Dose and schedule determine distinct molecular mechanisms underlying the efficacy of the p53-MDM2 inhibitor HDM201

Authors: Sébastien Jeay1†, Stéphane Ferretti1†, Philipp Holzer2†, Jeanette Fuchs1, Emilie A. Chapeau1, Markus Wartmann1, Dario Sterker1, Vincent Romanet1, Masato Murakami1, Grainne Kerr1, Eric Y. Durand1, Swann Gaulis1, Marta Cortes-Cros1, Stephan Ruetz1, Therese-Marie Stachyra1, Joerg Kallen3, Pascal Furet2, Jens Würthner4, Nelson Guerreiro5, Ensar Halilovic6, Astrid Jullion7, Audrey Kauffmann1, Emil Kuriakose8, Marion Wiesmann1, Michael R. Jensen1, Francesco Hofmann1†, William R. Sellers9†

Affiliations:
1Disease Area Oncology, Novartis Institutes for BioMedical Research, Basel, Switzerland; 2Global Discovery Chemistry, Novartis Institutes for BioMedical Research, Basel, Switzerland; 3Chemical Biology & Therapeutics, Novartis Institutes for BioMedical Research, Basel, Switzerland;
4Translational Clinical Oncology, Novartis Institutes for BioMedical Research, Basel, Switzerland;
5PK Sciences, Novartis Institutes for BioMedical Research, Basel, Switzerland;
6Translational Clinical Oncology, Novartis Institutes for BioMedical Research, Cambridge, MA, United States;
7Oncology Global Development, Basel, Switzerland;
8Translational Clinical Oncology, Novartis Institutes for BioMedical Research, East Hanover, NJ, United States;
9Disease Area Oncology, Novartis Institutes for BioMedical Research, Cambridge, MA, United States.

Running Title: Dosing regimen govern distinct HDM201 molecular mechanisms
Precis: Pulsed high doses versus sustained low doses of the p53-MDM2 inhibitor HDM201 elicit a pro-apoptotic response from wild-type p53 cancer cells, offering guidance to current clinical trials with this and other drugs that exploit the activity of p53.
Keywords: HDM201, p53, MDM2, intermittent dosing, mechanism of action

Additional information: †These authors contributed equally to this work.
Financial support: the study was founded by Novartis Pharma A.G.
Corresponding Authors: Stéphane Ferretti and Francesco Hofmann, Novartis Institutes for BioMedical Research, CH-4002, Basel, Switzerland; E-mail: [email protected] and [email protected].
Current address for S. Jeay: Idorsia Pharmaceuticals Ltd, Allschwil, Switzerland. Current address for M. Murakami: Daiichi-Sankyo, Tokyo, Japan.
Current address for J. Würthner: ADC Therapeutics Sarl, Epalinges, Switzerland.

Current address for W.R. Sellers: Broad Institute, 415 Main Street, Cambridge, MA 02142, USA.
Conflict of interest: All authors were employees thereof at the time the study was performed. Sebastien Jeay is now an employee of Idorsia Pharmaceuticals Ltd; Masato Murakami is now an employee of Daiichi-Sankyo; Jens Würthner is now an employee of ADC Therapeutics Sarl; William R. Sellers is now an employee of the Broad Institute
Data and materials availability: Researchers may obtain HDM201 with a material transfer agreement from Novartis. All reasonable requests for collaboration involving materials used in the research will be fulfilled provided that a written agreement is executed in advance between Novartis and the requester (and his or her affiliated institution). Such inquiries or requests should be directed to the corresponding authors.

Word count in title: 100
Characters count in running title: 58
Word count in abstract: 211 Word count in text: 4940 Number of references: 47
Number of Tables and Figures: 0 Table and 5 Figures
Supplementary Data (Figures and Tables):
⦁ Supplementary Materials and Methods
⦁ Supplementary References
⦁ 10 Supplementary Tables and 4 Supplementary Figures
⦁ 3 Supplementary Data Files:
⦁ Supplementary Data File S1. Summary of shRNA scores (separate Excel file).
⦁ Supplementary Data File S2. Summary of RSA scores (separate Excel file).
⦁ Supplementary Data File S3. Deep sequencing and gene centric common insertion sites landscapes in resistant tumors (separate Excel file).

Abstract

Activation of p53 by inhibitors of the p53-MDM2 interaction is being pursued as a therapeutic strategy in p53 wild-type cancers. Here we report distinct mechanisms by which the novel, potent, and selective inhibitor of the p53-MDM2 interaction HDM201 elicits therapeutic efficacy when applied at various doses and schedules. Continuous exposure of HDM201 led to induction of p21 and delayed accumulation of apoptotic cells. By comparison, high dose pulses of HDM201 were associated with marked induction of PUMA and a rapid onset of apoptosis. shRNA screens identified PUMA as a mediator of the p53 response specifically in the pulsed regimen. Consistent with this, the single high dose HDM201 regimen resulted in rapid and marked induction of PUMA expression and apoptosis together with down-regulation of Bcl-xL in vivo. Knockdown of Bcl-xL was identified as the top sensitizer to HDM201 in vitro, and Bcl- xL was enriched in relapsing tumors from mice treated with intermittent high doses of HDM201. These findings define a regimen-dependent mechanism by which disruption of MDM2-p53 elicits therapeutic efficacy when given with infrequent dosing. In an ongoing HDM201 trial, the observed exposure-response relationship indicates that the molecular mechanism elicited by pulse dosing is likely reproducible in patients. These data support the clinical comparison of daily and intermittent regimens of p53-MDM2 inhibitors.

Introduction

The primary response to a variety of cellular stresses is to activate and stabilize p53 which then drives transcriptional programs leading to cell cycle arrest, promotion of repair pathways, and in response to severe stress, the initiation of apoptosis (1, 2). Intracellular levels of p53 are
regulated by protein degradation through ubiquitin-dependent (3) and ubiquitin-independent mechanisms (4). Ubiquitination is the most important (3, 5, 6) and the E3 ligase MDM2 is the primary negative regulator of p53 (7, 8). MDM2 ubiquitination of p53 negatively regulates its transcriptional activity. Mono-ubiquitination triggers nuclear export, while poly-ubiquitination targets nuclear p53 for proteasomal degradation (9). MDM2 is transcriptionally up-regulated by p53 and this negative-feedback loop ensures that p53 levels remain low under normal conditions (10).
In the absence of loss-of-function mutations in TP53, MDM2 is often amplified and/or overexpressed in a number of cancers (11). MDM2 hence is a logical therapeutic target in cancer to increase wild-type (WT) p53 activity. Nutlin-3a (12) and subsequently, several p53-MDM2 inhibitors have been tested in preclinical and clinical studies (13, 14). In preclinical studies, increased p53, resulting from MDM2 inhibition, leads to a number of effects simplified into the categories of cell cycle arrest and apoptosis. The decision between these two pathways can be governed by the level and duration of p53 induction, in a context and /or tissue specific manner. Generally, but not always (15, 16), transient lower levels of p53 induce cell cycle arrest, while continuous low levels or elevated pulsed p53 promote death (17). A p53-mediated cell cycle arrest is mainly achieved through transcriptional activation of p53 target genes, primarily p21 and GADD45, which block cyclin-dependent kinases and induce G1/S (18) and G2 phase arrests, respectively (19). When damage or stress cannot be repaired, continued signaling (e.g. ATM/ATR, Chk1/2) leads to further accumulation of p53, the subsequent expression of pro- apoptotic p53 target genes including PUMA, Noxa, and Bax (20, 21) and the triggering of apoptosis (22, 23).

In this study, we sought to explore different doses and schedules to maximize the therapeutic impact of MDM2 inhibition by HDM201, a novel, selective and highly optimized p53-MDM2 inhibitor, currently in phase I testing (trial registration identifier: NCT02143635). We report that pulsed high dose treatment with HDM201 elicited a unique, rapid and substantial induction of PUMA and apoptosis that was sufficient to sustain tumor regressions without re-dosing. The emerging pharmacokinetic (PK) and pharmacodynamic (PD) measures of p53 activation in patients dosed with HDM201 strongly suggests that a differential pharmacologic response to low and high-dose regimens can be achieved in the clinic.
Materials and Methods

Chemical entity

HDM201 was synthesized by Global Discovery Chemistry at Novartis, Basel, Switzerland. For in vitro studies, 10 mM solutions were prepared in 100% dimethyl sulfoxide (DMSO). For in vivo experiments, HDM201 was freshly dissolved in methylcellulose 0.5% w/V in 50 mM phosphate buffer pH 6.8 for PO administration (5 and 10 ml/kg for rats and mice, respectively).

Cell lines and in vitro pharmacologic cell line profiling

All cell lines were obtained from ATCC (American Type Culture Collection), DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen), and HSRRB (Health Science Research Resources Bank) and cultured in RPMI or Dulbecco’s modified Eagle’s medium plus 10% FBS (Invitrogen, Carlsbad, CA) at 37°C, 5% CO2. Cell line identities were confirmed using a 48- variant SNP panel and all were confirmed as mycoplasma free. High-throughput cell viability assays were done as previously reported (24).For in vitro washout experiments, HDM201 or DMSO were added at the desired concentrations (final DMSO concentration: 0.1%), incubated for the time indicated at 37°C, and then removed by incubating cells with phosphate-buffered saline (PBS) for 5 min, followed by replacement with growth medium (GM) (RPMI-1640 with 10% fetal calf serum [FCS], 10 mM HEPES, 1 mM sodium pyruvate, 2 mM L-Glutamine, 1% Penicillin/Streptomycin). Seventy-two h following compound addition, effects of HDM201 on SJSA-1 growth and viability were measured using Cell Titer Glo™ (#G7570 Promega). The GI50 and GI80 values were calculated by curve fitting using XLfit (Fit Model #201).

shRNA Screen

SJSA-1 cells were transduced at 1000-fold coverage with a lentiviral library consisting of three pools of shRNAs targeting 7837 genes (20 shRNAs/gene) (25) and passaged for 15-19 days. Subsequently, pools were divided in 3 groups (0.01% DMSO for 72 h, 100 nM HDM201 for 72 h or 1.5 µM HDM201 for 8 h followed by 64 h 0.01% DMSO). After 3 days, >50×106 cells per pool and treatment group were harvested. Pool 1 and 2 are represented by two independent replicates, while pool 3 (containing shRNAs targeting BBC3) results are based on one replicate.
In vivo experiments

All animal studies were under the oversight of the Novartis Animal Welfare Organization (AWO) and were conducted in accordance with ethics and procedures covered by permit BS- 1763, BS-1975 and BS-2064 issued by the Kantonales Veterinäramt Basel-Stadt and in strict adherence to Swiss animal welfare law (Eidgenössisches Tierschutzgesetz and the Eidgenössische Tierschutzverordnung, Switzerland). All animals had access to food and water ad libitum and were identified with transponders. They were housed in a specific pathogen-fre facility with a 12-h light/12-h dark cycle. For PK/PD experiments in tumor-bearing rats, HDM201 was injected once at 5 or 27 mg/kg. Animals were randomized into groups of 2 and tissue samples collected at 0, 1, 3, 8, 10, 24 and 48 h. Blood samples were collected in EDTA coated tubes (Milian, #TOM-14C). Tumors were excised, weighed, frozen in liquid nitrogen and cryogenic dry pulverized with the CryoPrep™ system (model CP-02, Covaris). For efficacy experiments, rats were randomized into groups (n=6) for a mean tumor size of 500 mm3 and HDM201 was injected once at 27 mg/kg or daily at 5 mg/kg for 14 days. In mice, animals were randomized in groups with a mean tumor size of 200 mm3 and HDM201 was administrated at 40 mg/kg daily or 100 mg/kg twice a week on days 1 and 4. Tumor response and relapse are reported with the measures of tumor volumes from the treatment start. Conditional survival is defined as maximum tumor size of 1.5 cm diameter or sacrifice for morbidity including >15% body weight loss. Concentrations of HDM201 in plasma and tumors were determined by UPLC/MS-MS.

PK and PD assessments in patients

The primary objectives of the phase I study (trial registration identifier: NCT02143635) is to determine the maximum tolerated dose, the dose-limiting toxicities, and the safety profile of HDM201. The secondary objectives include evaluation of PK parameters and PD markers. Patients signed informed consent, and the study was conducted in accordance with the principles of the “Declaration of Helsinki” and Good Clinical Practice.
Patient blood samples were obtained to determine plasma HDM201 concentrations within 0.5 h pre-dose and at 0.5, 1, 2, 3, 4, 8, 24 and 48 h post-dose. Bioanalytical methods for measuring HDM201 in human plasma are detailed in Supplementary Materials and Methods. All 20 patients treated with daily dosing and 24 patients treated q3w were included in the statistical analyses. PKevaluations used actual sampling times and doses, and plasma concentration–time data were analyzed using non-compartmental methods.
Serum GDF-15 was measured using the Quantikine ELISA kit (R&D Systems #DGD150). Briefly, serum samples were diluted 1:4 and transferred to pre-coated plates. The assay was run with an 8-point standard curve, and absorbance read at 450 nM. All samples were assayed in duplicate and mean values are reported if within range (25-45000 pg/mL). A %CV <20 between duplicates was considered acceptable. The fold increase in GDF-15 at 24 h post-dose HDM201 (Cycle 1, Day 2) is relative to baseline.

Results

Pharmacological activity of HDM201 in cancer cell linesHDM201, based on an imidazolopyrrolidinone scaffold, emerged from optimization of a new class of pyrazolopyrrolidinone MDM2 inhibitors we recently reported (26) (Fig. 1A). HDM201 binds selectively to MDM2 (Supplementary Table S1) exploiting a “central valine” concept that relies on the placement of a planar unsaturated core within van der Waals distance of V93, a central residue in the p53 binding pocket of MDM2 (27) (Fig. 1B).

To investigate the cellular activity of HDM201, we determined the anti-proliferative activity of HDM201 in 291 cell lines from the Cancer Cell Line Encyclopedia (CCLE) (24). Cell lines were partitioned into sensitive and insensitive groups using a maximal effect (Amax) of <-50% and a growth inhibition (GI50) cut-off of 3 µM. Seventy-six cell lines (26.1%) were categorized as sensitive and 215 cell lines (73.9%) as insensitive (Fig. 1C and Supplementary Table S2). Consistent with the mechanism of action of HDM201, most of the sensitive cell lines harbored WT p53 (74/76; 97.4%) (Fig. 1C and Supplementary Table S2) (p-value=9.1×10-30)(Supplementary Table S3). Of note, the mutant p53 lines SNU-C4 and MOLT-16 categorized as sensitive carry heterozygous p53 hot-spot mutations (allele frequencies of 30% and 20% respectively (24)) suggesting the remaining WT p53 allele may be sufficient for sensitivity to HDM201.

To assess the relationship between HDM201 inhibitor sensitivity and genetic dependence on MDM2, we compared the anti-proliferative effects of HDM201 to the shRNA-mediated knockdown of MDM2. Here, we intersected the HDM201 sensitivity data with the results of a large-scale shRNA screen we performed targeting ~7837 genes with an average of 20 shRNAs per gene across 398 CCLE cell lines (25). Among the 261 cell lines tested in both screens, we compared MDM2 gene-level shRNA ATARiS values (28) to the GI50 of HDM201 (Fig. 1D and Supplementary Table S4). Here, 62/261 (23.8%) cell lines were sensitive to MDM2 shRNA [mean ATARiS ≤-0.5 (25)] (Fig. 1D and Supplementary Table S4), comparable to the 26.1% of cell lines sensitive to HDM201 (Fig. 1C). Among those, 45/62 (72.6%) were sensitive to HDM201 (GI50≤ 3 µM) (Fig. 1D and Supplementary Table S4). Conversely, among the 199 cell lines insensitive to MDM2 shRNA, 189/199 (95.0%) cells were insensitive to HDM201. Thus, shRNA-mediated MDM2 dependency and HDM201 sensitivity are well correlated (p- value=1.1×10-26 [Supplementary Table S5]). In addition to the selective biochemical profile across several Protein-Protein Interaction TR-FRET assays (Supplementary Table S1), these results strongly suggest that HDM201 is a highly selective inhibitor of MDM2.

A pulsed high dose treatment of HDM201 is associated with rapid PUMA-mediated induction of apoptosis in vitro
We next assessed whether HDM201 cell growth inhibition is dose- and/or time-dependent in SJSA-1 cells, a WT p53 and MDM2-amplified osteosarcoma cell line that is sensitive to MDM2inhibition. In the continuous treatment, the compound was added once and left for 3 days, while in the washout experiments, HDM201 was removed after an increasing period of incubation time, as indicated in Materials and Methods. After 3 days of continuous treatment, HDM201 inhibited SJSA-1 growth, with a GI50 of 38±15 nM and GI80 of 120±27 nM (Fig. 2A and Supplementary Fig. S1A). Washout experiments showed an inverse correlation between the concentrations of HDM201 required to achieve GI80 and time of treatment (Fig. 2A and Supplementary Fig. S1A). We then selected two distinct treatment paradigms: long-term exposure (48-72 h) of cells to HDM201, referred as “continuous low dose treatment”, which was associated with a GI80 of 0.1 µM (GI80-Cont), and short-term exposure (7-10 h), referred to as “pulsed high dose treatment”, which was associated with a GI80 of 1.5 µM (GI80-Pulse). Similarly, a GI80-Cont of 0.03 µM and GI80-Pulse of 0.4 µM could be determined for MOLM-13, a non-MDM2 amplified AML cell line (Supplementary Fig. S1B).

We next compared these regimens by evaluating markers of p53 activation i.e. p21, PUMA and GDF-15 (29) (Fig. 2B). Compared to controls, both the continuous and pulsed treatments resulted in rapid p21 mRNA induction that reached similar maximal levels of induction; 16.7- fold at 16 h for the continuous regimen and 25.7-fold at 8 h for the pulsed treatment. Both treatment regimens also induced comparable levels of GDF-15 with the expected delayed onset compared to p21 (Fig. 2B). Surprisingly, we observed a striking difference in the induction of PUMA where continuous HDM201 treatment at 100 nM for 48 h led to modest PUMA induction reaching 7-fold after 48 h, while pulsed treatment at 1.5 µM for 8 h resulted in marked induction of PUMA, reaching 53.5-fold 24 h post-treatment and 16 h after compound washout (Fig. 2B). Thus, different regimens have distinct patterns of p53 target gene induction.

We next compared the ability of both HDM201 dosing regimens to promote cell death in SJSA-1 cells. Both continuous low dose and pulsed high dose HDM201 treatments led to a significant increase in apoptosis, as measured by changes in the cleaved-caspase-3/7-positive SJSA-1 cell fraction over time (Fig. 2C). Under continuous low dose (GI80-Cont) treatment, apoptotic cells appeared first at 30 h and steadily accumulated at a constant rate over 70 h (Fig. 2C and Supplementary Fig. S2A). In contrast, high dose treatment with HDM201 for 8 h (GI80-Pulse) rapidly induced apoptosis with apoptotic cells appearing at 12 h and peaking 26 h after compound removal (Fig. 2C), in keeping with the observed upregulation of HDM201-induced PUMA mRNA (Fig. 2B). These data suggest that induction of apoptosis can be achieved with both regimens, however with different kinetics of onset and duration. Eventually, the low dose continuous treatment regimen can achieve the same cumulative magnitude of apoptotic events, when cells are continuously exposed to the compounds for multiple days (Supplementary Fig. S2A). Less potent MDM2 inhibitors including CGM097 and nutlin-3a were also used as comparators (Supplementary Fig. S2B). Interestingly, measuring cleaved-caspase-3/7 over time showed that only continuous low dose treatments with CGM097 and nutlin-3a led to a significant increase in apoptosis and this increase was dose-dependent (Supplementary Fig. S2C). These data suggest that a certain potency threshold exists and that highly potent MDM2 inhibition is required for the induction of apoptosis under pulsed dose regimens.
To investigate the mechanism of HDM201 induced apoptosis, a pooled shRNA screen was performed in SJSA-1 cells treated with continuous low dose (GI80-Cont) or pulsed high dose (GI80- Pulse) treatments. As expected, shRNA knockdown of TP53 led to robust rescue/resistance to both treatments (Fig. 2D). Likewise, BBC3 (PUMA) depletion rescued cell growth in pulsed high dose (Fig. 2D, upper right panel) and, to a lesser extent, in continuous low dose (Fig. 2D, upperleft panel) treatments. Indeed, in the pulse high dose treatment, all 20 shRNAs targeting PUMA were ranked much higher as top rescuers than in the continuous low dose treatment (rank 3 vs. rank 17, respectively; Supplementary Table S6). As PUMA is the only p53 target gene where knockout leads to resistance it appears to be the major mediator of the p53 response in the pulsed high dose treatment (Fig. 2D). Interestingly, shRNAs targeting BCL2L1 (Bcl-xL) significantly sensitized to cell growth inhibition in the presence of HDM201 in both dosing regimens (Fig. 2D, bottom panels, Supplementary Data Files S1 and S2). Taken together, these data strongly suggest that the robust PUMA induction observed with the pulsed high dose treatment is at least partly responsible for the rapid onset of apoptosis elicited by HDM201.

Single high dose treatment of HDM201 induces PUMA-associated tumor regression in vivo

Next, we assessed whether we could mimic continuous low dose and pulsed high dose regimens in vivo. We selected rat as the species for in vivo testing as the robust PK behavior of HDM201 (Supplementary Fig. S3A, B and Supplementary Table S7) in rats allowed exploration of different dose and schedules.
To investigate the PK/PD relationship of HDM201, the unbound plasma drug concentrations (plasma protein binding = 87.1% in rat) (Fig. 3A) mRNA levels of the 3 relevant PD markers, the p53 target genes p21, PUMA and GDF-15 were measured (Fig. 3B). In addition, protein levels of p21, PUMA, Noxa, Bax and Bcl-xL, were investigated (Fig. 3C). Of note, a single treatment of SJSA-1 tumor-bearing rats with low (5 mg/kg) or high (27 mg/kg) doses of HDM201 showed dose-proportional exposure in plasma (Fig. 3A). Interestingly, HDM201 unbound concentration remained above the GI80-Cont (100 nM) for the first ~20 h after a low dose (Fig. 3A, left panel). Moreover, the unbound concentration remained above the in vitro GI80-Pulse (1.5 µM) for 8 h after a high dose (Fig. 3A, right panel) indicating pulsed high and continuous12 dose treatments of HDM201 can be achieved in vivo. High dose HDM201 yielded a significantly different PD response compared to low dose treatment.

Induction of PUMA, p21 and GDF-15 mRNA was modest (Emax = 7-, 15- and 8-fold, respectively) after low dose HDM201 (Fig. 3B, left panel) compared to marked induction (Emax = 44-, 46- and 56-fold, respectively) after high dose treatment (Fig. 3B, right panel). Similarly, 27 mg/kg HDM201 induced robust increases in p21 and PUMA proteins (Fig. 3C, right panel) while only p21 protein was increased after low dose treatment (Fig. 3C, left panel). Bax protein remained unchanged after low and high dose treatments while Noxa slightly increased only after high dose HDM201. Interestingly, Bcl-xL protein levels were down-regulated in a time-dependent manner after high dose (Fig. 3C, right panel) but unchanged after low dose treatments. This observation is likely due to an indirect effect of HDM201 since induction of Bcl-xL mRNA levels remained low (Supplementary Fig. S3C). p53 was upregulated in tumors (Supplementary Fig. S3D) with a maximum reached 3 h post low dose treatment and 8 h post high dose treatment. Moreover, levels of p53 were high up to 24 h post high dose treatment. Apoptosis, measured as percent cleaved-caspase-3 positive pixels, increased in tumors (Supplementary Fig. S3E) at 24 and 48 h post high dose treatment to 6.6 and 9.7%, respectively (Fig. 3D, right panel), while at low dose, the increase was modest, 2.7 and 2.2% (Fig. 3D, left panel), recapitulating the distinct kinetics observed in vitro.
Following a daily treatment of HDM201 at 5 mg/kg to mimic the continuous low dose regimen, 55% tumor regression was observed in SJSA-1 tumor-bearing rats after 3 days and complete tumor regression after 9 days of treatment (Fig. 3E, left panel). However, 50% of tumors relapsed within 14 days after stopping treatment (Supplementary Fig. S3F, left panel). Interestingly, a single treatment of HDM201 at 27 mg/kg, resulted in 82% tumor regression inafter 3 days, consistent with a more rapid induction of apoptosis (Fig. 3D and Supplementary Fig. S3E). Complete tumor regression was reached after 9 days and was sustained in all animals for 30 days after stopping treatment (Fig. 3E and Supplementary Fig. S3F, right panel). Similar findings were seen in a MDM2 amplified well-differentiated liposarcoma patient-derived xenograft (HSAX2655) (Fig. 3F). Overall, while the degree of measurable tumor responses to both HDM201 dosing regimens was comparable, the response to the single high dose treatment was faster and more persistent in vivo.
Bcl-xL expression confers resistance to HDM201 treatment specifically with intermittent high dose scheduling

The observed difference in the mechanisms of action of HDM201 resulting from continuous vs. pulsed regimen raised the possibility that the mechanisms of resistance to HDM201 might also differ. To study this, we took advantage of tumor models described in a previous screen (30) where a constitutive PB transposon-based insertional mutagenesis system in an Arf-/- background was used to characterize resistance mechanisms to HDM201. Because of the more rapid clearance of HDM201 in mice, we previously used 100 mg/kg twice a week (2qw) in mice in order to mimic the rat intermittent high dose regimen (30). In mice, HDM201 unbound concentration remained above the GI80-Pulse (1.5 µM) threshold for only ~3 h (Fig. 4A). After single doses of 40 and 100 mg/kg of HDM201 we observed PUMA and p21 induction in tumors over 24 h (Supplementary Fig. S4A and B). These were followed by an increased apoptosis (Supplementary Fig. S4C and D) when both doses were applied, albeit not to the same extent as observed in SJSA-1 tumors implanted in rats.
To compare these prior results, we administered HDM201 by continuous dosing at 40 mg/kg daily in 6 of the Arf-/- PB transplanted tumor models (2 lymphoma and 4 medulloblastoma) thatwe previously showed were responsive to the pulsed regimen (30). As expected, after multiple dosing, similar response rates were seen across these 6 models for both regimens (Fig. 4B). For the purpose of comparison, we show here the extracted data from the previous study (30) for the same 6 models, out of 16 tumor models previously published (Fig. 4B, C, D).

We then identified continuous low dose regimen-specific insertional events linked to the development of resistance to MDM2 inhibition, by comparing the genomic DNA from 34 relapsing resistant and 53 vehicle-treated tumors. Tumor DNA was subjected to splinkerette PCR and deep sequencing and gene centric common insertion sites (gCIS) landscapes were obtained (Supplementary Data File S3). A differential integration analysis for each dosing regimen identified PB target genes significantly enriched in resistant compared to vehicle-treated tumors (Fig. 4C and E, and Supplementary Data File S3). Ten and 7 genes were targeted by PB in tumors that relapsed from the continuous low dose (Fig. 4E and F) or the intermittent high dose regimens (Fig. 4C and D), respectively. Most major resistance mechanisms were similar between the regimens, including genes directly regulating p53 (Trp53, MDM4, Trp63) (Fig. 4C, D, E and F). Interestingly, Bcl2l1, encoding Bcl-xL, was only significantly enriched in the resistant tumors treated with pulsed high dose HDM201, thus highlighting Bcl-xL re-expression as a unique mechanism of resistance to the pulsed high dose regimen, in line with previous observations (30). Taken together, these data suggest that preclinical mechanism of resistance to HDM201 is regimen-dependent and that modulation of Bcl-2 family members, such as PUMA and Bcl-xL drives anti-tumor efficacy in the pulsed high dose regimen.

HDM201 displays desirable PK and PD profiles in patients

Based on the preclinical findings described above, HDM201 entered a phase I study to explore and compare both dosing regimen in p53 WT patients with solid tumors (trial registrationidentifier: NCT02143635). In this study, HDM201 is administrated either continuously (daily for 2 weeks on a 4-week cycle [q24h 2 weeks on/off regimen]) or in a pulsed/intermittent manner (once in a cycle of 3 weeks [q3w regimen]) (31).
PK measurements were collected from 46 patients following a single administration of HDM201 (Day 1) during dose escalation (1 to 350 mg) (Fig. 5A). Following oral dosing, HDM201 was rapidly absorbed, with a median time to peak plasma concentration (Tmax) of 2.0 to 5.8 h across the dose range (2 to 350 mg) (Fig. 5A and Supplementary Table S8). With daily dosing, HDM201 steady-state was generally reached by Day 8, with <2-fold accumulation after 14 days, in keeping with the preclinical in vivo data (Supplementary Tables S8 and S9). Mean half-life (T1/2) after 50 to 350 mg HDM201 ranged from 11.8 to 16.3 h (Supplementary Table S10). HDM201 showed approximately a dose proportional increase in exposure (AUClast and Cmax) after a single dose on Day 1 (1 to 350 mg) and after multiple doses on Day 14 (1 to 20 mg qd) (Fig. 5A, Supplementary Tables S8 and S9).
Serum levels of GDF-15, a secreted protein strongly induced by activated p53 in both normal and tumor tissues (32, 33), were used to assess pharmacodynamic effects of HDM201. Serum levels of GDF-15 24 h post-treatment (Cycle 1, Day 2) increased in a dose-dependent manner in patients. Low doses from 7.5 to 20 mg HDM201 led to a modest increase of serum GDF-15 (from 1.6 to 6.1-fold) (Fig. 5B). In contrast, higher doses from 50 to 350 mg HDM201 robustly increased serum GDF-15 (up to 97.1-fold) in patients (Fig. 5B). The GDF-15 induction was strongly differentiated between the two regimens and was in line with the preclinical observation (Fig. 3B).
Taken together, the PK/PD behavior of HDM201 observed in patients appears to reflect the pre- clinical differences between pulsed and continuous regimens. In addition, these data indicate thatintermittent high dose regimen is much more likely to strongly induce p53 target genes in patients. This suggests that the regimen-dependent molecular mechanism leading to a robust and early onset of apoptosis in the tumor can be reproduced in patients and further support to explore the clinical activity of HDM201 administered in intermittent high dosing regimens.

Discussion

The anti-tumor activity of MDM2 inhibitors has been extensively reported in preclinical tumor models treated with “standard” daily dosing regimens (14, 34-37). Likewise, HDM201 achieves sustained tumor regressions in a rat xenograft model at a daily low dose of 5 mg/kg. In addition, our in vitro and in vivo results show that intermittent dosing HDM201 also induces sustained anti-tumor activity. Pulsed high dose HDM201 treatment in vitro is associated with robust PUMA induction and a rapid onset of apoptosis. Consistent with this observation, a single high dose of HDM201 promotes sustained tumor regression in SJSA-1 tumor-bearing rats. Tumor regressions upon pulsed high dose regimen were also demonstrated in a patient-derived liposarcoma xenograft model in rats and in Arf-/- derived allograft tumor models in mice. In line with a previous report (35), these observations support the notion that continuous suppression of the p53-MDM2 interaction is not required for optimal anti-tumor activity, provided that a certain level of target suppression is achieved over a defined time window that is sufficient to elicit a downstream response. Furthermore, our in vitro observations suggest that while induction of apoptosis can be achieved with both dosing regimens, this occurs with different kinetics of onset,
i.e. slow and cumulative for the continuous low dose regimen and more rapid for the pulsed high dose regimen.
The conditions that influence p53 to stimulate cell cycle arrest or apoptosis are not fully understood. The choice of particular subset of p53 target genes has been shown to make thedifference between life and apoptotic death of a cell (38). How p53 level of activity can differentiate amongst its target genes is a longstanding fundamental question. Regulation of transcription factors such as p53 can occur at multiple levels, depending on their abundance, posttranslational modifications, binding to cofactors and/or through temporal dynamics. The latter seems particularly critical for p53 regulation since previous studies revealed that following γ-irradiation, p53 undergoes a complex dynamical response to DNA damage that appears to be tightly controlled by negative feedback loops between p53 and Mdm2, and between p53, ATM and Chk2 (39). In addition, such pulses of p53 protein were shown to regulate distinct patterns of p53 target gene expression (40). Conversely, one could speculate that continuous low dose and pulsed high dose regimens of HDM201 may reactivate wild-type p53 with different dynamics, further leading to the expression of distinct sets on p53 target genes. Such a hypothesis is supported by our observations that continuous vs. intermittent dosing treatment with HDM201 leads to a dynamically different induction of p53 expression in the tumor (Supplementary Fig. S3D), and may warrant further evaluation. Elevated pulsed p53 expression was previously reported to trigger p53-mediated apoptosis (17, 41) via a “so-called” intrinsic apoptotic pathway, dominated by the Bcl-2 family of proteins (21), governing mitochondrial release of cytochrome c (42, 43). Intriguingly, a subset of the Bcl-2 family genes are p53 targets such as the “BH3-only” proteins PUMA, Bax and Noxa (20). Conversely, our findings show that a pulsed high dose of HDM201 is associated with robust induction of the mRNA and/or protein levels of PUMA in vitro and in vivo (Fig. 2B and 3C). Moreover, knockdown of PUMA rescued p53-mediated growth inhibition upon pulsed high dose treatment, suggesting that PUMA is the key mediator for the HDM201-induced apoptosis under this treatment condition. Hence, robust upregulation of PUMA appears to be part of the molecular mechanism explaining why intermittent high dosetreatment with HDM201 results in complete and sustained tumor regressions in vivo. Interestingly, intermittent treatment with HDM201 also leads to the down-regulation of Bcl-xL in vivo. Collectively, our data supports schedule-specific mechanisms in triggering p53- dependent apoptosis and anti-tumor activity following HDM201 treatment. In addition, our results suggest high intermittent dosing of HDM201 leads to a p53-dependent decrease of the Bcl-2-like/BH3-only protein ratio. Importantly, re-establishing this ratio by the up-regulation of Bcl-xL was specifically correlated with acquired resistance of spontaneous tumors from the Arf null mice to intermittent HDM201 treatments. Moreover, our findings that knockdown of Bcl-xL by shRNA sensitizes SJSA-1 cells to HDM201 are in keeping with the superior preclinical anti- tumor activity recently demonstrated for the novel combinations of MDM2 inhibitors and BH3 mimetics (44-47), further supporting the ongoing clinical investigation of such combinations (trial registration identifier: NCT02670044). This is consistent with the hypothesis that an imbalance between pro-survival and pro-apoptotic members of the Bcl-2 family may be critical for the HDM201-induced durable efficacy when dosed intermittently.

An intermittent dosing schedule might provide better tolerability for patients treated with MDM2 inhibitors. Adverse effects of p53 reactivating therapies are a major impediment to achieving a robust therapeutic index for such therapies. Hematological toxicities, in particular thrombocytopenia, have been the most commonly reported and dose-limiting toxicities of this class. HDM201 recently entered a comparative clinical study with regard to efficacy and tolerability under distinct dosing regimens (trial registration identifier: NCT02143635). Importantly, preliminary PK and PD profiles from this study demonstrate a desirable PK proportionality in patients, and a robust dose-dependent PD response assessed by GDF-15
plasma levels.

This ability to administer HDM201 on a Q21 day schedule may provide an opportunity to allow for hematologic recovery and eventually better tolerability for patients.
Overall, our findings generally suggest that chronic dosing of targeted therapeutics should be re- thought if apoptotic thresholds are not robustly covered. In addition, we previously observed that high intermittent dosing treatments of weaker MDM2 inhibitors like CGM097 did not optimally achieve in vivo anti-tumor activity when compared to chronic dosing perhaps due to insufficient on-target potency. These data suggest that the impact of intermittent and profound target inhibition might not be observable with early tool compounds or sub-optimal inhibitors. Finally, intermittent dosing regimens might allow versatile utilization in combinations with either intermittent or chronically administered partner therapeutics.

Acknowledgments: The authors thank Geneviève Albrecht, Joëlle Rubert, Marc Hattenberger, Kerstin Pollehn, Jacqueline Loretan and Andreas Hueber for technical assistance with cellular assays, Marjorie Berger, Ramona Rebmann, Francesca Santacroce, Claire Estadieu, Emeline Mandon and Ernesta Dammassa for technical assistance with in vivo profiling, Nirupama Biswal and Ramu Thiruvamoor for assistance in the clinical PK and PD data analyses, the Novartis teams who performed the high-throughput cell viability assay (24) and pooled short hairpin RNA (shRNA) screen (25) and the principal investigators David M. Hyman, Manik Chatterjee, Filip de Vos, Chia-Chi Lin, Cristina Suárez, David Tai, Philippe Cassier, Noboru Yamamoto, Vincent A. de Weger and Sebastian Bauer who are conducting the clinical studies with HDM201.

References:

1. Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000;408:307-10.

2. Harris SL, Levine AJ. The p53 pathway: positive and negative feedback loops. Oncogene

2005;24:2899-908.

3. Pant V, Lozano G. Limiting the power of p53 through the ubiquitin proteasome pathway.

Genes Dev 2014;28:1739-51.

4. Tsvetkov P, Reuven N, Shaul Y. Ubiquitin-independent p53 proteasomal degradation. Cell Death Differ 2010;17:103-8.
5. Meek DW, Anderson CW. Posttranslational modification of p53: cooperative integrators of function. Cold Spring Harb Perspect Biol 2009;1:a000950.10.1101.
6. Brooks CL, Gu W. p53 regulation by ubiquitin. FEBS Lett 2011;585-2803-9.

7. Kubbutat MH, Jones SN, Vousden KH. Regulation of p53 stability by Mdm2. Nature

1997;387:299-303.

8. Haupt Y, Maya R, Kazaz A, Oren M. Mdm2 promotes the rapid degradation of p53. Nature

1997:387:296-9.

9. Li M, Brooks CL, Wu-Baer F, Chen D, Baer R, Gu W. Mono- versus polyubiquitination: differential control of p53 fate by Mdm2. Science 2003;302:1972-5.
10. Shadfan M, Lopez-Pajares V, Yuan ZM. MDM2 and MDMX: alone and together in regulation of p53. Transl Cancer Res 2012;1:88-9.

11. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Donovan J, et al. The landscape of somatic copy-number alteration across human cancers. Nature 2010;463:899- 905.
12. Vassilev LT, Vu BT, Graves B, Carvajal D, Podlaski F, Filipovic Z, et al. In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science 2004;303:844-8.
13. Zhao Y, Aguilar A, Bernard D, Wang S. Small-molecule inhibitors of the MDM2-p53 protein-protein interaction (MDM2 Inhibitors) in clinical trials for cancer treatment. J Med Chem 2015;58:1038-52.
14. Jeay S, Gaulis S, Ferretti S, Bitter H, Ito M, Valat T, et al. A distinct p53 target gene set predicts for response to the selective p53-HDM2 inhibitor NVP-CGM097. eLife 2015;4:e06498.
15. Kastan MB, Zhan Q, el-Deiry WS, Carrier F, Jacks T, Walsh WV, et al. A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia. Cell 1992;71:587-97.
16. Lowe SW, Schmitt EM, Smith SW, Osborne BA, Jacks T. p53 is required for radiation- induced apoptosis in mouse thymocytes. Nature 1993;362:847-9.
17. Zhang XP, Liu F, Wang W. Two-phase dynamics of p53 in the DNA damage response. Proc Natl Acad Sci USA 2011;108:8990-5.
18. Reisman D, Takahashi P, Polson A, Boggs K. Transcriptional regulation of the p53 tumor suppressor gene in S-phase of the cell-cycle and the cellular response to DNA damage. Biochem Res Int 2012; 2012:808934.

19. Fischer M, Quaas M, Steiner L, Engeland K. The p53-p21-DREAM-CDE/CHR pathway regulates G2/M cell cycle genes. Nucleic Acids Res 2016;44:164-74.
20. Kuribayashi K, Finnberg N, Jeffers JR, Zambetti GP, El-Deiry WS. The relative contribution of pro-apoptotic p53-target genes in the triggering of apoptosis following DNA damage in vitro and in vivo. Cell Cycle 2011;10:2380-9.
21. Haupt S, Berger M, Goldberg Z, Haupt Y. Apoptosis – the p53 network. J Cell Sci

2003;116:4077-85.

22. Kracikova M, Akiri G, George A, Sachidanandam R, Aaronson SA. A threshold mechanism mediates p53 cell fate decision between growth arrest and apoptosis. Cell Death Differ 2013;20:576-88.
23. Khoo KH, Verma CS, Lane DP. Drugging the p53 pathway: understanding the route to clinical efficacy. Nat Rev Drug Discov 2014 ;13:217-36.
24. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 2012;483:603-7.
25. McDonald III ER, de Weck A, Schlabach MR, Billy E, Mavrakis KJ, Hoffman GR, et al.

Project DRIVE: a compendium of cancer dependencies and synthetic lethal relationships uncovered by large-scale, deep RNAi screening. Cell 2017;170:577-92.
26. Furet P, Masuya K, Kallen J, Stachyra-Valat T, Ruetz S, Guagnano V, et al. Discovery of a novel class of highly potent inhibitors of the p53-MDM2 interaction by structure-based design starting from a conformational argument. Bioorg. Med. Chem. Lett. 2016 ;26:4837-41.

27. Furet P, Chène P, De Pover A, Valat TS, Lisztwan JH, Kallen J, Masuya K. The central valine concept provides an entry in a new class of non peptide inhibitors of the p53-MDM2 interaction. Bioorg Med Chem Lett 2012;22:3498-502.
28. Shao DD, Tsherniak A, Gopal S, Weir BA, Tamayo P, Stransky N, et al. ATARiS: Computational quantification of gene suppression phenotypes from multisample RNAi screens. Genome Res 2013;23:665-78.
29. Riley T, Sontag E, Chen P, Levine A. Transcriptional control of human p53-regulated genes.

Nat Rev Mol Cell Biol 2008;9:402-12.

30. Chapeau EA, Gembarska A, Durand EY, Mandon E, Estadieu C, Romanet V, et al.

Resistance mechanisms to TP53-MDM2 inhibition identified by in vivo piggyBac transposon mutagenesis screen in an Arf-/- mouse model. Proc Natl Acad Sci USA 2017;114:3151-6.
31. Hyman D, Chatterjee M, Langenberg MHG, Lin CC, Suárez C, Tai D, et al. Dose- and regimen-finding phase I study of NVP-HDM201 in patients with TP53 wild-type advanced tumors, in Proceedings of the 28th EORTC-NCI-AACR Molecular Targets and Cancer Therapeutics Symposium, Munich, Germany, 29 November to 2 December 2016.
32. Yang H, Filipovic Z, Brown D, Breit SN, Vassilev LT. Macrophage inhibitory cytokine-1: a novel biomarker for p53 pathway activation. Mol Cancer Ther 2003;2:1023-9.
33. Andreeff M, Kelly KR, Yee K, Assouline S, Strair R, Popplewell L, et al. Results of the phase I trial of RG7112, a small-molecule MDM2 antagonist in leukemia. Clin Cancer Res 2016;22:868-76.

34. Tovar C, Graves B, Packman K, Filipovic Z, Higgins B, Xia M, et al. MDM2 small-molecule antagonist RG7112 activates p53 signaling and regresses human tumors in preclinical cancer models. Cancer Res 2013;73:2587-97.
35. Higgins B, Glenn K, Walz A, Tovar C, Filipovic Z, Hussain S, et al. Preclinical optimization of MDM2 antagonist scheduling for cancer treatment by using a model-based approach. Clin Cancer Res 2014;20:3742-52.
36. Canon J, Osgood T, Olson SH, Saiki AY, Robertson R, Yu D, et al. The MDM2 Inhibitor AMG 232 demonstrates robust antitumor efficacy and potentiates the activity of p53- inducing cytotoxic agents. Mol Cancer Ther 2015;14:649-58.
37. Wang S, Sun W, Zhao Y, McEachern D, Meaux I, Barrière C, et al. SAR405838: an optimized inhibitor of MDM2-p53 interaction that induces complete and durable tumor regression. Cancer Res 2014;74:5855-65.
38. Oren M. Decision making by p53: life, death and cancer. Cell Death Differ 2003;10:431-42.

39. Batchelor E, Loewer A, Lahav G. The ups and downs of p53: understanding protein dynamics in single cells. Nat Rev Cancer 2009;9:371-7.
40. Hafner A, Stewart-Ornstein J, Purvis JE, Forrester WC, Bulyk ML, Lahav G. p53 pulses lead to distinct patterns of gene expression albeit similar DNA-binding dynamics. Nat Struct Mol Biol 2017;24:840-7.
41. Chen X, Ko LJ, Jayaraman L, Prives C. p53 levels, functional domains, and DNA damage determine the extent of the apoptotic response of tumor cells. Genes Dev 1996;10:2438-51.
42. Cory S, Adams JM. The Bcl2 family: regulators of the cellular life-or-death switch. Nat Rev Cancer 2002;2:647-56.

43. Kuwana T, Mackey MR, Perkins G, Ellisman MH, Latterich M, Schneiter R, et al. Bid, Bax, and lipids cooperate to form supramolecular openings in the outer mitochondrial membrane. Cell 2002;111:331-42.
44. Lehmann C, Friess T, Birzele F, Kiialainen A, Dangl M. Superior anti-tumor activity of the MDM2 antagonist idasanutlin and the Bcl-2 inhibitor venetoclax in p53 wild-type acute myeloid leukemia models. J Hematol Oncol 2016;9:50-62.
45. Carter BZ, Mak PY, Mak DH, Ruvolo VR, Schober W, McQueen T, et al. Synergistic effects of p53 activation via MDM2 inhibition in combination with inhibition of Bcl-2 or Bcr-Abl in CD34+ proliferating and quiescent chronic myeloid leukemia blast crisis cells. Oncotarget 2015;6:30487-99.
46. Hoffman-Luca CG, Ziazadeh D, McEachern D, Zhao Y, Sun W, Debussche L, Wang S. Elucidation of acquired resistance to Bcl-2 and MDM2 inhibitors in acute leukemia in vitro and in vivo. Clin Cancer Res 2015;21:2558-68.
47. Gu D, Wang S, Kuiatse I, Wang H, He J, Dai Y, et al. Inhibition of the MDM2 E3 Ligase induces Siremadlin apoptosis and autophagy in wild-type and mutant p53 models of multiple myeloma, and acts synergistically with ABT-737. PLoS One 2014;9:e103015.