Introduction

Mastocytosis is a heterogeneous disease characterized by accumulation of mast cells (MCs) in various tissues, particularly skin and bone marrow. Recognized as a distinct hematological disorder since the WHO 2016 revised classification [1,2,3], it results from the clonal proliferation of MCs driven by gain-of-function mutations in the proto-oncogene SCF receptor KIT. A literature analysis indicates that ~70% of childhood-onset mastocytosis cases undergo complete or partial remission at puberty [4, 5]. In contrast, adult mastocytosis is a persistent and non-regressive disease [6]. We have previously shown that childhood mastocytosis is associated with activating KIT mutations in more than 85% of cases, confirming its clonal nature rather than a reactive condition, as previously suspected [7]. While both adult and pediatric mastocytosis are clonal diseases, the mechanism underlying clinical regression in children remains unknown. Notably, only half of childhood mastocytosis cases harbor an exon 17 KIT mutation in the intracellular phosphotransferase domain (PTD), whereas this mutation is present in over 90% of adult cases [8, 9]. Given these differences, we examined whether KIT mutations contribute to distinct disease outcomes. Our study, based on epidemiological data and cellular models, suggests that specific KIT mutations influence disease outcomes by activating distinct intracellular signaling pathways, particularly through telomere shortening and cellular senescence.

Patients, material and methods

Patients

We assessed patients with childhood-onset cutaneous mastocytosis and adult mastocytosis in a multicenter clinical study. Mastocytosis was suspected by the presence of a positive Darier’s sign and was confirmed by histological analyses. Mastocytosis diagnosis was defined by the WHO criteria. The children cohort was previously described by Bodemer et al. [7]. Parents of the children signed the informed consent and authorization form for the study protocol, including genetic analysis, and 2-mm biopsies were collected from a cutaneous lesion after informed consent was obtained. All patients were included in “mastocytosis pathophysiological study”, which started in 2003 and sponsored by AFIRMM (Association Française pour les Initiatives et la Recherche sur le Mastocyte et les Mastocytoses) promoter. The study was approved by Necker hospital ethical committee and carried out according to the Helsinki convention. Each patient provided informed consent. The result of KIT sequencing looking for D816V or other mutation in skin and/or bone marrow was available.

KIT mutation screening

Total RNA was extracted from cutaneous biopsy or bone marrow aspirate using the RNeasy Mini kit (Qiagen). RNA was reverse transcribed into cDNA using the AffinityScript Multiple Temperature cDNA Synthesis kit (Agilent Technologies), with random primers and oligodT as recommended by the manufacturer. To detect specifically the codon 816 mutation into exon 17, the c-KIT coding sequence was amplified in two independent PCR using the Phire Hot Start DNA Polymerase (Finnzymes) and the primers U2 and L1 (see Supplemental Material).

Telomere fluorescence in situ hybridization (FISH) and immuno-FISH

Immunofluorescence step: samples were fixed 10 min in PBS - 2% paraformaldehyde (PFA). Cells were permeabilized during 30 min in PBS - 1% BSA - 0,1% Triton 100X. Primary anti-KIT (polyclonal, R&D, AF332, dilution 1/20), anti-TRF2 (4A794.15, Imgenex, IMG-124A, dilution 1/20), anti-p-TRF2 (Thr188) (Polyclonal, ThermoFisher Scientific, PA5-105937, dilution 1/200) were incubated one hour at room temperature, secondary antibody donkey anti-IgG goat AF488 (Invitrogen, 1/200) was incubated 30 min at room temperature. Immunostaining was fixed in PBS - 2% PFA during 5 min at room temperature. Samples were then dehydrated in successive increasing concentrations of ethanol baths: 70%, 90% and 100% ethanol for 5 min each and dried at room temperature. Second step: in situ hybridization with PNA Cy3 Telo-C (CCCTAA)3 probe (Exiqon, South Korea). PNA probe was warmed 5 min at 70 °C and diluted in 70% deionized formamide - 0,5% blocking reagent - 10 mM Tris-HCL, 1/200. Hybridization was induced at 80 °C, 5 min, followed by 90 min incubation at room temperature. Samples were washed twice 15 min in 70% deionized formamide - 10 mM Tris-HCL - PBS and rehydrated finally in PBS. DNA staining was performed simultaneously with medium montage containing Dapi (Vectashield – Dapi).

Immunohistochemistry

Double immunostaining of cutaneous biopsies for p16INK4a (E6H4, CINtec Roche, 9511, Ready to use, dilution 1/2) and KIT (polyclonal, DAKO, A4502, dilution 1/400) were performed automatically in Leica Bond 3 automizer. In brief, after antigen retrieval and blocking endogenous peroxidase activity, the slides were incubated for 30 min with the p16INK4a antibody (clone E6H4) or with the Negative Reagent Control (isotype control antibody), both included in the CINtec Histology Kit. Secondary antibody reagent (polymer-based goat-anti-mouse antibody fragment conjugated to horseradish peroxidase) was applied for 30 min. After the chromogenic visualization step using the 3,3′-diaminobenzidine chromogens, slides were counterstained with hematoxylin and cover slipped.

Acquisition and processing of confocal microscopy image

Image acquisition was performed on LSM700 confocal microscope, with Zen 2010 software. Image’s treatment was performed using Image J software. For quantification, the background noise was removed from the negative control [10].

Lentiviral production and infection of HDF cells

Human Kit cDNAs were cloned in pRRL lentiviral vector as described by Yang et al. [11]. For infection of HDF or BMMC cells, 106 cells were incubated for 1 h in 1 ml of media and 8 μg/ml of polybrene (hexadimethrine bromide). Cells were then incubated for 3 h with 5.106 infectious particles and diluted in fresh media. GFP-positive infected cells were sorted 4 days later.

BMMC isolation and cell sorting

Bone marrow cells, from fresh bone marrow from mastocytosis patients, were stained with CD34-FITC, KIT-APC and CD45-PerCP (Beckton Dickinson). Bone marrow mature mast cells were KIT+CD34−CD45low and were sorted on a BD Aria 1® FACS sorter and cultured on 96 well plates in 200 µL of IMDM medium. Mature mast cells were cultivated after isolation and culture of the medullar precursor CD34+KIT+ in adequate medium for 8 weeks (as recommended by Dr M. Arock). After isolation, 2.105 cells were washed twice in 1X PBS and thus were lysed. QTRAP assay was performed on the protein extract as described by the Allied Biotech® Quantitative Telomerase Detection kit. The measure was done in triplicate and the mean was calculated. For each sample, the result was expressed in relative telomerase activity, calculated with positive control provided by the manufacturer.

Cell culture

Mice bone marrow mast cells (BMMC) were derived from femurs of 2 months old male mice. Bone marrow cells were cultured at a starting density of 2.105 cells/ml in Opti-Mem medium supplemented with 100U/mL-glutamine, 2 mM/mL penicillin-streptomycin, 10% fetal calf serum (FCS, Invitrogen), and 2 ng/ml murine recombinant IL-3 (Immugenex). The medium was renewed every 5–7 days. Cells were used after their differentiation into mast cells, which occurred after at least 28 days in culture. At monthly intervals, BMMCs were deprived of IL-3 to assess the acquisition of growth-factor independency.

Primary human dermal fibroblasts (HDF) were derived from a child foreskin and cultured in DMEM medium - 10% Fetal Bovin Serum –100U/mL-glutamine – 2 mM/mL penicillin-streptomycin. Then, the cells are treated for 21 days with 10 μm of p38 inhibitor (Cell Signaling: SB203580) or 10 µM Anisomycin (Cell Signaling #2222).

Senescence associated beta galactosidase assay

Equal number of BMMC control (no infection or infected with empty vector), BMMC infected with various KIT mutant (D816V, del417/418-D419Y, AY502/503), HDF control (no infection or infected with empty vector) and HDF infected with various KIT mutants (D816V, del417/418-D419Y, AY502/503), Rosa cell lines (WT, D816V and del417/418-D419Y), were plated in 48 wells plates for the senescence associated beta-galactosidase staining kit (Cell Signaling). According to instructions provided, cells were fixed and stained with X-gal for detection of beta-galactosidase activity. Percentage of cells exhibiting positive beta-galactosidase activity (turned blue) at pH 6.0 was quantified.

Real-time quantitative RT-PCR

We used a RQ-PCR assay based on TaqMan fluorescence methodology to quantify the full range of hTERT mRNA copy numbers. This method used a dual-labeled nonextendable oligonucleotide hydrolysis (TaqMan) probe in addition to the two amplification primers. For details, see supplemental methods.

Western blot

Cells were dried pellet and lysed with Laemmli buffer (100 mM Tris pH 6.8, 2% SDS, 10% Glycerol) warmed at 95 °C for 5 min. Extracts were diluted in beta-mercaptoethanol 10%. 30 μg of diluted proteins were migrated in 8.75% acrylamide gel and transferred on PVDF membrane overnight at 20 Volt. Membranes were blocked for one hour in TTBS – 3% BSA at room temperature under agitation. Primary antibody was incubated in TTBS – 1% BSA overnight at +4 °C under agitation (Cell Signaling: #9212 anti p38 MAPK – #9215 p-p38 MAPK (Thr180/Tyr182) – Imegenex : Img 124A anti-TRF2), secondary antibody HRP conjugated (anti rabbit, goat) was incubated in TTBS – 1% BSA overnight at room temperature under agitation. Revelation was performed with chemiluminescence kit (Thermo-scientific) and acquired with Chemidoc BioRad camera, treated with ImageLab Software 3.0, BioRad.

Phos-Tag

The procedure is the same as for SDS-PAGE and western blot described previously. However, a crucial step has to be carried out between the two techniques and the gels used are pre-cast (Wako: Phos-tagTM 50 µmol/L). Immediately after electrophoresis, the gel is transferred to a bath containing 5 mM EDTA 2 times 10 min (solution prepared with migration buffer). EDTA is a chelating agent that will help remove M2 + ions (Mn2+ for the gels used in this study). This step will increase the efficiency of the transfer, which will take place very slowly at 4 °C overnight or even the following day.

TRF2 mutagenesis

Wild type and mutated TRF2 genes were synthetized directly by full cDNA synthesis by GeneART technology (Life Technologies/Thermoscientific) and cloned in pDON R221. Sequences of both cDNA were verified and then were cloned in a pCDNA3 expression vector by gateway reaction using LR recombination sites. Mutations were directly introduced during cDNA synthesis on Ser and Thr of MK2 kinase motifs. Ser and Thr were replaced by an Ala residue (see supplemental methods).

Statistical analysis

Statistical comparisons between characteristics of children and adults with mastocytosis were based on unpaired t test. All reported p values were two tailed with confidence intervals of 95% and p value < 0.05 was considered significant. Data were tested using GraphPad Prism software version 5.01 (GraphPad Software Inc., San Diego, CA).

Results

Mastocytosis French epidemiology

In the data extraction of AFIRMM (Association Française pour les Initiatives de Recherche sur le Mastocyte et les Mastocytoses), we identified 1267 patients harboring a KIT mutation. Among them, data on the age of occurrence and persistence of mastocytosis into adulthood (after 18 years) were available for 886 patients, who were divided into three groups: childhood onset (before 15 years, n = 52), adolescent onset (15–18 years, n = 35) and adulthood onset (after 18 years, n = 799). Our results unambiguously showed that 98.4% of adulthood mastocytosis cases (872/886) harbored a PTD KIT mutation. In contrast, half of the 56 pediatric patients with mastocytosis followed at the Necker - Enfants Malades University Hospital carried KIT mutations in exons 8–11, affecting the juxta-membranous or extracellular domain (so called pediatric-type non-PTD KIT mutations) (see details in Table 1). In comparison, among adults with childhood-onset mastocytosis, only 3 patients (out of 52) harbored a non-PTD KIT mutation (Fisher’s Exact Test, p = 1.73.10−7, Odds Ratio: 0.0629, 95% confidence interval [0.0112; 0.2306], suggesting that patients with PTD mutations did not regress, in contrast to most patients with non-PTD mutations (See Fig. 1A–C).

Fig. 1: Main KIT mutations in adults and children mastocytosis.
Fig. 1: Main KIT mutations in adults and children mastocytosis.
Full size image

A Schematic representation of the Kit tyrosine kinase SCF receptor main mutations. B Main KIT mutations encountered among AFIRMM network patients: all patients are followed at adult age for persistent mastocytosis. They are separated into three groups according to the age of onset of the disease. *The three remaining patients harbored a Del419 KIT mutation. **The eleven remaining patients harbored dup502-503 (n = 3), dup504-505 (n = 1), E860G (n = 1), K786Q/D816V (n = 1), N822Y (n = 1), R804Q D816V (n = 1), S476C (n = 1), V560G (n = 2). C Schematic representation of mastocytosis KIT mutation panel according to age of onset and KIT mutation.D Schematic representation of mastocytosis evolution according to age of onset and KIT mutation.

Table 1 KIT mutations in the 56 mastocytosis pediatric patients.
Full size table

Mast cell senescence and telomere shortening

To better understand how mastocytosis could clinically regress in children, we hypothesized a differential telomere biology in childhood-onset mastocytosis with pediatric-type non-PTD, versus adult-type PTD KIT mutations. We thus wondered if senescence in the children’s mast cells could induce the disease regression in comparison to adults with persistent disease.

According to our hypothesis, which suggests that telomere shortening may be involved in the spontaneous regression of mastocytosis, and AFIRMM network epidemiologic data, we compared pediatric-type non-PTD KIT mutation (KIT Del 417/8+D419Y) and adult-type PTD KIT mutation (KIT D816V), the latter being predominantly found in adults. First, we investigated murine bone marrow mast cells (mBMMC) transduced with these different KIT mutations and measured the senescence-associated beta-galactosidase (SA-betagal) activity in in vitro culture. We found that KIT D816V mBMMC exhibited significantly less senescence than KIT Del 417/8+D419Y mBMMC (Fig. 2A). Then, we investigated whether this senescent phenotype was associated with telomere shortening. Our analysis confirmed that the telomeres of KIT Del 417/8+D419Y mBMMC were shorter than those of KIT D816V mBMMC and mBMMC transduced with the empty vector (Fig. 2B). These observations strongly suggest that KIT Del 417/8+D419Y cells become senescent because of telomere shortening.

Fig. 2: KIT Del 417/8+D419Y cells (primary human dermal fibroblasts and primary bone marrow mast cells) have short telomeres and a senescent phenotype.
Fig. 2: KIT Del 417/8+D419Y cells (primary human dermal fibroblasts and primary bone marrow mast cells) have short telomeres and a senescent phenotype.
Full size image

Primary murine bone marrow mast cells (mBMMC) beta-Galactosidase assay after 7 days of culture (A) and telomere length by FISH show telomere shortening in KIT Del 417/8+D419Y pBMMC compared with control and KIT D816V pBMMC (B). Primary human dermal fibroblasts (pHDF) beta-Galactosidase assay after 7 days of culture (C) and telomere length by FISH show telomere shortening in KIT Del 417/8+D419Y pHDF compared with control and KIT D816V pHDF (D). E Telomere length assay in cutaneous biopsies of mastocytosis patients, shows shorter telomeres in patients with KIT Del 417/8+D419Y mutation compared to patient with KIT D816V mutation. Pictures illustrate telomere length (red) in KIT positive mast cells (green) in KIT Del 417/8+D419Y (n = 4) and KIT D816V (n = 4) patients. F Cutaneous biopsies of mastocytosis patients with double staining immunohistochemistry: KIT positive mast cells (red) and p16 expression (brown) in KIT Del 417/8+D419Y (n = 4) and KIT D816V (n = 4) patients.

Validation in primary human dermal fibroblasts

We confirmed these results in a primary human dermal fibroblast (HDF) model, which does not physiologically express KIT or telomerase [12]. Primary HDF (pHDF) were transduced with the same KIT mutants as mBMMC: KIT D816V or KIT Del 417/8+D419Y. As in mBMMC, we measured the SA-beta-gal activity and observed increased senescence (Fig. 2C) and telomere shortening in KIT Del 417/8+D419Y pHDF (Fig. 2D). Moreover, after 52 days of culture, KIT Del 417/8+D419Y pHDF and control displayed increased senescence compared to the unchanged SA-beta-gal level observed for KIT D816V pHDF (Supplementary Fig. 1A). Telomere length after three weeks of culture was shorter in KIT Del 417/8+D419Y pHDF compared to the relative telomere length stability in KIT D816V pHDF (Supplementary Fig. 1B). Cell cycle analysis also reveals increased DNA content in KIT Del 417/8+D419Y pHDF, consistent with a senescent state (Supplementary Fig. 1C).

Validation in human skin biopsies of patients with mastocytosis in the skin

To validate these results in vivo, we examined telomere length in KIT positive mast cells from two series of formalin-fixed paraffin-embedded (FFPE) cutaneous biopsies obtained from age-matched children’s groups with KIT D816V or KIT Del 417/8+D419Y mastocytosis. The telomere length of the KIT D816V group (n = 4) was significantly longer compared to the KIT Del 417/8+D419Y group (n = 4; p < 0.001) (Fig. 2E). Moreover, in the KIT Del 417/8+D419Y biopsies, we observed elevated expression of p16 in KIT positive mast cells, a hallmark biomarker of senescence (Fig. 2F) [13].

Telomere length and TRF2 shelterin protein

To evaluate the role of telomerase in telomere maintenance in the KIT D816V model, we tested relative telomerase activity in primary HDF transduced with either the telomerase core protein hTERT (as a positive control), or with KIT D816V or KIT Del 417/8+D419Y. No telomerase activity was observed in pHDF transduced with either KIT mutant (Supplementary Fig. 2).

Since transduction of KIT D816V did not induce telomerase activity in pHDF, we hypothesized that this version of KIT may enhance telomere protection by acting on the shelterin protein complex, which caps and protects telomeres against the DNA repair machinery [14,15,16,17]. TRF2 is one of the main proteins bound to telomere DNA and plays a central role in telomere maintenance and protection against end-to-end fusion of chromosomes (Supplementary Fig. 3) [18]. Using immuno-FISH analysis targeting TRF2 protein and telomeric DNA, punctiform signals were detected and corresponded to the specific colocalization of TRF2 to telomeres (Fig. 3A). While total TRF2 protein level, quantified by western blot and immunofluorescence, was equivalent in both KIT D816V and KIT Del 417/8+D419Y pHDF, TRF2 was differentially localized within the nucleus (Fig. 3B, C). Indeed, we tested this hypothesis in FFPE skin biopsies of mastocytosis patients and observed a significantly increased TRF2 bright punctiform signal per cell in KIT D816V mast cells compared to a weak and diffuse signal in KIT Del 417/8+D419Y mast cells from mastocytosis patients (p = 0.0116). Similar results were observed in KIT D816V and KIT Del 417/8+D419Y pHDF (p = 0.0018) (Fig. 3C, D).

Fig. 3: Telomere length shortening correlates with senescence and TRF2 diffuse pattern expression and activation of the p38 MAP-kinase pathway.
Fig. 3: Telomere length shortening correlates with senescence and TRF2 diffuse pattern expression and activation of the p38 MAP-kinase pathway.
Full size image

A TRF2 punctiform pattern detected by immunofluorescence is colocalized with telomeric specific probe signal by fluorescent in situ hybridization (FISH) in control, KIT Del 417/8+D419Y and KIT D816V pHDF. Quantitative study shows a significant decreased percentage of colocalized TRF2 and telomeric signal in KIT Del 417/8+D419Y compared with control (p = 0.0003) and KIT D816V (p < 0.0001). B TRF2 quantitative expression in control, KIT Del 417/8+D419Y and KIT D816V pHDF by western blot (n = 3). C TRF2 expression by fluorescence intensity and TRF2 punctiform signal quantification in control pHDF (empty vector, EV) or KIT Del 417/8+D419Y (p = 0.0229) and KIT D816V (p = 0.0018). Picture highlights TRF2 differential signal topography (green) with diffuse (KIT Del 417/8+D419Y) or punctiform (KIT D816V) pattern. D Mast cells TRF2 expression in cutaneous biopsies from mastocytosis patients by fluorescence intensity and TRF2 punctiform signal quantification in patients affected by cutaneous mastocytosis with KIT Del 417/8+D419Y (n = 3) and KIT D816V (n = 3). Pictures highlight TRF2 differential signal topography (red) with diffuse (KIT Del 417/8+D419Y) or punctiform (KIT D816V) pattern in mast cells (green) (p = 0.0018).

Taken together, these findings indicate that telomere shortening and senescence of KIT Del 417/8+D419 patient mast cells result from the delocalization of TRF2 from the telomere ends.

MAP-kinase TRF2 phosphorylation and delocalization from telomere

KIT signaling pathway activates the MAP-kinase pathway [19], and we previously published a difference between KIT D816 and non-KIT D816 signaling pathway [11]. In order to identify the most significant and differentially expressed signaling pathways between the KIT D816V or non-KIT D816V models, we used a phospho-tyrosine kinase array (Supplementary Fig. 4A). We identified an increased activated phospho-p38 MAP-kinase in the KIT Del 417/8+D419 pHDF compared to the KIT D816V pHDF. Since the p38 MAP-kinase may interact directly with TRF2 [20], we specifically inhibited the p38 MAP-kinase pathway with SB203580, a specific and potent inhibitor of p38 MAP-kinase activity, over a 21-day period (Supplementary Fig. 4B) [21]. Our results showed that upon SB203580 treatment, total TRF2 expression was not modified (Fig. 4A). However, in agreement with our hypothesis, we observed a significant difference of signal distribution per cell in KIT Del 417/8+D419Y pHDF, with an increased bright and punctiform pattern, which was not modified in KIT D816V pHDF (Fig. 4A). Telomere length was maintained upon inhibition of the p38 MAP-kinase activity (Fig. 4B), and as expected, the TRF2 punctiform pattern corresponded to a colocalization with telomeres (Fig. 4C).

Fig. 4: MAP-Kinase pathway inhibition is associated with telomere maintenance and persistent TRF2 punctiform signal pattern.
Fig. 4: MAP-Kinase pathway inhibition is associated with telomere maintenance and persistent TRF2 punctiform signal pattern.
Full size image

A Quantitative TRF2 expression by western blot in control EV pHDF, KIT Del 417/8+D419Y pHDF and KIT D816V pHDF after 21 days of culture compared with SB03580 condition. TRF2 expression (A) in control EV pHDF, KIT Del 417/8+D419Y pHDF and KIT D816V pHDF after 21 days of culture compared with SB03580 condition (n = 3). B Telomere length in control EV pHDF, KIT Del 417/8+D419Y pHDF and KIT D816V pHDF after 21 days of culture compared with SB03580 condition. C Colocalization of telomere probes and TRF2 is significantly increased in KIT Del 417/8+D419Y pHDF (p < 0.0001) after SB03580 culture. D TRF2 PhosTag assay show increased high level phosphorylated TRF2 form in KIT Del 417/8+D419Y pHDF after 21 days of culture compared with SB203580 condition, decreased by SB203580, after treatment in all conditions, low level of phospho-TRF2 is increased. E Quantitative TRF2 expression by immunofluorescence in pHDF infected with wild type (WT) TRF2 and mutated 5S/A TRF2 on the five putative phosphorylation sites, treated by anisomycin to activate phosphorylation of MAPkinase and p38 MAP-kinase pathway. Activation of p38 pathway by anisomycin leads to a decrease signal/cell TRF2 in TRF2-WT pHDF, but no modification in the signal of TRF2 appears in mutant 5S/A TRF2 cells.

Thus, telomere length maintenance might be preserved in KIT D816V due to a modest activation of the p38 MAPK pathway. Furthermore, a three-week treatment of KIT Del 417/8+D419Y pHDF with the p38 inhibitor resulted in the maintenance of telomere lengths and punctiform signal of TRF2, while stopping this treatment led to a significant telomere length shortening after 10 days (Supplementary Fig. 5A–D). We also treated pHDF harboring wild-type KIT with anisomycin, a MAPK p38 activator, and as expected, the results showed a modification of the signal pattern of TRF2 in nuclei, similar to the KIT Del 417/8+D419Y pHDF model (Supplementary Fig. 6). Taken together, these results strongly suggest that p38 MAPK activation is involved in telomere shortening through TRF2 delocalization.

To confirm that TRF2 is phosphorylated, we used a PhosTag assay to assess its level of phosphorylation. Most of the p-TRF2 forms, called pp7 and pp8, were increased in the KIT Del 417/8+D419Y pHDF and decreased upon SB203580 treatment, reaching levels similar to those observed in KIT D816V pHDF. Accordingly, the less phosphorylated form of TRF2, called pp4 and pp5, were increased in KIT D816V pHDF and in KIT Del 417/8+D419Y pHDF upon SB203580 treatment (Fig. 4D). To demonstrate a direct phosphorylation of TRF2 dependent on p38 MAP-kinase, we first used p-TRF2 (Thr188) antibody and showed that the diffuse pattern of p-TRF2 in KIT Del 417/8+D419Y pHDF is decreased upon MAPKinase inhibition with SB203580 (Supplementary Fig. 7).

These results revealed an inverse correlation between TRF2 phosphorylation, the occupancy of TRF2 on telomeres, and telomere length. First, to ensure that decrease of occupancy was not due to short telomeres, we treated WT pHDF by Anisomycin [22] for a short period of time (thirty minutes), which does not allow telomere shortening. The result showed that TRF2 rapidly delocalized from telomeres independently of telomere shortening (Fig. 4E).

TRF2 MAP-kinase phosphorylation site mutagenesis

Second, to demonstrate the role of TRF2 phosphorylation in its delocalization, we performed a site-directed mutagenesis on all five potential phosphorylation sites (hprd.org database) and transduced pHDF with this non-phosphorylatable TRF2 mutant (5S/A TRF2). As expected, 5S/A TRF2 was unresponsive to anisomycin stimulation, as its telomere localization remained unchanged, in contrast to the wild-type TRF2 protein (Fig. 4E). These data strongly support a direct role of MAPKinase p38 dependent phosphorylation in TRF2’s telomere localization.

Discussion

Mastocytosis presents as a heterogeneous disorder with distinct clinical outcomes in adults and children. While adult-onset mastocytosis is chronic and non-regressive, pediatric cases frequently undergo spontaneous remission, a phenomenon that remains poorly understood. In this study, we uncover a novel mechanistic link between KIT mutations, telomere dynamics, and cellular senescence, which may explain the differential disease courses in pediatric and adult mastocytosis.

Our findings demonstrate that pediatric-type non-PTD KIT mutations, such as KIT Del 417/8+D419Y, induce telomere shortening in MCs, leading to a senescent phenotype that may drive disease regression. In contrast, adult-type PTD KIT mutations, primarily KIT D816V, are associated with preserved telomere integrity, which could contribute to disease persistence. This observation is consistent across our in vitro models, primary patient samples, and murine studies, supporting a direct role for telomere dysfunction in pediatric mastocytosis regression. Telomere shortening is a well-established mechanism of senescence, acting as a “biological clock” that limits excessive cellular proliferation [23, 24]. Under physiological conditions, telomere length is expected to be longer in children than in adults due to reduced cumulative cell divisions [25]. However, our analysis revealed that MCs from pediatric patients carrying KIT Del 417/8+D419Y mutations exhibit significantly shorter telomeres than those from adults, suggesting an accelerated shortening process. Given the absence of detectable telomerase activity in MCs [26], we hypothesized that alternative telomere regulatory mechanisms could be implicated in mastocytosis. Among these, the shelterin protein TRF2 plays a crucial role in protecting telomeres from DNA damage activation. TRF2 stabilization and its binding to telomere DNA can be modulated by post-translational modifications, which may, directly or indirectly, lead to its displacement from telomeres [27]. We observed that TRF2 localization to telomeres was disrupted in MCs carrying KIT Del 417/8+D419Y mutations, leading to telomere deprotection and increased senescence, as confirmed by p16 staining and SA-β-gal activity in patient-derived samples.

To identify potential upstream regulators of TRF2 delocalization and telomere instability, we investigated signaling differences between KIT D816V and KIT Del 417/8+D419Y mutations. Using phospho-kinase arrays, we identified a significant increase in p38 MAPK activation in KIT Del 417/8+D419Y MCs. This finding is particularly relevant given that p38 MAPK has been suggested to interact with TRF2 [20]. Further supporting this mechanism, we demonstrated that pharmacological inhibition of p38 MAPK using SB203580 restored TRF2 localization to telomeres, maintained telomere length, and prevented mast cell senescence. Conversely, activation of p38 MAPK with anisomycin led to TRF2 delocalization and telomere attrition, mimicking the effects of KIT Del 417/8+D419Y mutations. These data strongly suggest that p38 MAPK activation regulates TRF2 localization to telomeres, thereby influencing telomere maintenance in KIT Del 417/8+D419Y MCs. This mechanism may contribute to the distinct telomere dynamics observed in pediatric mastocytosis. In contrast to pediatric cases, adult-onset mastocytosis is strongly associated with KIT D816V mutations and a non-regressive phenotype. Our data suggest that KIT D816V-expressing MCs maintain telomere integrity, potentially prolonging their survival. Immuno-FISH analysis confirmed that TRF2 localization remained stable in KIT D816V MCs, indicating preserved telomere protection. Notably, despite lacking telomerase activity, adult MCs did not exhibit the same degree of telomere shortening or senescence as pediatric MCs. One potential explanation is that the maintenance of telomere length in KIT D816V MCs may be linked to a modest activation of the p38 MAPK pathway, which could limit TRF2 delocalization and prevent telomere deprotection. This mechanism could underlie the clonal persistence of MCs in adults and the chronic progression of the disease.

Our findings provide a mechanistic basis for the divergent clinical courses of adult and pediatric mastocytosis, linking KIT mutations with telomere stability and mast cell lifespan. Beyond mastocytosis, these insights may extend to other malignancies characterized by oncogenic KIT mutations or p38 MAPK activation, such as myeloid neoplasms and solid tumors.

Conclusion

Our study uncovers a previously unrecognized role of telomere biology in mastocytosis, demonstrating that KIT mutation subtypes dictate mast cell fate through differential effects on telomere stability and senescence. In children, mastocytosis is more frequently associated with non-KIT D816V mutations, such as KIT Del 417/8+D419Y. These mutations lead to p38 MAPK activation, which is associated with TRF2 delocalization and telomere deprotection, triggering mast cell senescence and spontaneous disease regression. In adults, mastocytosis is primarily driven by KIT D816V mutations, which are associated with preserved TRF2 localization and telomere integrity. This contributes to clonal mast cell persistence and the chronic nature of the disease. These findings provide a mechanistic framework for understanding the spontaneous resolution of pediatric mastocytosis, highlighting telomere dysfunction as a potential contributor to disease remission.

Finally, our work suggests that telomere-associated pathways could serve as novel intervention points for KIT-driven hematologic malignancies and could be extended to other neoplasms, especially those driven or associated with activation of the p38 MAPK pathway.