Recurrent Mutations in Epigenetic Modifiers and the PI3K/AKT/mTOR Pathway in Subcutaneous Panniculitis‐like T‐cell Lymphoma
Zhaoming Li,Lisha Li,Zhiyuan Zhou,Weili Xue,Yingjun Wang,Mengyuan Jin,Yajuan Qiu,Wei Sun,Xuefei Fu,Xudong Zhang,Yu Chang,Feifei Nan,Jiaqin Yan,Guannan Wang,Zhihua Sun,Xiaorui Fu,Ling Li,Xin Li,Xinhua Wang,Jingjing Wu,Lei Zhang,Mingzhi Zhang
DOI: https://doi.org/10.1111/bjh.14611
2017-01-01
British Journal of Haematology
Abstract:Subcutaneous panniculitis-like T-cell lymphoma (SPTCL) is a rare form of non-Hodgkin lymphoma accounting for approximately 1% of all primary cutaneous lymphomas (Willemze et al, 2005). To date, little is known about the pathogenesis of this lymphoma, partly due to its rarity. To obtain a genome-wide view of the molecular genetic alterations underlying SPTCL pathogenesis, we performed whole-exome sequencing (WES) and targeted sequencing for a series of SPTCL cases. Eighteen SPTCL cases with available formalin-fixed paraffin-embedded (FFPE) tumour tissue from four cancer centres between 2005 and 2016 were collected (see Supplementary Information). The demographics, clinical and immunohistochemical features of the patient cohort are summarized in Table S1. To identify somatic genomic variants associated with SPTCL, we first performed WES on 10 tumour samples and 3 matched whole blood samples from our cancer centre. The mean sequencing depth was 140·99× (ranging from 109·4 to 193·2×), and a mean of 98·10% of the target sequence covered a depth of at least 10× (Table S2). The mutation frequencies in the 10 patients with SPTCL varied remarkably, with a median of 345 somatic non-synonymous mutations per case (ranging from 104 to 2383, Figure S1A). We observed a preference for C>T/G>A alterations analogous to the somatic single-nucleotide variation (SNV) spectrum in other cancers (Figure S1B). Forty SNVs/Indels (insertion-deletions) were selected based on the variant frequency for validation by Sanger sequencing, and 92·5% of the variants were confirmed (Table S3). We identified substantial heterogeneity in the SPTCL somatic mutation landscape. There were 7032 somatic non-synonymous mutations affecting 4714 genes identified in the 10 SPTCL subjects. Interestingly, most common mutations described previously in other T-cell lymphomas (Forbes et al, 2015), such as TP53, DNMT3A, PLCG1, STAT3 and CDKN2A, were not frequently observed in SPTCL. To efficiently identify the potential driver mutations contributing to SPTCL pathogenesis, we selected genes included in the Cancer Gene Census (Futreal et al, 2004) and/or in the lists of cancer genes reported by Kandoth et al (2013); Vogelstein et al (2013) and Tamborero et al (2013) for further analysis. We detected 284 non-synonymous mutations affecting 202 genes in the 10 SPTCL patients (Table S4). Pathway analysis of these 202 mutated genes revealed that there was a significant enrichment of genes involved in the phosphoinositide 3-kinase (PI3K)/AKT/mechanistic/mammalian target of rapamycin (mTOR) and JAK-STAT signalling pathways (P < 0·001; Table S5). We further applied a comprehensive targeted sequencing approach to validate the above findings. The panel, comprising 560 genes selected for targeted sequencing, is listed in Table S6. Tumour and matched normal DNA samples extracted from 8 independent SPTCL cases in three other cancer centres (Data S1) were subjected to targeted sequencing. The mean sequencing depth was 674× (ranging from 505–1044×); 21 non-synonymous mutations were identified (Tables S7, S8). Collectively, mutations in epigenetic modifiers were found in 13 of 18 (72%) SPTCL cases (Fig 1). These modifiers are involved in the epigenetic machinery at almost every level, including key players in chromatin organization (ARID1B, 3/18; SMARCA4, 3/18; and CHD4, 3/18), DNA methylation (MBD1, 1/18), and histone modification (CREBBP, 2/18; KMT2D, 2/18; and DOT1L, 2/18). Recurrent mutations were identified in genes encoding components of the PI3K/AKT/mTOR signaling pathway in 8 of 18 (44%) SPTCL cases. Within the PI3K/AKT/mTOR cascade, most frequent mutations were noted in the MTOR and TSC1 genes (3/18 cases), followed by PIK3CA (2/18), PIK3CD (2/18) and PIK3CB, TSC2 and AKT2 (1/18 each) (Fig 1). These gene mutations have not been previously reported, except for the MTOR (p.A1124V) mutations, which have been reported in the Catalogue of Somatic Mutations in Cancer database (COSMIC, http://cancer.sanger.ac.uk/cosmic). Immunohistochemical staining showed that tumours with PI3K/AKT/mTOR mutations exhibited higher AKT1 and 4E-BP1 (also termed EIF4BP1) phosphorylation levels than that of tumours without the mutations (P < 0·05; Figure S2), suggesting that these gene mutations might participate in the activation of PI3K signalling. In addition, the mutations in the PI3K/AKT/mTOR cascade appeared to be mutually exclusive despite the low number of patients. Dysregulation of the JAK-STAT pathway is commonly implicated in the pathogenesis of T-cell lymphomas (Kiel et al, 2015). In this study, we identified somatic mutations affecting JAK3 (3/18), STAT3 (1/18) and IL7R (1/18), presenting in 28% (5 of 18) of our cases (Fig 1). We identified three novel mutations (p.V152M, p.G24S and p.G153D) in JAK3, an activating D661Y hotspot mutation in STAT3, and a p.T179M mutation in IL7R. All of these mutations appeared to be mutually exclusive. In our study, TP53 mutation was identified in only one SPTCL case, which is less frequent than that of other T-cell lymphomas. NAV3 (neuron navigator 3) deletion has been reported in approximately 40% of STCL samples (Hahtola et al, 2008). We also detected NAV3 mutations in two cases of SPTCL. These results therefore implicate that NAV3 might play a vital role in SPTCL pathogenesis, but further investigation is needed. To test the therapeutic potential of targeting the PI3K/AKT/mTOR pathway, one primary SPTCL sample (sample SP1) harbouring mutations in MTOR was subjected to PI3K and/or mTOR inhibition. Immunohistochemical analysis indicated that the PI3K/AKT/mTOR signalling pathway was activated (Fig 2A). Tumour cells (CD4−, CD8+ and CD56−) were enriched and purified from the primary SPTCL sample with magnetic beads, and the purity was approximately 85% (Fig 2B). Two inhibitors targeting mTOR (rapamycin and everolimus) and one inhibitor targeting mTOR and PI3K (apitolisib) reduced cell viability in a dose-dependent manner (Fig 2C). Moreover, these three inhibitors could inhibit the downstream signalling cascades by suppressing p70S6K, 4E-BP1 and AKT phosphorylation (Fig 2D). Primary SPTCL cells were resistant to an inhibitor for a non-relevant pathway, the NF-κB inhibitor BAY 11-7082; therefore, the observed effects were probably not attributable to general cytotoxicity. The primary tumour cells from SPTCL samples also exhibited high sensitivity to the histone deacetylase (HDAC) inhibitor romidepsin (Fig 2C), this is consistent with an earlier report that romidepsin was effective in SPTCL patients (Bashey et al, 2012). In summary, pharmacological targeting of PI3K/AKT/mTOR or HDAC signalling could inhibit the growth of primary SPTCL tumour cells. In conclusion, to the best of our knowledge this is the first study to comprehensively characterize the genome-wide mutational landscape of SPTCL. We identified frequent gene mutations in epigenetic modifiers and the PI3K/AKT/mTOR pathway in SPTCL. Targeting these two pathways may represent a promising therapeutic strategy in SPTCL. This study was supported by the Youth Innovation Funds Project of The First Affiliated Hospital of Zhengzhou University. The authors would like to thank Xinfeng Chen and Shuai Liu for their excellent technical assistance, and Virender Sachdeva, M.S. from Department of Paediatric Ophthalmology of Nimmagada, Prasad Children's Eye Care Centre, L V Prasad Eye Institute, GMRV Campus, India and Tian Tian from the Department of Neurology, the First Affiliated Hospital of Zhengzhou University, China for language editing. Z.L., L.L., and M.Z. designed the study, interpreted results, and wrote the manuscript; Z.L., L.L., Y.W., W.X., Z.Z., M.J., and Y.Q. performed the experiments; W.S. and X.F. performed the bioinformatics analysis; Y.C., F.N., J.Y., G.W., Z.S., X.F., L.L., X.L., X.W., J.W., and L.Z. provided samples and clinical data. Authors declare no conflict of interest. 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.