop. 42, 155–161 (1989). 63. Houlihan, D. D. et al. Isolation ofmousemesenchymal stemcells on the basis of expression of Sca-1 and PDGFR-α. Nat. Protoc. 7, 2103–2111 (2012). 64. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinfor- matics 29, 15–21 (2013). 65. Anders, S., Pyl, P. T. & Huber, W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015). 66. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold changeanddispersion for RNA-seqdatawithDESeq2.GenomeBiol. 15, 550 (2014). 67. Sanjana, N. E., Shalem, O. & Zhang, F. Improved vectors and genome-wide libraries for CRISPR screening. Nat. Methods 11, 783–784 (2014). Acknowledgements This work was supported by grant 2015_A238 to D.S.K. from the Else Kröner-Fresenius-Stiftungandby theDiscovery&DevelopmentProgram grants 2020 and 2021 from the Frankfurt Cancer Institute (FCI) and Project number 449828949 from the Deutsche For- schungsgemeinschaft to V.R.M. The results on human data published here are in part based upon data generated by the Therapeutically Applicable Research to Generate Effective Treatments (https://ocg. cancer.gov/programs/target) initiative, phs000464. The data used for this analysis are available at https://portal.gdc.cancer.gov/projects. We thank the Genomics and Proteomics Core Facility of the DKFZ for per- forming the RNA-sequencing. The staining of bone sections was per- formed by the Department of Pathology of the University Medical Center Mainz. Author contributions V.R.M. and D.S.K. designed the experiments. V.R.M., J.B. and C.K. per- formed the in vitro experiments. V.R.M., C.K., C.Z., R.K. and M.M. per- formed the in vivo experiments. R.P., N.T. and T.K. assisted with experiments, and P.L. assisted with CRISPR-CAS experiments. V.R.M. analyzed the data. M.P. and A.E. provided the ANXA2 overexpressing plasmid. S.H., K.B., V.M.-S. and S.L. provided the human samples. E.M. performed the bioinformatic analyses in the laboratory of B.J.P.H. W.R. provided input on experiments and critically reviewed the manuscript. V.R.M. wrote a first draft of the manuscript. D.S.K. supervised the research, analyzed data and wrote the manuscript. Funding Open Access funding enabled and organized by Projekt DEAL. Competing interests The authors declare no competing interests. Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-024-54361-4. Correspondence and requests for materials should be addressed to Daniela S. Krause. Peer review information Nature Communications thanks Eugenia Flores-Figueroa, Rui Yue and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is avail- able. Reprints and permissions information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jur- isdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2024 Article https://doi.org/10.1038/s41467-024-54361-4 Nature Communications | (2024) 15:10059 19 https://ocg.cancer.gov/programs/target https://ocg.cancer.gov/programs/target https://portal.gdc.cancer.gov/projects https://doi.org/10.1038/s41467-024-54361-4 http://www.nature.com/reprints http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ www.nature.com/naturecommunications Exploitation of the fibrinolytic system by B-cell acute lymphoblastic leukemia and its therapeutic targeting Results ANXA2-deficiency in the BMM leads to survival extension in BCR-ABL1+ leukemia Plasminogen activation contributes to ECM remodeling in the leukemic BMM B-ALL promotes hepatic generation of plasminogen Restricted access to growth factors in the BMM leads to reduced mTORC2 activation and proliferation of BCR-ABL1+ cells Differential sensitivity of leukemias to signaling events downstream of IGF1R Plasmin regulates IGF1 availability by contributing to breakdown of the ECM in an ANXA2-dependent manner Targeting fibrinolysis-associated pathways may extend survival of leukemic mice Fibrinolysis-associated pathways may influence leukemia progression in human patients Discussion Methods Mice Human samples Cell lines Microarray Retrovirus production Bone marrow transduction/transplantation In vivo drug treatment Homing experiment Analysis of diseased mice and tumor burden Progenitor analysis in normal and leukemic mice Colony-formation assay Immunohistochemistry Isolation of GFP+ stroma cells from the BMM Isolation of primary stromal cells Differentiation and colony-forming unit fibroblast (CFU-F) assays of MSC RNAseq Co-immunoprecipitation Immunoblotting Plasmin and MMP-9 activation assay CRISPR Immunofluorescence Immunofluorescence of bone sections Invasion assay Plasmid for ANXA2 overexpression Transient ANXA2 overexpression Quantitative real-time PCR Enzyme-linked immunosorbent assay (ELISA) Chromatin immunoprecipitation (ChIP) assay Proliferation assay Whitlock-Witte limiting dilution assay Statistics and reproducibility Reporting summary Data availability References Acknowledgements Author contributions Funding Competing interests Additional information