225Ac‐labeled CD33‐targeting antibody reverses resistance to Bcl‐2 inhibitor venetoclax in acute myeloid leukemia models

ABSTRACT Purpose Despite the availability of new drugs, many patients with acute myeloid leukemia (AML) do not achieve remission and outcomes remain poor. Venetoclax is a promising new therapy approved for use in combination with a hypomethylating agent or with low‐dose cytarabine for the treatment of newly diagnosed older AML patients or those ineligible for intensive chemotherapy. 225Actinium‐lintuzumab (225Ac‐lintuzumab) is a clinical stage radioimmunotherapy targeting CD33 that has shown evidence of single‐agent activity in relapsed/refractory AML. Increased expression of MCL‐1 is a mediator of resistance to venetoclax in cancer. Experimental design Here we investigated the potential for 225Ac‐lintuzumab‐directed DNA damage to suppress MCL‐1 levels as a possible mechanism of reversing resistance to venetoclax in two preclinical in vivo models of AML. Results We demonstrated that 225Ac‐lintuzumab in combination with venetoclax induced a synergistic increase in tumor cell killing compared to treatment with either drug alone in venetoclax‐resistant AML cell lines through both an induction of double‐stranded DNA breaks (DSBs) and depletion of MCL‐1 protein levels. Further, this combination led to significant tumor growth control and prolonged survival benefit in venetoclax‐resistant in vivo AML models. Conclusions There results suggest that the combination of 225Ac‐lintuzumab with venetoclax is a promising therapeutic strategy for the treatment of patients with venetoclax‐resistant AML. Clinical trial of this combination therapy (NCT03867682) is currently ongoing.


| 1129
GARG et Al. allogeneic hematopoietic cell transplantation (HCT) can result in long-term survival in approximately 30% to 40% of patients with relapsed AML, many patients are not suitable candidates for transplant due to age, comorbidities, or lack of a matched donor. 1 The prognosis for older patients is even worse, with a 5-year median survival rate of 5% for patients older than the age of 65. 2 Within the last few years, following decades of limited advancement in the treatment of AML, several new targeted therapies have been approved in the US. 3,4 Notably, the BCL-2 inhibitor venetoclax (ABT-199) was approved for use in combination with a hypomethylating agent (HMA) or with low-dose cytarabine (LDAC) for the treatment of newly diagnosed patients with AML who are age 75 years or older or are ineligible for intensive chemotherapy. 5,6 Approval was based on the M14-358 and M14-387 phase Ib/II trials. In M14-358, the combination of venetoclax with azacitidine led to a complete remission (CR) rate of 37% and a CR with partial hematological recovery (CRh) rate of 24%. 5 For the combination of venetoclax and decitabine, the rates were 54% and 7.7%, respectively. In the M14-387 trial, venetoclax in combination with lowdose cytarabine (LDAC) led to 21% CR and CRh rates. 6 For patients receiving all dosages of venetoclax in these studies, median overall survival was 17.5 months. 5 Despite the significant benefit to patients in this population, not all patients respond to initial therapy with venetoclax and most patients will eventually progress. Preclinical studies have investigated mechanisms of resistance, which are supported by clinical results in patients with progressive disease. Importantly, venetoclax as a B cell lymphoma 2 (BCL-2) selective inhibitor does not inhibit other BCL family members such as myeloid cell leukemia 1 (MCL-1) or B lymphoma extra-large (BCL-XL). Mechanistically, overexpression of these other antiapoptotic BCL-2 family members, in particular MCL-1, or their upregulation in response to venetoclax has been shown to mediate resistance to venetoclax in leukemia, lymphoma, and multiple myeloma. [7][8][9][10][11] Further, MCL-1 has been shown to be upregulated in AML patients at relapse following induction chemotherapy. 12 Strategies to reduce MCL-1 levels may therefore dramatically prolong the response to venetoclax and re-sensitize resistant tumors to venetoclax therapy. MCL-1 protein has a very short half-life of less than 1 hour and is therefore sensitive to changes in RNA or protein synthesis. 13 To that end, genotoxic stress as a result of DNA damage, for example, by UV or ionizing radiation, or chemo-induced, can effect a reduction in MCL-1 levels via inhibition of protein synthesis. 14,15 In turn, the combination with chemotherapy or ionizing radiation can increase sensitivity to BCL-2 inhibitors in preclinical tumor cell lines and patient samples by a reduction in MCL-1 levels. 7,14,16 225 Aclintuzumab is a clinical stage radioimmunotherapy targeting CD33 that has shown evidence of single-agent activity in relapsed/refractory AML [17][18][19] and is being actively investigated in combination clinical studies in the treatment of relapsed or refractory AML. 225 Ac-lintuzumab delivers the 225 Ac payload, a high energy, short path length, alpha emitting radionuclide directly to CD33-positive myeloid tumor cells creating lethal double-strand breaks in DNA and leading to selective tumor cell killing. 17 One hit of an alpha particle emitted by 225 Ac radionuclide can potentially kill a tumor cell, and the short path length focuses its radiation energy on targeted tumor cells, limiting exposure and damage to adjacent normal tissue. 20 While external beam radiation has been combined with venetoclax in preclinical models, there is the potential to damage normal tissue. Further, scheduling radiation treatment relative to venetoclax administration may also be a challenge. As a potent inducer of DNA damage in targeted tumor cells, it is anticipated that 225 Ac-lintuzumab may mediate effective down-modulation of MCL-1 leading to a sensitization of AML cells to venetoclax irrespective of inherent resistance to the BCL-2 inhibitor. Thus, the combination of potent targeted alpha radioimmunotherapy (RIT) plus venetoclax may be an effective combination strategy in AML. Here we describe the results of our in vitro and in vivo studies which evaluated 225 Ac-lintuzumab plus venetoclax in combination in established AML tumor cell lines exhibiting varying sensitivity to venetoclax. Unlike other strategies to deplete MCL-1 in combination with venetoclax, our results demonstrate that 225 Ac-lintuzumab-targeted internal alpha radiation exerts a dual mechanism of action, effecting potent single-agent tumor killing through DNA double-strand breaks, and the reduction in antiapoptotic proteins such as MCL-1, leading to re-sensitization of tumor cells to venetoclax and potent antitumor activity in preclinical models.

| Lintuzumab antibody conjugation and radiolabeling with 225 Ac
The lintuzumab was attached to the bifunctional chelating agent p-SCN-Bn-DOTA (Macrocyclics) by modified methodology reported in. 21 Five hundred microgram of lintuzumab was attached to the bifunctional DOTA employing a 50-fold molar excess of the chelating agent over antibody and incubated at 37°C for 1.5 h in 0.15 M sodium carbonate buffer, pH 8.5 (Cat. No.: 7527-04, VWR). The conjugation was followed by the purification of lintuzumab-DOTA via buffer exchange into 0.15 M ammonium acetate, pH 6.5 (Cat. No.: 11599, VWR) at 4°C. The number of DOTA molecules attached to the lintuzumab molecule in the final conjugate was measured by MALDI-TOF (University of Alberta) and determined to be approximately 10:1. The preservation of DOTA-lintuzumab immunoreactivity toward CD33 antigen was confirmed by ELISA.
Lintuzumab-DOTA conjugate was radiolabeled with 225 Ac by dissolving 225 Ac nitrate in 0.01 M HCl followed by 60 min incubation at 37°C with the conjugate at 37 kBq/ µg ratio. DTPA (0.05 M solution) was used to stop the radiolabeling reaction and radiolabeling yield was determined by performing instant thin layer chromatography (iTLC) on Silica Gel-Glass paper strips (Cat. No.: SG10001, Agilent Technologies) immediately after radiolabeling and counting the radioactivity on the iTLC strips 24 h after. At that time, 225 Ac is in equilibrium with its daughter isotopes. A 2470 Wizard2 Gamma counter (Perkin Elmer) calibrated for the 225 Ac radioactive daughter 213 Bi emission spectrum was utilized to count radioactivity of iTLC strips. The radiolabeled antibody is found at the point of application on the iTLC strips, while 225 Ac in form of 225 Ac-DTPA travels with the solvent front. Radiolabeling yields were greater than 99% and no further purification was required. The stability of 225 Aclintuzumab (formally known as 225 Ac-HuM195) in vitro and in vivo was determined previously 22 and showed no leakage of Ac from the product.

DNA double-strand breaks
The presence of phosphorylated H2A.X was measured by flow cytometry according to the manufacturer's instructions (Cat. No.: 17-344, Upstate Cell Signaling, Millipore Sigma). Briefly, U937 and OCI-AML3 cells were treated with media, venetoclax alone (0.5 µM), 225 Ac-Lintuzumab (1.48 kBq), or combination with venetoclax. After 72 h treatment, cells (2 × 10 6 cells/ml) were washed twice with PBS and then resuspended in fixation solution (provided in kit). Cells were incubated for 20 min on ice and then washed twice with PBS. Subsequently, cells were resuspended in permeabilization solution (provided in kit) and stained with 3.5 µL of FITC-conjugated anti-phospho-H2A.X (Ser139) or isotype control mouse IgG (provided in kit). After brief incubation for 20 min on ice, the cells were washed and resuspended in FACS buffer (1X PBS containing 0.2% BSA and 0.03% sodium azide). Stained samples were analyzed with a CytoFLEX flow cytometer (Beckman Coulter, Mississauga, ON) and data were analyzed using FlowJo software (Version 10, Tree Star Inc.).

| In vivo efficacy studies
Animal studies were approved by the University of Saskatchewan's Animal Research Ethics Board and followed to the Canadian Council on Animal Care guidelines for humane animal use. Tumor xenografts were established in 6-week-old female SCID mice (Strain code: 236, CB17/Icr-Prkdc scid /IcrIcoCrl, obtained from Charles River Laboratories, Saint-Constant, QC, Canada) by subcutaneous injection of 2x10 6 OCI-AML3 or U937 cells into the right flank. Tumor growth was measured with electronic calipers every 3 days (volume = length × width 2 /2). When tumors reached an volume of ~200 mm 3 , tumor-bearing mice were randomized into five treatment groups of five animals and were treated with on day 1: IP injection of 0.4 μg unlabeled lintuzumab; IP injection of 7.4 kBq 225 Ac-lintuzumab; venetoclax (200 mg/kg) by oral gavage; combination of venetoclax and 7.4 kBq 225 Ac-lintuzumab; or left untreated. Mice in venetoclax alone and in combination treatment groups continued to receive venetoclax via oral gavage once daily over a period of 20 days. Mice body weights and tumor volumes were recorded three times per week. The animals were humanely euthanized when they experienced excessive weight loss (≤20%), or any tumor reached 4000 mm 3 volume or became necrotic. The study was terminated at 38 days when the difference in survival between the treatment and control groups became significant. For safety study, blood was collected from euthanized mice and analyzed for liver toxicity (aspartate aminotransferase; AST and alanine transaminase; ALT) and kidney toxicity (creatinine and blood urea nitrogen; BUN).

| Statistical analysis
Power analysis for the in vivo studies was estimated using PASS version 11 (NCSS, Inc.) using simulations of different tumor volumes based on pilot data and conservative assumptions regarding the groups treated with the radiolabeled antibodies. All simulations showed power of at least 83% with only five animals per group because of the large differences between treated and untreated animals. Thus, five mice per group were utilized in the in vivo studies. GraphPad Prism 7 was used to analyze all the data (GraphPad Software, Inc.). Differences between the treated and untreated groups in vitro were assessed using nonparametric Kruskal-Wallis test with Dunn's correction for multiple comparisons. Error bars represent ±standard deviation (SD). Kaplan-Meier data were analyzed by logrank (Mantel-Cox) test.

| Single-agent cytotoxicity of venetoclax and 225 Ac-lintuzumab in AML cell lines
Initially, two venetoclax resistant (OCI-AML3 and U937) and one sensitive (MOLM-13) AML cell lines were screened for CD33 expression. Flow cytometry analysis revealed high expression of CD33 in all three AML cell lines ( Figure S1). To determine the optimal dose for in vitro combination studies, we evaluated the sensitivity of these cell lines to singleagent venetoclax or 225 Ac-lintuzumab. The AML cell lines were treated individually with increasing concentrations of venetoclax or 225 Ac-lintuzumab and cell viability measured after 72 h by XTT assay. U937 and OCI-AML3 cells began to show evidence of cytotoxicity at 0.5 µM venetoclax treatment, whereas MOLM-13 cells were shown to be sensitive to concentrations as low as 0.005 µM ( Figure 1A). The IC 50 for venetoclax in MOLM-13 was determined to be 0.0085 µM in comparison to venetoclax-resistant lines U937 (IC 50 s 0.663 µM) and OCI-AML3 (IC 50 s 1.115 µM). Treatment of U937 and OCI-AML3 cells with 225 Ac-lintuzumab showed modest sensitivity following 1-hour exposure to the antibody radio-conjugate. MOLM-13 was shown to be more sensitive to 225 Ac-lintuzumab at concentrations as low as 1.48 kBq ( Figure 1B).

| 225 Ac-lintuzumab and venetoclax combination shows synergetic cytotoxicity toward both sensitive and resistant AML cell lines
Based on results of single-agent cytotoxicity of venetoclax and 225 Ac-lintuzumab, we performed studies to assess the potential for enhanced cell killing with the combination of 225 Ac-lintuzumab with venetoclax in both sensitive and resistant AML lines. In vitro combination cell cytotoxicity studies utilized IC 50 concentrations determined from the single-agent analyses: venetoclax, 0.05 µM (MOLM-13) or 0.5 µM (U937 and OCI-AML3) and 1.48 kBq ( 225 Ac -lintuzumab). The cells were first exposed to 225 Ac-lintuzumab in the presence of venetoclax, then washed, and subsequently incubated in medium containing fresh venetoclax. Inhibition of cell growth was measured after 72 h posttreatment by XTT assay. Treatment with 225 Ac-lintuzumab combined with venetoclax resulted in greater tumor cell killing as compared to individual treatment in all three cell lines ( Figure 2). Importantly, the enhancement of cell killing in the two venetoclax-resistant lines-4.5-fold increase in cell growth inhibition by combination treatment in comparison with single treatments for U937 cells and 2-fold for OCI-AML3 cells-suggests that the combination of venetoclax with 225 Ac-lintuzumab may affect re-sensitization of these lines to venetoclax.

| Initiation of double-strand breaks in AML cells exposed to 225 Ac-lintuzumab
The 225 Ac-lintuzumab carries the 225 Ac radionuclide generating potent high linear energy transfer alpha particles which can cause lethal DNA double-strand breaks (DBSs). Upon induction of DNA damage, the nucleosomal histone protein H2A.X is rapidly phosphorylated at serine 139 to γ-H2A.X at the DSBs site, making phosphorylated γ-H2A.X a sensitive marker for detection of DNA DSBs. 25 To define the potential mechanisms by which 225 Ac-lintuzumab induces significantly enhanced cellular cytotoxicity in AML cell lines, we screened for phosphorylated H2A.X following exposure to the antibody radio-conjugate. Treatment of both U937 and OCI-AML3 cells with 225 Ac-lintuzumab alone or its combination with venetoclax induced high levels of γ-H2A.X as compared to venetoclax with synergistic effect observed for combination treatment (Figure 3 and Figure S2).

1, BCL-2, and BCL-XL levels in treated AML cells
It has been reported that genotoxic stress may inhibit the expression of BCL-2 family proteins such as MCL-1, 14,15 a protein implicated in resistance to BCL-2 inhibitors such as venetoclax. To evaluate the potential for 225 Ac-lintuzumab to modulate MCL-1 levels, we assessed the levels of antiapoptotic proteins in OCI-AML3 cells via western blot. We determined that transient exposure of OCI-AML3 cells to 225 Ac-lintuzumab significantly reduced the levels of not only MCL-1, but also BCL-XL, and to a lesser extent BCL-2, in a dose-dependent manner. This decrease in BCL-2 family protein levels may serve as a mechanism by which this AML cell line can become re-sensitized to venetoclax following combination treatment with 225 Ac-lintuzumab (Figure 4 and Figure S3).

| Robust antitumor efficacy and survival of combination therapy in venetoclaxresistant xenografts
Since significant antitumor cell cytotoxicity was observed with combination treatment of 225 Ac-lintuzumab and venetoclax in AML cell lines, we evaluated this combination therapy in AML subcutaneous transplantation models using OCI-AML3 and U937 xenografts in SCID mice. Vehicletreated, venetoclax-treated, and single dose of naked unlabeled lintuzumab mAbs (0.4 µg) were unable to control tumor burden in mice in both models OCI-AML3 ( Figure 5A-C) and U937 ( Figure 6A-C). Single-agent venetoclax treatment showed no antitumor activity, although a modest survival benefit was observed in these resistant models compared to vehicle control and unlabeled lintuzumab antibody. Notably, both single-agent 225 Ac-lintuzumab and 225 Ac-lintuzumab in combination with venetoclax significantly reduced tumor burden resulting in increased survival in both OCI-AML3 ( Figure 5D and E) and U937 ( Figure 6D and E) models. Furthermore, at the end of the OCI-AML3 animal study, on day 38 posttreatment, the group treated with 7.4 kBq 225 Aclintuzumab alone demonstrated 80% survival, with complete response in two of five mice (Table 1), while the group treated with the combination of 225 Ac-lintuzumab with venetoclax showed 100% survival and CR in three of five mice ( Figure 5F). We chose to stop the study at 38 days' time point as by that time the difference in survival between the control and treatment groups became significant. Interestingly, in the U937 animal model, there was 80% survival with the combination of 7.4 kBq 225 Ac-lintuzumab with venetoclax and CR in two of five mice (Table 2), while the animals treated with the 225 Ac-lintuzumab alone demonstrated 20% survival ( Figure 6F) with CR in one of five mice. Altogether these F I G U R E 2 Combination treatment with 225 Ac-Lintuzumab and venetoclax (veneto) induces enhanced cytotoxicity in AML cell lines. MOLM-13 (A), U937 (B) and OCI-AML3 (C) cell lines were pretreated with 225 Ac-Lintuzumab for 1 h, washed, and then incubated with 0.05 µM (MOLM-13) or 0.5 µM (U937 and OCI-AML3) venetoclax for 72 h. Cell viability was measured by XTT assay. Data are shown as mean with SD of percentage of inhibition of three technical replicates per cell type. Experiment was repeated thrice with similar results. Statistics was calculated using nonparametric Kruskal-Wallis test with Dunn's correction for multiple comparisons. results demonstrate robust antitumor control and survival of the 225 Ac-lintuzumab combination with venetoclax in AML models refractory to single-agent venetoclax.

Ac-lintuzumab and venetoclax combination in tumor-bearing mice
A safety evaluation was performed on mice treated with venetoclax or 225 Ac-lintuzumab single agent or in combination by measuring the weight, liver toxicity (AST and ALT), and kidney toxicity (creatinine and BUN). 225 Ac-lintuzumab alone and the combination with venetoclax demonstrated transient but reversible weight loss ( Figure 7A). There were no significant changes in AST ( Figure 7B) and ALT ( Figure 7C) levels in any of the treatment groups, though one of five mice in the venetoclax single-agent cohort exhibited elevated enzyme levels. Furthermore, we did not find any evidence of increased creatinine ( Figure 7D) and BUN ( Figure 7E) levels indicating the absence of kidney toxicity in mice in any of the treatment groups.

| DISCUSSION
Acute myeloid leukemia (AML) is a complex hematological disease often occurring in older patients. While several new targeted therapies have been recently approved, most patients eventually relapse and succumb to their disease. Strategic clinical approaches applying rational and scientifically supported therapeutic combinations are a logical step to further extend the response to, and benefit of, these novel targeted agents. To the best of our knowledge, this is the first study describing the combination of the recently approved venetoclax agent and radioimmunotherapy with an alpha emitter in AML. In this study, we demonstrated in vitro and in vivo that the combination of venetoclax with 225 Ac-lintuzumab, an antibody armed with the potent alpha emitting radionuclide Actinium-225, had a synergistic effect with venetoclax on tumor cell killing in vitro and the combination provided a pronounced survival advantage in AML tumor models shown to be resistant and refractory to venetoclax.
BCL-2 family proteins can promote tumor cell survival, making them attractive drug targets. 3,26,27 Antagonists targeting BCL-2/BCL-XL (ABT-737 and ABT-263/ navitoclax) or BCL-2 only (ABT-199/GDC-0199/venetoclax) have shown clinical benefit, notably leading to the recent regulatory approval of venetoclax in CLL and AML. Not surprisingly, one mechanism of resistance to venetoclax is mediated by upregulation or overexpression of other BCL-2 family members, including MCL-1 and BCL-XL. 7-11 225 Ac-lintuzumab targets the myeloid specific marker CD33 found overexpressed on most tumor cells in AML and MDS and less frequently in MM. 17 malignances. 18,19 Importantly, 225 Ac-lintuzumab internalizes on binding to its CD33 antigen into the cells and effectively carries 225 Ac inside the cells 22 thereby limiting off-target toxicities and contributing to a high degree of patient tolerability in clinical trials. [17][18][19] In the present study, we demonstrated robust antitumor activity and survival from the combination of 225 Ac-lintuzumab with venetoclax in AML lines refractory to the BCL-2 inhibitor. It has been shown that DNA damage can reduce the expression of MCL-1 protein. 15,30 Therefore, the use of targeted therapies that increase tumor-specific DNA damage may be a promising approach to enhance the potency of venetoclax and overcome resistance mechanisms. As anticipated, exposure of AML cell lines to 225 Ac-lintuzumab potently induced DNA double-strand breaks and demonstrated single-agent activity in vitro and in vivo. Importantly, the combination with venetoclax had an apparent synergistic effect on DNA damage that in turn reduced the expression of MCL-1 protein leading to a significant enhancement of antitumor potency in AML cell lines resistant to single-agent venetoclax. To investigate the possible mechanisms by which 225 Ac-lintuzumab may potentiate or re-sensitize resistant AML lines to the venetoclax, we assessed the impact of the antibody radio-conjugate on cellular levels of antiapoptotic proteins MCL-1, BCL-2, and BCL-XL. Previously, Niu et al. 31 have screened 11 AML cell lines and showed wide range of venetoclax sensitivity. Based on that study, we have selected two highly venetoclax resistant (OCI-AML3 and U937) and one sensitive (MOLM-13) AML cell lines to cover the wide range of sensitivities to single-agent venetoclax-the differences between IC50 for the sensitive MOLM-13 and resistant OCI-AML3 and U937 were 131and 78-fold, respectively. Interestingly, OCI-AML3 cells exposed to 225 Ac-lintuzumab demonstrated significantly reduced levels of both MCL-1 and BCL-XL compared to controls. Mechanistically, the reduction of both MCL-1 and  Mice were sacrificed if they experienced > 20% weight loss tumor volume reached 4000 mm 3 , or tumors became necrotic. b Anti-tumor response was determined by change in tumor volume following treatment relative to tumor starting volume: SD, less than 30% change in tumor volume; PR. greater than 30% but less than 100% reduction in tumor volume; and CR. greater than 100% reduction in tumor volume.
T A B L E 2 Viability a and antitumor response b to single agent and combination treatment in U937 model BCL-XL was likely the result of genotoxic stress induced in tumor cells because of alpha particle-mediated DNA damage. The reduction of BCL-2 family proteins by 225 Aclintuzumab effectively mitigates resistance to venetoclax in AML tumor lines. There might be additional mechanisms involved in the combination treatment such as radiosensitizing effect of venetoclax. In this regard, O'Steen et al., who combined radioimmunotherapy with beta-emitter 90 Y and venetoclax, and observed the synergistic results of such combination, refer to "multiple mechanisms" of action. 16 The radiosensitizing nature of venetoclax warrants future studies. As a result, the combination induced high DNA damage which in turn caused a pronounced antitumor response and survival benefit in vivo in mice bearing venetoclax-resistant AML xenografts. In these models, a single dose of F I G U R E 7 Combination treatment with 225 Ac-Lintuzumab and venetoclax showed tolerable safety profile in tumor-bearing mice. OCI-AML3 tumor-bearing mice were treated as described in the legend for Figure 5. (A) Body weights were measured and recorded three times per week. Blood was collected from euthanized mice and analyzed for liver and kidney toxicity, that is, aspartate aminotransferase; AST (B), alanine transaminase; ALT (C), creatinine (D) and blood urea nitrogen; BUN (E). Data represent mean with SD of five animals per treatment group. The grey area represents normal ranges of the analytes for SCID mice.