Unveiling a novel role for p19Arf (alternative reading frame) in mESC differentiation toward the pancreatic lineage Authors: De Marino Elena, Napolitano Giuliana, Roscigno Giuseppina, Lucci Valeria, Falco Geppino, Calabrò Viola, Vivo Maria, Angrisano Tiziana, Antonini Dario, Pollice Alessandra Published: 2026 ABSTRACT The tumor suppressor ARF (p14 in human, p19 in mouse), has traditionally been characterized by its pivotal role in tumor surveillance. However, its involvement in an expanding range of cellular processes reveals that its functions are broader and more complex than initially appreciated. Here, we uncover a previously unrecognized role of p19ARF in endodermal differentiation, specifically in pancreatic lineage specification using an in vitro differentiation model of mouse embryonic stem cells (mESCs). Using CRISPR/Cas9-mediated mutagenesis, we show that mESCs with mutations in p19Arf are unable to efficiently differentiate towards pancreatic endoderm. Transcriptomic profiling reveals substantial alterations in gene networks associated not only with lineage commitment but also with cytoskeletal organization and cell morphology. These changes correlate with a disruption in stem cell architecture and suggest the persistence of pluripotency feature when only one copy of functional p19Arf is present. Taken together, our findings highlight a role for p19Arf in modulating endodermal differentiation while maintaining the essential cellular properties of stem cells, thus expanding its relevance beyond tumor suppression.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-53355-0. Introduction The p14/p19Arf (Alternative Reading Frame), encoded by the Cdkn2a locus, is recognized as one of the most critical tumor suppressors due to its well-established role in cancer suppression1–3. It is transcribed from the INK4a/ARF locus located on mouse chromosome 4 (human chromosome 9p21.3) which presents a unique organization in eukaryotes, with two intertwined tumor suppressor genes, p16INK4a and p14ARF (human)/p19Arf (mouse), rendering this locus central to tumor defense. While p16INK4a directly regulates cell cycle progression by inhibiting CDK4/6 and enforcing Rb-mediated G1 arrest, Arf acts primarily as a sensor of oncogenic and cellular stress signals-such as those triggered by viral infection, hypoxia, or hyperproliferative stimuli mainly by stabilizing p53 through inhibition of MDM2, thereby inducing either cell cycle arrest or apoptosis4–7. In addition, Arf is also a key regulator of the irreversible growth arrest associated with senescence8 and appears to play a role in modulating innate immunity response9,10. Additionally, Arf can modulate cell proliferation in a p53-independent manner11,12 by regulating cell growth through direct interaction with B23 (nucleophosmin/NPM) thereby contributing to the control of cell mass and the maintenance of cellular homeostasis under stress conditions13. Intriguingly, Arf is a small, basic, and largely unstructured protein whose ability to interact with a wide variety of binding partners14,15 enables it to adopt more defined conformations upon binding, thus acquiring functional specificity. Through these interactions, Arf not only gains structural organization but also expands its functional repertoire. More recent studies have highlighted the multifaceted role of the Arf tumor suppressor in regulating cell-fate decisions during development16 including mammary gland development17, angiogenesis18 and tissue regeneration19. Notably, p19Arf has been implicated in controlling cell adhesion and differentiation of pluripotent stem cells into extraembryonic endoderm20 as well as in transiently regulation of male germ cell and eye formation21,22. Furthermore, Arf also contributes to epidermal differentiation by promoting the SUMOylation and degradation of ΔNp63α in human keratinocytes23. Moreover, Arf plays critical roles in ESCs and iPSCs, where the INK4a/ARF locus must remain epigenetically silenced to preserve pluripotency24,25. We previously observed that, during the induction of mouse ESC commitment toward the Posterior Foregut Endoderm (PFE), both Ink4a and Arf are expressed at Day 826. These findings, together with the evidence that physiological expression of Arf occurs within precise developmental windows across multiple tissues (including spermatogonia, the developing eye and the yolk sac-see above), prompted us to further investigate the role of Arf in pancreatic lineage specification. Pancreatic development is a complex, tightly regulated process that originates from the definitive endoderm, requiring coordinated transcriptional programs and dynamic cellular rearrangements. Mouse embryonic stem cells (mESCs) offer a powerful in vitro model system to dissect the molecular mechanisms governing lineage specifications. In this context, understanding the interplay between cell cycle regulators and differentiation cues is crucial, especially given the impact of cellular architecture and identity in guiding fate decisions. This study reveals a novel function of p19Arf in directing the differentiation of mESCs towards Posterior Foregut Endoderm. Using a gene editing approach via CRISPR/Cas9-mediated mutagenesis, we demonstrate that altering only one copy of p19Arf hinders cell differentiation. Transcriptomic analyses uncover a significant role for p19Arf in gene regulatory networks governing both cell-fate determination and cytoskeletal organization, two key determinants of stem cell identity and differentiation competence. Overall, our findings suggest that p19Arf plays a dual role in differentiation signaling and structural maintenance, thereby expanding its functional role in stem-cell biology. Materials and methods Generation of p19Arf mutant clones The LentiCRISPRv2 vector (Addgene #52961) was used for CRISPR/Cas9 editing of the p19Arf gene. A single guide RNA was specifically designed with the necessary 5’ and 3’ ends for cloning after vector digestion with BsmBI (#R0580, New England Biolabs Ipswich, MA, USA)27. The gRNA was chosen according to28 Fw 5’-CACCGTGGTGAAGTTCGTGCGATCC-3’; Rev5’-AAACGGATCGCACGAACTTCACCAC-3’. Paired oligos were cloned into the LentiCRISPRv2 vector that was used to stably transfect mouse embryonic stem cells (E14ES-E14TG2a) by using Lipofectamine ® 2000 (Invitrogen , Carlsbad, CA, USA) according to manufacturer’s protocol. In parallel, E14TG2a cells were transfected with an empty lentiCRISPRv2 vector as control. Forty-eight hours after transfection, cells were detached and plated in 100 mm dishes adding 0.8 µg/mL puromycin for selection (Gibco, Waltham, MA, USA) to obtain single clones. Eight days post transfection, single colonies were picked and expanded sequentially in 96, 24 and 12 well culture dishes (reducing puromycin to 0.4 µg/mL in 12 well dishes). Isolated clones were then expanded for subsequent genotyping by Sanger sequencing (Eurofins Genomics, Turin, Italy) using the following primers to amplify the region of interest respectively mapping to the 5’UTR and in Exon 1β of Arf, respectively (Fw 5’-TTCTCACCTCGCTTGTCAC-3’; Rev 5’-TTCTCAAGATCCTCTCTAGCC-3’). Control cells transfected with empty vector (Ctrlv2) were subjected to the same puromycin selection without single-clone selection and subjected to subsequent genotyping by Sanger sequencing. Cells and differentiation protocol mESCs (E14ES-E14TG2a-CRL-1821) were purchased from ATCC cell bank and cultured on 0,1% gelatine (Sigma-Aldrich, St. Louis, MO, USA) in complete GMEM medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 15% fetal bovine serum (FBS) (HyClone, USA), 1,000 U/mL leukaemia inhibitory factor (LIF) (Millipore, St. Charles, MO, USA), 1.0 mM sodium pyruvate (Invitrogen, Carlsbad, CA, USA), 0.1 mM non-essential amino acids (Invitrogen, Carlsbad, CA, USA), 2.0 mM L-glutamine (Invitrogen, Carlsbad, CA, USA), 0.1 mM β-mercaptoethanol (Sigma-Aldrich, St. Louis, MO, USA), 500 U/mL penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA) at 37 °C in 5% CO2. Cell culture medium was changed daily, and cells were routinely splitted every 2–3 days. Differentiation toward the pancreatic progenitor lineage was performed according to29. Briefly, cells were grown in DMEM Low Glucose (Euroclone, Milan, Italy) depleted from LIF and supplemented with 5% fetal bovine serum, 0.1 mM 2(β)-Mercaptoethanol, 1 mM non-essential amino acids, 2 mM glutamine, 20 ng/ml activin A (R&D Systems, Minneapolis, MN, USA), overlaid with 200 µg/ml Matrigel (Corning, NY, USA). After 4 days 5 µM all-trans retinoic acid (RA) (Sigma-Aldrich, St. Louis, MO, USA) and 10 ng/ml Fibroblast Growth Factor 10 (FGF10) (R&D Systems, Minneapolis, MN, USA) replaced with activin A for the subsequent 4 days. Medium was changed every 2 days till day 8. RNA extraction and qPCR analysis Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to manufacturer’s protocols. Reverse transcription of total RNA was performed using iScript cDNA Synthesis kit (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s instructions using an equal amount of total RNA per sample (1 µg). Quantitative real-time PCR analyses were performed using 20ng or 100ng of cDNA per well with the iTaq Universal SYBR Green Supermix (Bio-rad, Hercules, CA, USA). Experiments were conducted in triplicate and normalized respect to the expression of housekeeping glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene. Reactions were run on the Aria Mx Real-Time PCR system (Agilent Technologies, Santa Clara, CA, USA). Fold induction was calculated using 2^−ΔCT method. Each data point was analyzed performing using Agilent software. All Primers used are listed in Sup. Table S1. Phalloidin staining 4 × 104 or 1.4 × 105 cells (for analyses, respectively, at day 0 and day 8, respectively) were plated in 3 well chambers (ibidi GmbH) either induced to differentiate or left undifferentiated (day 0). Cells were fixed with 3.7% PFA (Sigma-Aldrich St. Louis, MO, USA) permeabilized with 0.5% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) and incubated with 1:1000 rhodamine-phalloidin (Thermo Fisher Scientific, Waltham, MA, USA) to visualize F-actin. Nuclei were counterstained with 1:10000 4′,6-diamidin-2-phenilindol (DAPI) (Thermo Fisher Scientific, Waltham, MA, USA). The fluorescence intensity of phalloidin was measured by calculating the corrected total cell fluorescence (CTCF) using ImageJ software as previously described30. Western blot analysis Total proteins were extracted using RIPA buffer supplemented with 1XProtease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO, USA) and processed as previously described31. Proteins were separated by SDS-PAGE and transferred onto a Protran® nitrocellulose membranes (Sigma-Aldrich, St. Louis, MO, USA) using a Mini trans-blot system (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s instructions. Membranes were blocked with 5% non-fat dry milk and incubated overnight at 4 °C with gentle shaking and the following primary antibodies: rabbit anti-p19Arf (1:500; R&D Systems, Minneapolis, MN, USA, NB200-106), rabbit anti-Oct4 (1:500; Santa Cruz Biotech, Dallas, TX, USA sc5279), rabbit anti-Sox2 (1:1000; Millipore, St. Charles, MO, USA, AB5603), rabbit anti-Sox9 (1:1000; Millipore, St. Charles, MO, USA, AB5535), mouse anti-Gapdh (1:10.000; Santa Cruz Biotech, Dallas, TX, USA sc32233). HRP-conjugated secondary antibodies (anti-mouse 7076, anti-rabbit 7074, Cell Signaling Technologies, Danvers, MA, USA) were used at 1:3000 or 1:6000 for 1 h at room temperature (RT). Proteins were detected using enhanced chemiluminescence (ECL) (Bio-Rad, Hercules, CA, USA) and visualized with the ChemiDoc TM XRS system using Quantity One software of (Bio-Rad, Hercules, CA, USA). For relative protein levels analysis, p19Arf band intensities were quantified using ImageJ software, normalized to Gapdh, and expressed as fold change relative to the Ctrlv2 at Day 0. Bars represent the mean ± SD (n = 3). RNA sequencing and functional annotation analysis RNA sequencing and data analysis, including library preparation, were performed by Novogene UK. Detailed descriptions are available in the Novogene methods report. Specifically, RNA samples collected from 3 independent experiments were extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA), library preparation was performed using the Novogene NGS RNA Library Prep Set. Library quality was assessed using an Agilent Bioanalyzer 5400 system (Agilent Technologies, Santa Clara, CA, USA). Libraries were sequenced on an Illumina NovaSeq X Plus platform (Paired-End, 150 bp). Read mapping, quantification, and differential gene expression analyses were performed using: HiSat2, FeatureCounts, and DESeq2, respectively. Functional annotation of significantly deregulated genes (FDR ≤ 0.01; Log2(FC) ≤ -1; Log2(FC) ≥ 1) was performed using ShinyGO32. Statistical analysis All data are expressed as the mean of at least three biological replicates ± standard deviation (SD). Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc testing for multiple comparisons. A p-value p < 0.05 was considered statistically significant. All analyses were conducted using GraphPad Prism 10 software (San Diego, CA, USA). Results Generation of p19Arf mutated mESCs by RNA-Guided CRISPR/Cas9 To uncover a potential role of p19Arf in the pancreatic endoderm differentiation program, we used the CRISPR/Cas9 genome editing technology to specifically target p19Arf in mESC. To this end, due to the unique arrangement of the INK4a/Arf locus which contains two overlapping genes, we selected a gRNA specifically directed against Exon 1β of p19Arf, thereby avoiding mutations in p16Ink4aA, located 70 bp downstream of the Arf Start codon ATG28 (Fig. 1A and “Materials and methods”). Two independent selected clones (named Arf-B and Arf-C) were isolated and sequenced, and both exhibited mutations in the target region (Fig. 1B–D). In each clone, the mutation consisted of an A insertion (Fig. 1B, C), causing a frameshift at amino acid 30 that generates a premature stop codon at residue 65 in one allele (Fig. 1D). This leads to a truncated peptide of 64 amino acids (Fig. 1D), while the remaining allele retained the wild-type sequence. Fig. 1Generation of p19Arf mutated mESCs by RNA-Guided CRISPR/Cas9. (A) Schematic representation of the INK4a/Arf locus and the location of CRISPR/Cas9 gRNA targeting Arf Exon 1β. (B) DNA sequence of the CRISPR/Cas9 target region (PAM sequence highlighted in red) and the resulting mutation introduced by the system (bottom part: see the insertion of an A/T highlighted in grey). (C) Chromatogram of the wild-type Ctrlv2 (left part) and mArf mutant clones (right part). The interpretation shows one mutated allele (see T/A insertion) and one wild type allele in mArf clones. (D) DNA and aminoacidic sequences of wild type and Arf mutant alleles showing the frameshift from amino acid 30 (indicated in red). Generation of p19Arf mutated mESCs by RNA-Guided CRISPR/Cas9. (A) Schematic representation of the INK4a/Arf locus and the location of CRISPR/Cas9 gRNA targeting Arf Exon 1β. (B) DNA sequence of the CRISPR/Cas9 target region (PAM sequence highlighted in red) and the resulting mutation introduced by the system (bottom part: see the insertion of an A/T highlighted in grey). (C) Chromatogram of the wild-type Ctrlv2 (left part) and mArf mutant clones (right part). The interpretation shows one mutated allele (see T/A insertion) and one wild type allele in mArf clones. (D) DNA and aminoacidic sequences of wild type and Arf mutant alleles showing the frameshift from amino acid 30 (indicated in red). p19ARF mutated mESCs show an altered differentiation process To investigate the potential role of p19Arf in Posterior Foregut Endoderm, we adopted an in vitro protocol that recapitulates embryonic development through the sequential exposure of cells to specific growth factors and signaling molecules29(See Fig. 2A for the experimental schedule). As a first attempt, we analyzed the behavior of wild type mESCs and Ctrlv2 cells (generated by transfecting of E14TG2a cells with the empty vector, see Materials and Methods) at different time points during differentiation (day 0, day 4 and day 8). Phase contrast microscopy, Real Time RT-qPCR and Western Blot analyses (Fig. 2B, C and E) showed that wild-type and Ctrlv2 cells exhibited comparable behavior both for the acquisition of the characteristic morphology and for the proper expression of key selected markers. In fact, as expected, during differentiation the pluripotency marker Pou5f1/Oct4 becomes downregulated while Sox17, Foxa2, Onecut1/Hnf6, Nkx6-1 and Sox9 are properly expressed. In agreement with our previous observations26, p19Arf expression becomes detectable at day 8 at both RNA and protein levels in wild-type and Ctrlv2 cells (Fig. 2D, E). Based on these results, Ctrlv2 was used as the reference control for subsequent experiments. Strikingly, phase contrast microscopy revealed marked morphological differences in the Arf mutant clones during differentiation (Fig. 3A). While control cells tend to aggregate and acquire a morphology typical of posterior foregut endoderm33, Arf-B and Arf-C mutant clones fail to do so and tend to maintain a rounded shape at day 4 with even more prominent abnormalities at day 8, indicating defects in the differentiation program. Interestingly, the observed morphological defects were accompanied by alterations in gene expression. As expected, in Ctrlv2 control cells, Pou5f1/Oct4 RNA levels were downregulated at day 8, while Nkx6-1, Sox9 and p19Arf are expressed (Fig. 3B-D). In contrast, although Arf-mutant clones showed a similar overall trend, the changes appeared much less pronounced (Fig. 3B-D). Day 4 was not included in these analyses, as the most evident phenotype occurs at day 8. These observations suggest that Arf may play a role in the differentiation of the Posterior Foregut Endoderm, thereby, prompting further investigation of this hypothesis. Fig. 2mESC and Ctrlv2 differentiation toward PFE. (A) Schematic overview of the in vitro pancreatic endoderm differentiation protocol. (B) Phase-contrast microscopy analysis of cell morphology during differentiation toward the pancreatic lineage. Images are representative of the behavior of wild-type and Ctrlv2 cells. Images were acquired using a Leica DMIL LED phase contrast microscope using a 20X objective. (C) RT-qPCR analysis of the pluripotency marker Pou5f1/Oct4, and the canonical markers of PFE differentiation at day 0, day4 and day 8. (D) mRNA levels of Cdkn2a/p19Arf evaluated by RT-qPCR. Each graph represents three biological replicates (n = 3). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Asterisks indicate significant increase or decrease relative to day 0: *(p ≤ 0.05), ** (p ≤ 0.01), *** (p ≤ 0.001) and ****(p < 0.0001). (E) Representative western blot at day 0, day 4 and day 8 of differentiation analyzing the pluripotency Oct4 marker and p19Arf. Unnecessary lanes were cropped out. mESC and Ctrlv2 differentiation toward PFE. (A) Schematic overview of the in vitro pancreatic endoderm differentiation protocol. (B) Phase-contrast microscopy analysis of cell morphology during differentiation toward the pancreatic lineage. Images are representative of the behavior of wild-type and Ctrlv2 cells. Images were acquired using a Leica DMIL LED phase contrast microscope using a 20X objective. (C) RT-qPCR analysis of the pluripotency marker Pou5f1/Oct4, and the canonical markers of PFE differentiation at day 0, day4 and day 8. (D) mRNA levels of Cdkn2a/p19Arf evaluated by RT-qPCR. Each graph represents three biological replicates (n = 3). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Asterisks indicate significant increase or decrease relative to day 0: *(p ≤ 0.05), ** (p ≤ 0.01), *** (p ≤ 0.001) and ****(p < 0.0001). (E) Representative western blot at day 0, day 4 and day 8 of differentiation analyzing the pluripotency Oct4 marker and p19Arf. Unnecessary lanes were cropped out. Fig. 3Differentiation perturbation in Arf mutant clones. (A) Phase contrast microscopy analysis of cell morphology during differentiation toward the pancreatic lineage. Images are representative of the behavior of control and p19Arf mutant clones (Arf-B and Arf-C). Images were acquired using a Leica DMIL LED phase-contrast microscope with a 20× objective. (B) RT-qPCR analysis of Pou5f1/Oct4, Nkx6-1 and Sox9 at day 0 and day 8 of differentiation. (C) mRNA levels of Cdkn2a/p19Arf evaluated by RT-qPCR. Each graph represents three biological replicates (n = 3). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Asterisks indicate significant increase or decrease relative to day 0: *(p ≤ 0.05), **(p ≤ 0.01), ***(p ≤ 0.001), ****(p ≤ 0.0001). (D) Representative western blot analysis of Oct4 (pluripotency marker), Sox9 (PFE marker), and p19Arf at day 0 and day 8 of differentiation in Ctrlv2 and Arf-B and Arf-C clones. Unnecessary lanes were cropped out (left panel). p19Arf relative protein levels respect to control (Gapdh) from three independent replicates (n = 3) (right panel). Differentiation perturbation in Arf mutant clones. (A) Phase contrast microscopy analysis of cell morphology during differentiation toward the pancreatic lineage. Images are representative of the behavior of control and p19Arf mutant clones (Arf-B and Arf-C). Images were acquired using a Leica DMIL LED phase-contrast microscope with a 20× objective. (B) RT-qPCR analysis of Pou5f1/Oct4, Nkx6-1 and Sox9 at day 0 and day 8 of differentiation. (C) mRNA levels of Cdkn2a/p19Arf evaluated by RT-qPCR. Each graph represents three biological replicates (n = 3). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Asterisks indicate significant increase or decrease relative to day 0: *(p ≤ 0.05), **(p ≤ 0.01), ***(p ≤ 0.001), ****(p ≤ 0.0001). (D) Representative western blot analysis of Oct4 (pluripotency marker), Sox9 (PFE marker), and p19Arf at day 0 and day 8 of differentiation in Ctrlv2 and Arf-B and Arf-C clones. Unnecessary lanes were cropped out (left panel). p19Arf relative protein levels respect to control (Gapdh) from three independent replicates (n = 3) (right panel). Transcriptomic profiles of the Arf mutant clones support the observed differentiation defects To gain further insights into the role of Arf during Posterior Foregut endoderm differentiation, we performed global RNA expression profiling by comparing the transcriptome of Ctrlv2 cells with those of Arf-B and Arf-C mutants at day 0 and day 8 of differentiation. At the global level, differentiation induced widespread transcriptional changes. When comparing day 8 to day 0, 3383 genes were upregulated and 2570 genes were downregulated (FDR ≤ 0.01; Log2(FC) ≤ -1 ≥ 1) in Ctrlv2 (Fig. 4A left panel), confirming a well-established differentiation in PFE at day 8 (Supplementary Fig. S1). Similar changes in gene expression were observed in the Arf-B and Arf-C mutants, with 2207 upregulated and 3835 downregulated genes in Arf-B mutant, and 2050 upregulated and 3673 downregulated genes in Arf-C mutant (FDR ≤ 0.01; Log2(FC) ≤ -1 ≥ 1) (Fig. 4A, middle and right panels). However, although the overall number of differentially expressed genes was similar among Ctrlv2 and Arf mutant clones, functional annotation analysis revealed marked differences in the biological processes and pathways enriched during differentiation. Notably, in Arf mutants there is a reduced enrichment of downregulated genes associated with adhesion molecules and pluripotency signaling pathway, as well as a reduced enrichment of upregulated genes involved in cell cycle regulation and metabolic processes (Fig. 4B). Importantly, Pearson’s correlation analysis revealed a strong positive linear relationship (R2 = 0.96) between the transcriptomes of Arf-B and Arf-C mutants at day 8, highlighting the highly similar behaviour of the two mutant clones during the differentiation process, consistent with the presence of the same frameshift mutation (Fig. 4C). The shared differentially expressed genes (DEGs) at day 8 were significantly enriched in pathways related to stem cell pluripotency (e.g. Dppa5a, Sox2, Nanog and Pou5f1, Hoxb1, Esrrb, Pax6), focal adhesion and cell-cell adhesion molecules (Col4a1, Col4a2, Col6a1, Lama1, Lamc1, Ocln, Cldn3, Cldn7, Jam2, Cdh1), and cell differentiation (Gata4) (see chord plot and heatmap in Fig. 4D and also volcano plots in Supplementary Fig S2). Real Time RT-qPCR analysis confirmed these findings showing a significantly weaker induction of Gata4 and Gata6 differentiation markers, Col4a2 cell-cell adhesion molecule and Cdkn1a/p21 (key differentiation player) (Fig. 5B) in Arf mutants at day 8 compared with control cells, as well as a markedly reduced downregulation of Nanog, Sox2 and Dppa5a expression (Fig. 5A). These data prompted us to examine in greater detail the morphology of mutant cells during differentiation. We therefore analyzed cytoskeletal organization by staining F-actin filaments with fluorescent phalloidin, (Fig. 6). At day 0, Arf-mutant cells exhibited a rounded morphology similar to that of control cells. However, by day 8, while Ctrlv2 cells acquired the characteristic basket-like structure, Arf-B and Arf-C clones largely retained a rounded, stem-like appearance. These observations suggest that Arf may also contribute to changes in cell shape and substrate adhesion during ESC differentiation. Fig. 4Arf-B and Arf-C mutants show altered expression of genes associated with adhesion molecules, pluripotency signaling pathway, cell cycle regulation and metabolic processes. (A) Volcano plot of DEGs in Ctrlv2, Arf-B and Arf-C obtained by comparing gene expression at day 8 versus the respective day 0. Red dots indicate up-regulated genes (FDR ≤ 0.01 Log2FC ≥ 1), blue dots indicate down-regulated genes (FDR ≤ 0.01 Log2FC ≤ -1), and grey dots indicate non-significant changes. (B) Bubble plot showing the enrichment analysis of DEGs using the KEGG pathway database34–36. Color and x-axis represent the negative logarithm of the FDR. Bubble size indicates the numbers of genes enriched in each pathway and bubble shape represents the genotype. (C) Linear regression analysis comparing DEGs of Arf-B and Arf-C at day 8 relative to Ctrlv2 at day 8 (R² = 0.9601), demonstrating a strong overall similarity in transcriptional gene expression alterations between the two mutants compared with Ctrlv2. (D) Chord plot (left panel) showing the relationship between DEGs (Arf mutants versus Ctrlv2) at day 8 and their associated KEGG pathways34–36. Colors indicate up- (red) or down-regulated (blue) genes. The right panel shows a heatmap of the expression of genes enriched in the indicated pathways at day 8 in each sample. Arf-B and Arf-C mutants show altered expression of genes associated with adhesion molecules, pluripotency signaling pathway, cell cycle regulation and metabolic processes. (A) Volcano plot of DEGs in Ctrlv2, Arf-B and Arf-C obtained by comparing gene expression at day 8 versus the respective day 0. Red dots indicate up-regulated genes (FDR ≤ 0.01 Log2FC ≥ 1), blue dots indicate down-regulated genes (FDR ≤ 0.01 Log2FC ≤ -1), and grey dots indicate non-significant changes. (B) Bubble plot showing the enrichment analysis of DEGs using the KEGG pathway database34–36. Color and x-axis represent the negative logarithm of the FDR. Bubble size indicates the numbers of genes enriched in each pathway and bubble shape represents the genotype. (C) Linear regression analysis comparing DEGs of Arf-B and Arf-C at day 8 relative to Ctrlv2 at day 8 (R² = 0.9601), demonstrating a strong overall similarity in transcriptional gene expression alterations between the two mutants compared with Ctrlv2. (D) Chord plot (left panel) showing the relationship between DEGs (Arf mutants versus Ctrlv2) at day 8 and their associated KEGG pathways34–36. Colors indicate up- (red) or down-regulated (blue) genes. The right panel shows a heatmap of the expression of genes enriched in the indicated pathways at day 8 in each sample. Fig. 5Real Time RT-qPCR analysis of key pluripotency (A), differentiation (B) and matrix remodeling (C) markers in Ctrlv2 and mArf-B/ C clones at day 0 and day 8 of differentiation. Each graph represents three biological replicates (n = 3). Significance was assessed using one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Asterisk* indicate significance in reduction/increase relative to day 0, *(p ≤ 0.05), **(p < 0.01), ***(p < 0.001) and ****(p < 0.0001); # Hash symbols indicate significance at day 8 between clones ## (p < 0.01) and #### (p < 0.0001). Real Time RT-qPCR analysis of key pluripotency (A), differentiation (B) and matrix remodeling (C) markers in Ctrlv2 and mArf-B/ C clones at day 0 and day 8 of differentiation. Each graph represents three biological replicates (n = 3). Significance was assessed using one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Asterisk* indicate significance in reduction/increase relative to day 0, *(p ≤ 0.05), **(p < 0.01), ***(p < 0.001) and ****(p < 0.0001); # Hash symbols indicate significance at day 8 between clones ## (p < 0.01) and #### (p < 0.0001). Fig. 6Mutations in Arf influence cell shape and architecture during ESCs differentiation. Upper panel: Confocal microscopy analysis of cell morphology of control and p19Arf mutant clones (Arf-B and Arf-C) at day 0 and day 8 of differentiation toward the pancreatic lineage. Panels show representative images stained with phalloidin to visualize F-actin filaments (red) and DAPI label nuclei (blue). Images were acquired using a 60X objectives of an Olympus FV3000 confocal microscope. Lower panel: Violin plot representing the quantification of phalloidin fluorescence signal in three independent experiments including Ctrlv2, Arf-B and Arf-C clones at Day 0 and Day 8. (Total number of analyzed cells n = 20). One-way ANOVA followed by Tukey’s post hoc reveals that at day 8 in each condition phalloidin staining is not statistically significant. Mutations in Arf influence cell shape and architecture during ESCs differentiation. Upper panel: Confocal microscopy analysis of cell morphology of control and p19Arf mutant clones (Arf-B and Arf-C) at day 0 and day 8 of differentiation toward the pancreatic lineage. Panels show representative images stained with phalloidin to visualize F-actin filaments (red) and DAPI label nuclei (blue). Images were acquired using a 60X objectives of an Olympus FV3000 confocal microscope. Lower panel: Violin plot representing the quantification of phalloidin fluorescence signal in three independent experiments including Ctrlv2, Arf-B and Arf-C clones at Day 0 and Day 8. (Total number of analyzed cells n = 20). One-way ANOVA followed by Tukey’s post hoc reveals that at day 8 in each condition phalloidin staining is not statistically significant. Discussion While p14/19Arf is well known for its critical function as a tumor suppressor, recent evidence has provided new insights regarding its role in cells fate control, placing it at a pivotal crossroads between oncogenesis and differentiation. During development, Arf is expressed within a tightly regulated temporal window, emerging at specific stages of differentiation across multiple tissue types, such as in differentiating keratinocytes23, in spermatogonia during the first synchronized wave of spermatogenesis and in the developing eye21,22, in the late-stage of differentiation of the extraembryonic endoderm (ExEn)20 and during pregnancy and lactation, in response to progesterone stimulation17. Notably, the Ink4a-Arf locus is epigenetically silenced in stem cells to preserve their self-renewal capabilities24, yet its expression becomes modulated during the generation of iPSCs24,25. We previously demonstrated that Ink4a/Arf mark pancreatic progenitor specification during mESC differentiation toward the endodermal lineage26. However, the mechanisms through which Arf contributes to differentiation processes remain unclear and only partially elucidated. Given Arf’s established role in cancer and the importance of progenitor cell’s (PCs) homeostasis for tissue maintenance, clarifying Arf’s involvement in these contexts is particularly important. In fact, the homeostasis of PCs could be involved not only in defects of differentiation but also in aberrant cell growth and cancer. Pancreatic progenitor cells (PPCs) in particular, are essential for pancreas maintenance and regeneration, an organ characterized by extensive plasticity37,38. Of note, Arf mutations are highly prevalent in pancreatic cancer, second only to mutations in p53, underscoring its crucial role in tumorigenesis within this tissue type39–41. Here, we demonstrate a role for p19Arf in guiding mESCs specification toward the pancreatic lineage. Specifically, p19Arf mutants retaining only one copy of the wild type gene and a truncated allele comprising the first N-ter 30 amino acids display profound defects in differentiation. Mutant Arf clones tend to maintain stem-like features despite the ongoing differentiation suggesting that Arf contributes to the loss of pluripotency. Indeed, transcriptomic profiles reveal differences between mutants and control cells not only in the expression of differentiation markers -such as Gata factors, essential for the multipotent progenitor cell identity42–46- but also in pluripotency markers including Oct4, Sox2, Nanog and Dppa5a/Esg-147–51 which show an anomalously prolonged expression in the mutants. Under physiological conditions, differentiation requires a metabolic transition from glycolysis, typical of ESCs, toward oxidative phosphorylation, accompanied by changes in metabolite levels, biomass, cell cycle dynamics and redox state52–54. Instead, Arf mutant clones exhibit marked differences relative to control in the expression of genes involved in metabolic pathways, consistent with impaired differentiation. Similarly, differentiation entails cytoskeletal reorganization leading to major changes in cell morphology, culminating in the acquisition of a flattened, epithelial-like shape characteristic of definitive endoderm55–57. In contrast, Arf mutant clones display a pronounced defect, retaining a rounded morphology by day 8 of differentiation, consistent with their ongoing expression of pluripotency markers. Accordingly, transcriptomic analysis reveals differences in genes involved in focal and cell adhesion, which may be consistent with a role for Arf in cytoskeletal and adhesion related processes. Interestingly, knockdown of p14ARF in HaCat and HeLa cells levels by siRNA has been shown to alter actin cytoskeleton organization, producing a rounded phenotype probably due to the loss of Arf-interaction with FAK and their co-localization at focal adhesion points during cell spreading58. Importantly, re-expression of either p14ARF or p19ARF rescues the spreading defects, indicating conservation of this function58. Taken together, our data underline the importance of Arf in pancreatic endodermal differentiation and suggest that Arf gene dosage may also be relevant. In fact, mutant clones retain a single wild-type Arf allele and one allele retaining the first 30 N-terminal amino acids, which is likely insufficient to support full Arf functionality59–61. Indeed, p19Arf has been shown to exert effects even in a single-copy context, although this observation has been demonstrated primarily in tumor settings62. Differentiation, however, is a finely tuned process that may be far more sensitive than normal cellular physiology to even modest variations in gene dosage63,64. Therefore, the observation that the presence of only one functional Arf allele is sufficient to give a clear differentiation defect underscores its importance during this process. On the other hand, we cannot exclude that the Arf truncated allele may have some effect on the phenotype through a gain of function and/or a dominant negative activity. It would be interesting to investigate the phenotype associated with a complete Arf knockout, as the total absence of Arf might result in a more severe impairment of differentiation or potentially affect cell viability. In conclusion, our findings provide valuable insights on a new function of ARF, paving the way for future investigation to explore the undiscovered role of ARF in physiological/pathological networks regulations. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2