Pharmacological activation of p53 induces dose-dependent changes in endothelial cell fate during angiogenic sprouting

ARTICLE OPEN Pharmacological activation of p53 induces dose-dependent changes in endothelial cell fate during angiogenic sprouting Omayma Al-Radi 1 , Katrine Ingelshed 1,2 , Lisa Eichhorn 1 , Heidi Josefsson 1 , Martin Krkoska 1,3 , Lars Bräutigam 4 , Susanne Lindström 4 , Ákos Végvári 5 , Sania Kheder 1 , Carmine P. Cerrato 1 , Suzon Fermé 1 , Cecilia Bosdotter 1 , Amin Allalou 6,7,8 , Fredrik Levander 9 , Borivoj Vojtesek 3 , David P. Lane 1 and Pavitra Kannan 1 ✉ © The Author(s) 2025 The cell cycle is a key regulator of endothelial cell specification into tip and stalk cell phenotypes, which are essential for angiogenesis in both normal development and pathological conditions. While the tumor suppressor p53 is known to regulate the cell cycle and influence cell fate, its role in modulating the cell fate of these phenotypes remains unclear. Using non-genotoxic small molecule and stapled peptide compounds to pharmacologically activate p53 via MDM2 inhibition, we demonstrate that graded levels of p53 induce distinct cellular fates in normal endothelial cells. Low levels of p53 induce reversible cell cycle arrest by reducing DNA replication, while high levels induce senescence and cell death. Surprisingly, all tested levels of p53 activation reduced the growth of venous blood vessels in vitro and in zebrafish embryo models. This reduction in sprouting may stem from distinct cellular responses in tip-like and non-tip-like cells to pharmacological p53 activation: low p53 levels primarily reduced proliferation in non-tip-like cells, whereas high levels decreased the frequency of tip-like cells and the expression of genes associated with tip and stalk cell identities. Our findings show for the first time that pharmacological p53 activation modulates endothelial cell fate in a dose-dependent manner during sprouting angiogenesis. They also highlight the potential of using graded p53 modulation as a therapeutic strategy to target abnormal tip or stalk cell development in pathological angiogenesis, such as in cancer. Cell Death and Disease (2025) 16:883 ; https://doi.org/10.1038/s41419-025-08292-7 INTRODUCTION Angiogenesis is a fundamental biological process during which endothelial cells undergo several phenotypic transitions to form new blood vessels from existing vessels [1,2]. When exposed to growth stimuli, a single tip cell extends from the existing vessel and migrates toward the growth signal using filopodia, while adjacent stalk cells elongate the sprout [3]. The specification of tip and stalk cell phenotypes is carefully regulated during development and tissue repair. However, it becomes dysregulated in cancer, age-related macular degeneration, and ischemic conditions, leading to excessive or insufficient vessel growth that can enable tumor growth or cause tissue damage, respectively. Understanding the mechanisms that regulate the specification of endothelial cells into these phenotypes is crucial for improving treatment strategies for these diseases [3]. Preclinical studies have shown that the cell cycle plays a critical role in tip and stalk cell specification during angiogenesis [4,5]. In zebrafish embryo models of sprouting angiogenesis, tip cells emerge from the dorsal aorta in S/G2/M phases [6] or from the posterior cardinal vein in G1 phase [7]. Vessel growth subsequently occurs through tip cell division, stalk cell proliferation, or tip and stalk cell proliferation [6–9]. In contrast, these mechanisms differ in mouse retinal models of sprouting angiogenesis. Tip cells emerge from the G1 phase and rarely divide, while stalk cells proliferate at the angiogenic front [10]. Despite the differences between these models, chemical or genetic disruption of cell cycle phases leads to vascular abnormalities in both models [6–10], indicating that cell cycle regulators are integral to specification of endothelial cell phenotypes during angiogenesis. A key regulator of cell cycle progression is the tumor suppressor protein p53. It is kept at low levels through proteosomal degradation under unstressed conditions. However, p53 is stabilized under stressed conditions, resulting in the transcriptional activation of its downstream targets that mediate cell cycle arrest, senescence, or apoptosis [11]. While p53 deletion in endothelial cells does not impair vascular development [12], its activation in normal endothelial cells has been shown both to promote the initiation of sprouting through increased cell cycle arrest [7] and to reduce sprouting through Received: 4 February 2025 Revised: 21 October 2025 Accepted: 24 November 2025 1 Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Stockholm, Sweden. 2 Department of Immunology, Genetics and Pathology, Uppsala University, Uppsala, Sweden. 3 RECAMO, Masaryk Memorial Cancer Institute, Brno, Czech Republic. 4 Comparative Medicine, Karolinska Institutet, Stockholm, Sweden. 5 Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Stockholm, Sweden. 6 Department of Information Technology, Uppsala University, Uppsala, Sweden. 7 DanioReadout, Immunology Genetics and Pathology, Uppsala University, Uppsala, Sweden. 8 SciLifeLab BioImage Informatics Facility, Uppsala University, Uppsala, Sweden. 9 Department of Immunotechnology, National Bioinformatics Infrastructure Sweden, Science for Life Laboratory, Lund University, Lund, Sweden. ✉email: [email protected] Edited by Gerry Melino www.nature.com/cddis Official journal of CDDpress 1234567890();,: decreased proliferation [13] or enhanced apoptosis [14]. However, in vascular diseases, the activation of p53 in endothelial cells appears to be primarily associated with decreased vessel formation [15]. In cancer and other normal cells, the level of p53 activation can lead to distinct cell fates [16]. In the context of angiogenesis, these data raise the possibility that p53 levels may determine angiogenic outcome, and/or may differentially affect tip and stalk cells. Yet, the effects of p53 activation on the fate of these endothelial phenotypes during sprouting angiogenesis are not known. O. Al-Radi et al. 2 Cell Death and Disease (2025) 16:883 We hypothesized that different levels of p53 influence endothelial cell fate by triggering distinct responses in tip vs stalk cells, thereby resulting in varied phenotypic effects during angiogenesis. Specifically, we predicted that low levels of p53 activation could promote the initiation of sprouting by inducing cell cycle arrest, thereby increasing the proportion of tip cells. In contrast, we predicted that high levels of p53 activation could reduce sprouting by inducing cell death in both tip and stalk cells. To test this hypothesis, we studied the molecular and phenotypic effects by which different levels of p53 activation affect endothelial cells using cellular assays, proteomics, and in vitro and in vivo sprouting assays. We used a pharmacological approach to induce different levels of p53 by applying graded concentrations of a new generation of small molecule and stapled peptide compounds that transcriptionally activate p53 via inhibition of its negative regulator MDM2 [17–21]. Our findings reveal that different levels of p53 activation trigger distinct cellular responses in each population, but surprisingly lead to the same angiogenic outcome of reducing vessel growth. RESULTS Graded p53 activation by MDM2 inhibitors induces distinct cell fates in endothelial cells The molecular consequences of activating p53 in normal endothelial cells were investigated using two small molecule MDM2 inhibitors and one stapled peptide MDM2/MDMX inhibitor, which stabilize p53 by preventing it from being targeted for degradation by MDM2. Cellular growth of human umbilical vein endothelial cells (HUVEC) decreased by nearly 100% in a concentration-dependent manner upon treatment with the small molecules navtemadlin and nutlin-3a, and the stapled peptide sulanemadlin (Fig. 1A). Cellular growth was more potently inhibited by navtemadlin (IC 50 =0.010 µM [95% CI =0.007–0.013]) than by nutlin-3a (IC 50 =0.340 µM [95% CI =0.266–0.413]) or sulanemadlin (IC 50 =0.046 µM [95% CI =0.028–0.064]). In contrast, treatment with a control stapled peptide did not affect HUVEC growth at any concentration (Fig. 1A). The growthreducing effects were not specific to HUVEC, as growth reductions of 60–75% after navtemadlin treatment were also observed in normal human dermal microvascular endothelial cells (IC 50 =0.319 µM [95% CI =0.171–0.468]) and normal human dermal fibroblasts (IC 50 =0.015 µM [95% CI =0.006–0.024], Fig. 1B). Since navtemadlin was the most potent inhibitor among those tested, it was used in subsequent experiments. Within the experimental time frame, the growth-reducing effects induced by p53 activation in HUVEC could be recovered at the population level at low concentrations, but not at high concentrations of navtemadlin treatment (Fig. 1C). While HUVEC treated with 0.01 µM navtemadlin for 24 h had no significant effects on regrowth (P adj at 72h =0.70), cells treated with 0.05 µM (P adj at 72h =0.03) or 0.1 µM navtemadlin (P adj at 72h =0.007) for 24 h had significant delays in regrowth. In contrast, cells treated with 0.5 µM navtemadlin for 24 h did not regrow in the experimental time frame (Fig. 1C). Moreover, in brightfield images of HUVECs, increasing concentrations of navtemadlin were associated with cell rounding after 72 h (Fig. 1D), indicative of cell death. Cellular rounding was also observed after 72 h treatment in dermal microvascular endothelial cells, but not in dermal fibroblasts (Fig. 1D), highlighting the remarkable sensitivity of endothelial cells to pharmacological p53 activation. We reasoned that the observed morphological changes in endothelial cells after navtemadlin treatment might reflect a switch from cell cycle arrest to cell death. To capture the molecular changes that precede morphological changes, we measured the expression of key proteins in the p53 pathway following 24 h treatment with low (0.05 μM) and high (1 μM) concentrations of navtemadlin. In western analysis of HUVEC lysates, the expression of MDM2 (negative regulator of p53), p53, and p21 (marker of cell cycle arrest) proteins each increased more than 1.7-fold at 0.05 μM and more than 4-fold at 1 μM. In contrast, the expression of p53upregulated modulator of apoptosis (PUMA) did not measurably increase at 0.05 μM but increased by 4-fold at 1 μM(Fig.1E). Similar results were found using immunofluorescence staining in HUVECs. The expression of p53 increased at 0.05 μM with weak evidence (P adj =0.05) and at 1 μM with strong evidence (P adj =0.002), while p21 increased at both concentrations (Fig. 1F–H; P adj for 0.05 μM = 0.002; P adj for 1 μM =0.0006). The expression of Ki67 (marker of proliferation found in active phases of the cell cycle) decreased only at 1 μM(Fig.1F, I; P adj for 0.05 μM =0.26; P adj for 1 μM =0.003); given that Ki67 levels decline gradually during G 0 /G 1 arrest as a function of time, transient decreases in proliferation may not be reflected by a measurable drop in Ki67 levels at 0.05 μM[22,23]. Treatment with 1μM navtemadlin was also associated with increased cell death as measured by increased PUMA expression (Fig. 1F, J; P adj for 0.05 μM = 0.11; P adj for 1 μM =0.008) and by live imaging of dead cells (Fig. 1F, Fig. 1 Pharmacological activation of p53 induces cell cycle arrest, cell death, and senescence in endothelial cells in a concentrationdependent manner. A Concentration-dependent growth inhibition of human umbilical vein endothelial cells (HUVEC) by three p53-activators (MDM2 inhibitors navtemadlin and nutlin-3a; MDM2/MDMX inhibitor sulanemadlin), but not by non-specific control peptide. Cell growth was measured by live-cell imaging as the percent confluence normalized to untreated wells following 72 h treatment. Data points show averaged value from one experiment (n=3 experiments) and are fitted with a best fit model for concentration-growth response. BGrowth inhibition of human dermal microvascular endothelial cells (HDMEC) and normal human dermal fibroblasts (NHDF) by navtemadlin, as measured by livecell imaging. Data points show averaged value from one experiment (n=3 experiments) and are fitted with a best fit model for concentrationgrowth response. CGrowth inhibition of HUVEC recovers within 48 h following an initial 24 h treatment using ≤0.1 μM navtemadlin. Cell growth was measured by live-cell imaging and quantified over 72 h as cell counts per well and normalized to counts at time 0. Data points show mean ± SD (n=3 independent experiments). *P adj < 0.05, **P adj < 0.01 using 1-way, repeated measures ANOVA with adjustment using Dunnett’s correction. DMorphological abnormalities are visible in phase-contrast images of venous ECs (HUVEC) and capillary ECs (HDMEC), but not in those of fibroblasts (NHDF) after 72 h of navtemadlin treatment, but not after 24 h. Scale bar =200 μm. EIncreased expression of proteins involved in p53 signaling (MDM2, clone IF2, 0.5 μg/mL; p53, clone DO-1, 0.4 μg/mL), cell cycle arrest (p21, clone 12D1, 0.24 μg/mL), and apoptosis (PUMA, clone D30C10, 0.96 μg/mL) in HUVEC following 24 h navtemadlin treatment, as determined by western blot analysis. Total protein controls correspond to distinct membranes (L1 or L2). Images are cropped from full-length blots of one biological experiment (see ‘Full Length Western Blots’) and are representative of at least two experiments. Further details regarding antibodies used are shown in SI Table 1. FIncreased expression of p53 (clone DO-1, 12 μg/mL), cell cycle arrest (p21, clone 12D1, 1.22 μg/mL), apoptosis (PUMA, clone D30C10, 4.8 μg/mL), dead cells (Sytox green, 100 nM), and senescence (β-galactosidase), and reduced expression of active cell cycle (Ki67, clone SP6, 0.12 μg/mL), following 24 h treatment of HUVEC using navtemadlin as visualized by immunofluorescence, live-cell fluorescence imaging, and colorimetric staining. Scale bar =50 μm. Further details regarding antibodies and concentrations used are shown in SI Table 1. G–K Quantification of fluorescence and colorimetric levels of protein markers, showing increased expression of p53, cell cycle arrest (p21), apoptosis (PUMA), cell death (Sytox green), and senescence, and reduced activity in cell cycle (Ki67). Data points indicate value from one experiment (n=3 experiments). **P adj < 0.01, ***P adj < 0.001 using one-way ANOVA with adjustment using Dunnett’s correction. Horizontal black line indicates the mean value. O. Al-Radi et al. 3 Cell Death and Disease (2025) 16:883 K; P adj for 0.05 μM =0.43; P adj for 1 μM =0.003), as well as with induction of senescence as measured by β-galactosidase (Fig. 1F, L; P adj for 0.05 μM =0.18; P adj for 1 μM =0.001). Taken together, these results suggest that the low concentrations of pharmacological p53 activation in endothelial cells mainly result in cell cycle arrest, while higher concentrations result in cell death and senescence. The cellular effects induced by navtemadlin are largely p53dependent To test whether the molecular and phenotypic changes induced by navtemadlin are p53-dependent, we performed molecular assays using HUVECs transfected with Dicer-substrate siRNA (DsiRNA) targeting TP53 or a non-human control sequence. Since the phenotypic changes observed at 24 h could confound O. Al-Radi et al. 4 Cell Death and Disease (2025) 16:883 interpretation of p53 transcriptional activity, we measured gene expression after 6 h of treatment with 1 μM navtemadlin—a concentration at which off-target effects are more likely to be detected. Following navtemadlin treatment, TP53 mRNA levels were 77% lower in TP53-knockdown cells compared to controlknockdown cells (Fig. 2A; P adj =0.007); no significant changes were measured in DMSO-treated cells. Similarly, upon navtemadlin treatment, the mRNA levels of canonical p53 targets CDKN1A and MDM2 were >70% lower in TP53-knockdown cells than in control-knockdown cells (Fig. 2B, C; P adj for CDKN1A =0.01; P adj for MDM2 =0.006). Despite the partial knockdown, we observed a rescue of phenotypic cell rounding at 1 μM navtemadlin, consistent with reduced cell death triggered by p53 activation (Fig. 2D). The blunted activity of navtemadlin in TP53-knockdown cells was also evident at the protein level, as measured by western blotting and immunofluorescence staining. In western analysis of lysates treated for 24 h with 0.05 μM navtemadlin, the expression of p53 and p21 were reduced by more than 1.3-fold in TP53knockdown cells compared to control-knockdown cells. At 1 μM navtemadlin, the expression of p53 and p21 were reduced by more than 1.7-fold in TP53-knockdown cells, while expression of PUMA was 2.4-fold lower than in control-knockdown cells (Fig. 2E). In immunofluorescence staining, a 44% reduction in p53 and a 35% reduction in p21 levels were also measured in TP53-knockdown cells, but only at 1 μM navtemadlin (Fig. 2F–H; P adj for p53 = 0.005; P adj for p21 =0.02). When taken together, these results provide strong evidence that the molecular and phenotypic effects exerted by navtemadlin are mediated through p53. Different protein networks are affected by increasing levels of p53 activation in HUVEC To identify broader molecular changes associated with graded levels of p53 activation, we performed mass spectrometry-based proteomics on HUVEC lysates treated for 24 h with 0 µM, 0.05 µM, or 1 µM navtemadlin (Fig. 3A). After confirming similarities in the distributions of protein abundance across samples (SI Fig. 1A), we selected mean-normalized data based on dendrogram separation (SI Fig. 1B) to perform principal component analysis. Samples primarily clustered by treatment group (Fig. 3B). Of the 2800 total detected proteins, 87 total proteins at 0.05 µM and 632 total proteins at 1 µM were differentially abundant between the DMSOtreated and navtemadlin-treated samples (two-way ANOVA, FDR < 0.03, SI Fig. 2A). For example, treatment with 0.05 µM navtemadlin resulted in downregulation of proteins involved in cell cycle regulation, such as MCM2, CDK1, and H4C8. In contrast, treatment with 1 µM led to downregulation of many proteins, including THBS1 and PECAM (two proteins involved in angiogenesis), as well as, upregulation of proteins, such as TP53I3 (a p53induced protein involved in oxidative stress response). Treatment at both concentrations resulted in differential expression of 79 proteins (SI Fig. 2A), including upregulation of proteins such as TIGAR (p53-induced protein involved in apoptosis) and downregulation of proteins such as KPNA2 (a nuclear export protein) (Fig. 3C; SI Fig. 2B–D). Unlike in the navtemadlin-treated tumor cell proteome [24], we did not detect MDM2 or p21 in the navtemadlin-treated endothelial proteome by mass spectrometry; the proteins were, however, detected by western analysis (Fig. 1E). The proteomics results were confirmed by immunoblotting: the expression of THBS1, PECAM and KPNA2 decreased with navtemadlin treatment, while that of TP53I3 and TIGAR increased with treatment (SI Fig. 2E). Given that THBS1 and PECAM are proteins involved in angiogenesis, we also tested whether the decrease measured upon navtemadlin treatment was p53dependent and confirmed that their expression was rescued in TP53-knockdown cells (SI Fig. 2F). Functional overrepresentation and network analyses were then applied to identify the biological processes enriched within the differentially expressed proteins (Fig. 3D–F). Enriched proteins at the low concentration were associated with Gene Ontology (GO) terms of Biological Processes such as DNA replication (q-value =6.3 × 10 −10 ) and cell cycle (q-value =2.5 × 10 −5 ), while those at the high concentration were associated with cytoplasmic translation (q-value =1×10 −19 ) and ribosome biogenesis (q-value =1.6 × 10 −4 ). By analyzing the network of proteinprotein interactions, we observed that dense clusters of proteins within these pathways were affected by navtemadlin treatment (SI Fig. 2B–D). Given that TMT-based MS captures only a subset of the proteome, we ran functional overrepresentation analysis on all the detected proteins to identify which protein families are generally enriched in the endothelial proteome. Among the 52 significant pathways (P< 0.05) were those involved in cytoskeletal regulation, integrin signaling, fibroblast growth factor (FGF) signaling, metabolism, cell cycle, and angiogenesis. Together, these results corroborate our cellular assays showing that navtemadlin induces different effects in a concentration-dependent manner: while both low and high concentrations of navtemadlin reduce DNA replication and cell cycle proteins, high concentrations of navtemadlin also affect translation and ribosome biogenesis. All tested levels of p53 activation reduce vessel growth and integrity in vitro and in vivo Given that disruption of cell cycle progression has previously been shown to increase sprouting and to impair vascular sprout formation, we predicted based on our in vitro and proteomics data that different levels of p53 activation would likely have a similar effect on vessel growth. To test this hypothesis, we used an in vitro sprouting angiogenesis assay (Fig. 4A) in the first instance to measure the sprouting of endothelial spheroids upon exposure to vascular endothelial growth factor (VEGF) and Fig. 2 Knockdown of TP53 reduces activation of p53 pathway induced by navtemadlin in HUVEC. A–CKnockdown of TP53 reduces the activation of TP53 and its target genes CDKN1A and MDM2 induced by 6 h treatment with 1 μM navtemadlin in HUVEC, as measured by RTqPCR. Cells were transfected for 24 h using Dicer-substrate short interfering RNA (DsiRNA) targeting a negative control sequence (“Ctrl”)ora pool of three DsiRNAs targeting TP53 before navtemadlin treatment in fresh medium. Data points represent values from one experiment (n=3 independent experiments) and are plotted as log 10 fold change (FC). Bar height indicates mean value. Statistical analysis was performed on the delta Ct values, after normalization to housekeeping gene B2M.*P adj < 0.05, **P adj < 0.01, ***P adj < 0.001 using two-way ANOVA with adjustment using Šidák’s correction. DPhase-contrast images of HUVEC following knockdown of TP53 show phenotypic rescue of morphological changes observed after 24 h treatment using 1 μM navtemadlin. Scale bar =300 μm. EIn HUVEC treated with TP53 DsiRNA, protein levels of p53, p21, and PUMA are reduced following 24 h navtemadlin, as determined by western blot analysis. Images are cropped from full-length blots of one biological experiment (see ‘Full Length Western Blots’) and are representative of at least three experiments. FIn HUVEC treated with TP53 DsiRNA, protein levels of p53 and p21 are reduced following 24 h navtemadlin treatment, as visualized by immunofluorescence staining. Scale bar =50 μm. G,HQuantification of fluorescence levels of p53 (G) and p21 (H) shows reduced expression in HUVEC treated with TP53 DsiRNA compared to those treated with control DsiRNA. Data points indicate averaged value from one experiment (n=3 experiments). *P adj < 0.05, **P adj < 0.01, ***P adj < 0.001 using two-way ANOVA with adjustment using Dunnett’s correction. Bar height indicates mean value. Statistical analysis was performed on log-transformed data, but plotted on linear scale to show differences more clearly. O. Al-Radi et al. 5 Cell Death and Disease (2025) 16:883 increasing concentrations of navtemadlin (Fig. 4B). Unexpectedly, all tested concentrations of navtemadlin decreased the length of VEGF-induced sprouting (baseline-corrected) in HUVEC spheroids by > 67% (Fig. 4C) and in microvascular HDMEC spheroids by > 80% (SI Fig. 3A, B). In HUVEC spheroids, the fraction of sprouts that were still associated with the spheroid body also decreased (Fig. 4D), as did the total number of sprouts emerging from each spheroid (Fig. 4E). To confirm that the reduced sprouting induced by navtemadlin is p53-dependent, we then performed the sprouting assay in TP53or controlknockdown cells (Fig. 4F). Although TP53-knockdown alone reduced sprouting compared to the control knockdown (P adj for TP53 KD vs control KD =0.02), treatment with navtemadlin did not further reduce sprouting in TP53-knockdown spheroids (Fig. 4G, H; P adj for TP53 KD DMSO vs TP53 KD navt =0.48). Given that migrated sprouts in navtemadlin-treated spheroids still appeared to extend filopodia (Fig. 4B), we subsequently assessed whether navtemadlin treatment led to reduced cell migration using scratch-wound assays on monolayer HUVEC. While VEGF (20 ng/mL) significantly increased migration of HUVEC at 10 h (P adj =0.0003), navtemadlin did not significantly reduce VEGF-induced migration at either of the tested concentrations (SI Fig. 4A, B); further time points were not measured to avoid confounding effects of cell proliferation. We confirmed that navtemadlin treatment for 10 h induced p53 and p21 levels using both western blotting and immunofluorescence staining, excluding the possibility that the drug had no effect at this time point (SI Fig. 4C, D). Because sprout stability is typically maintained by stalk cells through formation of tight and adherent junctions in the growing vessel [25], we reasoned that navtemadlin may affect the connectivity between endothelial cells. Using immunofluorescence staining and an endothelial permeability assay on confluent endothelial monolayers, we found that both concentrations of treatment led to a reduction in the expression of zona occludens-1 (a tight junction marker) and VE-cadherin (an adherent junction marker; SI Fig. 5A) after 24 h. Additionally, both concentrations also increased the permeability (or leakiness) of a fluorescent dextran across the endothelial barrier (SI Fig. 5B, C). Although the increased permeability measured at 1 µM navtemadlin may have resulted from an increase in cell death (SI Fig. 5D–F), the increased permeability measured at 0.05 µM navtemadlin without significant Fig. 3 Low levels of p53 activation alter protein networks in DNA replication while high levels alter those involved in stress responses and ribosome assembly. A Schematic of proteomics workflow in which HUVEC were treated with navtemadlin for 24 h, digested, and quantified using TMT-based mass spectrometry. BProtein abundances (mean-normalized) separate by treatment (DMSO vs navtemadlin), as assessed by principal component analysis. Percentage of variance is shown in parentheses. CVolcano plots showing differentially expressed proteins with decreased and increased abundance between DMSO-treated and navtemadlin-treated samples. Downregulated proteins include CDK1 (involved in cell cycle), THBS1 and PECAM (associated with angiogenesis), and KPNA2 (important for nuclear export), while upregulated proteins include those TIGAR and TP53I3 (involved in p53 signaling and apoptosis). FDR on P adj < 0.03. DHeat map depicting the top 20 enriched terms within ontological biological processes following p53 activation by navtemadlin. Enriched terms are involved in p53 signaling, DNA replication, and ribosome assembly. Proteins were mapped to their corresponding gene symbols in Metascape for functional enrichment analysis using various ontologies (GO biological processes, GO molecular functions, KEGG pathways, Reactome, and canonical pathways). The color gradient represents P-values (as -log 10 ). ECircos plot depict partial overlap in ontology terms between low and high levels of p53 activation induced by navtemadlin. Dark orange segments indicate shared ontology terms between treatment groups, while light orange segments indicate ontology terms unique to each treatment. Proteins were mapped to their corresponding gene symbols for enrichment analysis using Metascape. FDifferent biological functions are affected at low and high levels of p53 activation induced by navtemadlin. After protein hits were converted to their respective gene symbols (using Metascape), functional analysis was performed to identify enrichment in ontology terms. The top 20 clusters of enriched ontology terms were then selected using a P-value < 0.01, a minimum count of 3 genes, and an enrichment factor > 1.5 (i.e., pathway is represented 1.5 more frequently in list than would be expected by chance). Each node (circle) in the network represents an ontology term comprising a group of genes that share a common biological function. The size of the node indicates the number of genes comprising that group, and the color of the pie charts within each node represents treatment groups. Nodes are grouped into clusters using the Kappa similarity score, so that clustered nodes are functionally related. Each cluster is labeled with a representative term summarizing its main biological function. O. Al-Radi et al. 6 Cell Death and Disease (2025) 16:883 cell death suggests that even low levels of p53 activation may affect vessel integrity. When combined, these in vitro results indicate that all tested levels of p53 activation impair vessel growth and integrity, although effects at the higher concentration may be confounded by cell death. To determine whether these in vitro results using venous endothelial cells translated in vivo, we then measured the growth of angiogenic venous vessels in zebrafish embryos [9] treated with sulanemadlin (Fig. 4I). This stapled peptide was used in vivo because we previously found that it had higher in vivo efficacy O. Al-Radi et al. 7 Cell Death and Disease (2025) 16:883 than small molecule MDM2 inhibitors [26]. By imaging a transgenic zebrafish model in which endothelial cells are fluorescently labelled (Tg(fli1:eGFP)), we found that the growth of the vessels in the subintestinal venous plexus was reduced by > 30% following 48 h treatment of sulanemadlin or sunitinib (positive control) compared to solvent-treated controls (Fig. 4J). In contrast, no significant changes were measured following treatment with the control stapled peptide (Fig. 4J). Previous studies showed that ectopic sprouts form in the subintestinal vessels prior to vessel elongation and subsequent retraction during vessel maturation [9]. In embryos treated with sulanemadlin or sunitinib, these leading sprouts remained during treatment, instead of retracting into the vessel basket (Fig. 4I, K). Reductions in the growth of subintestinal vessels could not be explained by abnormal growth, as none of the treatments resulted in a significant change in fish size (Fig. 4L). When combined, the results from the in vitro and in vivo sprouting assays indicate that p53 activation reduces the growth of venous vessels. Increasing levels of p53 activation induce different effects in tip-like and non-tip-like cells We hypothesized that reduced vessel growth could result from decreased proliferation of stalk cells upon p53 activation within growing vessels. Indeed, p53 can be stabilized more readily in proliferating tissues than in quiescent tissues, likely reflecting its role in regulating the maturation and genomic integrity of fastrenewing cells [27]. To test this hypothesis, we first measured whether p53 activation reduced the formation of tip (i.e., leader) or stalk (i.e., follower) cells (Fig. 5A). In HUVEC spheroids, treatment with low concentrations of navtemadlin had no significant effect on tip or stalk cell numbers, while high concentrations significantly reduced the number of tip cells and the number of stalk cells by > 33% to 45%, respectively (Fig. 5B, C). Because the reduction in formation of tip cells was unexpected, we subsequently delineated p53 activity and cell cycle changes in each population using flow cytometry detection of CD34, a molecular marker expressed by cultured endothelial cells that exhibit tip-like features [28,29] (SI Figs. 6, 7). While low concentrations of navtemadlin had no measurable effect, high concentrations of navtemadlin reduced the frequency of CD34+ (tip-like) cell formation in the presence of VEGF (Fig. 5D), corroborating the results observed in spheroids. To assess whether this reduction was due to increased p53 activity, we measured the levels of p53 and that of p21 (as a marker for p53 transcriptional activity) in CD34+ (tip-like) and CD34−(non-tip-like) cells. After navtemadlin treatment, p53 levels increased in both CD34+ and CD34−cells in the absence (P adj for CD34+cells =0.007; P adj for CD34cells =0.02) and presence of VEGF (P adj for CD34+cells =0.0003; P adj for CD34cells =0.009) (Fig. 5E).Incontrast,significant increases in p21 levels were measured only in CD34−cells in the absence (P adj for CD34+cells =0.09; P adj for CD34cells =0.0006) and presence of VEGF (P adj for CD34+cells =0.14; P adj for CD34cells =0.003) (Fig. 5F). However, it is possible that changes in CD34+ could not be detected with sufficient statistical power due to the lower abundance of this population. Given that p21 induction was significantly increased in CD34− cells, we subsequently assessed whether p53 might differentially affect their cell cycle distribution. While p53 activation had no significant effect on the proliferation or G 0 G 1 phases of CD34+ cells (Fig. 5G), it altered the cell cycle phases of CD34−cells depending on the level of p53 activation (Fig. 5H). At low concentrations, it reduced the proliferation of CD34−cells and increased G 0 G 1 arrest, but at high concentrations, it reduced proliferation and increased G 2 M arrest. At high concentrations, p53 activation also significantly increased G 2 M arrest in CD34+ cells. Thus, although cell cycle changes were strongly detected in CD34−cells, we cannot exclude the possibility that additional effects occurred in CD34+ cells but were not detected due to lower abundance. Based on these data, we reasoned that p53 activation in CD34+ cells may affect their formation through mechanisms other than cell cycle regulation. One possibility is that p53 activation leads to reductions in the expression of key tip cell genes needed for tip cell formation. Using RT-qPCR, we found that treatment with low concentrations of navtemadlin alone reduced expression of two tip cell genes ANGPT2 and CXCR4, while treatment with high concentrations reduced the expression of all tip and stalk cell genes (Fig. 5I). However, in the presence of VEGF, p53 activation reduced all the tip cell genes but only one of the stalk cell genes (Fig. 5J). Taken together, these results suggest that p53 activation influences the cell fate of tip-like and non-tip-like cells in a concentration-dependent manner: low levels primarily affect proliferation in non-tip-like cells, while high levels impair both cell cycle regulation and angiogenic gene expression in both populations. Fig. 4 Pharmacological activation of p53 reduces growth of venous vessels in vitro and in vivo at all tested concentrations. A Schematic of the endothelial spheroid sprouting assay. HUVEC were embedded in a fibrin gel matrix and treated for 24 h using solvent (DMSO) or increasing concentrations of navtemadlin (navt). Vascular endothelial growth factor (VEGF, 20 ng/mL) was used to stimulate angiogenic sprouting. BFluorescence microscopy images show reduced sprouting in HUVEC spheroids treated with increasing navtemadlin concentrations; the images are brightness and contrast adjusted for clarity. Insets show cell filopodia in sprouts. Scale bars =100 µm. C–EAll tested concentrations of navtemadlin reduce (C) total sprout length, (D) fraction of sprouts associated with spheroid body (a measure of connectivity), and (E) total sprout number of HUVEC spheroids. Each data point in violin plot indicates one spheroid (n=64 spheroids for baseline; 61 spheroids for VEGF; 66 spheroids for 0.1 and 1 μM navt; 84 spheroids for 10 μM navt; pooled from three experiments). *P adj < 0.05, ** P adj < 0.01, ***P adj < 0.001, **** P adj < 0.0001 using Kruskal-Wallis with Dunn’s correction. Note y-axis is shown on log scale. FSchematic of endothelial sprouting assay following 24 h transfection with control or TP53-targeting DsiRNAs (pool). After embedding, spheroids were treated for 24 h using DMSO or navtemadlin in the presence of VEGF. GFluorescence microscopy images show that navtemadlin does not further reduce sprouting of HUVEC spheroids after TP53-knockdown; the images are brightness and contrast adjusted for clarity. Scale bars = 100 µm. HQuantification of total sprout length in drug-treated HUVEC spheroids following transfection with control or TP53-targeting DsiRNAs (pooled). Each data point in violin plot indicates one spheroid (n=66 spheroids for DMSO and 82 spheroids for 1 μM navt in Ctrl DsiRNA; 90 spheroids for DMSO and 65 spheroids for 1 μM navt in TP53 DsiRNA spheroids/group pooled from three experiments). *P adj < 0.05, ** P adj < 0.01, using Kruskal-Wallis with Dunn’s correction. Note y-axis is shown on log scale. IA zebrafish embryo (Danio rerio) model expressing fluorescent vasculature (Tg(fli1:eGFP)) was used to measure effects of p53 activation in developing subintestinal vessels (SIV) in vivo. Embryos were treated with vehicle (DMSO), sunitinib (positive control), sulanemadlin (stapled peptide activator of p53), or control peptide for 48 h, starting at 20 h post-fertilization (hpf). Cropped maximum intensity projections of the subintestinal vessels and corresponding vessel masks (binary images below) are shown; the images are brightness and contrast adjusted for clarity. Arrows (orange) on the binary masks indicate ectopic sprouts. Scale bar of cropped images =150 µm. J–LTreatment with p53 activator sulanemadlin leads to (J) reduction in subintestinal vessel area, (K) an increase in number of endothelial cell tip extensions, and (J) no measurable change in fish size. Each data point in violin plot represents one embryo (n=26 embryos for DMSO; 30 embryos for sunitinib; 32 embryos for sulanemadlin; 31 embryos for control peptide; pooled from two independent experiments). ***P adj < 0.001, ****P adj < 0.0001 for vessel areas using BrownForsythe ANOVA with Dunnett’s T3 correction and *P adj < 0.05, **P adj < 0.01 for sprout number using Kruskal-Wallis non-parametric test with Dunn’s correction for multiple testing. O. Al-Radi et al. 8 Cell Death and Disease (2025) 16:883 DISCUSSION Effects of pharmacological p53 activation on distinct cell fates in endothelial cells In this study, we evaluated whether different levels of p53 could modulate endothelial cell fate by triggering distinct cellular responses during angiogenesis. By measuring the molecular and phenotypic effects induced by low and high levels of p53 activation, we found that increasing p53 levels transitioned endothelial cells from cell cycle arrest to cell death. Notably, increasing levels of p53 also triggered distinct effects in tip-like O. Al-Radi et al. 9 Cell Death and Disease (2025) 16:883 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creativecommons.org/licenses/by/4.0/. © The Author(s) 2025 O. Al-Radi et al. 16 Cell Death and Disease (2025) 16:883