KT-253, a Novel MDM2 Degrader and p53 Stabilizer, Has Superior Potency and Efficacy than MDM2 Small-Molecule Inhibitors

Abstract

Murine double minute 2 (MDM2) is an E3 ligase that inhibits the tumor suppressor protein p53. Clinical trials employing small-molecule MDM2/p53 interaction inhibitors have demonstrated limited activity, underscoring an unmet need for a better approach to target MDM2. KT-253 is a highly potent and selective heterobifunctional degrader that overcomes the MDM2 feedback loop seen with small-molecule MDM2/p53 interaction inhibitors and induces apoptosis in a range of hematologic and solid tumor lines. A single intravenous dose of KT-253 triggered rapid apoptosis and sustained tumor regression in p53 wild-type acute myeloid leukemia and acute lymphoblastic leukemia xenograft models. Additionally, a single intravenous dose of KT-253 in combination with standard-of-care venetoclax overcame venetoclax resistance in an acute myeloid leukemia xenograft model. The data herein define the therapeutic potential of KT-253 and support its clinical development in a range of hematologic and solid p53 wild-type malignancies, as a monotherapy and in combination with standard-of-care agents.

Graphical Abstract

Introduction

The transcription factor p53 is a potent tumor suppressor that plays a key role in cancer prevention. Through its downstream target genes, p53 regulates cell stress responses and guides cell fate decisions including cell cycle arrest and apoptosis (1, 2). The tumor suppressor function of p53 can be altered by various mechanisms, including loss-of-function mutations, deletions, or downregulation of expression (3, 4). A predominant mechanism leading to altered wild-type (WT) p53 function is through proteasomal degradation triggered via ubiquitination, catalyzed by the E3 ligase murine double minute 2 (MDM2; ref. 5). Additionally, MDM2 abolishes p53 transcriptional activity by occluding its transactivation domain and promoting its translocation from the nucleus to cytoplasm (6–9). The destabilization of p53 by MDM2 enables cell survival by blocking cell cycle arrest and apoptotic response. MDM2 has been found to be overexpressed or amplified in a variety of p53 WT cancers (10). The direct dependency of p53 WT on MDM2 has been demonstrated by genetic knockdown across multiple cell lines (11). Therefore, there is a strong rationale for targeting MDM2 to increase the activity and stability of p53 WT and cause either cell stasis or cell death. This has made MDM2 an attractive target for cancer therapies in p53 WT solid tumors and hematologic malignancies (9–11).

Targeting the MDM2/p53 interaction with MDM2 reversible small-molecule inhibitor (SMI) has garnered significant clinical attention as a therapeutic approach for p53 WT malignancies (12). Several SMIs have advanced into clinical development for the treatment of human cancers, including acute myeloid leukemia (AML) and solid tumor indications (13). Although these molecules are able to stabilize p53 and induce apoptosis in hematologic and solid tumors with p53 WT, their biological activity is limited because of their occupancy-driven mechanism of action and initiation of an autoregulatory feedback loop, which upregulates MDM2 (14). As a result, SMIs require multiday dosing regimens that result in dose-limiting toxicities, including neutropenia and thrombocytopenia, which limits the therapeutic index and the majority of clinical applications (13, 15).

Targeted protein degradation is a promising therapeutic approach to eliminate pathogenic proteins that have remained undrugged or inadequately drugged using conventional approaches (16). Targeted protein degraders employ a cell’s own naturally occurring protein degradation system, the ubiquitin proteasomal system, to eliminate a target protein of interest (POI; ref. 17). Heterobifunctional degraders are bivalent targeted protein degraders that consist of two ligands tethered together by a linker. One of the ligands binds to the POI, whereas the other binds to an E3 ligase, resulting in the potential for formation of a ternary complex (18). The induced proximity catalyzes ubiquitination of the POI by the E3 ligase and thereby tags it for degradation via the proteasome. This process is catalytic in nature as a single heterobifunctional degrader molecule is capable of degrading multiple copies of the POI, which in turn can enable the identification of exquisitely potent degraders. Conversely, SMIs rely on an occupancy-driven mechanism, in which a 1:1 stoichiometry is required to drive efficacy.

KT-253 is a bivalent targeted protein degrader that recruits cereblon (CRBN) E3 ligase to degrade MDM2. KT-253 catalyzes the degradation of MDM2 and disrupts the acute autoregulatory feedback loop, a phenomenon inherent to MDM2 SMIs. The SMI-induced feedback loop drives rapid MDM2 upregulation posttreatment. This feedback loop drives the need for increased dosing frequency and higher concentrations of SMIs to maintain stoichiometry for sustained MDM2 inhibition (19, 20). In contrast, with acute genetic knockdown–like removal of MDM2, KT-253 can act in a catalytic manner to achieve the high potency necessary to send malignant cells into acute apoptosis with less frequent dosing than SMIs. This biology associated with pharmacology has been well established (21, 22). Such an intermittent dosing schedule is expected to allow for the recovery of nonmalignant/normal cells, which will provide the opportunity for improved efficacy and therapeutic index.

Here, we describe the characterization of KT-253, the first in-clinic highly potent and selective MDM2 degrader for the treatment of solid and liquid tumors with p53 WT. KT-253 induces an irreversible apoptotic response in a wide range of hematologic and solid tumor lines, including acute lymphoblastic leukemia (ALL) and AML xenograft models. It offers broad application potential across a variety of hematologic and solid tumor malignancies. KT-253 has advanced into a phase 1 clinical study to evaluate its safety, tolerability, pharmacokinetics (PK)/pharmacodynamics (PD), and clinical activity in adult patients with relapsed or refractory high-grade myeloid malignancies, ALL, lymphoma, and solid tumors (NCT05775406; ref. 23).

Materials and Methods

Cell culture

RS4;11 and MOLM-13 cells were cultured in RPMI1640 (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum. MV4;11 cells were cultured in Iscove’s modified Dulbecco’s medium (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum at 37°C and 5% CO₂. All other cell lines used in this study were obtained from ATCC or DSMZ and cultured in recommended growth media at 37°C and 5% CO₂ (see Supplementary Table S1). All lines were confirmed to be mycoplasma negative, and cell line authentication was performed using a human nine species-specific short-tandem repeat marker profile. Cell growth inhibition and apoptosis were tested using CellTiter-Glo (CTG, Promega) and Caspase-Glo 3/7 (Promega) assays according to the manufacturer’s instructions.

Cell line screening CTG assay and Caspase-Glo assay data were analyzed using R v4.2 (24) and the R drc package v3.0-1 (25) to fit growth curves (using the drm function) and calculate absolute IC₅₀ (using the ED function). An increase in caspase signal by 50% over baseline was considered positive. Plots were generated using the R ggplot2 package v2.3.4 (26). Cell line annotation was downloaded from DepMap (RRID:SCR_017655; ref. 27).

MDM2-HiBiT assay

HEK293T cells were grown in DMEM supplemented with 10% fetal bovine serum at 37°C and 5% CO₂. Cells were transfected with lentiviral vector construct expressing C-terminal HiBiT-tagged MDM2 and puromycin selection marker. Stably transfected cells were maintained in complete media supplemented with 1 μg/mL puromycin. Degradation was evaluated as the loss of luciferase signal posttreatment detected using the Nano-Glo HiBiT lytic detection system (Promega) following the manufacturer’s protocol.

Western blotting

Cells treated per treatment condition were lysed using Mammalian Protein Extraction Reagent (M-PER, Thermo Fisher Scientific, Catalog # 78501) supplemented with a 1× protease/phosphatase inhibitor cocktail. Equal quantities of lysates were electrophoresed on a 4% to 12% gradient Bis-Tris midi gel followed by transfer onto polyvinylidene difluoride (PVDF) membrane using iBlot2 PVDF transfer stacks. Protein levels were detected using anti-MDM2 antibody (sc-965, Santa Cruz) and anti-β-actin antibody (926-42210, LICORbio).

p53 meso scale discovery

p53 levels were quantified using total p53 assay kit [K150DBA-3, meso scale discovery] following the manufacturer’s protocol. Cells were lysed 2 hours posttreatment with 1 to 10,000 nmol/L KT-253 or DS-3032. The meso scale discovery signal was measured using a SECTOR imager. p53 levels in treated samples were calculated relative to dimethyl sulfoxide (DMSO) control. Dose–response data were fitted using four-parameter nonlinear-regression model in GraphPad Prism version 10.3.0.

qRT-PCR

For in vitro experiments using cell lines, cells were treated for indicated timepoints followed by a washout and allowed to grow in drug-free growth media for a total of 24 hours. RNA from cell pellets from in vitro studies and from xenograft tissues from in vivo studies was extracted using PureLink RNA Mini Kit (Thermo Fisher Scientific). Reverse transcription was performed using High Capacity RNA-to-cDNA Kit (Thermo Fisher Scientific). Quantitative PCR was performed using TaqMan Gene Expression Master Mix (Thermo Fisher Scientific).

Relative change in mRNA levels was calculated using the ΔΔCt method (28). As reported previously (29–32), the stable expression gene IPO8 was used as reference gene. TaqMan assay IDs for all genes used in this study are summarized in Supplementary Table S2.

Flow cytometry

Cell cycle distribution after compound treatment was assessed using the Click-iT EdU Alexa Fluor 647 Flow Cytometry Assay Kit (Thermo Fisher Scientific, C10419) protocol. Cellular apoptosis was assessed using a fluorescein isothiocyanate (FITC) Annexin V/PI kit (V13242, Invitrogen). Briefly, cells were treated with the indicated compounds or combinations thereof for 4 hours followed by two washouts using PBS and were allowed to grow in complete media for an additional 20 hours. Data were acquired using an iQue3 or Attune NxT flow cytometer. For cell cycle analysis, the cells were gated by fluorescence intensity for percent dead (PI⁻EDU⁻), G0/G1 (PI⁺Edu⁻), S (PI⁺⁺⁺Edu⁺), and G2/M (PI⁺⁺Edu⁻). For apoptosis, cells were gated by fluorescence intensity to assess the percentage of live (Annexin V⁻/PI⁻), early apoptotic (Annexin V⁺/PI⁻), and late apoptotic (Annexin⁺/PI⁺) cells by fluorescence intensity as described previously (33).

Discovery proteomics and data analysis

The cells were lysed using the iST sample preparation kit (PreOmics). After tryptic digestion, the peptides were desalted and labeled using tandem mass tag pro reagents (Thermo Fisher Scientific). Pooled samples were fractioned offline using basic reversed-phase chromatography and recombined using a noncontinuous pooling scheme, as described previously (34). Each peptide fraction was separated using Easy-nLC 1200 nanohigh-performance liquid chromatography over a 150-minute gradient and analyzed online using an Orbitrap Eclipse (Thermo Fisher Scientific) mass spectrometer in data-dependent mode with MS2-based reporter quantification. Raw data were processed with MaxQuant (version 1.6.14.0; ref. 35) and searched using the Andromeda (36) search engine against a comprehensive Swiss-Prot database release for human. Peptide spectral matches were filtered for a precursor intensity fraction of >0.75 to be considered for quantification. Protein identifications were collapsed at the gene level, and at least two quantified razor or unique peptides were required for proteins to be reported. A paired statistical analysis was performed using the limma R package (37). Significant degradation was determined by application of a weighted cutoff incorporating both negative logarithmic P value and log₂ fold change using the following equation (34):

$$f\left(x\right)=\hspace{0.17em}-\text{log}\hspace{0.17em}10\left(0.05\right)+\hspace{0.17em}\frac{1}{|x^2-0.5|}$$

LC-PRM-MS analysis

Sample preparation

Tumor samples were homogenized and denatured using Biognosys’ Denature Buffer with Precellys Evolution homogenizer device. Samples were reduced and alkylated using Biognosys’ Reduction and Alkylation solution. Samples were digested overnight with sequencing grade trypsin (Promega) at a protein/endoproteinase ratio of 50:1. Purification for mass spectrometry was carried out using an Oasis HLB μElution 30-μm plate (Waters) according to the manufacturer’s instructions. Peptides were dried down to complete dryness using a SpeedVac system and dissolved in liquid chromatography solvent A (1% acetonitrile in water with 0.1 formic acid) containing Biognosys’ iRT peptide mix for retention time calibration. Peptide concentrations in mass spectrometry–ready samples were measured using the mBCA assay (Thermo Fisher Scientific). Fifty-four stable isotope-labeled reference peptides were spiked into the final peptide samples at known concentrations (Biosynth ± 10% quantification precision, >95% purity).

PRM-LC-MS data acquisition and analysis

For each sample, 1 μg of peptides was injected onto a custom-packed reversed-phase column [PicoFrit 75 μm × 60 cm column with a 10-μm emitter (New Objective)] packed with 1.7-μm Charged Surface Hybrid C₁₈ particles (Waters) installed onto an Easy-nLC 1200 nanoliquid chromatography system (Thermo Fisher Scientific) connected to a Q Exactive HF mass spectrometer (Thermo Fisher Scientific) equipped with a Nanospray Flex Ion Source (Thermo Fisher Scientific). Liquid chromatography solvents were (A) 0.1% formic acid and (B) 80% acetonitrile/0.1% formic acid. The liquid chromatography gradient was 1% to 59% B for 55 minutes in nonlinear increments and 90% B for 7.5 minutes (total gradient 67 minutes). A standard data-independent acquisition was performed for retention time–based scheduling using Biognosys’ high-precision normalized iRT concept (38). For analytical data acquisition, the scheduling window was 6.7 minutes for each peptide. Data extraction was carried out using SpectroDive software version 11 (Biognosys), and absolute quantities (fmol/μg) were calculated applying a q-value of 0.01 using a single point calibration according to the following formula: Target/Reference × SIS amount on column (fmol). Graphical analysis and interpretation were subsequently performed using GraphPad Prism 10.0.3.

Preparation of KT-253

KT-253 was prepared as described in MDM2 degraders and uses thereof patent WO2021188948 (39). KT-253 is example #174.

DS-3032

DS-3032, also known as milademetan, was procured from MedChemExpress (Catalog # HY-101266). See Supplementary Fig. S1C for chemical structure.

In vivo studies

The animal studies described in this manuscript were conducted at Pharmaron Inc. and Champions Oncology. All procedures related to animal handling, care, and treatment in this study were approved by the Institutional Animal Care and Use Committee of Pharmaron (Ningbo, P.R. China). All procedures related to animal handling, care, and treatment were pursuant to the guidance of the Association for Assessment and Accreditation of Laboratory Animal Care and were approved by the Institutional Animal Care and Use Committee of Champions Oncology.

Female Balb/c nude (MV4;11; purchased from GemPharmatech Co., Ltd.), CB-17 SCID (OCI-LY10, TMD8; purchased from Vital River Laboratory Animal Co.), or NOD-SCID (RS4;11, MOLM-13; purchased from GemPharmatech Co.) mice were used in the experiments. Tumor cells were implanted subcutaneously in the hind flank of the animals with 2 × 10⁶ (TMD8), 5 × 10⁶ (MOLM-13), or 10 × 10⁶ (OCI-LY10, RS4;11, MV4;11) cells in 0.2-mL PBS with Matrigel. When tumors reached approximately 100 mm³ (MOLM-13) or 250 to 500 mm³ (RS4;11, MV4;11), the mice were sorted into treatment groups of five to six mice per group based on their tumor volume to ensure that all the groups had comparable tumor burden. For patient-derived xenograft studies, cells were inoculated intravenously to enable a systemic model. KT-253 was formulated in 20% (w/v) 2-hydroxylpropyl-β-cyclodextrin (HPβCD), 0.025 mol/L hydrochloric acid (HCl), and 0.025 mol/L acetate, pH 4, to the appropriate concentration for each dose level before dosing at 5 mL/kg i.v.. To enable induction of apoptotic response in a hit-and-run manner as well as to mirror and model clinical application and clinical PK profile, KT-253 was formulated and administered intravenously. DS-3032 was dissolved in 0.5% methylcellulose at the appropriate concentration prior to each dose. The components of R-CHOP [rituximab, 10 mg/kg, intraperitoneal (IP); doxorubicin, 3 mg/kg, i.v.; vincristine, 0.25 mg/kg, i.v.; cyclophosphamide, 20 mg/kg, i.v.; prednisone, 0.5 mg/kg, orally; all compounds formulated in PBS] were dosed sequentially, starting with rituximab, followed by doxorubicin, vincristine, cyclophosphamide, and prednisolone, in order, with 15 minutes between each compound. R-CHOP was administered only once during the experiment. In the MOLM-13 model, mice were dosed with either KT-253 or cytarabine (25 mg/kg, IP, 5 days on/2 days off, formulated in saline) alone or in combination with venetoclax [100 mg/kg, oral daily dose, formulated in 10% ethanol, 30% polyethylene glycol (PEG) 400, 60% Phosal 50 PG]. Body weights and tumor volumes were recorded twice per week using the formula V = 0.5 (a × b²), in which a and b are the long and short diameters of the tumor, respectively. Endpoint was reached when tumor volumes reached 10% of the mouse’s weight (e.g., 3,000 mm³ for a 30 g mouse) or when body weight loss reached 20% of initial weight. A complete response to therapy was defined as no measurable tumor for at least any 10 days during the experiment.

For PK/PD experiments, mice bearing RS4;11 tumors were dosed intravenously with KT-253 at the indicated doses and euthanized at 1, 8, and 24 hours post dosing. Blood was collected by cardiac puncture and processed for plasma, and tumors were collected, cut into four equal pieces, fixed in 10% formalin, and snap frozen for analysis or placed into RNAlater for qRT-PCR analysis of pathway genes.

Patient-derived xenograft studies

Female NOG mice were sublethally irradiated with 150 cGY using an RS 2000 X-ray Biological Irradiator (Rad Source Technologies Inc). Four hours after irradiation, 2 × 10⁶ primary, non–mouse passaged AML blast samples derived from patient leukapheresis or peripheral blood mononuclear cells collection were injected intravenously to enable a systemic model. AML cells were thawed at 37°C, an aliquot was diluted 1:10 in PBS, and viable cells were counted using a Cellometer. AML cells were diluted to 2 × 10⁶ human AML cells per 0.2-mL volume and stored on ice (10 × 10⁶ cells/mL). CD3⁺ T cells were depleted by adding anti-CD3 antibody UCHT1 (Bio X Cell, 1 μL per million cells) and incubating for 30 minutes, on ice, prior to intravenous injection.

Mice were monitored daily for clinical signs of AML burden, and body weights were recorded twice during the week post implant and once weekly until the start of treatment. For each model, AML burden was determined by flow cytometric analysis of whole blood and bone marrow in a surrogate cohort of animals (n = 3–5). Sampling times were based upon previously known engraftment kinetics.

Once surrogate animals had sufficient engraftment levels in the bone marrow (% hCD45⁺ of live cells ≥20% on average), the remaining pre-study animals were randomized based on body weight.

KT-253 was prepared as described above and dosed at 1 mg/kg, i.v., at 10 mL/kg. The vehicle control mice were dosed with 20% (w/v) HPβCD, 0.025 mol/L HCl, and 0.025 mol/L acetate, pH 4, at 10 mL/kg.

On day 42 after the start of dosing, the mice received a second dose of KT-253 or vehicle. Sixteen hours after the second dose, the animals were euthanized, and whole blood was collected via cardiac puncture and femurs collected. Bone marrow was isolated by flushing femurs with MACS media. Whole blood and bone marrow were stained for AML-associated surface marker expression. Human CD123 (Clone 6H6, BioLegend), human CD45 (Clone HI30, BD Biosciences), human CD3 (Clone UCHT1, BD Biosciences), CD34 (Clone 561, BioLegend), CD33 (Clone P67.6, BD Biosciences), CD117 (Clone 104D2, BioLegend), and mouse CD45 (Clone 30-F11, BioLegend) were used to identify specific cell populations, and YG2 FVS620 Viability live/dead dye (BD Biosciences) was used to eliminate dead cells from analysis. Samples were analyzed on a flow cytometer and cell populations representing total human CD45⁺ cells or AML blasts (CD123⁺, CD117⁺, or CD33⁺) quantified for evaluation of antitumor activity. Objective response to treatment was defined using the following criteria: mice with tumor burden decrease of >50% from initial hCD45⁺ engraftment levels represent a complete response to treatment, a decrease of 25% to 50% was considered a partial response, and a decrease or increase of <25% in hCD45⁺ levels was considered no response.

Statistical analysis

At least two biological replicates were performed for experimental data shown, unless otherwise indicated. Each data point indicates mean values, and error bars indicate standard deviation. Cell line screening CTG assay and Caspase-Glo assay data were analyzed using R v4.2 (24) and the R drc package v3.0 1 (25) to fit growth curves (using the drm function) and calculate absolute IC₅₀ (using the ED function). For comparisons between two groups, an unpaired t test using the Holm–Sidak method was used. Statistical significance is as shown in which *, P < 0.05, and **, P < 0.01.

Time-to-endpoint analysis was performed by long-rank Kaplan–Meier survival analysis (Sigmaplot 14.5, Systat Software, Inc.) on multiple. Animals that reached tumor endpoint were scored as “event = 1”; animals that did not reach the endpoint or were removed from study for reasons unrelated to tumor or treatment were “censored = 0.” If the log-rank test statistic for the survival curves was significantly different (P < 0.05), all pairwise multiple comparison procedure (Holm–Sidak method) was performed. Kaplan–Meier survival plots (percent animals remaining on study vs. days posttreatment initiation) were generated in GraphPad Prism (GraphPad Software, LLC).

Data availability

The data generated in this study are available within the article and its Supplementary Materials. The mass spectrometry proteomic data have been deposited to the ProteomeXchange Consortium via the PRIDE (40) partner repository with the dataset identifier PXD057859. Derived data can be available from the corresponding author upon reasonable request.

Results

KT-253 is a potent and selective degrader of MDM2

KT-253 is a heterobifunctional molecule consisting of a potent MDM2 ligand connected through a linker to a high-affinity ligand for CRBN (Fig. 1A). The selectivity of KT-253 for MDM2 was investigated using tandem mass tag–based deep proteome profiling of approximately 9,000 human proteins. Treatment of RS4;11 ALL cells with 20 nmol/L KT-253 [10-fold 90% inhibitory concentration (IC₉₀)] for 2, 4, and 8 hours revealed a high level of selectivity (Fig. 1B). KT-253 treatment induced selective and temporal upregulation of p53 and its downstream targets. Despite the lack of direct MDM2 quantification in discovery proteomics owing to its very low abundance, the presence of MDM2 was confirmed in RS4;11 cells by Western blotting and a more sensitive targeted proteomic approach (Fig. 2A and B; Fig. 4C; Supplementary Fig. S2). These results demonstrate that KT-253-mediated degradation of MDM2 promotes stabilization of transcriptionally active p53 and its downstream signaling pathway.

Figure 1.

KT-253 is a potent and selective heterobifunctional MDM2 degrader with superior activity to MDM2/p53 small-molecule inhibitors. A, Chemical structure of KT-253. B, Volcano plots showing deep tandem mass tag proteomic analysis in RS4;11 ALL cells treated with KT-253 for indicated timepoints. No off-target protein degradation was observed. All significantly upregulated proteins are tumor suppressor protein p53 and its target genes as shown. A limma statistical package was used, and a weighted score was calculated that incorporates both fold change of replicate groups of KT-253-treated and DMSO-treated cells and significance for each protein. Proteins were deemed statistically significant on either tail when they fell outside of the cutoff function defined as described in the Methods section. C, HEK293-HiBiT cells treated for 4 hours at indicated concentrations of KT-253 or DS-3032. Nano-Glo HiBiT lytic detection assay was used to demonstrate picomolar MDM2 degradation potency with KT-253 relative to DMSO control. The half maximal degradation concentration (DC₅₀) was obtained by fitting the data with a four-parameter nonlinear-regression model in GraphPad Prism version 10.3.0 (n = 3–4 biological replicates). D, CellTiter-Glo (CTG) assay was used to measure growth inhibition in RS4;11 cells treated with indicated concentrations of KT-253, DS-3032. KT-253 shows picomolar growth inhibition potencies compared with the five other SMIs (see Supplementary Table S3). The half maximal inhibitory concentration (IC₅₀) values were obtained by fitting the data using the four-parameter nonlinear-regression model in GraphPad Prism version 10.3.0 (n = 3 biological replicates). E, CTG assay measuring growth inhibition of RS4;11 cells treated with KT-253, Compound 1 [a cereblon (CRBN)-dead KT-253 analog], or Compound 2 (KT-253 warhead; see Supplementary Fig. S1 for chemical structures). Right shift in potencies with CRBN-dead analog and KT-253 warhead indicates that growth inhibition potency by KT-253 is driven by the recruitment of CRBN and MDM2 (n = 3 biological replicates). Dots indicate mean values, and error bars indicate standard deviation.

Figure 2.

KT-253 overcomes p53/MDM2 feedback loop and shows differential pharmacology to MDM2 SMIs. The degrader advantage demonstrated using Western blot and targeted proteomic data. A, In RS4;11 cells, MDM2 levels are kept at undetectable levels with KT-253 treatment (IC₉₀ = 1.8 nmol/L) as indicated by Western blot analysis. However, when treated with DS-3032 SMI (IC₉₀ = 581.4 nmol/L), MDM2 levels are upregulated by activation of a positive feedback loop, potentially impairing p53 stabilization. β-Actin (ACTB) was used as a loading control. DMSO, dimethyl sulfoxide; IC₉₀ = 90% inhibition concentration. B, Targeted proteomic analysis of MDM2 levels in RS4;11 cells shows that KT-253 [150 nmol/L KT-253] can achieve greater than 90% degradation of MDM2 within 1 hour posttreatment, whereas MDM2 levels continued to increase 1 hour after treatment with DS-3032 (1 mmol/L DS-3032; N = 2 replicates per observation except where noted by § n = 1, n.d., not detected). C, Experimental design for washout experiments performed in RS4;11 cells. D, Caspase 3/7 activation relative to DMSO controls measured at indicated timepoints and concentrations post a washout at 4-hour treatment with KT-253 or DS-3032. Dots indicate mean values, and error bars indicate standard deviation (n = 3 replicates per concentration; CTG, CellTiter-Glo; DMSO, dimethyl sulfoxide).

Figure 4.

A single dose of KT-253 drives sustained tumor regression in ALL xenografts. A, Six cohorts (n = 6) of size-matched RS4;11 xenograft model tumors (∼400 mm³ at the start of treatment) were treated with KT-253 [administered intravenously (IV)] or DS-3032 [administered orally (PO)] or vehicle (administered intravenously). Tumor volumes were measured at indicated timepoints. Complete responses were observed with a single dose (SD) of KT-253 at 1 and 3 mg/kg. No complete responses were observed with clinically relevant dosing of DS-3032. Dots indicate mean values, and error bars indicate standard deviation. B, Kaplan–Meier survival analysis shows the median survival after an SD of KT-253 at 3 mg/kg was 50 versus 12 days for the clinically equivalent dosing regimen of DS-3032. QW × 3, once weekly for 3 weeks. C, Targeted proteomic analysis of tumors demonstrates robust degradation of MDM2 1 hour post dosing with KT-253. ACTB, β-Actin. This is associated with (D) activation of the p53 pathway as evidenced by a corresponding upregulation of proteomic biomarkers p53, p21, and PHLDA3. E, Immunohistochemical analysis of RS4;11 tumors demonstrates more robust activation of the p53 pathway and induction of cleaved caspase-3 (CC-3) following an SD of KT-253 (3 or 1 mg/kg) than following exposure-matched weekly dosing (h, hours; QD × 3, once daily for three consecutive days). Induction of CC-3 was not observed following treatment with the SMI DS-3032 (n = 3 per group; scale bar, 100 μm). F, Plasma concentrations of KT-253 achieved with noted dosing regimens and the resulting tumor stasis or regression. Time over a certain threshold exposure of KT-253, shown with a dotted line, seems to be critical to achieving apoptotic cell fate commitment measured as accumulation of CC-3 in the xenografts (right). PK, pharmacokinetic. **, P < 0.01.

To investigate the mechanism of action of KT-253, the potency of MDM2 degradation was compared with a number of MDM2-targeting SMIs, including DS-3032, in HEK293T-HiBiT assays. Our results showed that KT-253 degrades MDM2 with subnanomolar potency [half maximal degradation concentration (DC₅₀) = 0.4 nmol/L], whereas SMI DS-3032 treatment upregulated MDM2 (Fig. 1C).

RS4;11 ALL cells were also used to investigate the effect of KT-253 compared with SMIs on MDM2-dependent cell viability. The half maximal inhibitory concentration (IC₅₀) for KT-253-treated cells was 0.3 nmol/L, whereas the IC₅₀ values for SMIs were all significantly higher and ranged from 67 nmol/L for DS-3032 to 620 nmol/L for SAR405838 (Fig. 1D; Supplementary Table S3). Strong growth inhibition by KT-253 was driven by the recruitment of both CRBN and MDM2. The former was established by evaluating an analog of KT-253, referred to as Compound 1 (Supplementary Fig. S1A), in which a small modification was made to ablate the affinity of the molecule to CRBN. This led to a significant right shift in the IC₅₀, supporting the hypothesis that CRBN plays a key role in the observed pharmacology (Fig. 1E). The growth inhibition observed with Compound 1, which was unable to engage CRBN, was similar to that observed with the MDM2 ligand-only warhead of KT-253, referred to as Compound 2 (Supplementary Fig. S1B), further supporting the hypothesis that the potent activity of KT-253 was a result of its engagement with both MDM2 and CRBN (Fig. 1E).

KT-253, unlike small-molecule inhibitors, overcomes the MDM2/p53 autoregulatory feedback loop and shows differential pharmacology to MDM2 SMIs

An important advantage of MDM2 degraders over MDM2 SMIs is their ability to overcome the acute MDM2/p53 autoregulatory feedback loop. In RS4;11 cells, as shown using Western blot, 4 hours of treatment with KT-253 led to potent and sustained MDM2 degradation to undetectable levels, whereas DS-3032 treatment resulted in increased MDM2 protein levels after 4 hours (Fig. 2A).

Targeted proteomic analysis further confirmed the degradation of MDM2 to undetectable levels upon KT-253 treatment. Treatment of RS4;11 cells with 150 nmol/L KT-253 for 15 minutes led to an 84% knockdown of MDM2 protein levels, whereas treatment with 1 μmol/L of DS-3032 for the same time period led to an 18% increase in protein levels. Further reduction in MDM2 protein levels was observed 1 hour posttreatment with KT-253 (91% MDM2 degradation), whereas MDM2 concentration continued to increase 1 hour posttreatment with DS-3032 (155%; Fig. 2B).

We hypothesized that the genetic knockdown-like removal of MDM2 after brief exposure to KT-253 would be sufficient to trigger apoptotic cell fate commitment. As outlined in Fig. 2C, RS4;11 cells were treated for 4 hours with varying concentrations of KT-253 or DS-3032, followed by a washout. Cells were allowed to recover in drug-free media, and caspase activity was monitored 24 and 48 hours after the start of treatment. The results show that the 4-hour treatment with KT-253 was sufficient to induce apoptosis over a 48-hour time frame. This trend was maintained across a concentration range of 1 nmol/L to 1 μmol/L KT-253, whereas cell death resulting from DS-3032 treatment was insignificant after 24 hours and minimal after 48 hours (Fig. 2D; Supplementary Fig. S3). These data suggest that brief exposure to KT-253 is sufficient to trigger apoptotic cell fate commitment and supports the use of an intermittent dosing schedule for KT-253 treatment.

KT-253 leads to potent p53 stabilization and induction of apoptosis

We measured the extent of p53 stabilization, the associated upregulation of downstream targets, and apoptotic pathway activation in cellular models. As shown using a mesoscale discovery platform, within 2 hours posttreatment, KT-253 achieved potent p53 stabilization at subnanomolar concentrations [0.05 nmol/L half maximal effective concentration (EC₅₀)]. In contrast, the stabilization achieved by SMI DS-3032 was >1,500-fold less potent (81.4 nmol/L EC₅₀) than that observed with KT-253 (Fig. 3A). An 8-hour treatment with KT-253 showed potent induction of p53 transcriptional gene targets, including MDM2, GDF15, CDKN1A, GADD45A, TNFRSF10B, FAS, and BBC3 (see Supplementary Table S2 for gene names). All these targets showed an approximate fourfold or higher increase in relative mRNA levels at KT-253 concentrations as low as 10 nmol/L. This trend continued in a concentration-dependent manner, reaching an eightfold or greater increase in the relative mRNA levels for all targets at the highest KT-253 treatment concentration evaluated (1 μmol/L). A modest twofold induction of p53 downstream targets was observed upon treatment with DS-3032 only at the highest concentration of 1 µmol/L (Fig. 3B). Furthermore, we observed that treatment with KT-253 led to potent induction of apoptosis in a panel of AML, ALL, and diffuse large B-cell lymphoma (DLBCL) cell lines. Significant growth inhibition was observed in all 10 cell lines tested, with subnanomolar IC₅₀ in multiple lines (Fig. 3C; Supplementary Table S4).

Figure 3.

KT-253 potently stabilizes p53, leading to induction of apoptosis in vitro in hematologic and solid tumor lines. A, Mesoscale discovery platform analysis of p53 levels in RS4;11 ALL cells 2 hours posttreatment with KT-253 or DS-3032 at indicated concentrations, relative to DMSO-treated cells, shows rapid and potent stabilization of p53 protein levels with KT-253. Dose–response data were fitted using a four-parameter nonlinear-regression model in GraphPad Prism version 10.3.0. Dots indicate mean values, and error bars indicate standard deviation (n = 2 biological replicates). B, RS4;11 cells treated for 8 hours with indicated compounds and indicated concentrations, followed by qRT-PCR analysis, show dose-dependent, potent, and acute induction of p53 target gene (MDM2, GDF15, CDKN1A, GADD45A, TNFRSF10B, FAS, and BBC3). Fold change in mRNA levels for each target gene was calculated relative to housekeeping gene IPO8 using the ΔΔCt method (28). Dose–response data were fitted using a four-parameter nonlinear-regression model in GraphPad Prism version 10.3.0. Dots indicate mean values, and error bars indicate standard deviation (n = 3 replicates). C, CellTiter-Glo (CTG) assay shows potent growth inhibition across a panel of p53 WT hematologic cell lines treated with increasing concentrations of KT-253. The half maximal inhibitory concentration (IC₅₀) values, shown in Supplementary Table S4, were calculated by fitting the data using a four-parameter nonlinear-regression model in GraphPad Prism version 10.3.0 (n = 2 biological replicates). Caspase activation was measured using Caspase-Glo 3/7 assay. D, Cells were treated for 4 hours followed by two washouts using PBS and allowed to grow in complete media for additional 20 hours. Quantification of flow cytometry data on cell cycle distribution and apoptosis in RS4;11 and MV4;11 cells show depletion of S-phase cells and potent induction of apoptosis with KT-253 treatment versus DS-3032. Scatter plots indicate that a large subset of p53 WT (E) hematologic and (F) solid tumor cell lines show potent growth inhibition (96-hour CTG assay) and induction of apoptosis (48-hour caspase 3/7 activity assay). Cell panels were treated with KT-253 across a concentration range of 1 to 10,000 nmol/L. Induction of apoptosis in a given cell line is shown as a green circle at its corresponding IC₅₀ value. Lack of caspase activation is shown in red.

To investigate the effect of KT-253 treatment on cell cycle progression, we quantified cell cycle distribution and viability in RS4;11 cells and a human AML cell line, MV4;11, treated with KT-253 or DS-3032 (Fig. 3D). In both cell lines, treatment with 1,000 nmol/L KT-253 resulted in a significantly lower percentage of cells in S-phase (0.97% and 0.45% in RS4;11 and MV4;11, respectively) than with 1,000 nmol/L DS-3032 treatment (30.6% and 12.0% in RS4;11 and MV4;11, respectively). Additionally, significantly more cells were in the late apoptotic stage following treatment with 1,000 nmol/L KT-253 (75.8% in RS4;11 cells and 62.1% MV4;11 cells) than with 1,000 nmol/L DS-3032 treatment (6.9% in RS4;11 cells and 9.3% MV4;11 cells). Viability data also confirmed a significantly higher percentage of live 1,000 nmol/L DS-3032-treated cells (90.4% in RS4;11 cells and 87.4% MV4;11 cells) compared with 1,000 nmol/L KT-253-treated cells (17.1% in RS4;11 cells and 23.8% MV4;11 cells; Fig. 3D).

In vitro screening of KT-253 indicated that multiple p53 WT hematologic and solid tumor cell lines were similarly sensitive to KT-253 treatment. A wide range of these lines showed subnanomolar growth inhibition potency and apoptotic cell fate commitment as measured by growth inhibition and apoptosis assays, respectively. AML, T-cell lymphomas, mantle cell lymphoma, and DLBCL were among the sensitive hematologic indications (Fig. 3E), whereas sensitive solid tumor indications included prostate, brain, ovarian, soft tissue sarcoma, neuroblastoma, lung, and colorectal cancer (Fig. 3F). Notably, KT-253 required an expression of functional p53. Cell lines expressing mutated p53 (p53 MUT) did not respond to KT-253 at concentration as high as 10,000 nmol/L (Supplementary Fig. S4). These data demonstrate the potential for KT-253 development across a wide variety of p53 WT heme and solid tumor indications.

A single dose of KT-253 induces apoptosis and sustained tumor regression in ALL

To assess the in vivo activity of KT-253, we evaluated its activity against the RS4;11 ALL xenograft model. The extent and duration of KT-253 antitumor activity against well-established (approximately 400 mm³) RS4;11 tumors were compared with DS-3032 activity at clinically achievable DS-3032 exposure (30 mg/kg) and at a preclinically reported (mouse) efficacious dose (100 mg/kg). To test the hypothesis that treatment with a relatively high dose of KT-253 on an intermittent schedule would have a superior efficacy and safety profile than frequent dosing at a relatively low dose, we compared the activity of a single dose (SD) to the activity of an AUC-matched fractionated dose on a weekly schedule. Mice were treated with either a single i.v. dose of KT-253 at 3 mg/kg, a single i.v. dose at 1 mg/kg, or an i.v. dose of 0.3 mg/kg weekly for 3 weeks. This dosing was selected to match the AUC exposure to the 1 mg/kg dose. DS-3032 was administered orally at 30 and 100 mg/kg for 3 days on and 11 days off.

A SD of KT-253 at 1 or 3 mg/kg induced tumor regressions. The lower exposure-matched weekly dosing regimen (0.3 mg/kg once weekly for 3 weeks) induced transient stable disease. In contrast, DS-3032 had no effect on tumor growth at the clinically relevant dose and led to modest tumor growth inhibition at the higher dose (Fig. 4A). The median survival after a SD of KT-253 at 3 mg/kg was 50 versus 12 days (P = 0.007) for the clinically equivalent dosing regimen of DS-3032 in RS4;11 tumor–bearing mice (Fig. 4B).

Targeted proteomic analysis of RS4;11 tumors demonstrated robust degradation of MDM2 1 hour post dosing with an SD of KT-253 at 1 or 3 mg/kg (94% and 92% MDM2 degradation, respectively). The lower exposure-matched weekly dosing regimen of KT-253 resulted in slightly less, but still significant, MDM2 degradation (81% MDM2 degradation) compared with SD treatments. MDM2 degradation was not observed with the clinically relevant or high dose of DS-3032 (Fig. 4C). The trend of MDM2 degradation observed with KT-253 treatments corresponds to the activation of the p53 pathway, as evidenced by an upregulation of proteomic biomarkers p53, cyclin-dependent kinase inhibitor 1 (p21), and pleckstrin homology-like domain family A, member 3 (PHLDA3; Fig. 4D).

Consistent with the efficacy data, immunohistochemical analysis of RS4;11 tumors revealed robust upregulation of p53 and induction of apoptosis, as indicated by cleaved caspase-3 (CC-3) staining, following an SD of KT-253 at 1 or 3 mg/kg. In contrast, neither weekly dosing of KT-253 at 0.3 mg/kg (which led to stasis) nor treatment with DS-3032 induced apoptosis (Fig. 4E).

Dose fractionation data demonstrate that tumor regressions in RS4;11 are driven by an SD of KT-253 at 1 or 3 mg/kg to achieve plasma concentrations ≥10 nmol/L for 12 to 22 hours. The lower exposure-matched weekly dosing regimen of KT-253 resulted in tumor stasis. This time-over-threshold-concentration exposure profile is critical to achieve rapid induction of the apoptotic pathway (Fig. 4F). Additionally, a single 1 mg/kg dose of KT-253 led to robust activation of mRNA markers of both the intrinsic (BBC3) and extrinsic (TNFRSF10B and FAS) apoptotic pathways. In contrast, the exposure-matched weekly dosing regimen led to a modest induction of these markers (Supplementary Fig. S5; see Supplementary Table S2 for gene names).

A single dose of KT-253 induces apoptosis and sustained tumor regression in AML models

To extend the analysis of KT-253 antitumor activity relative to DS-3032 in an AML model, MV4;11 tumor–bearing animals were also investigated. Similar to the RS4;11 ALL model, AML tumor–bearing animals also demonstrated sustained tumor regressions after a single i.v. dose of KT-253 at 3 mg/kg (Fig. 5A). Complete responses with a single i.v. dose of KT-253 at 3 mg/kg were achieved in five of six animals, and four of the six remained tumor-free and survived for the duration of the study (6 months), despite no additional dosing. Median survival was also determined for a single i.v. dose of KT-253 at 1.0 mg/kg, DS-3032 100 mg/kg oral dose for 3 days on 11 days off, DS-3032 30 mg/kg oral dose for 3 days on 11 days off, and a vehicle control, with median survival durations of 22, 31, 16, and 11 days, respectively (Fig. 5B). The MV4;11 AML model treated with an exposure-matched KT-253 weekly regimen resulted in transient stable disease, whereas DS-3032 treatment at the clinically relevant dose resulted in transient tumor growth inhibition, and no complete responses were achieved (Fig. 5A and B).

Figure 5.

A single dose of KT-253 drives sustained tumor regression in MV4;11 AML xenografts and demonstrates strong antileukemic activity in patient-derived xenograft models of systemic AML. A, Five cohorts (n = 6) of size-matched MV4;11 AML xenograft model tumors (∼300 mm³ at the start of treatment) were treated with KT-253 (administered intravenously) or DS-3032 (administered orally) or vehicle (administered intravenously). Tumor volumes were measured at indicated timepoints. A single 3 mg/kg dose of KT-253 led to complete responses in five out of six animals. No complete responses were observed with clinically relevant dosing of DS-3032. B, Kaplan–Meier survival analysis shows that the median survival after a single dose (SD) of KT-253 at 3 mg/kg was >180 versus 16 days for the clinically equivalent dosing regimen of DS-3032 (QW × 3, once weekly for 3 weeks). C, Immunohistochemical analysis of MV4;11 tumors demonstrates robust activation of the p53 pathway and induction of cleaved caspase-3 (CC-3) following an SD of KT-253 but not following exposure-matched weekly dosing. Induction of CC-3 following treatment with the SMI DS-3032 was modest (QD × 3, once daily for three consecutive days, scale bar = 100 μm). D, AML patient-derived xenograft model CTG-2227 showing percent hCD45⁺ cells in peripheral blood (PB) and bone marrow (BM), and hCD34⁺ leukemic stem blasts in peripheral blood 42 days posttreatment initiation indicate robust reduction in tumor burden. Animals were dosed on days 0, 21, and 41, and samples were collected 24 hours post the last dose (*, P < 0.05; **, P < 0.01).

ALL and AML models also showed similar PD effect on p53 pathway and apoptosis. In the MV4;11 AML model, immunohistochemical analysis showed robust upregulation of p53 and induction of CC-3. An SD of KT-253 at 3 mg/kg (which caused tumor regression) led to robust upregulation of p53 and CC-3, whereas AUC exposure-matched weekly dosing of KT-253 or the DS-3032 dosing regimen did not (Fig. 5C).

The activity of KT-253 in AML was next evaluated in a panel of four AML patient–derived xenograft models with systemic disease: CTG-2227, CTG-2700, CTG-2240, and CTG-2235. KT-253 (1 mg/kg) was administered to tumor-bearing mice. In the CTG-2227 model, KT-253 led to reduction of human CD45⁺ (hCD45⁺) tumor burden in peripheral blood and bone marrow (P = 0.0057, unpaired t test) at day 41 relative to vehicle and a complete response (P = 0.009, unpaired t test; Fig. 5D; Supplementary Table S5). A partial antitumor response, characterized by a significant reduction in hCD45⁺ cells in the peripheral blood, was observed in the CTG-2700 and CTG-2240 models, whereas the CTG-2235 model was nonresponsive (Supplementary Table S5).

Finally, the activity of KT-253 was also investigated in p53 WT and mutant activated B-cell (ABC)-subtype DLBCL models. This evaluation demonstrated that KT-253 was highly active against the p53 WT ABC-subtype DLBCL (OCI-LY10) xenograft model, with responses similar to R-CHOP, a combination therapy of drugs containing rituximab, cyclophosphamide, doxorubicin hydrochloride (hydroxydaunorubicin), vincristine sulfate (Oncovin), and prednisone, but was not active against the p53 mutant ABC-subtype (TMD8) model (Supplementary Fig. S6).

Significant combination benefit of KT-253 with standard-of-care treatments in AML

Next, we sought to assess KT-253 activity in combination with SoC agents for AML. In the p53 WT MOLM-13 AML model in vitro, KT-253 alone was less effective at driving the cell cycle to early and late apoptosis than the combination treatment of KT-253 with venetoclax. Specifically, when KT-253 (at either 1 or 10 nmol/L) was combined with venetoclax at 1.6 μmol/L, there was a significantly higher percentage of cells (50% higher) in the early or late apoptotic stage than with venetoclax alone at the same concentration (1.6 μmol/L; Fig. 6A). This advantageous combination effect was also observed when cell growth inhibition was monitored in MOLM-13 cells. The most significant cell growth inhibition was observed when KT-253 was combined with venetoclax at a concentration above its IC₅₀ value (Fig. 6B).

Figure 6.

KT-253 shows significant combination benefit with venetoclax in a venetoclax-resistant AML model. A, Analysis of apoptosis by flow cytometry in an AML cell line, MOLM-13, treated in vitro at indicated concentrations of KT-253 alone or in combination with venetoclax shows that KT-253 in combination with venetoclax enhances apoptotic cell fate commitment (n = 2 replicates per condition). DMSO, dimethyl sulfoxide; IC₅₀, half maximal inhibitory concentration. B, CellTiter-Glo (CTG) assay on MOLM-13 AML cells also shows maximal cell growth inhibition with KT-253 in combination with venetoclax (n = 3 replicates per concentration). C, MOLM-13 xenograft tumor growth curves showing that a single 3 mg/kg dose of KT-253 in combination with specific dosing regimens of venetoclax can induce durable complete responses, whereas either agent alone cannot. i.v., intravenous; PO, orally; QD × 1W, once daily for 1 week; QD × 3, once daily for three consecutive days; QD × 3W, once daily for 3 weeks; SD, single dose. D, Kaplan–Meier survival analysis of MOLM-13 tumor-bearing mice treated with KT-253 alone, venetoclax alone, or a combination, as indicated. Results show all animals treated with the three different combination regimens tested [subclinical 50 mg/kg dose of venetoclax QD × 3W (blue) or clinical 100 mg/kg dose level for shorter duration (QD × 3 or QD × 1W for light green and red traces, respectively)] achieved complete responses (CRs) and prolonged survival, whereas none of the mice treated with either agent alone (teal and orange traces) achieved CR (n = 6 animals per group) at end of study (EoS).

To determine whether a combination benefit could also be achieved in vivo, we evaluated the activity of KT-253 alone or in combination with venetoclax in the MOLM-13 subcutaneous xenograft model. MOLM-13 tumors were insensitive to the clinically relevant dosing regimen of venetoclax alone (100 mg/kg oral daily dose) and demonstrated transient (approximately 7 days) regression with KT-253 alone (3 mg/kg). However, when an SD of KT-253 was combined with venetoclax, durable complete responses were achieved (Fig. 6C and D). Although the clinical dosing regimen for venetoclax is continued daily administration, as little as 1 week of venetoclax administration in combination with a SD of KT-253 was sufficient to induce complete responses in six out of six animals. Remarkably, administration of venetoclax for only 3 days also induced complete responses in six out of six animals when combined with KT-253, but these responses were less durable (median survival of 68 days vs. >150 days for 1 week of venetoclax). In addition, half of the clinically equivalent dose of venetoclax (50 mg/kg) administered daily for 3 weeks in combination with an SD of KT-253 was sufficient to induce durable complete responses in all animals (Fig. 6D). In addition to the combination benefit, KT-253 treatment also showed strong single-agent activity in a primary systemic model of AML resistant to venetoclax treatment (Supplementary Fig. S7).

Overall, these data suggest that KT-253 can potentially be effective in patients with venetoclax-resistant AML as a monotherapy or in combination with venetoclax. Moreover, both the dose level and duration of venetoclax dosing can be reduced when venetoclax is combined with KT-253, thereby minimizing the risk of toxicity.

Discussion

Despite promising preclinical data, the clinical development of multiple MDM2 SMIs in heme and solid tumor indications has faced dose-limiting gastrointestinal and hematologic toxicities, which contribute to their limited success in achieving favorable outcomes in the clinic (41). Maintaining the 1:1 stoichiometric interaction required to drive occupancy-driven efficacy, inherent to protein–protein interaction inhibitors, has likely been self-limiting in the presence of the MDM2/p53 autoregulatory feedback loop, resulting in failure to enable the necessary pharmacology to induce the acute apoptotic response. Studies have demonstrated differential dynamics of cell fate commitment pathways triggered by the pulsatile activation of p53 in combination with posttranscriptional modifications and differential cofactor recruitment, which lead to apoptosis over cell cycle arrest (21, 42, 43). To preferentially achieve the activation of apoptotic pathways in cancerous cells, attempts have been made to optimize dosing schedules (21, 22). However, suboptimal p53 activation potencies of MDM2 SMIs were insufficient to achieve the desired efficacy in the clinic.

In this study, we characterized KT-253, a highly selective and potent MDM2 heterobifunctional degrader (Fig. 1). Because of catalytic degradation of MDM2 at picomolar degradation potency, KT-253 can overcome the acute MDM2/p53 autoregulatory feedback loop to achieve genetic knockdown-like depletion of MDM2 (Fig. 2), which results in picomolar growth inhibition potencies. This contrasts with two previously reported MDM2-targeting heterobifunctional degraders recruiting either CRBN (19) or von Hippel–Lindau (VHL) E3 ligase (44), which show low to high micromolar growth inhibition potencies. With unknown selectivity profiles or molecular glue-like potential, the contribution from possible off-target mechanisms remains unclear for these molecules.

It is unclear precisely how p53 preferentially activates a certain set of genes to favor the desired cell fate commitment. However, there are distinct downstream gene profiles when p53 activity is pulsatile versus when it is continuous (42, 43). Key transcription factors, such as p53, which regulate multiple cellular programs likely follow a hit-and-run transcription model (45, 46) for rapid and transient activation of unique gene regulatory networks. Akin to this model, we show that short-term exposures with KT-253 are sufficient to activate caspase 3/7–mediated cell death (Fig. 2). Potent growth inhibition and apoptotic activity following short-term KT-253 treatment were observed across a panel of hematologic and solid tumor cell lines (Fig. 3). Even though the extent of p53 upregulation seem similar, albeit at >1,500-fold difference in potency between KT-253 and DS-3032 (Fig. 3A), the feedback loop-mediated elevation of MDM2 levels by SMI treatment likely hinders the ability of p53 to transcriptionally activate its downstream targets (6–9). These observations support a hit-and-run hypothesis for KT-253, as demonstrated in vivo in xenograft models, in which we show that an SD of KT-253 intravenously was able to drive sustained tumor regression. Exposure-matched (AUC) weekly dosing of KT-253 resulted in cell cycle arrest and tumor stasis, showing pharmacology influenced biological outcomes (Figs. 4 and 5), as demonstrated previously (21). Similarly, sustained AUC-driven pharmacology led to cell stasis. In contrast, time over a certain threshold could initiate an acute apoptotic response. Because these apoptotic effects can be achieved through intermittent dosing, this also supports our hypothesis that KT-253 treatment has the potential to result in an improved efficacy and safety profile compared with the more frequent dosing needed for MDM2 SMIs. Our data support an intermittent dosing regimen for KT-253 and the potential to overcome safety and efficacy limitations of MDM2 SMIs shown in the clinic.

Relapsed and refractory AML remains a daunting clinical challenge owing to limited therapeutic options combined with poor outcomes and emergent resistance mechanisms (47, 48). Several studies have demonstrated synergistic combination benefit with MDM2 SMIs and venetoclax, an SoC agent in AML (49–53). We show that an SD of KT-253 in combination with SoC not only achieves durable complete responses at clinically relevant doses but also at subclinical doses of venetoclax, resensitizing models that have become refractory and/or are resistant to venetoclax (Fig. 6). In addition, we show that inherently venetoclax-resistant models can respond to KT-253 monotherapy. The combination of both sets of data show the promise of KT-253 in addressing the currently unmet need for treatment options for venetoclax-resistant tumors.

Overall, KT-253 shows promise for clinical development in p53 WT tumors with high sensitivity to degrader mechanisms in a variety of solid tumors and hematologic malignancies, with superiority over SMIs. As of this publication, KT-253 has advanced into an ongoing phase 1 clinical study to evaluate the safety, tolerability, PK/PD, and clinical activity in adult patients with relapsed or refractory high-grade myeloid malignancies, ALL, lymphoma, and solid tumors (NCT05775406).

Supplementary Material

Supplemental Figure 1

Supplemental Figure 1: Chemical structures of (A) Compound 1, a KT 253 analog that lacks ability to engage cereblon (CRBN), (B) Compound 2, the warhead of KT 253 that engages MDM2, and (C) DS-3032, a small-molecule MDM2/p53 interaction inhibitor.

Supplemental Figure 2

Supplemental Figure 2: Targeted proteomics analysis of MDM2 levels in MV4;11 AML cells shows that 150 nM KT 253 can achieve degradation of MDM2 within 1 hour posttreatment.

Supplemental Figure 3

Supplemental Figure 3: Short term exposures with KT-253 sufficient for growth inhibition compared with MDM2 SMIs.

Supplemental Figure 4

Supplemental Figure 4: A functional p53 is required for growth inhibition with KT-253.

Supplemental Figure 5

Supplemental Figure 5: A single dose of KT-253 more robustly induces p53 targets compared with exposure matched weekly dosing regimen.

Supplemental Figure 6

Supplemental Figure 6: KT 253 activity in p53 WT ABC subtype DLBCL model but not p53 mutated ABC subtype DLBCL model.

Supplemental Figure 7

Supplemental Figure 7: A single dose of KT 253 shows strong single-agent activity in a venetoclax resistant patient-derived AML model.

Supplemental Table 1

Supplemental Table 1: Cell lines and reagents

Supplemental Table 2

Supplemental Table 2: Genes and TaqMan Gene Expression Assay IDs used for RT-qPCR assays

Supplemental Table 3

Supplemental Table 3: KT-253 shows picomolar growth inhibition potency compared with other MDM2 small molecule inhibitors.

Supplemental Table 4

Supplemental Table 4: KT-253 shows potent growth inhibition and caspase activation across a panel of p53 wild-type hematologic cell lines.

Supplemental Table 5

Supplemental Table 5: Activity of KT-253 in AML patient-derived xenograft models.

Authors’ Disclosures

Y.K. Chutake reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. N. Dumont reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. S.B. Breitkopf reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. D. Chen reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. E.W. Kuhn reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. S. Bollinger Martinez reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. A.A. McDonald reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. J. Qi reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. K.-N. Hu reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. R. Karnik reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. J.D. Growney reports other support from Kymera Therapeutics outside the submitted work. K. Sharma reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. A.M. Gollerkeri reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. N. Mainolfi reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. J.A. Williams reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US11932624 issued. M.M. Weiss reports other support from Kymera Therapeutics outside the submitted work, as well as a patent for US 11932624 issued. No disclosures were reported by the other authors.

Authors’ Contributions

Y.K. Chutake: Conceptualization, data curation, formal analysis, supervision, visualization, methodology, writing–original draft, writing–review and editing. M.F. Mayo: Conceptualization, formal analysis, supervision, visualization, methodology. N. Dumont: Conceptualization, data curation, formal analysis, supervision, visualization, methodology, writing–original draft, writing–review and editing. J. Filiatrault: Data curation, formal analysis, methodology. S.B. Breitkopf: Conceptualization, formal analysis, supervision, methodology, writing–original draft. P. Cho: Data curation, formal analysis, methodology. D. Chen: Conceptualization, formal analysis, methodology. V.S. Dixit: Conceptualization, formal analysis, methodology. W.R. Proctor: Conceptualization, visualization, methodology. E.W. Kuhn: Conceptualization, data curation, formal analysis, methodology. S. Bollinger Martinez: Formal analysis, visualization, methodology. A.A. McDonald: Conceptualization, supervision, visualization, methodology. J. Qi: Conceptualization, visualization, methodology. K.-N. Hu: Conceptualization, methodology. R. Karnik: Data curation, software, formal analysis, visualization, methodology. J.D. Growney: Data curation, supervision, methodology, writing–original draft. K. Sharma: Conceptualization, formal analysis, supervision, visualization, methodology. S.S. Schalm: Conceptualization, data curation, supervision, methodology. A.M. Gollerkeri: Conceptualization, supervision, methodology. N. Mainolfi: Conceptualization, supervision, methodology, writing–review and editing. J.A. Williams: Conceptualization, formal analysis, supervision, visualization, methodology, writing–original draft, writing–review and editing. M.M. Weiss: Conceptualization, formal analysis, supervision, methodology, writing–original draft, writing–review and editing.