DT2216—a Bcl-xL-specific degrader is highly active against Bcl-xL-dependent T cell lymphomas

Background Patients with advanced T cell lymphomas (TCLs) have limited therapeutic options and poor outcomes in part because their TCLs evade apoptosis through upregulation of anti-apoptotic Bcl-2 proteins. Subsets of TCL cell lines, patient-derived xenografts (PDXs), and primary patient samples depend on Bcl-xL for survival. However, small molecule Bcl-xL inhibitors such as ABT263 have failed during clinical development due to on-target and dose-limiting thrombocytopenia. Methods We have developed DT2216, a proteolysis targeting chimera (PROTAC) targeting Bcl-xL for degradation via Von Hippel-Lindau (VHL) E3 ligase, and shown that it has better anti-tumor activity but is less toxic to platelets compared to ABT263. Here, we examined the therapeutic potential of DT2216 for TCLs via testing its anti-TCL activity in vitro using MTS assay, immunoblotting, and flow cytometry and anti-TCL activity in vivo using TCL cell xenograft and PDX model in mice. Results The results showed that DT2216 selectively killed various Bcl-xL-dependent TCL cells including MyLa cells in vitro. In vivo, DT2216 alone was highly effective against MyLa TCL xenografts in mice without causing significant thrombocytopenia or other toxicity. Furthermore, DT2216 combined with ABT199 (a selective Bcl-2 inhibitor) synergistically reduced disease burden and improved survival in a TCL PDX mouse model dependent on both Bcl-2 and Bcl-xL. Conclusions These findings support the clinical testing of DT2216 in patients with Bcl-xL-dependent TCLs, both as a single agent and in rational combinations.


Introduction
T cell lymphomas (TCLs) are a rare and heterogeneous group of lymphoid malignancies that account for approximately 10% of all non-Hodgkin lymphomas (NHLs) in high-income countries [1] and a higher fraction in lower-and middle-income countries [2]. About two thirds of TCLs are peripheral TCLs (PTCLs) and the remainder are cutaneous TCLs (CTCLs). The prognosis for patients with PTCLs and advanced-stage CTCLs are very poor because the majority of these TCL patients either fail to achieve a remission or experience relapse within 2 years after completion of first-line therapy [3]. Agents approved for relapsed/refractory TCL induce responses in a minority of patients with relapse-free survival typically less than 4 months [4,5]. Therefore, new therapeutic agents that can more effectively treat TCL patients are urgently needed.
Like many other cancers, TCLs achieve resistance to therapy in part through overexpression of anti-apoptotic proteins in the Bcl-2 family, including Bcl-2, Bcl-xL, and Mcl-1 [6,7]. Recently utilizing the BH3 profiling, a functional assay that directly interrogates the dependence on Bcl-2 anti-apoptotic proteins, we showed that nearly all CTCLs and a subset of PTCLs, including some patientderived xenografts (PDXs) and primary patient samples, are dependent on Bcl-xL, which correlated with sensitivity to the Bcl-2/Bcl-xL inhibitor ABT263 (also known as navitoclax) but not the Bcl-2 inhibitor ABT199 (also known as venetoclax) [8]. These findings suggest that targeted inhibition of Bcl-xL could offer therapeutic benefits for some patients with TCL. Several Bcl-xL inhibitors, including the orally bioavailable ABT263, have been developed as new molecular targeted antitumor agents. However, these small molecule Bcl-xL inhibitors have failed during clinical development due to on-target and dose-limiting thrombocytopenia [9][10][11], as platelets are solely dependent on Bcl-xL for survival [9,12].
We recently reported the development of DT2216 as a selective Bcl-xL degrader that has the potential to be developed as a safe antitumor agent because it spares platelets [13]. DT2216 targets Bcl-xL to the Von Hippel-Lindau (VHL) E3 ligase for degradation by the proteasome [13]. Because platelets express minimal levels of VHL, DT2216 can potently degrade Bcl-xL in various tumor cells but not in platelets, in a VHL-and proteasome-dependent manner. DT2216 was highly potent against various Bcl-xL-dependent leukemia and cancer cells because it effectively inhibited the growth of several xenograft tumors as a single agent (e.g., MOLT-4 T-ALL xenograft) or in combination with other therapeutic agents (e.g., H146 SCLC xenograft, MDA-MB231 breast cancer xenograft, and T-ALL PDX models) [13]. These previous findings suggest that DT2216 has great clinical potential than ABT263 or other Bcl-xL inhibitors. Before moving this promising agent to clinic, more cell and animal models are urgently needed to evaluate its efficacy and safety. However, the effect of DT2216 on TCL was unknown. Therefore, here we examined the therapeutic potential of DT2216 against different TCL cell lines in vitro and in TCL xenograft and PDX models. The results from our studies show that targeting Bcl-xL using DT2216 can selectively kill Bcl-xLdependent TCL cells without causing significant platelet toxicity. Moreover, the combination of DT2216 with ABT199 may have broad therapeutic utility against TCLs that depend on both Bcl-xL and Bcl-2 for survival.

MyLa TCL xenograft mouse model and treatment
The animal work described in this study was approved by and done in accord with the Institutional Animal Care and Use Committee of the University of Florida. Female NOD-scid IL2Rg null (NOD/SCID) mice were purchased from The Jackson Laboratory (Stock No. 005557; The Jackson Laboratory, Bar Harbor, ME, USA) at 5 weeks of age. They received food and water ad libitum and were allowed to acclimatize for 1 week before being used for experiments at the age of 6 weeks. One million (1 × 10 6 ) MyLa cells were suspended in 100 μL of regular RPMI 1640 Medium (Cat. No. 22400089, Thermo Fisher Scientific, Waltham, MA, USA): Matrigel (Cat. No. 356231, Corning, New York, USA) 1:1 mixture and s.c. implanted in the right flank of NOD/SCID mice. Tumor growth was monitored daily and tumors were measured twice a week using Vernier caliper or digital calipers. Tumor volume was determined using the formula [(L × W 2 ) × 0.5], where L is the length/long dimension in millimeter (mm) and W is the width/short dimension in millimeter. The treatment started once the average tumor volume reached 100 mm 3 . The animals were randomly assigned to treatment groups (n = 8 mice per group at the start of treatment) such that each group had a nearly equal starting average tumor volume. Mice were weighed twice a week, and the treatments were given according to average mouse weight within each group before initiation of treatment. . Blood was collected 1 day after the first dose of the drugs for complete blood cell count (CBC) analysis. The mice were euthanized when the maximum tumor size reached the humane endpoint according to institutional policy concerning tumor endpoints in rodents. In addition, to prevent excessive pain or distress, the mice were euthanized if the tumors became ulcerated or the mice showed any signs of ill health. Mice were euthanized by CO 2 suffocation followed by cervical dislocation, and various tissues including tumors were harvested for further analyses.

DFTL-28776 T cell prolymphocytic leukemia (T-PLL) PDX model and treatment
Four-week-old female NOD/SCID mice from The Jackson Laboratory were allowed to acclimatize for 1 week before being used for experiments at the age of 5 weeks. DFTL-28776 cells were purchased from PRoXe (https://www. proxe.org/) and expanded once in NOD/SCID mice according to the instructions from PRoXe. They were injected into NOD/SCID mice via the tail veins at 1 × 10 6 cells/mouse. Mice were weighed twice a week and DFTL-28776 cell engraftment in blood was monitored weekly by measuring human CD45 + CD2 + cell engraftment in blood as described below. For the initial efficacy and survival study, DFTL-28776 cells were freshly harvested from the spleen of a host mouse with 60% of blood engraftment of DFTL-28776 cells and were immediately engrafted into NOD/SCID mice for further study. Mice with substantial engraftment of DFTL-28776 cells in blood (> 1.0%) were randomly assigned to different treatment groups (n = 5 mice for VEH and n = 4 mice for other groups at the start of treatment). They received treatment with vehicle, ABT199 (50 mg/kg, qd/p.o.), DT2216 (15 mg/kg, q4d/ i.p.), or DT2216 plus ABT199. The mice were euthanized when they became moribund according to institutional policy concerning tumor endpoints in rodents by CO 2 suffocation followed by cervical dislocation. The spleens were harvested from the mice for immunoblotting analyses of Bcl-xL, Bcl-2, and Mcl-1 expression. To assess additive vs. synergistic activity of DT2216 + ABT199 in DFTL-28776 TCL PDX mice, the Bliss independence model [17] was adapted to survival analysis as previously described [8]. A second study was performed to further characterize the effects of ABT199 and/or DT2216 on DFTL-28776 T-PLL PDX, in which previously frozen DFTL-28776 cells were used for the xenografts. Mice with a detectable engraftment of DFTL-28776 cells in blood (≥ 0.1%) were randomly assigned to different treatment groups (n = 6 mice per group); the average blood engraftment of DFTL-28776 cells per group was around 0.1%. They received treatment with vehicle, ABT263 (50 mg/kg, qd/p.o.), ABT199 (50 mg/kg, qd/p.o.), DT2216 (15 mg/kg, q4d/i.p.), or DT2216 plus ABT199. Vehicle-treated mice were euthanized 27 days after the initiation of the treatment due to their deteriorating condition, while the other groups were euthanized 7 days later. The spleen, liver, bone marrow, and blood were harvested. Age-matched female NOD/SCID control (CTL) mice (n = 5) without engraftment or any other treatments were included for comparative purposes. The weight of the spleens and livers from each mouse was measured and recorded. DFTL-28776 cell engraftment in the spleen, bone marrow, and blood were analyzed as described below.
Analysis of DFTL-28776 cell engraftment in the spleen, bone marrow, and blood

DFTL-28776 PDX cell purification and analyses
When DFTL-28776 cell engraftment in blood reached approximately 60%, DFTL-28776 PDX mice were euthanized by CO 2 suffocation followed by cervical dislocation. The spleens were harvested from the mice for the preparation of single splenocyte suspension. DFTL-28776 cells were isolated using the EasySep Mouse/Human Chimera Isolation Kit (Cat. No. 19849; StemCell Technologies, Vancouver, BC, Canada) according to the manufacturer's instructions. Cell viability of isolated DFTL-28776 cells after treatment with different agents was measured by tetrazolium-based MTS assay as described above with a slight modification. Briefly, DFTL-28776 cells were seeded into 96-well plates at 2 × 10 5 / well and treated for 24 h with various agents, the absorbance was read after adding the MTS reagents and incubation with cells overnight. All other assays for DFTL-28776 cells were done in a way similar to the assays for other TCL cell lines described above.
Hematoxylin & eosin staining and anti-hCD45 immunostaining of the spleens and livers of DFTL-28776 PDX mice The formalin-fixed tissues of mice from different groups were embedded in paraffin, sectioned into 5-μM slices, and mounted onto glass slides. Hematoxylin & eosin (HE) staining was done by the Molecular Pathology Core of the University of Florida. A set of slides was used for immunostaining of human CD45 (Cat. No. M0701, DAKO, Agilent Technologies, Inc., Santa Clara, CA, USA). Slides were immersed twice (5 min each time) in xylene solution followed by serial washes with 100, 95, and 75% ethanol, respectively (2 min each time). Then, the slides were dipped in water twice and then incubated in trilogy solution (Cat. No. 920P-10, Cell Marque Corporation, Rocklin, CA, USA) at 95°C for 25 min. After this incubation, the slides were washed under running water for 1 min and then washed with 1X TBST for 5 min. After incubation in 1.5% horse serum (Cat. No. ZE0122, Vector Laboratories, Inc., Burlingame, CA, USA) for 30 min, the slides were washed with TBST again and then sequentially incubated with avidin and biotin blocking solutions in a commercial kit (Cat. No. ZE0919, Vector Laboratories, Inc., Burlingame, CA, USA) following the manufacturer's instructions. The slides were washed with TBST and incubated with mouse monoclonal anti-human CD45 primary antibody (1:50) overnight followed by three washes with TBST. Slides were incubated with anti-mouse secondary antibody (1:200) for 30 min, washed again with TBST, followed by incubation with reagents in the Vectastain Alkaline Phosphatase Standard AK-5100 kit (Cat. No. ZE0628, Vector Laboratories, Inc., Burlingame, CA, USA). Finally, the slides were washed with TBST and mounted with a coverslip using the mounting medium with DAPI (Cat. No. ZD0421C, Vector Laboratories, Inc., Burlingame, CA, USA). The slides were visualized under a fluorescent microscope (Olympus BX43 with Olympus DP80 camera) at a × 200 magnification to obtain images.

Complete blood cell counts
Approximately 50 μL of blood was collected from each mouse in EDTA tubes via the submandibular plexus. The blood was immediately used for CBC analysis using the HEMAVET 950FS hematology analyzer (Drew Scientific, Miami Lakes, FL, USA). The data were expressed as the number of different blood cells or platelets per microliter of blood.

Statistical analysis
Most of the graphs presented in this manuscript were made and statistical analyses were performed using GraphPad Prism 7 software (GraphPad Software, San Diego, CA, USA). For analysis of means of three or more groups, analysis of variance (ANOVA) tests were performed. In the event that ANOVA justified post hoc comparisons between group means, the comparisons were conducted using Tukey's multiple comparison test. Two-sided unpaired Student's t test was used for comparisons between the means of two groups. Kaplan-Meier test was used for the survival rate analysis, and the data were statistically analyzed using log-rank (Mantel-Cox) test. The Bliss independence model was adapted to assess additive vs. synergistic activity of DT2216 + ABT199 in the DFTL-28776 TCL PDX mice as we previously described [8]. p < 0.05 was considered statistically significant.

DT2216 selectively kills Bcl-xL-dependent TCL cell lines by degrading Bcl-xL in a VHL-dependent manner
Using the BH3 profiling assay, our previous studies revealed that a fraction of human TCL cell lines are exclusively dependent on Bcl-xL for survival and are highly sensitive to Bcl-xL-specific and dual Bcl-2/Bcl-xL inhibitors [8]. We tested four Bcl-xL-dependent TCL cell lines (i.e., MyLa, MJ, MAC2A, and L82), and all of them expressed high levels of Bcl-xL and low levels of Bim (Fig. 1a) and were relatively resistant to conventional chemotherapy drugs, but were highly sensitive to DT2216 (EC 50 , 5-280 nM) (Fig. 1b and Table 1). In contrast, TCL lines (DL40, SMZ1, and FEPD) found to depend on Mcl-1 for survival by BH3 profiling [8] expressed relatively higher levels of Mcl-1 and were resistant to DT2216 (EC 50 > 10,000 nM) (Fig. 1a and Table 1). Notably, the three lines resistant to DT2216 lacked expression of Bcl-xL (Fig.  1a). All 7 cell lines expressed variable levels of Bax, Bak, and Noxa and similar levels of VHL (Fig. 1a). As we previously showed [13], human platelets were highly sensitive to ABT263 but very resistant to DT2216 compared with Bcl-xL-dependent TCL cells because platelets express minimal levels of VHL to degrade Bcl-xL (Fig. 1b, Table 1, and Supplementary  Fig. S1A, B).
To validate the mechanism of action of DT2216 against TCL cells, we used MyLa cells as a model system as they are solely dependent on Bcl-xL for survival and are highly sensitive to DT2216 (Table 1). We first examined if DT2216 kills MyLa cells via degrading Bcl-xL and inducing apoptosis. We found that DT2216 dose-and time-dependently decreased the expression of Bcl-xL but had no effect on the expression of BCL2L1 (the gene that encodes Bcl-xL) mRNA in MyLa cells (Fig. 1c, d and Supplementary Fig. S1C-E). The decrease in Bcl-xL expression was associated with the activation of caspase-3, cleavage of poly ADP ribose polymerase (PARP), and induction of apoptosis, all of which were abrogated by pretreating the cells with QVD-OPh (QVD, a pancaspase inhibitor) (Fig. 1c, d Fig. 1e-h, MyLa cells were pretreated with indicated compounds for 2 h, and proteins were measured 16 h after treatment with DT2216. k Pretreatment with VHL-L blocked the effect of DT2216 on cell viability. MyLa cells were pretreated with VHL-L for 2 h, and then treated with DT2216 for 72 h. The data are mean ± SD (n = 6 replicates). a p < 0.05 vs. DT2216. l DT2216, but not DT2216 NC, reduced the cell viability that was assayed after treatment with DT2216 or DT2216 NC for 72 h. The data are mean ± SD (n = 6 replicates). m DT2216-induced Bcl-xL degradation persisted up to 48 h upon the removal of DT2216 after 24 h DT2216 treatment. n The effect of DT2216, but not ABT263, on reducing MyLa cell viability, persisted up to 48 h upon the removal of ABT263 or DT2216 after 24 h treatment. The data presented are mean ± SD (n = 3 replicates) DT2216 also reduced the levels of Mcl-1, this effect was likely attributable to the activation of caspase-3 resulting from DT2216-induced Bcl-xL degradation. This suggestion is supported by the observation that inhibition of caspase activity with QVD abrogates the effect of DT2216 on Mcl-1 (Fig. 1c). Mcl-1 is a known caspase-3 substrate [18], and we previously showed that DT2216 does not bind to Mcl-1 and cannot degrade Mcl-1 directly [13].
Next, we tested whether DT2216 degrades Bcl-xL in a VHL-and proteasome-dependent manner in TCL cells. First, we examined the effects of ABT263 and VHL ligand (VHL-L) alone and in combination on Bcl-xL levels in MyLa cells and found that neither affected the level of Bcl-xL (Fig. 1e). In addition, we found that preincubation of MyLa cells with an excess amount of ABT263 or VHL-L blocked the degradation of Bcl-xL induced by DT2216 (Fig. 1f, g). Addition of the proteasome inhibitor MG-132 had the same effect (Fig. 1h). DT2216 had no effect on the levels of Bcl-xL in VHLnull cells (Fig. 1i). A negative control molecule, DT2216 NC, which binds Bcl-xL but does not bind VHL [13] was incapable of reducing the level of Bcl-xL and activating caspase-3 in MyLa cells (Fig. 1j). Moreover, the effect of DT2216 on MyLa cell viability was also dependent on its proteolysis targeting chimera (PRO-TAC) activity because pre-incubation of MyLa cells with an excess amount of VHL-L not only inhibited DT2216induced Bcl-xL degradation but also reduced the cytotoxicity of DT2216 (Fig. 1g, k). Finally, DT2216 NC showed no cytotoxicity against MyLa cells (Fig. 1l). Similar results were observed in other Bcl-xL-dependent TCL cell lines (Supplementary Fig. S2). Collectively, these data confirm that DT2216 acts as a PROTAC that depends on the VHL E3 ligase and proteasome to degrade Bcl-xL and induce apoptosis in Bcl-xLdependent TCL cells.
Because DT2216 functions as a Bcl-xL PROTAC, it can induce protein degradation and then be recycled and thereby function in a sub-stoichiometric manner. Unlike traditional occupancy-driven protein Bcl-xL inhibitors such as ABT263, its effects should not be limited by equilibrium occupancy [19][20][21][22][23]. Therefore, DT2216 should be more potent and have longeracting activity after washout than ABT263 [24][25][26][27][28][29][30]. This hypothesis is supported by the finding that DT2216 reduced the levels of Bcl-xL in MyLa cells even after DT2216 was removed from the culture for up to 48 h compared to cells without DT2216 treatment (Fig. 1m). Furthermore, MyLa cells resumed proliferation within 24 h after the removal of ABT263 in the culture, whereas the proliferation of MyLa cells remained to be suppressed even 48 h after the removal of DT2216 (Fig. 1n).
DT2216 is more potent against TCL but less toxic to platelets than ABT263 in vivo Next, we compared DT2216 to ABT263 in vivo by xenografting MyLa cells s.c. into immunodeficient mice. MyLa-engrafted mice were randomized into vehicle, ABT263 (50 mpk/qd/p.o.), or DT2216 (15 mpk/q4d/i.p.) treatment groups when their tumors reached~100 mm 3 (Fig. 2a). The dose of ABT263 used for this experiment was selected to avoid the ABT263 treatment-induced severe thrombocytopenia we observed in our recent study [13]. One day after the first treatment, the blood was collected from each mouse for a CBC analysis. ABT263treated mice had~73% reduction in blood platelet counts compared to vehicle-treated mice, whereas DT2216-treated mice had~16% reduction (Fig. 2b).
The results were calculated from a representative experiment with 6 replicates per concentration for each compound. Similar results were also got in two or more assays EC 50 the half maximal effective concentration, CTCL cutaneous TCL, ALK -ALCL anaplastic lymphoma kinase negative anaplastic large cell lymphoma, ALK + ALCL anaplastic lymphoma kinase positive anaplastic large cell lymphoma, PTCL-NOS peripheral TCL not otherwise specified, PLT platelets Neither treatment significantly affected other blood cell counts (Fig. 2c, d). Mice were continued on treatment until they became moribund or their tumors were close to 1000 mm 3 or ulcerated. Neither ABT263 nor DT2216 significantly affected body weight (Fig. 2e). ABT263 treatment slightly slowed the tumor growth and moderately extended the median survival time of the mice (from 8 to 14.5 days) compared to vehicle treatment (Fig. 2f, g). In contrast, DT2216 treatment markedly suppressed the primary tumor growth and extended the median survival time of the mice to 29.5 days (Fig. 2f, g). More importantly, the marked tumor suppression induced by DT2216 was associated with a more than 95% reduction in Bcl-xL levels, substantial decreases in Bcl-2 and Mcl-1 levels, and significant activation of caspase-3 and cleavage of PARP in the tumors collected from the mice (Fig. 2h-m). The decreases in tumor expression of Bcl-2 and Mcl-1 after DT2216 treatment were likely due to the activation of caspase-3 and induction of MyLa cell apoptosis (as shown in Fig. 1c) [18]. In contrast, ABT263 treatment had no significant effect on the tumor expression of these Bcl-2 family proteins, nor did it induce significant activation of caspase-3 and cleavage of PARP in the tumors (Fig. 2h-m). These findings demonstrate that DT2216 is more potent against MyLa TCL cells but less toxic to platelets than ABT263 in mice.

DT2216 can synergistically kill TCL PDX cells in combination with ABT199 in vitro
We previously established a panel of TCL PDX models and demonstrated that these TCL PDX models capture various aspects of TCL biology and found that multiple TCL models co-depend on more than one Bcl-2 family member [8,31]. We showed that a PDX of T cell prolymphocytic leukemia (T-PLL; DFTL-28776) is co-dependent on Bcl-xL and Bcl-2 [8]. In fact, treatment of DFTL-28776 PDX mice with the Mcl-1 inhibitor AZD5991 failed to induce any benefit compared to vehicle [8]. T-PLL is a rare but aggressive disease that involves the peripheral blood, bone marrow, lymph nodes, liver, and spleen. T-PLL is considered refractory to conventional chemotherapy and requires allogeneic stem cell transplantation for cure [32]. As expected, DFTL-28776 cells were highly resistant to doxorubicin, etoposide, and vincristine ( Fig. 3a and Supplementary Table S1). They were also relatively resistant to ABT199, A-155463 (a selective Bcl-xL inhibitor), and S63845 (a selective Mcl-1 inhibitor) (Fig. 3b and Supplementary Table S1). However, ABT263 potently induced cell death, confirming dependence on both Bcl-xL and Bcl-2 for survival ( Fig. 3b and Supplementary Table S1). Of note, DFTL-28776 cells express high levels of Bcl-xL, Bcl-2, and Mcl-1 (Fig. 3c), indicating that protein levels are not predictive of response. DT2216 induced Bcl-xL degradation in DFTL-28776 cells in vitro in a dose-dependent fashion (Fig. 3d). Both VHL-L and MG-132 blocked Bcl-xL degradation induced by DT2216, consistent with the effects being dependent on VHL and the proteasome (Fig. 3e, f). Because DFTL-28776 cells depend on both Bcl-2 and Bcl-xL, we examined whether DT2216 could more effectively kill the cells in combination with ABT199. The combination of DT2216 and ABT199 (ratio 1:1) synergistically killed DFTL-28776 cells in vitro via induction of apoptosis (Fig. 3g-i) and their combined effects were similar to ABT263 (Fig. 3b, g).

DT2216 is more effective against T-PLL PDX cells when combined with ABT199 in vivo
We injected DFTL-28776 PDX cells via the tail vein of NOD/SCID mice, which results in orthotopic engraftment in the bone marrow, spleen, liver, and blood [8,31]. Mice with > 1% engraftment of DFTL-28776 PDX cells in blood were randomly assigned to four different treatment groups (Supplementary Fig. S3). They received treatment with vehicle, DT2216, ABT199, or DT2216 plus ABT199 (Fig. 4a). ABT199 had minimal effects on blood T-PLL burden and modestly increased the median survival (17 days vs. 26 days in vehicle-treated mice); DT2216 was slightly more effective than ABT199 in inhibiting DFTL-28776 PDX cell blood engraftment and prolonging survival (Fig. 4b, c and Supplementary Fig.  S3). In contrast, the combination of DT2216 and   (Fig. 4b, c and Supplementary Fig. S3). More importantly, the survival analysis of the mice using the Bliss independence model reveals that DT2216 plus ABT199 achieved synergy based on a significantly superior survival (p = .020) compared with that expected from additivity. These effects of DT2216 were associated with a significant reduction in Bcl-xL expression in spleen DFTL-28776 PDX cells (Fig. 4d, e). Of note, ABT199 treatment led to Mcl-1 upregulation in spleen T-PLL cells, which was blocked by the addition of DT2216 (Fig.  4d, g). None of the treatments affected the body weight of the mice (Fig. 4h).
To better characterize the effects of ABT199 and/or DT2216 on DFTL-28776 PDX, we repeated the study presented in Fig. 4a but initiated the treatments when the mice had a mean blood DFTL-28776 PDX cell engraftment of around 0.1%. The vehicle-treated mice were humanely euthanized 27 days after the initiation of the treatments when the mice became nearly moribund due to a high burden of the disease (Fig. 5a and Supplementary Fig. S4). The remaining groups of mice were terminated 7 days later. This experiment allowed us to analyze the effects of different treatments on not only blood engraftment of DFTL-28776 PDX cells, but also the engraftment of DFTL-28776 PDX cells in the spleen, liver, and bone marrow. In addition, we added the ABT263 treatment group into the study to compare the effects of Bcl-xL inhibition vs. degradation on DFTL-28776 PDX. The results from this study confirmed that the combination of DT2216 and ABT199 was more effective than either of these agents alone in suppression of DFTL-28776 PDX cell blood engraftment as shown in the previous study (Figs. 4b and 5b). ABT263 was also more effective than ABT199 or DT2216, but less effective than DT2216 plus ABT199, in inhibition of DFTL-28776 PDX cell blood engraftment ( Fig. 5b and Supplementary Fig. S4). ABT199 had no significant effect on the size/weight of the spleen and liver and DFTL-28776 PDX cell engraftment in the spleen and bone marrow, whereas DT2216 only moderately reduced DFTL-28776 PDX cell engraftment in the spleen but had no significant effect on all other parameters measured (Fig. 5b-j). In contrast, DT2216 plus ABT199 normalized the size/weight of the spleen and liver and substantially reduced DFTL-28776 PDX cell engraftment in the spleen and bone marrow. ABT263 also normalized the size/ weight of the liver but was less effective in suppression of DFTL-28776 PDX cell engraftment in the spleen and had no effect on DFTL-28776 PDX cell engraftment in the bone marrow ( Fig. 5b-j). More importantly, the combination treatment with DT2216 and ABT199 did not result in further reduction in blood platelets and white blood cells (WBCs) than the treatment with either agent alone (Supplementary Fig. S4B, C). In fact, the DFTL-28776 PDX mice treated with DT2216 plus ABT199 exhibited less reduction in blood platelets and WBCs than the mice treated with ABT199. This effect may be attributed to a better control of Data are presented as mean ± SEM (n = 5 for VEH and n = 4 for other groups at the start of treatment). c Survival curve showing the survival of mice after the TCL PDX cell engraftment. Median survival time within each treatment group is presented along with statistical analysis results (n = 5 for VEH and n = 4 for other groups at the start of treatment). *p < 0.05 and **p < 0.01 in indicated comparisons; ns, not significant. The orange line indicates a calculated additive model, with 95% confidence intervals indicated as a yellow shaded range. p value of synergy of DT2219 + ABT199 over additivity: 0.020. d Immunoblot analysis of Bcl-xL, Bcl-2, and Mcl-1 in the spleen (n = 3 mice in each group). Proteins were extracted from the spleens of mice after euthanization when they reached the experimental endpoint as described in the "Materials and methods" section. e-g Quantification of Bcl-xL, Bcl-2, and Mcl-1 levels in d. The data presented are mean ± SEM (n = 3 mice per each group). a, b, and c, p < 0.05 vs. VEH, DT2216, and ABT199, respectively. h Body weight changes over time in DFTL-28776 PDX cell-bearing mice after DT2216 and/or ABT199 treatment as shown in Fig. 5. The data presented are mean ± SEM (n = 5 for VEH and n = 4 for other groups at the start of treatment) TCL by the combination treatment, which likely further lessens the bone marrow suppression caused by TCL bone marrow infiltration to promote the production of platelets and WBCs. Collectively, these results demonstrate that the combination of DT2216 and ABT199 is an effective treatment for TCLs that rely on both Bcl-xL and Bcl-2 for survival. The synergy between DT2216 and ABT199 is potentially greater than dual inhibition of Bcl-xL and Bcl-2 by ABT263, which also induces dose-limiting thrombocytopenia in humans.

Discussion
Our findings demonstrate that targeting Bcl-xL using DT2216 can selectively kill Bcl-xL-dependent TCL cells without causing significant platelet toxicity. Moreover, the combination of DT2216 with a Bcl-2 inhibitor can also effectively target TCLs that depend on both Bcl-xL and Bcl-2 for survival. The ability to combine DT2216 with ABT199 in vivo suggests that DT2216 could be used in many different combinations to synergistically increase TCL killing without inducing thrombocytopenia, which cannot be achieved by the use of either Bcl-xlspecific inhibitors such as A-1155463 or the Bcl-xL/2dual inhibitor ABT263 because Bcl-xL is essential for platelet survival and these inhibitors causes severe thrombocytopenia [33]. The improved potency of DT2216 against TCLs compared to Bcl-xL inhibitors is likely attributed to its PROTAC property. As conventional Bcl-xL inhibitors, ABT263 or A-1155463 requires constant occupancy of the BH3-binding sites to sequester Bim and other proapoptotic Bcl-2 family proteins [34]. The effectiveness of ABT263 or A-1155463 in vivo against TCLs depends on the maintenance of long-term and high-level drug exposure, which is difficult to achieve without causing significant toxicity [9,12]. In contrast, DT2216 acts as a Bcl-xL PROTAC that can catalytically induce Bcl-xL degradation in a sub-stoichiometric manner in TCL The data presented are mean ± SEM (n = 6 mice per each group). c Spleen DFTL-28776 PDX cell burden in the PDX xenograft mice after treatment as shown in a. The data presented are mean ± SEM (n = 6 mice per each group). a, b, c, and d, p < 0.05 vs. VEH, ABT263, DT2216, and ABT199, respectively. d Bone marrow DFTL-28776 PDX cell burden in the DFTL-28776 PDX xenograft mice after treatment as shown in a. The data presented are mean ± SEM (n = 6 mice per each group). a, b, c, and d, p < 0.05 vs. VEH, ABT263, DT2216, and ABT199, respectively. e Liver images of DFTL-28776 PDX cell-engrafted mice. f Liver weight in DFTL-28776 PDX xenograft mice after treatment as shown in a. The data presented are mean ± SEM (n = 5 for the CTL group and n = 6 mice for other groups). a, b, c, d, and e, p < 0.05 vs. CTL, VEH, ABT263, DT2216, and ABT199, respectively. g Spleen images of DFTL-28776 PDX cell engrafted mice. h Spleen weight in DFTL-28776 PDX xenograft mice after treatment as shown in a. The data presented are mean ± SEM (n = 5 for the CTL group and n = 6 mice for other groups). a, b, c, d, and e, p < 0.05 vs. CTL, VEH, ABT263, DT2216, and ABT199, respectively. i Hematoxylin & eosin (HE) staining of the spleens and livers from DFTL-28776 PDX cell-engrafted mice. j Human CD45 (hCD45) staining of the spleens and livers from DFTL-28776 PDX cell-engrafted mice. Red color, hCD45 staining with vectastain; blue color, nuclear staining with DAPI cells but not in platelets because VHL is abundantly expressed in TCL cells but is poorly expressed in platelets [13]. As such, the effect of DT2216 is not limited by equilibrium occupancy and does not cause thrombocytopenia. In addition, the catalytic nature of DT2216 as a PROTAC results in an extended duration of action beyond the plasma pharmacokinetic half-life. Therefore, its therapeutic effect is longlasting and requires only once or twice a week dosing to be effective in vivo [13], which makes DT2216 not only therapeutically more effective, but also less toxic than ABT263 or other Bcl-xL-specific inhibitors. Furthermore, after ABT263 is converted into DT2216, the PROTAC molecule becomes a specific Bcl-xL degrader without inducing Bcl-2 degradation. Similar phenomena have been reported previously [35,36]. The lack of efficient degradation of Bcl-2 by DT2216 may be attributed to the significant reduction of its binding affinity to Bcl-2 [14], lack of the lysine for ubiquitination, and inability of DT2216 to form ternary complex with VHL and Bcl-2 in cells [13]. Since Bcl-2 inhibition with ABT-263 can lead to neutropenia and lymphocytopenia [37], the inability of DT2216 to induce Bcl-2 degradation may confer to DT2216 additional advantages by reducing the ontarget toxicity of ABT263 on Bcl-2.
As expected, single-agent DT2216 was unable to cure mice engrafted with TCLs. The mechanisms by which TCL cells evade DT2216 (or combination with ABT199) have yet to be elucidated. One possibility is that that specific microenvironments can serve as sanctuaries. This could involve the suppression of pro-apoptotic factors through various soluble factors, cell adhesion molecules, and extracellular matrix molecules. In fact, previous studies have implicated a range of factors, including CXCL12/CXCR4, NOTCH ligands and receptors, VCAM-1/VLA-4 and 5, fibronectin/VLA-4 and 5, hyaluronate and collagen/CD44 [38,39]. Alternatively, TCL cells may upregulate other Bcl-2 anti-apoptotic proteins to compensate for the loss of Bcl-xL, as observed in patients who developed adaptive resistance after long-term responses to venetoclax [40,41]. Further studies are needed to clarify the mechanisms, first in preclinical models and then in humans, in order to design rational combinations that obviate the development of resistance to DT2216.

Conclusions
In summary, DT2216 is highly active both in vitro and in vivo against TCL cells with dependence on Bcl-xL and does not cause significant thrombocytopenia. Combinations that target additional dependences do not increase the toxicity of DT2216, suggesting that it could enhance the efficacy of a broad range of therapeutics.
Additional file 1: Supplementary Figure 1. Effect of DT2216 on BCL2L1, BCL2, and MCL1 mRNA expressions in TCL cells. Supplementary  Figure 2. Effect of DT2216 on MJ, MAC2A and L82 TCL cells. Supplementary Figure 3. Effect of DT2216 and/or ABT199 on TCL PDX cell blood engraftment and body weight in TCL PDX mice. Supplementary Figure 4. The combination therapy of DT2216 and ABT199 is more effective than either agent alone or ABT263 against TCL PDX in mice. Supplementary Table 1. Effects of chemotherapy drugs, Bcl-2 family protein inhibitors and DT2216 on DFTL-28776 TCL PDX cells in vitro.