Graduation Date

Fall 12-18-2020

Document Type


Degree Name

Doctor of Philosophy (PhD)


Pathology & Microbiology

First Advisor

Kai Fu


Diffuse large B-cell lymphoma (DLBCL) is an aggressive lymphoid malignancy. About 30% of DLBCL cases respond poorly to initial treatment and eventually relapse. For these patients, the current treatment regimen is quite limited, and the prognosis is poor. Gene mutations and genetic alterations play an important role in lymphomagenesis. However, the genetic alterations or gene mutations underlying the disease resistance/relapse in DLBCL are still unknown. The clonal evolution during the process of disease progression is elusive as well. Our goal is to study the genetic alterations in DLBCL, particularly paired diagnostic and relapsed/refractory DLBCL, to better understand the mutation landscape of relapsed/refractory DLBCL and identify potential novel targets for effective therapy.

We performed whole-exome sequencing (WES) on 16 pairs of diagnostic and relapsed/refractory DLBCL samples. In addition, the WES data of seven diagnostic and relapsed/refractory pairs were downloaded, as well as another 30 relapsed cases, each with a paired germline. In total, 23 pairs were analyzed all together to discover the mutation landscape for relapsed/refractory DLBCL. These results were validated in the 30 relapsed cases when needed. We classified our DLBCL cases into MCD, BN2, N1, and EZB subtypes according to the recently developed classification based on genomic abnormalities. The MCD subgroup was found to be enriched in our relapsed/refractory cohort, suggesting the MCD subgroup was prone to relapse. Consistently, MYD88 mutations, a hallmark for the MCD subgroup, were enriched in our relapsed/refractory cohort as well. Moreover, two different evolution patterns (Pattern 1 and 2) were revealed. Pattern 1 cases carried more original mutations to relapse, whereas Pattern 2 cases had a considerable number of novel variants only present in the relapsed/refractory sample. Interestingly, Pattern 2 was correlated to worse prognoses, such as shorter relapse-free survival. With further validation, these data can assist in precise clinical decisions; for example, patients with Pattern 2 may have worse prognoses predictions and thus be subjected to more intense therapies. Additionally, we discovered that the canonical NF-κB signaling pathway mutations probably play a critical role in DLBCL relapse. The novel potential target genes in the NF-κB pathway included SHARPIN, PTPN14, and TMSB4X. Finally, we identified a rare fusion between MYC and its non-IG partner Activation-induced Cytidine Deaminase (AICDA) in our relapsed/refractory cohort by RNA-sequencing. Notably, this de novo fusion significantly evolved after treatment, indicating it was resistant to the R-CHOP (rituximab, cyclophosphamide, doxorubicin hydrochloride, vincristine, and prednisolone) immunochemotherapy and expanded over the course of treatment. These results suggest the AICDA-MYC fusion is a potential prognostic marker and may mediate DLBCL relapse.

Available for download on Saturday, August 28, 2021