High Frequency of Germline ATM Variants in Familial Philadelphia‐Negative Myeloproliferative Neoplasms

Familial Philadelphia-negative myeloproliferative neoplasms (MPNs) represent a recognized subset of these clonal hematopoietic disorders, accounting for approximately 7%–8% of all MPN cases 1. Population-based studies have demonstrated a high risk of MPNs among first-degree relatives of affected individuals, particularly in families with multiple affected members or in relatives of younger patients 2, 3. Notably, familial and sporadic cases show no significant differences in clinical phenotype, disease complications, or risk of progression, suggesting that familial MPNs share the same disease biology and prognosis as their sporadic counterparts 4. However, as in other familial cancers, second-generation individuals experience earlier disease onset or more aggressive phenotypes 5. In addition, the spectrum of somatic mutations closely mirrors that of sporadic MPNs, with JAK2 V617F being the most common acquired driver mutation. Nevertheless, discordant somatic mutations are frequently observed among related individuals, suggesting that while shared germline factors may confer susceptibility to clonal hematopoiesis, the acquisition of specific driver mutations occurs stochastically and independently 6. In this context, we conducted a retrospective, single-center study aimed at identifying germline variants (GVs) in individuals with familial MPNs through next-generation sequencing (NGS). We included all patients diagnosed with MPNs at Fondazione Policlinico Gemelli IRCCS between 1976 and 2024 who had at least one relative affected by an MPN. All diagnoses were confirmed according to WHO 2022/ICC criteria 7, 8. DNA was extracted from cryopreserved peripheral blood samples, and clinical data were retrieved from medical records. All participants provided written informed consent per the Declaration of Helsinki. The study was conducted under protocol NCT06923670 (ClinicalTrials.gov). A custom enrichment panel targeting 69 genes (Illumina, San Diego, CA), selected based on established evidence for hereditary cancer syndromes, was designed using Illumina Design Studio. The full gene list is provided in Table S1. Germline variants were classified into 5 categories according to American College of Medical Genetics and Genomics guidelines: pathogenic (P), likely pathogenic (LP), variant of uncertain significance (VUS), likely benign, and benign 9. The SIFT (Sorting Intolerant From Tolerant) tool was used to predict the functional impact of amino acid substitutions, with variants scoring < 0.05 considered deleterious 10. Among a cohort of 1515 MPN patients, 107 (7.1%) met the inclusion criteria and were included in the analysis. The median age at diagnosis was 47.5 years (range 1–82 years), with a predominance of female patients (56.1%). In 57.9% of cases, the proband had at least one first-degree relative with a myeloid neoplasm; 28.9% had affected second-degree relatives, and 13.1% had affected third-degree relatives. In 16 cases (14.9%), more than one relative was affected by a myeloid neoplasm: two affected relatives in 12 cases and three affected relatives in 4 cases. The most common MPN subtype was Essential Thrombocythemia (ET, 49.5%), and the majority of patients (72.9%) harbored the JAK2 V617F mutation. Complete follow-up data were available for 105 patients. Among these, 100 (95.2%) received antithrombotic therapy, 69 (65.7%) underwent at least one cytoreductive treatment, and 30 (28.6%) experienced a thrombotic event. Twenty-eight (28.9%) of the 97 patients with ET, polycythemia vera (PV), pre–primary myelofibrosis (pre-PMF), or myeloproliferative neoplasm unclassifiable (MPN-U) progressed to overt-PMF (patients with overt-PMF at diagnosis were not included), and three (2.9%) of the entire cohort developed acute leukemia. At the time of analysis, 13 patients (12.4%) had undergone allogeneic stem cell transplantation, 8 (7.6%) died, and 16 (15.2%) were lost to follow-up. Notably, 22 patients (20.9%) had a history of another malignancy, including three (2.9%) who developed an additional hematologic neoplasm (two chronic lymphocytic leukemia CLL and one chronic myeloid leukemia CML). Demographic and clinical characteristics of the cohort are summarized in Table 1. A total of 106 GVs were identified across 45 distinct genes included in our panel. Sixty-eight patients (63.6%) carried at least one GV (benign and likely benign GVs were excluded from the analysis): 38.3% had a single variant, 16.8% two, 6.5% three, and 1.9% four variants (Figure 1A). Most patients (46.7%) carried only VUS, while 9.3% harbored a LP variant, and 7.5% carried a pathogenic variant. JAK2 V617F was the most frequent somatic mutation in patients with GVs, detected in 76.5% of these individuals. There was no significant difference in GVs prevalence between JAK2-mutated and wild-type patients (χ2 = 0.35, p = 0.55), and GVs presence was independent of MPN subtype (Fisher's exact test p = 0.65). However, a significant difference in JAK2 V617F variant allele frequency (VAF) was observed between patients with and without GVs (Wilcoxon rank-sum test p = 0.016), with a higher median VAF in patients harboring GVs (39% vs. 14%). Thirty-eight patients (35.5%) in the cohort were directly related, forming 19 family pairs. Among these, 11 pairs (57.9%) harbored the same driver mutation—JAK2 V617F in 9 pairs (47.4%) and MPL S505N in 2 pairs (10.5%)—while only 8 pairs (42.1%) presented with concordant clinical phenotypes, all diagnosed as ET. In addition, six other individuals with MPN (5.6%) had a family member affected by a different myeloid malignancy (4 with acute myeloid leukemia, 1 with CML, and 1 with chronic myelomonocytic leukemia), all followed at our institution. A sub-analysis was therefore conducted on 25 pairs (50 individuals) of related subjects with available genetic material. The median age at diagnosis in the second generation was significantly lower than in the first generation (43 years vs. 61 years, p = 0.006). Among these 50 individuals, 36 (72.0%) carried at least one GV, most of which (n = 33, 66.0%) were classified as VUS. Nine (18%) presented a P/LP variant and in 12 pairs (48%), both individuals shared at least one identical GV. Figure S1. ATM was the most frequently mutated gene, with GVs identified in 12 patients (11.2%) (Figure 1B). Detailed information on the identified ATM variants is provided in Table 2. Among these 12 patients, 8 were from 4 familial pairs, and in each case, both members of the pair carried the identical mutation; therefore, the incidence of the ATM GVs among families was 9.1% (8 of 88 families). One individual was found to harbor two distinct ATM GVs. ATM variants were classified as pathogenic in 3 patients (3.1%), LP in 1 patient (1.0%), and VUS in 8 patients (8.2%). However, when evaluated using the SIFT algorithm, 5 of the 9 VUS were predicted to be deleterious (SIFT score < 0.05). Patients with identified ATM variants were further investigated to confirm the germline origin of the mutation. Specifically, NGS analysis was performed on DNA samples obtained from saliva, which confirmed the germline origin of the mutation. These patients were subsequently also examined via WES, which corroborated the variants identified by the NGS panel and did not reveal any additional variants associated with hematopoietic abnormalities of note. In this single-center study, approximately 7% of MPN cases met criteria for familial disease, consistent with prior reports 1. We confirmed anticipation, with second-generation individuals showing earlier disease onset. JAK2 V617F was the most frequent somatic driver mutation, and patients harboring GVs exhibited higher median VAFs, suggesting that inherited alterations may facilitate clonal expansion. In familial pairs, concordance of phenotype and driver mutation was uncommon, whereas nearly half shared at least one GV, highlighting the role of inherited factors beyond canonical drivers in familial MPN predisposition. ATM emerged as the most frequently mutated germline gene. ATM, located on chromosome 11, encodes a serine/threonine kinase essential for DNA damage response and oxidative stress regulation, coordinating cell-cycle arrest, apoptosis, and metabolic reprogramming through TP53 activation to preserve genomic integrity 12. The contribution of ATM to cancer predisposition is well established in patients with ataxia–telangiectasia, who exhibit a markedly increased risk of early-onset malignancies, predominantly hematologic 13. Moreover, several studies have reported germline ATM variants in familial hematologic neoplasms—including CLL and myeloid diseases as well as in hereditary solid tumors 14-18. An Italian study recently reported P/LP germline mutations in approximately 7% of individuals with familial or early-onset (< 27 years) MPNs, a frequency lower than that observed in our cohort (18.4%) 19. This difference likely mirrors the fact that our study population consisted solely of individuals with affected family members. Notably, 2.2% of subjects in this study harbored CHEK2 mutations, which interact with ATM and have been associated with an increased risk of second malignancies and leukemic transformation 19; in our cohort, only one subject (0.9%) carried a CHEK2 LP variant. Beyond the ATM GVs, several identified variants affect genes involved in DNA interstrand cross-link repair. Biallelic mutations in these genes cause Fanconi anemia, while heterozygous variants increase cancer susceptibility 20. We identified 18 GVs across 9 genes. BRCA2 and FANCD2 were most frequently affected 3 (2.8%) patients each with VUS while 6 (5.6%) subjects carried P/LP variants in FANCA, FANCE, FANCL, and XRCC2. Our findings support a potential role for ATM, along with other genes predominantly involved in DNA repair pathways, as predisposing factors in familial MPNs, underscoring the critical importance of genomic stability in cancer predisposition. A limitation of our study is the use of a targeted gene panel; however, WES performed in the subset of patients with ATM mutations did not reveal any additional pathogenic variants. Family history should be an integral component of the diagnostic workup for MPNs, and further studies are needed to clarify germline contributions to disease susceptibility, with potential implications for therapeutic strategies and donor selection in patients undergoing hematopoietic stem cell transplantation. P.C., S.S., F.R., T.M., and L.D.M. conceived and designed the study. P.C., F.R., L.D.M., R.M., and S.B. collected the data. T.M., M.R., and G.M. performed the NGS analyses, while A.M., M.D.B., and R.B. conducted the WES analysis. P.C. and F.R. wrote the manuscript. E.R., S.S., V.D.S., and P.C. critically reviewed and revised the manuscript for important intellectual content. This monocentric study serves as the pilot phase of a larger multicenter project titled “Prevalence of Germline Gene Mutations in Patients with Myeloproliferative Neoplasms and a Family History of Hematological Neoplasms.” This research is conducted under the patronage of GIMEMA (Italian Group of Adult Hematological Diseases). The authors acknowledge the support of the “Centro di Ricerca sulle Cellule Staminali Emopoietiche e le Terapie Cellulari, Università Cattolica del Sacro Cuore, Roma” research funding. We sincerely thank Prof. Paola Guglielmelli for her valuable advice. This study was conducted with the support of the research funding from the “Centro di Ricerca sulle Cellule Staminali Emopoietiche e le Terapie Cellulari, Università Cattolica del Sacro Cuore, Roma”. The data supporting the findings of this study were collected as part of a registered clinical trial (ClinicalTrials.gov Identifier: NCT06923670). The authors declare no conflicts of interest. Original data will be made available upon reasonable request to the corresponding author, in accordance with institutional and ethical guidelines. The raw sequencing reads have been deposited in the NCBI Sequence Read Archive (SRA; accession no. SRP631076) and BioProject (accession no. PRJNA1312845). Table S1: List of genes included in NGS-panel and their cellular function. Figure S1: Distribution of germline variants per gene identified among 25 pairs of related individuals. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.