A novel CD7-directed antibody-drug conjugate targeting BCL-XL with potent anti-leukemic activity in T-cell acute lymphoblastic leukemia Authors: Mariana L Oliveira, Kanokporn Nuantang, Grégoire Huré, Andrea Ávila-Ávila, Laura Bresson, Sandra Haumont, Vesela Kostova, Tibor Novak, Gaëtane Le Toumelin-Braizat, Julien Guerlesquin, Didier Demarles, Mira Merdas, Nicolas Cauquil, Oumou Goundiam, Francesca Rocchetti, Sophie Courtade-Gaiani, Damien Valour, Marwa Zerhouni, Robin Artus, Emmanuelle Douillet-Michel, Isabelle Laurent, Alice Denis, Alexandra Ovreiu, Boris Duvauchelle, Matyas Ecsedi, Frédéric Colland, Olivier Geneste, Ana Leticia Maragno, Jacques Ghysdael, Christine Tran Quang Published: 2026 ABSTRACT AbstractCurrent treatments of T-cell acute lymphoblastic leukemia (T-ALL) are based on intensive chemotherapy regimens which provide overall survival rates of ~85% in children and <50% in adults. Therefore, there is an unmet need for novel therapeutic options in T-ALL. Pre-clinical studies and clinical trials have demonstrated that inhibitors of BCL-XL and/or BCL-2, two anti-apoptotic proteins of the BCL-2 family, are anti-leukemic in T-ALL. However, BCL-XL inhibitors (BCL-XLi) efficacy is undermined by severe, on-target thrombocytopenia. We report here the design of a novel anti-hCD7 mAb-based ADC carrying a BCL-XL-selective inhibitor (ADC-CD7-BCL-XLi) that circumvents this significant limitation. We show that ADC-CD7-BCL-XLi efficiently kills most T-ALL cell lines. Using T-ALL PDXs we further show that (i) ADC-CD7-BCL-XLi displays potent anti-leukemic activity and is devoid of toxicity to platelets; (ii) ADC-CD7-BCL-XLi acts synergistically with venetoclax, a BCL-2 selective antagonist, to prolong leukemia remission and mouse survival; (iii) the anti-leukemic effect of the ADC-CD7-BCL-XLi+venetoclax combination can lead to cure when combined with chemotherapy. These pre-clinical data strongly support the evaluation of ADC-CD7-BCL-XLi in T-ALL patients, including as a potential bridging option to curative hematopoietic stem cell transplantation (HSCT).Subject terms: Acute lymphocytic leukaemia, Targeted therapies Introduction T-cell lineage acute lymphoblastic leukemia (T-ALL) is an aggressive disease characterized by proliferation of T-cell precursors abnormally arrested in differentiation. Current first-line chemotherapy regimens provide an overall survival rate of 85-90% in children and about 50% in adults. Refractory and relapsed cases (r/r T-ALL) have limited salvage options and a dismal prognosis [1–3]. Several biological functions and molecular pathways essential to leukemia maintenance are currently being explored to identify novel, alternative targets to expand and optimize therapeutic options to treat T-ALL patients [4]. The intrinsic mitochondrial pathway to apoptosis is initiated by activation of BAX/BAK, leading to pore formation in the outer mitochondrial membrane and cytosolic release of proteins that subsequently trigger caspases activation and cell death [5]. BAX and BAK belong to the BCL-2 family of proteins, characterized by the conservation of up to 4 homology domains (BH1-4 domains) involved in critical interactions between these proteins. Besides effectors of apoptosis, this family encompasses pro-survival proteins including BCL-2, BCL-XL and MCL-1 as well as pro-apoptotic proteins characterized by only a BH3 domain. In normal cells, pro-survival BCL-2 proteins interact with BAX/BAK to block their activation and function as sink for BH3-only pro-apoptotic proteins. In response to an array of stresses, expression of BH3-only proteins increases and - through binding with differential affinity and selectivity to anti-apoptotic BCL-2 proteins - neutralizes their activity [5]. Evasion from apoptosis is a hallmark of cancer cells, with over-expression of anti-apoptotic or silencing of pro-apoptotic BCL-2 proteins frequently observed in hematologic malignancies [6]. This led to the development of small molecules BH3 mimetic drugs able to disrupt the interactions between pro-apoptotic BH3-domain and anti-apoptotic BCL-2 family proteins. These drugs include venetoclax, a selective BCL-2 inhibitor, the dual BCL-2/BCL-XL inhibitor navitoclax and A-1331852, a BCL-XL selective inhibitor [6]. In vitro profiling of the apoptotic response of T-ALL primary samples using specific BH3 peptides and BH3 mimetics demonstrated that early T cell progenitors (ETP) T-ALL, a subtype with an immunophenotype and gene expression pattern akin to that of early thymocytes, shows survival dependence on BCL-2 while non-ETP T-ALL rather depends on BCL-XL [7–10]. Pre-clinical studies in T-ALL patient-derived xenografts (PDXs) showed delayed leukemia progression in mice treated with navitoclax [7, 8, 11]. However, as platelet survival critically depends on BCL-XL [12–14], the full anti-leukemic potential of BCL-XL-selective inhibitors could not be explored because of dose-limiting thrombocytopenia. Antibody-drug conjugates (ADCs) represent a promising approach that exploits the specificity of monoclonal antibodies (mAbs) linked to a cytotoxic drug (“payload”) allowing direct delivery of payload to tumor cells while avoiding unwanted side effects [15]. CD7 is a cell surface glycoprotein belonging to the immunoglobulin family that is expressed on normal human T and NK cells and over-expressed in >99% of diagnostic and relapse T-ALL cases [16]. Aiming to circumvent the on-target, dose-limiting toxicity of BCL-XL inhibitors, we report here the design of an anti-CD7-mAb-based ADC carrying a BCL-XL inhibitor payload (ADC-CD7-BCL-XLi) and show that (i) treatment with ADC-CD7-BCL-XLi is pro-apoptotic in T-ALL cell lines and has potent, dose-dependent anti-leukemic activity in PDX models of T-ALL while sparing platelets; (ii) ADC-CD7-BCL-XLi synergizes with venetoclax to extend leukemia remission and improve survival of leukemic mice; (iii) the anti-leukemic efficacy of the ADC-CD7-BCL-XLi+venetoclax therapeutic regimen can lead to cure when combined with chemotherapy. Materials and methods ADC generation The antibody to human CD7 used in this study is a chimera between the constant light and heavy chains of human IgG1 (hIgG1) and the variable light and heavy chains of the anti-mouse anti-CD7 monoclonal antibody (mAb) TH69 (sequence obtained from patent WO218098306A1). A similar construct was produced for the Isotype control anti-CMV B12 (sequence from PDB_1HZH, thereafter named Iso). Cysteine mutations were introduced in the constant k light chain at valine 205 and in the CH3 domain of the heavy chain at serine 400 of both mAbs. In addition, the hIgG1 Fc domain of both mAbs carried L234A/L235A mutations, which abolish complement activity and reduce binding to Fcγ receptors [17]. For linker and payload attachment, see Supplementary Materials and Methods. Mice xenotransplantations NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) female mice (2-month-old; Charles River Laboratories) were maintained under specific pathogen-free conditions at Institut Curie (IC) and Servier Research Institute (SRI) animal facilities according to national and european regulations. Mice were tail vein injected (106 cells/mouse) with leukemic cells (defined as FSChi; hCD45+; hCD7+) obtained from the bone marrow (BM) of donor NSG mice previously xenografted with diagnostic T-ALL cases (Supplementary Materials and Methods). Leukemia expansion overtime in blood, BM and spleen was monitored as in Supplementary Materials and Methods. After randomization, unless otherwise stated, treatments were started when tumor burden reached around 1% in blood meaning over 10% in BM. Control mice were treated with the respective solvents or control antibodies. For experiments using ALL-SIL cell line, mice were grafted subcutaneously (s.c.) with 5×106 cells (100% Matrigel). Tumor volume and body weight were monitored regularly post cell inoculation and animals were randomized into treatment groups (n = 6) when reaching a mean tumor volume of 150–250 mm3. Tumor size was measured using an electronic caliper and calculated using the formula (length × width2)/2. Mice were euthanatized when tumor size reached 2000 mm3. In vivo inhibitors treatments Leukemic mice were tail vein-intravenously injected with either 2.5, 7.5 or 15 mg/kg of either ADC-CD7-BCL-XLi, Iso-BCL-XLi or anti-CD7-naked antibody as detailed in legends. Venetoclax was administered by oral gavage (50 mg/kg) for 3 cycles of 5 consecutive days followed by 2 days without administration. S55746 (a selective BCL-2 inhibitor [18]) was administered by oral gavage (100–200 mg/kg as detailed in legends) daily during 5 consecutive days, followed by 2 days off (2 cycles). A-1331852 (selective BCL-XL inhibitor) was administered by oral gavage at doses ranging from 2.5 to 20 mg/kg as indicated in legends. Chemotherapy consisted of intraperitoneal administration of vincristine (Sigma-Aldrich) 0.5 mg/kg on day 1 and dexamethasone (Sigma-Aldrich) 15 mg/kg on days 1–5, followed by 2 days off. This cycle was repeated over 3 weeks. Navitoclax (50 mg/kg) was administered by oral gavage for 4 consecutive days the first week followed by 2 days off and 5 consecutive days the second week (See also Supplementary Materials and Methods). Ethics approval All methods were performed in accordance with the relevant guidelines and regulations. Experiments using mice at IC were carried out in accordance with the European Union and French National Committee recommendations, under agreement APAFIS #49667-2024043015406385 v4. Experiments at IRS facility (license number C78-100-2) were in accordance with European Union and national regulations for the protection of vertebrate animals (directives 2010/63/EU and Decree n°2013-118) and the requirements of the Servier Research Institute Animal Ethics Committees (CE031). Statistics Leukemic load statistical analysis was performed using Prism (version 10.0; GraphPad Software). The data are expressed as mean +/- SEM of n = 4 or more determinations. In short-term in vivo and in vitro assays, we used ANOVA and Tukey’s Multiple Comparison Test or Student’s T Test. Kaplan-Meier survival curves were compared using the log-rank test. Differences were considered statistically significant at * p < 0.05, ** p < 0.01, *** p < 0.001, or **** p < 0.0001. Results Previous studies have shown that survival of ETP T-ALL primarily depends upon BCL-2 whereas that of non-ETP cases rather relies upon BCL-XL [7, 10]. Accordingly, treatment of NSG mice xenografted with a series of non-ETP diagnostic T-ALL samples with A-1331852, a selective BCL-XL inhibitor [19], induced a clear anti-leukemic effect while a weaker response was observed upon treatment with S55746, a BCL-2-selective inhibitor [18] (Supplementary Fig. 1A). In contrast, an ETP T-ALL PDX showed a reverse response pattern to these drugs, being more sensitive to S55746 (Supplementary Fig. 1A). We next investigated the therapeutic response of non-ETP T-ALL PDXs to the A-1331852 + S55746 combination. In this configuration, and as expected from the well characterized dependence of platelets survival on BCL-XL [12, 13], combination of 100 mg/kg of S55746 with either 20 or 10 mg/kg of A-1331852, proved rapidly toxic, inducing hemorrhages and subsequent death of mice (Supplementary Fig. 1B). However, a regimen of 3 days pre-treatment with 2.5 mg/kg A-1331852 followed by daily treatment with 7.5 mg/kg A-1331852 in presence of 100 mg/kg S55746 alleviated toxicity (Supplementary Fig 1B), in line with previously published data using a similar regimen for navitoclax [20]. Under these conditions, inhibition of BCL-XL and BCL-2 showed synergistic anti-leukemic effects (Supplementary Fig. 1C), translating into improved host survival (Supplementary Fig. 1D). Higher dosage of A-1331852 (2.5 mg/kg pre-treatment followed by 10 mg/kg instead of 7.5 mg/kg) showed better therapeutic activity of the BCL-XL/BCL-2 inhibitors combination as evidenced by improved leukemia free survival of some mice but was associated with lethal hemorrhages in a subset of mice (Supplementary Fig. 1E). This data suggests that finding a strategy to prevent the on-target thrombocytopenia associated with high dose systemic BCL-XL inhibitors could unleash the therapeutic potential of combined BCL-XL/BCL-2 inhibition for T-ALL treatment. Design of ADC-CD7-BCL-XLi To circumvent the toxicity of systemic administration of BCL-XLi and selectively target leukemic lymphoblasts, we generated an anti-CD7-BCL-XLi ADC (ADC-CD7-BCL-XLi, Fig. 1A and Supplementary Fig. 2 and Materials and Methods for details). We used a cathepsin-cleavable dipeptide linker (Mc-Val-Cit-PABC) to conjugate the BCL-XL inhibitor payload 11, an analog of A-1331852, which was specifically designed to possess high affinity and reduced cell penetration, thereby minimizing toxicity [21]. As a control, we conjugated the same linker-payload to an hIgG1 control antibody (thereafter named Iso-BCL-XLi). The unconjugated anti-CD7 antibody (CD7 Naked) was used as a second control. ADC-CD7-BCL-XLi shows potent in vitro pro-apoptotic activity in a panel of T-ALL cell lines. A Schematic representation of ADC-CD7-BCL-XLi. B The 20 indicated T-ALL cell lines were treated with increasing doses of either ADC-CD7-BCL-XLi (3 to 300 nM) or BCL-XLi payload (ranging from 0.1 to 30 µM) and cell viability was measured 5 days after treatment using Cell Titer Glo (CTG). IC50 corresponds to the compound concentration at which the CTG signal is reduced to 50% of that measured for the control. Measurements were done in triplicates and mean of 3 independent experiments is shown. Blue intensity scale refers to cell surface CD7 expression, measured as described in Materials and Methods and reported in Supplementary Table 1. C Spearman’s correlation analysis between ADC-CD7-BCL-XLi IC-50 and CD7 cell surface expression using data shown in B and in Supplementary Table 1. D Caspase activation measurement in ALL-SIL cells treated for 6 h with increasing doses of either ADC-CD7-BCL-XLi (0.3, 3, 10, 100 nM; blue bars), or BCL-XLi payload (1.2, 12, 40, 400 nM; yellow orange bars), using Caspase-GLO® 3/7 assay Kit. Doses have been adapted to be able to compare ADC-CD7-BCL-XLi having a DAR of 4 (Drug to Antibody Ratio) with BCL-XLi payload. Iso-BCL-XLi (grey bars) and naked CD7 antibody (dashed grey bars) at the maximum dose (100 nM) are used as controls. E Loss of mitochondrial membrane potential measured in ALL-SIL cells treated as in D. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. ADC-CD7-BCL-XLi has pro-apoptotic activity in vitro We first analyzed the effect of ADC-CD7-BCL-XLi on the viability of a series of T-ALL cell lines. Out of the twenty T-ALL cell lines evaluated, fifteen exhibited specific, high sensitivity to ADC-CD7-BCL-XLi (IC50 values ranged from 10pM to 4 nM, Fig. 1B), whereas four displayed an intermediate response (IC50 from 50 to 215 nM) or poor response (SUPT11, IC50 > 300 nM; Fig. 1B). Importantly, IC50 values of ADC-CD7-BCL-XLi were found to be much lower as compared to IC50 values of the non-conjugated payload (Fig. 1B, Supplementary Table 1) or those of Iso-BCL-XLi (Supplementary Table 1) and correlated with CD7 expression (Fig. 1C). While IC50 of ADC-CD7-BCL-XLi did not correlate with BCL-XL expression, it correlated with the BCL-XL to MCL-1 plus BCL-2 ratio [22], this correlation being stronger when this ratio was multiplied by the number of CD7 receptors per cell (Supplementary Fig. 3). CD7-negative non-T-ALL cell lines, such as H146 and SUDHL8 - although sensitive to BCL-XLi payload - failed to respond to ADC-CD7-BCL-XLi (Supplementary Table 1). As expected, both the payload and ADC-CD7-BCL-XLi induced apoptotic cell death in T-ALL cells as evidenced by increased activation of caspases 3 and 7 (Fig. 1D) and loss of mitochondrial membrane potential (Fig. 1E). ADC-CD7-BCL-XLi is anti-leukemic in T-ALL We next evaluated ADC-CD7-BCL-XLi anti-tumor potential in NSG mice subcutaneously xenografted with the ALL-SIL T-ALL cell line (schematized in Fig. 2A) and investigated its anti-tumor activity and toxicity on platelets as compared to the dual BCL-2/BCL-XL inhibitor navitoclax. Engrafted NSG mice were administered with either 15 mg/kg ADC-CD7-BCL-XLi (Fig. 2A, blue arrow), or with 50 mg/kg navitoclax (purple arrows) and compared to untreated mice (black tracing) for tumor growth. A single dose of ADC-CD7-BCL-XLi resulted in pronounced tumor regression and showed faster, deeper, and prolonged anti-tumoral effects as compared to navitoclax administration (Fig. 2B). Pharmacokinetic experiments evaluated the half-life of ADC-CD7-BCL-XLi to range from 12.2 days (2.5 mg/kg) to 14.1 days (15 mg/kg) (Fig. 2C and Supplementary Table 2) while only a small amount of free payload was detected in plasma of treated mice (Fig. 2C and Supplementary Table 2). Navitoclax showed the expected pharmacokinetic profile (Fig. 2C and Supplementary Table 2). Most importantly, thrombocytopenia was only observed in navitoclax-treated mice (Fig. 2D, purple bars), while ADC-CD7-BCL-XLi had no effect on murine platelet counts (Fig. 2D, blue bars) nor on body weight (Supplementary Fig. 4B). These findings were confirmed using a hCD7/mCD7 chimeric knock-in mice (See Supplementary Fig. 5A for description), in which no deleterious effect was observed on platelets, white blood cell, lymphocytes or red blood cell counts upon ADC-CD7-BCL-XLi administration (Supplementary Fig. 5B, C) nor on body weight (Supplementary Fig. 4C). The non-toxic effect of ADC-CD7-BCL-XLi at concentrations cytotoxic to T-ALL cells was confirmed ex vivo on platelets obtained from buffy coat of healthy donors (Supplementary Fig. 6). ADC-CD7-BCL-XLi also proved to be non-toxic to both naïve and activated T and NK cells from healthy donors (Supplementary Fig. 6). Taken together, these in vivo and ex vivo results show the lack of non-specific toxicity of ADC-CD7-BCL-XLi. ADC-CD7-BCL-XLi has potent anti-tumor activity in ALL-SIL xenografted mice. A Schematic representation of the experiment. NSG mice were subcutaneously (s.c.) inoculated with 5×106 ALL-SIL cells and monitored for tumor growth over time. Once tumor reached 150–250 mm3, mice (n = 6 per group) were administered with either ADC-CD7-BCL-XLi (15 mg/kg, blue arrow) intravenously or navitoclax (50 mg/kg, purple arrow) by oral gavage. B Tumor size for each mouse in the three arms over time. Black line: untreated; blue line: ADC-CD7-BCL-XLi; purple line: navitoclax. C Left: Pharmacokinetics of ADC-CD7-BCL-XLi antibody (plain lines) and free payload (dotted lines) in mice (n = 4/arm) treated with either 2.5 mg/kg or 15 mg/kg of ADC-CD7-BCL-XLi. Right: Pharmacokinetics of navitoclax in mice (n = 4) treated with 50 mg/kg navitoclax. D Platelet counts measured 6 h and 48 h post-treatment in mice treated as described in A. **** p < 0.0001. We next assessed the anti-leukemic potential of ADC-CD7-BCL-XLi using six distinct non-ETP T-ALL PDXs, which all express homogenous levels of CD7 (Supplementary Fig. 7A, B) - albeit at different levels (Supplementary Fig. 7C) - as well as BCL-2 and BCL-XL (Supplementary Fig. 7D). Groups of leukemic mice were treated with a single injection of different doses (ranging from 2.5 mg/kg to 15 mg/kg) of either Iso-BCL-XLi (white bars), the CD7-naked antibody (grey bars), or ADC-CD7-BCL-XLi (blue bars) and compared for leukemia burden to mice treated with vehicle (black bars) as schematized in Fig. 3A. As shown in Fig. 3B and Supplementary Fig. 8, ADC-CD7-BCL-XLi treatment resulted in a significant dose-dependent reduction in circulating leukemic cells for five out of six T-ALL PDXs tested as compared to vehicle-treated, Iso-BCL-XLi-treated or CD7-naked antibody-treated arms. T-ALL PDX #5 showed only a poor response that was linked to its low CD7 cell surface expression, since over-expression of hCD7 improved its response to ADC-CD7-BCL-XLi (Supplementary Fig. 9). ADC-CD7-BCL-XLi also exhibited potent anti-leukemic activity in bone marrow (BM, Fig. 3C), an organ often less responsive to therapeutic intervention (for review see [23]). These anti-leukemic effects translated into improved mouse survival (Supplementary Fig. 10). Of note, whereas evolution of leukemia burden in Iso-BCL-XLi-treated mice was similar to that of the control arm, administration of the CD7-naked antibody showed minor but significant anti-leukemic effects in a subset of T-ALL cases (Fig. 3B, C, Supplementary Figs. 8 and 10). Treatment of leukemic mice with ADC-CD7-BCL-XLi was devoid of deleterious effect on platelets whereas, as expected, administration of A-1331852 induced thrombocytopenia (Supplementary Fig. 11). The anti-leukemic effects of ADC-CD7-BCL-XLi were improved by an iterative treatment regimen, in which mice that had been treated with ADC-CD7-BCL-XLi were individually monitored for relapse with each mouse being re-treated with ADC-CD7-BCL-XLi only when 1-5% of CD7-positive leukemic cells were detected in blood (Fig. 4A for a schematic representation of experiment). As shown in Fig. 4B, leukemic cells remained sensitive to ADC-CD7-BCL-XLi for up to three treatments. However, mice ultimately experienced relapse while under treatment with two out of three mice showing no decreased leukemia burden at the fourth ADC-CD7-BCL-XLi injection while the remaining mouse relapsed under treatment at the fifth injection (Fig. 4B, C). Relapsed leukemic cells from these mice were CD7-positive (Fig. 4D) and associated with enhanced expression of BCL-2/BCL-XL as well as activation of the protein tyrosine kinase ACK1, a recently identified mechanism of resistance to BCL-XLi in T-ALL [10] (Fig. 4E). Of note, co-administration of ADC-CD7-BCL-XLi and dasatinib, which inhibits ACK1 [10] proved to be synergistic in T-ALL PDXs (Supplementary Fig. 12). ADC-CD7-BCL-XLi has potent anti-leukemic activity in xenografts from diagnostic T-ALL cases (PDXs). A Schematic representation of the experiment. NSG mice infused with 106 leukemic cells from PDXs derived from the indicated T-ALL cases were monitored for leukemia expansion over time. Once leukemic, mice (n = 4 per group) were injected with either vehicle, a single injection of Iso-BCL-XLi, naked CD7 mAb, or ADC-CD7-BCL-XLi. B Percentage of leukemic cells (hCD45+) in blood at the time leukemia burden in control group reached >50% (black bars) in leukemic mice treated with the indicated increasing doses of either Iso-BCL-XLi (white bars), naked CD7 antibody (grey bars) or ADC-CD7-BCL-XLi (blue bars). C Percentage of leukemic cells (hCD45+) in bone marrow of mice, two weeks after the start of treatment. ns non-significant, * p < 0.05, ** p < 0.01, *** p < 0.001, ****p < 0.0001. Iterative ADC-CD7-BCL-XLi treatment prolongs mice survival, resistance being associated with increased BCL-2 and BCL-XL expression and ACK1 activation. A Schematic representation of the experiment. NSG mice infused with 106 leukemic cells from T-ALL#2 were monitored for leukemia expansion over time. Once leukemic, mice (n = 3 per group) were injected with either vehicle (group 1) or a single injection of 15 mg/kg ADC-CD7-BCL-XLi (group 2) or successive injections of 15 mg/kg ADC-CD7-BCL-XLi (group 3). In this last group, each mouse was re-treated with ADC-CD7-BCL-XLi only when 1-5% of CD7-positive leukemic cells were detected in blood. Leukemia burden was followed over time for all mice. B Before-After plot showing the percentage of leukemic cells (hCD45+) in blood of each mouse of group 3, before and three days after 15 mg/kg ADC-CD7-BCL-XLi treatment. Mouse 1 (schematized in green) was treated 20, 56, 78 and 84 days post-injection; mouse 2 (schematized in blue) at days 20, 49, 60 and 66, while mouse 3 (red) was treated at days 20, 49, 63, 78 and 84. Interestingly, the Before-After plot of these mice shows that all mice from arm 3 responded to the anti-leukemic effect of ADC-CD7-BCL-XLi up to the third treatment and became resistant to treatment thereafter despite expressing CD7 at their cell surface as shown in D. C Kaplan-Meier survival curves of leukemic mice described in A and B. Single administration of ADC-CD7-BCL-XLi arm was compared to iterative administration of ADC-CD7-BCL-XLi arm. *p < 0.05. D Flow cytometry analysis of CD7 cell surface expression in leukemic cells (hCD45+) from blood of mice treated as depicted in A at the time mice were terminally ill and euthanized. E Western blot analysis of BCL-2, BCL-XL, activated ACK1 and total ACK-1 expression in leukemic cells from BM of mice treated as depicted in A when terminally ill. Results from three independent mice in each group are shown. hCD45 and actin expressions are used as loading controls (bottom panels). ADC-CD7-BCL-XLi cooperates with venetoclax to strongly inhibit leukemia expansion We next explored the potential synergy between combined administration of ADC-CD7-BCL-XLi and venetoclax, a selective inhibitor of BCL-2. Synergy between these compounds was observed in T-ALL cell lines in vitro (Supplementary Fig. 13). To investigate this in vivo, leukemic mice xenografted with T-ALL #2 were treated with either vehicle, a single administration of ADC-CD7-BCL-XLi, venetoclax, or the ADC-CD7-BCL-XLi+venetoclax combination. Leukemia burden and leukemic cell apoptosis were analyzed 12, 36 and 60 h post-treatment. As shown in Fig. 5, ADC-CD7-BCL-XLi decreased leukemia burden in blood, BM and spleen over time, a phenotype strongly augmented by venetoclax co-treatment, while venetoclax alone had minor effect. This was associated with synergy in apoptosis induction (Fig. 5D), that was detected as soon as 12 h post-treatment and was still observed at 60 h, a time-point at which leukemic cells were hardly detectable in mice. To investigate whether this anti-leukemic effect translated into improved mouse survival, leukemic mice carrying two independent T-ALL PDXs (T-ALL #2 and #3) received either vehicles (black tracings), Iso-BCL-XLi+venetoclax (dark grey tracings), CD7-naked antibody either alone (plain light grey tracings) or combined with venetoclax (dashed light grey tracings), ADC-CD7-BCL-XLi either alone (blue tracings) or combined with venetoclax (pink tracings) (Fig. 6A for a scheme). As shown in Fig. 6B, while ADC-CD7-BCL-XLi monotherapy significantly delayed leukemia progression as compared to controls, venetoclax exerted minimal anti-leukemic effects. Strikingly, dual administration of ADC-CD7-BCL-XLi+venetoclax profoundly inhibited leukemia expansion and substantially prolonged time to relapse as compared to the ADC-CD7-BCL-XLi alone (Fig. 6B). Furthermore, the combinatorial regimen enhanced survival outcome, with T-ALL #2 and T-ALL #3 xenografted mice surviving up to 140 days, compared to 83 days (T-ALL #2) and 102 days (T-ALL #3) in the groups treated with only ADC-CD7-BCL-XLi (Fig. 6C). Continuous treatment of leukemic mice with ADC-CD7-BCL-XLi+venetoclax (Supplementary Fig. 14A for a scheme) improved inhibition of leukemia expansion but eventually resulted in relapses showing severe downmodulation of surface CD7 expression, increased expression of BCL-2/BCL-XL and inability to respond to ADC-CD7-BCL-XLi when re-injected into a new cohort of NSG mice (Supplementary Fig. 14B, G). ADC-CD7-BCL-XLi synergizes with venetoclax to induce leukemic cell apoptosis in vivo. NSG mice infused with 106 leukemic cells from T-ALL #2, were monitored for leukemia expansion over time. Once leukemic load reached about 10–20% in blood, mice (n = 3 per group) were treated with either vehicles (black columns), a single injection of 15 mg/kg of Iso-BCL-XLi + 50 mg/kg venetoclax (red columns), a single injection of 15 mg/kg of ADC-CD7-BCL-XLi (blue column) or co-injection of 15 mg/kg of ADC-CD7-BCL-XLi + 50 mg/kg venetoclax (pink column). Leukemic load (hCD45+) was measured in blood A, bone marrow B and spleen C at 12 h, 36 h and 60 h post-treatment. D Apoptosis measurement in BM leukemic cells collected from mice treated as described above. * p < 0.05, ** p < 0.01, *** p < 0.001, ****p < 0.0001. ADC-CD7-BCL-XLi acts synergistically with venetoclax in T-ALL PDXs. A Schematic representation of the experiment. NSG mice infused with 106 leukemic cells from PDXs derived from the 2 indicated T-ALL cases (#2 and #3), were monitored for leukemia expansion over time. Once leukemic, mice (n = 4 per group) were treated with either vehicles (black arrow), or a single injection of 15 mg/kg of Iso-BCL-XLi + 50 mg/kg of venetoclax (dark grey arrow), venetoclax being delivered daily for 5 days separated by two days off-treatment (3 cycles) as indicated by red arrows, or a single injection of 15 mg/kg of naked CD7 (light grey arrow) antibody plus or minus 50 mg/kg venetoclax administered as above, or 15 mg/kg ADC-CD7-BCL-XLi plus or minus 50 mg/kg venetoclax administered as above (pink and blue arrows respectively). B Leukemic load over time (% hCD45+ cells) in blood of individual mice in each arm injected with T-ALL #2 and #3 treated as described in A. Single administration timepoint of ADC-CD7-BCL-XLi is indicated as a blue dotted line; administration period of venetoclax is shadowed in red. C Kaplan-Meier survival curves of leukemic mice described in A and B. ADC-CD7-BCL-XLi+venetoclax arm was compared to ADC-CD7-BCL-XLi arm, ** p < 0.01. Note that time to relapse between ADC-CD7-BCL-XLi alone and ADC-CD7-BCL-XLi+venetoclax changes from 29 to 76 ± 3.35 days for T-ALL #2; 44 to 78 ± 2.87 days for T-ALL #3. Chemotherapy and ADC-CD7-BCL-XLi+venetoclax combination can induce cure in T-ALL PDXs Recent clinical data reported that dual BCL-2/BCL-XL inhibition associated with chemotherapy is feasible but limited by navitoclax toxicity, providing a compelling therapeutic avenue in B- and T-ALL r/r cases [24]. To evaluate the therapeutic potential of the ADC-CD7-BCL-XLi+venetoclax combination with chemotherapy, T-ALL xenografted mice were treated with 3 cycles of vincristine plus dexamethasone followed or not by ADC-CD7-BCL-XLi+venetoclax administration (see Fig. 7A for a scheme). As shown in Fig. 7B, chemotherapy reduced leukemia burden (grey tracing), followed by relapse upon cessation of drug administration. Strikingly, administration of ADC-CD7-BCL-XLi+venetoclax post-chemotherapy considerably improved outcome and survival, with several mice being cured (T-ALL #2, green tracing, three out five mice remained disease-free up to the end of the study at 491 days; T-ALL #3, green tracing, all mice remained disease free up to the end of the study at 300 days). Chemotherapy combined with dual ADC-CD7-BCL-XLi+venetoclax treatment is profoundly anti-leukemic in T-ALL. A Schematic representation of the experiment. NSG mice infused with 106 leukemic cells from PDXs derived from the 2 indicated T-ALL cases (#2 and #3), were monitored for leukemia expansion over time. Once leukemic, mice (n = 5 per group) were either untreated (control) or treated with vincristine (0.5 mg/kg) + dexamethasone (15 mg/kg) for 3 cycles, as detailed in Materials and Methods, followed by treatments with either a single injection of 15 mg/kg Iso-BCL-XLi (indicated by the arrow) -/+ 50 mg/kg venetoclax or a single injection of 15 mg/kg ADC-CD7-BCL-XLi (indicated by the arrow) -/+ 50 mg/kg of venetoclax and compared for leukemia burden and survival to the untreated control group (black). Venetoclax administration was for 3 cycles of 5 days separated by 2 days off. B Leukemia burden over time in blood of mice injected with T-ALL #2 and #3 treated as described in A. Treatment period with chemotherapy is shadowed in grey; single administration of ADC-CD7-BCL-XLi is indicated by a blue dotted line; treatment period with venetoclax is shadowed in red. C Kaplan-Meier survival curves of leukemic mice transplanted with T-ALL #2 and #3, and treated as described in A and B. VD/ADC-CD7-BCL-XLi arm was compared to VD/ADC-CD7-BCL-XLi + venetoclax, * p < 0.05, ** p < 0.01. Discussion Our results using BCL-XL and BCL-2-selective inhibitors show that survival of PDXs derived from non-ETP T-ALL primarily relies on BCL-XL, but also on BCL-2 as evidenced by the anti-leukemic synergy observed between BCL-2- and BCL-XL-selective inhibitors (Supplementary Fig. 1). This agrees with previously published data [8, 25]. However, the severe on-target platelets toxicity associated with the systemic in vivo administration of BCL-XL-selective or dual BCL-2/BCL-XL inhibitors has so far limited their therapeutic efficacy in T-ALL and other malignancies. We thus investigated the concept that targeted delivery of a selective BCL-XL inhibitor to T-ALL cells would circumvent platelet cytotoxicity and its sequelae while allowing efficient BCL-XL inhibition in leukemic cells, thus improving therapeutic index. Anti-CD7 ADCs carrying the anti-mitotic drug Monomethyl Auristatin E (MMAE) or the Topoisomerase 1 inhibitor Deruxtecan (DxD) have previously been generated but only induced a transient anti-leukemic effect in T-ALL cell line models [26, 27]. We thus exploited a similar strategy, but used a potent BCL-XL inhibitor [21] as payload. We found that a single administration of ADC-CD7-BCL-XLi has clear anti-leukemic effects in T-ALL PDXs, preserves platelets and does not induce any detectable toxicity. Several administrations of ADC-CD7-BCL-XLi show improved anti-leukemic effects but still led to relapse. Single administration of ADC-CD7-BCL-XLi synergizes with venetoclax, a BCL-2-selective inhibitor, to prolong leukemia remission while combination of the ADC-CD7-BCL-XLi+venetoclax regimen with chemotherapy leads to protracted mouse survival and even cure in a subset of mice. Resistance to iterative ADC-CD7-BCL-XLi treatment either alone or in combination with venetoclax was associated with compensatory over-expression of BCL2/BCL-XL, activation of the protein tyrosine kinase ACK1 and downmodulation of surface CD7 expression, the latter being often observed for antigens targeted by other immunotherapies. Adults with newly diagnosed T-ALL and relapsed or refractory (r/r) cases have poor outcome. Recent clinical trials show that inclusion of venetoclax or the combination of venetoclax with navitoclax, a dual BCL-2/BCL-XL inhibitor, to standard of care therapeutic regimens can improve depth and duration of the therapeutic response in both diagnostic and r/r T-ALL patients [24, 28–31]. Based upon the promising results of the association of chemotherapy with venetoclax and navitoclax (at low dose to avoid thrombocytopenia) in pediatric and adult r/r T-ALL cases [24], it is reasonable to anticipate that replacement of navitoclax with ADC-CD7-BCL-XLi in this therapeutic regimen could provide a larger safety window, further improvement in patient outcome and a superior bridge to curative HSCT. This clinical trial also revealed that several T-ALL cases initially found to be dependent upon BCL-2 for leukemic cell survival, a common feature in ETP T-ALL, switch to BCL-XL dependency during treatment [24]. The ADC-CD7-BCL-XLi+venetoclax combination is expected to address this compensatory mechanism and therefore be beneficial also for treatment of ETP T-ALL patients. Of note, clinical development of navitoclax has been recently discontinued, calling for the development of novel, non-toxic BCL-XL inhibitors, ADC-CD7-BCL-XLi fully fitting this requirement. Although of small sample size, other clinical trials reported promising efficacy of venetoclax inclusion in standard of care chemotherapy in newly-diagnosed B- and T-ALL patients [28, 29, 31]. This, together with the strong synergy observed in T-ALL PDXs between chemotherapy and ADC-CD7-BCL-XLi+venetoclax call for investigation of this combination in front line T-ALL treatment. Other BCL-2/BCL-XL-dependent hematopoietic malignancies are characterized by cell surface expression of CD7, including T prolymphocytic leukemia (T-PLL [32, 33]), rare cases of B-ALL [34], about 30% of AML patients [35–38] and a subset of T cell lymphomas. These patient populations could also benefit from the ADC-CD7-BCL-XLi+venetoclax combination. Novel BCL-XL or dual BCL-XL/BCL-2 systemic antagonists with decreased on-target toxicity in platelets and with potentially broader spectrum as cancer therapeutics have recently been developed and investigated to some extent in pre-clinical studies in hematological malignancies. These include second generation dual inhibitors with reduced affinity towards BCL-XL and controlled delivery rates [25, 39, 40] and proteolysis targeting chimera (PROTAC) agents. In PROTACs, BH3 mimetics are chemically linked to a ligand that binds a specific E3 ligase expressed at low levels in platelets, including VHL and cereblon, leading to BCL-XL ubiquitination and proteolytic degradation preferentially in tumor cells [41–43]. Limited studies of these inhibitors in T-ALL PDXs are available [25, 41] but studies in T-ALL cell lines [43, 44] show lower potency on a molar basis as compared to ADC-CD7-BCL-XLi. The ability to utilize distinct targeting approaches to inhibit BCL-XL/BCL-2 is important and complementary because resistance mechanisms to ADC-based, PROTAC-based, or systemic inhibitors can be distinct [45, 46]. Immunotherapeutic approaches, such as inotuzumab ozogamicin (a CD22-targeting ADC), Blinatumomab (a bispecific CD19/CD3 antibody), and CD19-directed CAR-T cells, have significantly improved outcomes for patients with r/r B-ALL and are beginning to be integrated into frontline therapy [47]. In contrast, immunotherapeutic strategies for T-ALL lag behind despite preclinical data supporting the potential of targeting several cell surface antigens, including CD38, CD3, CD5, and CD7, either as antibody targets or through CAR-T-based cellular therapies [4]. ADC-CD7-BCL-XLi does provide a novel immunotherapeutic tool in T-ALL treatment and its effectiveness in our pre-clinical studies call for its translation into clinical trials in combination with venetoclax and standard of care chemotherapy. Supplementary information Supplementary Material and Methods Supplementary Figures Acknowledgements This work was supported by funds from Servier and Institut Curie. MLO was supported by a post-doctoral fellowship from Servier, KN and GH were supported by funds from Servier, CTQ is supported by Institut National de la Santé et de la Recherche Médicale (INSERM), JG is Research Director Emeritus from CNRS (Centre National pour la Recherche Scientifique). The authors thank C Lasgi (Institut Curie cytometry platform), E Belloir, C Alberti, S. Bourgery and P. Dubreuil (Institut Curie animal facility) for expert technical assistance. They also thank M. Merdas, M. Verdier, A. Pay, and Y. Fahmi for their support with the bioanalysis assays; H. Darville and S. Catesson-Dromaint for their contributions to genomic and molecular analysis; and A. Bonnin and her team for their assistance with hematology parameters. Vernalis R&D Ltd (Cambridge, UK) participated on the characterization of the payload and other small molecules used in this work. Author contributions MLO, CTQ, JG, ALM, LB, GLTB TN, VK, DD, FR, A-AA, SH, MZ, RA, EDM, IL, AD, AO, BD, JGuerlesquin, NC carried out the experiments, data analysis and interpretation, and organized the data; KN, GH, DD, SCG, DV, O. Goundiam and ME participated to the experimental work and data analysis; MLO, LB, OG, FC, ALM, JG and CTQ designed the studies, carried out data analysis and interpretation, organized the data and wrote the manuscript. Funding This work was supported by Servier Laboratories. Data availability All original data are available, please contact christine.tran-quang@curie.fr. Competing interests ALM, LB, GLTB, TN, VK, DD, A-AA, SH, JGuerlesquin, NC, SCG, DV, OGoundiam, ME, MM, MZ, RA, EDM, IL, AD, AO, BD are employees of Servier. FC and OG have been employees of Servier and are currently employees of Pierre Fabre Laboratories. FR has been employee of Servier and is currently employee of Boehringer Ingelheim. MLO was supported by a postdoctoral fellowship from Servier. KN and GH were supported by funds from Servier. Consent to participate statement All PDXs were derived from patients’ T-ALL diagnostic samples included in the GRAALL-2005 study or in the FRALLE protocol and informed consent was obtained from all patients at trial entry (trial registration ID: NCT00327678). Studies were conducted in accordance with the Declaration of Helsinki and approved by local and multicenter research ethical committees. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Ana Leticia Maragno, Jacques Ghysdael, Christine Tran Quang. Supplementary information The online version contains supplementary material available at 10.1038/s41375-026-02998-z. References Associated Data Supplementary Materials Supplementary Material and Methods Supplementary Figures Data Availability Statement All original data are available, please contact christine.tran-quang@curie.fr.