Genetic background of relapsing lymphoid neoplasms : comprehensive genetic characterization of primary - relapse pairs by chromosomal copy number and mutational analysis

Diffuse large B-cell lymphoma (DLBCL) is the most frequent tumor of the lymphoid system.
Standard first line therapy is successful in the majority of cases, however about 30% of DLBCL
are either refractory to it or relapse after a period of remission. Therapy options for such patients
are very limited and disease outcomes are often unfavorable. Recent years of research have
considerably advanced our understanding of molecular processes that drive lymphomagenesis of
primary DLBCL. However, due to lack of focused studies, knowledge on the genetic basis of
DLBCL relapse remains scarce. It is largely unknown how tumors evolve under treatment, which
genetic events lead to recurrence, how resistance emerges and whether relapses can be reliably
predicted at the timepoint of initial diagnosis.
We sought to address these questions by comprehensive genetic analysis of two retrospective
DLBCL cohorts. One consisted of paired primary and relapse samples from 20 relapsing
DLBCL cases, another of 11 non-relapsing primary DLBCL samples. We performed histological
characterization, investigated genome-wide DNA copy number aberrations and most common
nucleotide-level alterations, and tested clonal relationships between paired tumors of the same
patient. Our study provided important novel findings and strengthened some previous
observations.
Clonally-unrelated DLBCL relapses were previously equivocally documented by demonstration
of distinct immunoglobuln gene (IG) rearrangements between the first and the second lymphoma
occurrence. We support this observation and provide high quality evidence showing not only
distinct IG rearrangements, but also lack of unifying genetic alterations throughout the whole
tumors’ genomes of clonally-unrelated relapses. Therefore we propose to call such
manifestations second de novo DLBCL occurring in the same individual.
Genetic evolution of DLBCL at relapse is largely unexplored except of a single study, but much
is assumed from results and conclusions in other lymphoid neoplasms. Analysis of th e dynamics
of shared and private alterations in paired tumors suggested existence of at least two distinct
patterns of genome evolution at relapse. About one third of investigated clonally-related cases
followed an early-divergent/branching pattern of a relapse, characterized by a small number of
shared mutations between the two occurrences and a large number of primary tumor-specific
alterations. The rest of the cases relapsed via a late-divergent/linear pattern. They were
characterized by sharing most mutations between the respective primaries and relapses and very
small numbers, if any, of primary tumor-specific genetic alterations.
The identification of precise genetic drivers and prognostic factors of relapse is challenged by an
extreme heterogeneity of DLBCL. Together with the small sample sizes of analyzed collectives,
this is probably the main reason why there is only a limited overlap in findings between studies,
addressing this issue up to date. Our analysis identified two genes, KMT2D and MEF2B, which
recurrently gained additional mutations at relapse. Further, we identified recurrently shared
alterations of CD79B, KMT2D, MYD88, PIM1, which represented early drivers of
lymphomagenesis. Finally, comparing genetic data between relapsing and non-relapsing DLBCL
cohorts we identified group-specific DNA lesions (recurrent gains of 10p15.3-13 containing
GATA3 and PRKCQ in primary relapsing DLBCL and mutations of SOCS1 and RELN in nonrelapsing DLBCL). The prognostic power of them could be further investigated in dedicated
large-scale studies.
We are committed in further analyzing paired lymphoma samples, also in difficult settings, as
demonstrated by two small-scale studies included in this dissertation. We developed and applied
a technique to enrich rare lymphoma cells from formalin-fixed paraffin-embedded tissues. It
enables meaningful genetic analysis of relapses in lymphoid neoplasms, such as classical
Hodgkin lymphoma, which are important but understudied, due to their research-limiting
morphological properties.