Genomics Analysis Of Leukaemia Predisposition In X-Linked Agammaglobulinaemia

X-linked agammaglobulinemia (XLA) is an inborn error of immunity caused by pathogenic variants in Bruton’s tyrosine kinase (BTK). BTK plays an important role in B-cell differentiation, proliferation and survival, and XLA patients display a B-cell differentiation arrest at the pro-B cell to pre-B cell transition.1 Several XLA patients have been diagnosed with malignancies in three large cohorts (summarized in Table SI).2-4 It is biologically plausible that XLA patients may have higher rates of acute leukaemia. In genetic models, BTK was found to function as a tumour suppressor: genetic ablation of Btk in Blnk (Slp65) knockout mice leads to an accelerated development of pre-B-cell leukaemia.5 The prevalence of XLA with acute leukaemia is approximately 2 per 300 XLA patients in Japan. The epidemiological data, including previous reports, is shown in Data S1. Furthermore, in a large cohort of chronic lymphoid leukaemia patients receiving BTK inhibitors, the occurrence of secondary cancers, including haematological malignancies, were observed.6 We previously reported two XLA patients with B-cell precursor acute lymphoblastic leukaemia (BCP-ALL).7, 8 In order to identify the contribution of the germline BTK mutation in malignant predisposition, we here analyzed the leukaemic cells for somatic mutations, copy number alterations and the DNA methylation profile using next-generation sequencing in these two patients and one newly identified XLA patient with acute megakaryoblastic leukaemia (AMKL). The clinical and laboratory data of the three patients are described in Table SII, Figure S1 and Data S1. First, we analyzed non-synonymous somatic gene variants of leukaemic cells. Following whole-exome sequencing (WES) on paired tumour and genomic DNA samples, 13, 7 and 5 somatic non-synonymous variants were identified in patients 1, 2 and 3 respectively (Table I). A heterozygous missense variant of KMT2D/MLL2 (c.8740delC, p.H2914Tfs*29) (variant allele frequency (VAF): 0·47) and a heterozygous splice site variant of CDKN2A (c.193+5G>A) (VAF: 0·36) were identified as pathologic variants in patient 1. The KMT2D variant has been reported as an inactive and recurrent variant in non-Hodgkin lymphomas.9 A CDKN2A splice site variant was reported as pathogenic in hereditary cancer-predisposing syndrome.10 In patient 2, heterozygous missense variants of KMT2C (c.2294A>G, p.E765G) (VAF: 0·15), TP53 (c.818G>A, p.R273H) (VAF: 0·81) and TCF3 (c.359T>C, p.L120P) (VAF: 0·31) were identified. The variants of KMT2C and KMT2D were reported as oncogenic abnormalities of histone methylation.11 TP53 is one of the most common tumour suppressor genes. R273H variant of TP53 has been reported as a pathogenic dominant negative variant.12 TCF3 plays a crucial role in lymphocyte development and a case of severe hypogammaglobulinemia and BCP-ALL with TCF3 variant has been reported.13 Heterozygous missense variants of FOXM1 (c.1208G>A, p.R403H) (VAF: 0·03) and NSD1 (c.6656G>A, p.R2219H) (VAF: 0·17) were identified in patient 3. FOXM1 is a transcription factor associated with cell cycle and proliferation.14 NSD1 variants were also reported as oncogenic alterations of histone methylation.11 All somatic gene variants, except for TCF3, were involved in a signaling pathway, epigenetic regulation and tumour suppression, but not in cell maturation. All patients acquired somatic loss of function (LOF) of tumour suppressors (CDKN2A and TP53 variants and monosomy 7). To evaluate if the somatic variant of TCF3 (L120P) was potentially involved in leukaemogenesis, functional analysis of target DNA binding was performed with a luciferase assay. Although a previously identified disease control variant (E555K) showed lower luciferase activity compared with WT (P < 0·01), the L120P mutant was as efficient in luciferase activity as the WT (P = 0·07) (Figure S2). Thus, the L120P does not appear to result in LOF of TCF3. We further analyzed somatic copy number variants of leukaemic cells of patients 1 and 2. Somatic amplification of ERG in 21q22.2 was detected as a pathogenic copy number variant in patient 1 (Figure S3A). ERG amplification was reported as a pathogenic event of various acute leukaemias.15 Consistent with the finding of the G-banding karyotype, somatic amplifications were observed of chromosomes 1, 6, 10, 11, 12, 14, 18, 19, 21, 22, X and Y in patient 2 (Table SII and Figure S3B). DNA methylation profiles were analyzed for patients 1 and 2. BCP-ALL of the XLA patients did not show a distinctive methylation profile. No pathogenic hypermethylation status (e.g. CDKN2A and CASP8AP2) was identified in BCP-ALL of the XLA patients (Figure S4). In recent years, leukaemogenic abnormalities have been reported in five classes of proteins: signaling pathway components, transcription factors, epigenetic regulators, tumour suppressors and components of the spliceosome.16 Somatic BTK mutations of leukaemia were registered in the COSMIC database (https://cancer.sanger.ac.uk/cosmic). Furthermore, the aberrant BTK splice variants were detected in BCP-ALL without BTK mutations and encoded a truncated BTK kinase domain.17 Reconstitution of functional BTK expression induced B-cell differentiation and apoptosis in these leukaemic cells. BTK mutations, which impair the pre-BCR signaling pathway in B cells and GM-CSF-mediated differentiation in myeloid progenitor cells, might induce maturation arrest of pre-B cells in patients 1 and 2 and megakaryoblasts in patient 3 respectively.18 Moreover, BTK mutation impairs the signaling pathways for proliferation. Therefore, we presume that somatic driver mutations, as second hits, promote leukaemogenesis in XLA patients. As previously described, Btk/p53-deficient mice showed an increased proliferative capacity of immature B cells.19 In our study, leukaemic cells of all patients acquired LOF variants in tumour suppressors, suggesting that germline BTK mutation-induced maturation arrest as a first hit and LOF variants of tumour suppressors as a second hit, ultimately leading to evolution to acute leukaemia. The clonal evolution models of the XLA patients are shown in Fig 1. These findings would be important for BTK inhibitor treatment. We suggest the need for closer and more systematic follow-up for malignancies in XLA patients. From the perspective of precision medicine, XLA patients who acquire somatic variants may be candidates for haematopoietic cell transplantation as a pre-emptive therapy. The limitation of this study is that we evaluated only three patients and did not perform expression analysis and mutation analysis of non-coding regions because of limited samples. Future studies in which many patients with XLA with leukemia are enrolled may be able to clarify the more detailed characteristics of leukaemic cells in XLA. The authors thank the patients and families of patients. The authors also thank Ms. Mika Nagase, Mr. Wataru Iguchi, and Ms. Naomi Terada-Takahashi for their technical assistance. This work was supported by MEXT/JSPS KAKENHI (Grant Number: JP17K10099) to HK, and AMED practical Research for Innovative Cancer Control (19ck0106467h0001) to TM. Akira Nishimura, Menno C. van Zelm and Hirokazu Kanegane designed the study and wrote the manuscript. Akira Nishimura, Takuya Naruto, Akihiro Hoshino and Osamu Ohara performed genomics analysis. Akira Nishimura and Satoshi Miyamoto performed luciferase assay. Andrew Grigg, Julian J. Bosco, Keishiro Amano, Shotaro Iwamoto, Masahiro Hirayama and Masahiro Migita provided clinical care. Masatoshi Takagi and Tomohiro Morio provided critical discussion. Data S1. Supplementary Data. Table SI. Literature review of haematological malignancies with XLA. Table SII. Clinical data of the XLA 3 patients with acute leukaemia. Fig S1. Pathological and immunophenotypic findings of patient 3. Fig S2. Luciferase assay of TCF3. Fig S3. Copy number analysis of somatic variants. Fig S4. DNA methylation analysis. 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.