Toward a Molecular Classification of Peripheral T-Cell Lymphomas: The Role of Gene Expression Profiling

Peripheral T-cell lymphomas-not otherwise specified (PTCL/NOS) are the most common T-cell neoplasms. This study sought to reshape the PTCL/NOS sub-classification (including its two main morphological variants, Lennert lymphoma, LL, and Follicular variant, F-PTCL) based on the correspondence between their molecular features and those of different functional T-cell subsets, also assessing the clinical impact of such an approach. We found that PTCLs/NOS could be divided into groups corresponding to T-cell subsets differently reliant on transcription regulators including mTOR and FOXP3, and identified minimal gene sets discriminating among these groups. Notably, by grouping tumors according to their dependency on master regulators of T-lymphocyte fate, we identified three groups (T-cytotoxic, Treg/TFH, and other-T-helper) characterized by specific genetic patterns and significantly different clinical outcomes. Immunohistochemistry partially substituted for the molecular analysis by consistently recognizing only Treg and TFH cases. Finally, targeted inhibition of MTOR in T-helper cases (that were characterized by genetic lesions targeting the pathway) was proved to be effective ex vivo. We conclude that PTCL/NOS can be divided into subgroups corresponding to different cellular counterparts, characterized by different genetic patterns and possibly sensitivity to specific therapeutic approaches. Furthermore, we identified different gene and microRNA signatures for LL capable of differentiating it from other PTCL/NOS and enriched in cytotoxic function. Moreover, PI3K/Akt/mTOR pathway emerged as novel therapeutic targets for LL. Additionally, LL showed some differences with other PTCL/NOS in terms of clinical features, all supporting its recognition as a distinct entity. Besides, we found that F-PTCL has a distinct molecular signature more similar to PTCL/NOS rather than AITL, and therefore cannot be included among AITLs at least based on GEP, although this necessities more genetic studies. Overall, these results may impact on PTCL classification as well as on future studies aimed to define the more appropriate therapeutic strategy for each identified subgroup/entity.