Next-generation sequencing has enabled us to detect driver mutations in a sensitive manner. By whole-exome sequencing, we previously identified a somatic mutation of the plant homeodomain finger 6 (PHF6) gene (p.G291X) in 1 out of 29 cases with myelodysplastic syndromes (MDS).1 Initially, germline mutations of PHF6, located at Xq26.2, are reported to cause congenital Börjeson–Forssman–Lehmann syndrome (BFLS) with X-linked recessive inheritance.2 BFLS is characterized by mental deficiency, epilepsy, hypogonadism, obesity and dysmorphic features.3 Recently, it was found that germline PHF6 mutations are also responsible for the female cases with a congenital disorder similar to Coffin–Siris syndrome.4 Moreover, somatic PHF6 mutations were reported in hematological neoplasms, including T-acute lymphocytic leukemia (38% of cases were positive for mutations)5 and acute myeloid leukemia (AML) (3%).6 According to recent studies, somatic PHF6 mutations were identified in 3% of cases with de novo AML7 and in ~3% of those with MDS.8, 9 Nevertheless, pathophysiology due to PHF6 defects in myeloid neoplasms remains to be fully elucidated.
In this study, we clarified the implications of somatic PHF6 mutations in the cases with various myeloid neoplasms (N=1760), including the cohort of MDS and AML that we previously reported.8, 10, 11 To identify somatic mutations, we applied whole-exome sequencing to 49 cases. Subsequently, targeted sequencing (SureSelect, Agilent, Santa Clara, CA, USA) and PCR-based pool sequencing were performed in 1428 and 356 cases, respectively, 73 of which were subjected to both methods (Supplementary Table 1). Detailed methods of the sequencing were previously reported.1, 8 Written consent forms were obtained from all the patients. Genetic analysis was approved by the ethical review board in each institution. Somatic mutations were confirmed by paired DNA from tumor and germline samples (buccal smear or CD3-positive cells). In case of non-paired DNA, the nonsense and frameshift mutations were classified to be somatic, and the missense mutations were classified as somatic if they were already reported as somatic in the Catalogue of Somatic Mutations in Cancer database (http://www.sanger.ac.uk/genetics/CGP/cosmic/).
In total, we identified 62 somatic mutations of PHF6 in 54 cases (Table 1). By copy number analysis,1, 8 deletions affecting the PHF6 locus were identified in five cases, while no focal amplifications of PHF6 locus were identified. Among somatic mutations, 17 were missense, 16 frameshift, 23 nonsense and 6 affecting splice sites. Therefore, mutations leading to truncated transcripts were dominant (63%, 39/62). While PHF6 mutations were distributed to the whole coding region, 14 out of 17 (82%) missense mutations were located at the PHD2 domain and 8 (47%) were recurrent (p.R274Q) (Figure 1a). The PHD2 domain of PHF6 is rich in positively charged amino acids including arginine and lysine, which were confirmed to be essential for the DNA-binding capacity of PHF6 as recently reported.12 Consequently, missense mutations affecting these amino acids in the PHD2 domain (p.R274Q and p.K235E) (Figure 1a) might result in loss of PHF6 function. Together with highly frequent truncating mutations and dominant deletions, most of the PHF6 mutations (87%; 53/61) might be pathogenic in myeloid neoplasms due to loss of function.
Clinically, PHF6 mutations were detected in the cases with AML with myelodysplasia-related changes (AML/MRC) (4/26, 15.4%), de novo AML (11/340, 3.2%), chronic myelomonocytic leukemia (CMML) (4/86, 4.7%), MDS (34/1139, 3.0%) and chronic myelogenous leukemia (CML) (1/64, 1.6%) (Table 1). In MDS, PHF6 mutations were significantly more frequent in the cases with refractory anemia with excess blasts (RAEB) (5.3%) than those with the other phenotypes (1.3%, P<0.001). Such predominant mutations in the cases with AML/MRC and RAEB suggest that PHF6 mutations might have an important role in leukemic evolution. To the contrary, PHF6 mutations were rare in the cases with myeloproliferative neoplasms (0.7%), including CML, which is similar to the findings reported in a recent paper from International Cancer Genome Consortium.13
Although frequent PHF6 mutations in male cases with AML were reported,6 the unbiased comparisons of subgroups with matched clinical backgrounds did not confirm the male dominance (Supplementary Table 2). While another paper showed shorter overall survival in de novo AML cases with PHF6 mutations,7 PHF6 mutations were not significantly associated with poor outcomes in any phenotype of our data set (data not shown).
Next, we investigated an impact of PHF6 mutations on clinical parameters. By univariate analyses on several variables including bone marrow blast ratio (%), peripheral blood neutrophil counts, hemoglobin levels, and platelet counts, PHF6 mutations were significantly associated with higher bone marrow blast ratio (P=0.008) and lower platelet counts (P=0.027) (Supplementary Table 3a). A following multivariate analysis of these significant factors revealed that PHF6 mutations were significantly associated with high blast ratio but not with platelet counts (Supplementary Table 3b). These findings were compatible to the high prevalence of PHF6 mutations in advanced categories of myeloid neoplasms. In addition, we assessed a relationship between PHF6 mutations and cytogenetics (in 944 patients with MDS and 156 with de novo AML whose information on karyotyping was available), which revealed that PHF6 mutations were not significantly associated with abnormal karyotype (Supplementary Table 4). Subgroup analyses on the specific chromosomal abnormalities most frequently identified in cases with PHF6 mutations (+8, t(8;21), and complex karyotype) showed that neither of them was significantly associated with PHF6 mutations. These results suggest that cytogenetic abnormalities are not likely related to PHF6-defective myeloid neoplasm.
Therefore, we further investigated concomitant mutations of 20 commonly evaluated genes in patients with PHF6 mutations, in whom several driver genes were frequently mutated (Supplementary Table 5). In our whole cohort, PHF6 mutations were significantly associated (false discovery rate <0.1) with those of RUNX1, U2AF1, SMC1A, ZRSR2, EZH2, and ASXL1. In contrast, SF3B1 mutations were exclusive with PHF6 mutations, which is compatible with the fact that PHF6 mutations were very rarely detected in cases with refractory anemia with ringed sideroblasts (0.9%). For subtypes of AML, RUNX1 mutations significantly co-occurred with PHF6 mutations in AML/MRC cases (P=0.009), and EZH2 (P=0.008), SMC1A (P=0.007) and RUNX1 (P=0.04) mutations were also significantly frequent in de novo AML cases with PHF6 mutations. In MDS cohort, PHF6 mutations were significantly coincident with U2AF1 (P=0.0005), RUNX1 (P=0.0009), IDH1/2 (P=0.01), and ASXL1 (P=0.01) mutations, while SF3B1 mutations were exclusively identified (P=0.002). Out of these, RUNX1 mutations were significantly frequent in multiple disease subtypes with PHF6 mutations, suggesting this gene might have a pathogenic role in PHF6-related myeloid neoplasm. With regard to copy number alterations, two out of five deletions at PHF6 locus were focal lesions, two others represented whole-chromosomal losses and the fifth was an i(Xq). These two focal deletions of PHF6 also included the loci of STAG2 and SMC1A. These findings suggest that the deletions and mutations of these genes with PHF6 on the same chromosome might have a synergetic effect on leukemic evolution in myeloid neoplasms.
To further clarify the PHF6 pathophysiology in MDS, we focused our study on the well-annotated MDS cohort (N=944).8 First, the mutational number of the genes except for PHF6 on average per case was significantly higher in the cohort with PHF6 mutations (3.8) compared to that with wild-type PHF6 (2.8) (P=0.018). To further elucidate an impact of PHF6 mutations on the increased number of other mutations, a two-way factorial analysis of variables including disease risks was performed. The number of mutations was significantly associated with high-risk MDS (P<0.001), but not with PHF6 mutations. Accordingly, the association between PHF6 mutations and a high number of other mutations is indirect, and most likely the reason why cases with PHF6 mutations are associated with a high number of mutations is that PHF6 mutations are more frequent in high-risk MDS harboring more mutations (Supplementary Figure 1). Second, the ratio of the cases with intra-tumor heterogeneity was significantly higher in the cohort with PHF6 mutations (P=0.006), which was evaluated by comparing variant allele frequencies (VAFs) between PHF6 and the other genes as previously described.8 VAFs of PHF6 mutations tended to be low, and more than half of them (54%. 13/24) were acquired in subclonal populations (Figure 1b). VAFs of mutations in the genes associated with RNA splicing and DNA methylation were higher than those of PHF6 mutations, suggesting that PHF6 mutations might be acquired in the later phase of the clinical course.
In 7 out of 53 cases with PHF6 mutations, multiple PHF6 mutations were identified. Among these, five mutations and a deletion of PHF6 were detected in an index female patient. By targeted sequencing, p.C297S and p.R319X mutations were observed on the independent sequence reads, suggesting that these two mutations might be acquired in the distinct clones concomitant with the deletion of the wild-type PHF6 allele (parallel mutations) (Supplementary Figure 2).
PHF6 encodes a transcription factor with two PHD-type zinc-finger domains. Wang et al.14 showed that PHF6 in the nucleolus regulates the cell cycle by controlling synthesis of ribosomal RNA, and that the loss of PHF6 increases DNA damage. This suggests the tumor suppressor potential of PHF6, which dovetails with our results that most of the PHF6 defects probably resulted in the loss of function of the gene. Furthermore, PHF6 mutations were often identified in the subclone and sometimes in a parallel manner, which suggests that these mutations were acquired during disease progression. PHF6 mutations were frequent in the cases with AML/MRC and significantly associated with RUNX1 mutations. This result recapitulates the previous report showing RUNX1 mutations are frequent in the cases with AML/MRC,15 suggesting a synergistic effect of these mutations on leukemic evolution.
In summary, PHF6 defects most likely result in their loss of function and have a substantial effect on the evolution into the aggressive types of myeloid neoplasms, associated with other concomitant genetic defects including RUNX1 mutations.
Yoshida K, Sanada M, Shiraishi Y, Nowak D, Nagata Y, Yamamoto R et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature 2011; 478: 64–69.
Lower KM, Turner G, Kerr BA, Mathews KD, Shaw MA, Gedeon AK et al. Mutations in PHF6 are associated with Borjeson-Forssman-Lehmann syndrome. Nat Genet 2002; 32: 661–665.
Borjeson M, Forssman H, Lehmann O . An X-linked, recessively inherited syndrome characterized by grave mental deficiency, epilepsy, and endocrine disorder. Acta Med Scand 1962; 171: 13–21.
Wieczorek D, Bogershausen N, Beleggia F, Steiner-Haldenstatt S, Pohl E, Li Y et al. A comprehensive molecular study on Coffin-Siris and Nicolaides-Baraitser syndromes identifies a broad molecular and clinical spectrum converging on altered chromatin remodeling. Hum Mol Genet 2013; 22: 5121–5135.
Van Vlierberghe P, Palomero T, Khiabanian H, Van der Meulen J, Castillo M, Van Roy N et al. PHF6 mutations in T-cell acute lymphoblastic leukemia. Nat Genet 2010; 42: 338–342.
Van Vlierberghe P, Patel J, Abdel-Wahab O, Lobry C, Hedvat CV, Balbin M et al. PHF6 mutations in adult acute myeloid leukemia. Leukemia 2011; 25: 130–134.
Patel JP, Gonen M, Figueroa ME, Fernandez H, Sun Z, Racevskis J et al. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 2012; 366: 1079–1089.
Haferlach T, Nagata Y, Grossmann V, Okuno Y, Bacher U, Nagae G et al. Landscape of genetic lesions in 944 patients with myelodysplastic syndromes. Leukemia 2014; 28: 241–247.
Papaemmanuil E, Gerstung M, Malcovati L, Tauro S, Gundem G, Van Loo P et al. Clinical and biological implications of driver mutations in myelodysplastic syndromes. Blood 2013; 122: 3616–3627.
Kihara R, Nagata Y, Kiyoi H, Kato T, Yamamoto E, Suzuki K et al. Comprehensive analysis of genetic alterations and their prognostic impacts in adult acute myeloid leukemia patients. Leukemia 2014; 28: 1586–1595.
Kon A, Shih LY, Minamino M, Sanada M, Shiraishi Y, Nagata Y et al. Recurrent mutations in multiple components of the cohesin complex in myeloid neoplasms. Nat Genet 2013; 45: 1232–1237.
Liu Z, Li F, Ruan K, Zhang J, Mei Y, Wu J et al. Structural and functional insights into the human Borjeson-Forssman-Lehmann syndrome-associated protein PHF6. J Biol Chem 2014; 289: 10069–10083.
Nangalia J, Massie CE, Baxter EJ, Nice FL, Gundem G, Wedge DC et al. Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. N Engl J Med 2013; 369: 2391–2405.
Wang J, Leung JW, Gong Z, Feng L, Shi X, Chen J . PHF6 regulates cell cycle progression by suppressing ribosomal RNA synthesis. J Biol Chem 2013; 288: 3174–3183.
Harada H, Harada Y, Niimi H, Kyo T, Kimura A, Inaba T . High incidence of somatic mutations in the AML1/RUNX1 gene in myelodysplastic syndrome and low blast percentage myeloid leukemia with myelodysplasia. Blood 2004; 103: 2316–2324.
This work was supported by Grant-in-Aids from the Ministry of Health, Labor and Welfare of Japan and KAKENHI (23249052, 22134006, 21790907, 15km0305018h0101, 16H05338, 26115009, 26890016 and 15H05668) (Kyoto; SO, HM, MS and KY), project for development of innovative research on cancer therapies (p-direct) (Kyoto, SO), The Japan Society for the Promotion of Science (JSPS) through the ‘Funding Program for World-Leading Innovative R&D on Science and Technology’, initiated by the Council for Science and Technology Policy (CSTP) (Kyoto, SO) and NHRI-EX100-10003NI Taiwan (Taipei LYS).
TM, YN, HidM, MS, SO and KY performed study design and interpretation of the data. TM, YN, MS, YusS, AK, TY, ASO, KK and RK performed molecular experiments and data analysis. YuiS, KC, HT and SaM performed bioinformatics. KI, ShM, HirM, TsN, RK, HK, HPK, LYS, ToN, CH, WK and TH collected samples and clinical data, contributed to the interpretation of the data and critically reviewed the draft.
TH, CH and WK declare equity ownership of MLL Munich Leukemia Laboratory GmbH.
Supplementary Information accompanies this paper on the Leukemia website
About this article
Cite this article
Mori, T., Nagata, Y., Makishima, H. et al. Somatic PHF6 mutations in 1760 cases with various myeloid neoplasms. Leukemia 30, 2270–2273 (2016). https://doi.org/10.1038/leu.2016.212
Distinctive phenotypes in two children with novel germline RUNX1 mutations - one with myeloid malignancy and increased fetal hemoglobin
Pediatric Hematology and Oncology (2021)
Current Opinion in Hematology (2020)
Journal of Molecular Neuroscience (2020)
Proceedings of the Japan Academy, Series B (2020)
Leukemia & Lymphoma (2020)