New Recurrently Mutated Genes in Classical Hodgkin Lymphoma Revealed By Whole-Exome Sequencing of Microdissected Tumor Cells

Abstract

Abstract Dissecting the genetics of classical Hodgkin lymphoma (cHL), a common lymphoma derived from germinal center B (GCB) cells, has remained a difficult task compared to other lymphomas, due to the rarity (usually <5%) of tumor cells (the Hodgkin and Reed/Sternberg - HRS - cells) in the lymph node. Thus, we microdissected 1200-1800 single HRS cells per case from frozen biopsies of 36 patients, along with a similar number of adjacent non-neoplastic cells, and performed in duplicate whole-genome DNA amplification (WGA) followed by independent whole-exome sequencing, thus allowing to control for possible biases introduced by the WGA step. The putative somatic protein-changing mutations identified in a major tumor clone (i.e., ≥20% variant allele frequency) were then validated by targeted deeper sequencing of the same WGA DNA, as well as by ultra-deep sequencing and digital PCR of unamplified whole-biopsy DNA. We found highly recurrent (11/36 patients; 31%), heterozygous, STAT6 missense mutations in exons encoding for the DNA binding domain, identical to the activating STAT6 variants known in primary mediastinal B-cell lymphoma (PMBCL), diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma. ShRNA-induced STAT6 knockdown caused marked apoptosis in the STAT6-mutated cHL cell line L1236, which expresses phosphorylated STAT6 (pSTAT6), while it showed little or no effect in the STAT6-unmutated/pSTAT6+ cHL lines L428 and HDLM2, and in the STAT6-unmutated/STAT6-unphosphorylated cHL line L540. Thus, STAT6 mutations in cHL appear to confer a survival advantage distinct from (and beyond that of) STAT6 phosphorylation, possibly due to their aberrant DNA binding activity following pSTAT6-dependent entry into the nucleus. Interestingly, STAT6 mutations were often associated with disruptive somatic mutations of the JAK-STAT signaling inhibitor SOCS1 (frequency in the evaluable STAT6-mutated versus unmutated patients: 6/7, i.e. 87%, versus 4/14, i.e. 29%; p=0.02), suggesting a genetic interaction. Indeed, transduction of wild-type SOCS1 induced apoptosis of STAT6-mutated L1236 cells to a much greater extent than STAT6-unmutated L428 cells, despite both carry a SOCS1disruptive mutation and express pSTAT6. Furthermore, the presence in additional cases of genetic lesions of other JAK-STAT pathway members (mutations activating STAT3, STA5B, JAK1 or inactivating the inhibitor PTPN1; JAK2 copy number gains) attests to the pervasive role of JAK-STAT signaling in the genetics of cHL (up to ~70% of patients affected). GNA13, encoding Gα13, carried mutations, mostly heterozygous, in 8/36 (22%) patients. By transmitting signals from the S1PR2 and P2RY8 receptors, Gα13 ensures the confinement of proliferating GCB cells within secondary lymphoid follicles, while constraining at the same time their expansion by facilitating apoptosis in this potentially dangerous niche (Nature 2014;516:254). GNA13 mutations in cHL included disruptive (frame-shifting and non-sense) variants and were similar to those identified in GC B cell-derived Burkitt lymphoma and DLBCL of the GCB-type. The apparent histogenetic selectivity of inactivating GNA13mutations, almost never found in post-germinal center ABC-DLBCL, suggests they may facilitate cHL lymphomagenesis by rescuing from apoptosis crippled GC B cells (the proposed HRS cell precursors) and by promoting their dissemination outside the follicles, where HRS cells are indeed observed in lymph nodes with early, partial involvement by cHL (Am J Surg Pathol 1983;7:145). Heterozygous missense mutations of XPO1 occurred in 6/36 patients (17%) at the hot-spot amino acid residue E571, which is recurrently targeted also in chronic lymphocytic leukemia and PMBCL. XPO1encodes exportin-1, which shuttles outside the nucleus more than 200 proteins, including tumor suppressors involved in cHL pathogenesis and functioning in the nucleus (e.g., NFKBIA, FOXOs, TP53) (Blood 2012;120:4621). XPO1 inhibition by the clinical compound selinexor caused growth inhibition and apoptosis of cHL cell lines in a mutation-dependent manner. By providing a global characterization of the cHL coding genome, this works uncovered several new genes implicated in the genetics of this lymphoma, and most prominently in the JAK-STAT pathway, that likely play a significant role in its biology. #Shared first authorship °Shared last authorship Disclosures No relevant conflicts of interest to declare.