New insights into the molecular pathogenesis of T-cell lymphomas

T-cell lymphomas are highly heterogenous diseases, encompassing over 30 different subsets according to the World Health Organization (WHO)¹. Peripheral T-cell lymphomas, not otherwise specified (PTCL-NOS) along with angioimmunoblastic T‐cell lymphoma (AITL) are the most frequent of entities of mature T‐ and Natural Killer (NK)‐cell neoplasms.² Chemotherapy, autologous stem cell transplantation (ASCT), and allogeneic stem cell transplantation (alloSCT), remain the main treatment options due to a lack of effective, more targeted therapies. Therefore, there is an unmet medical need to develop novel treatments to improve patient outcomes and reduce treatment-related toxicity. Understanding of the disease biology, as well as knowledge of differences and similarities between the entities, could guide the drug development process.

Over the last decade, whole-genome studies revealed that PTCLs originate from subsets of helper T (Th) cells and oncogenic events. However, the role of the tumor environment, immune evasion and genetic susceptibility remain to be fully elucidated. At the 2019 annual meeting of Society of Hematologic Oncology (SOHO), Houston, US, Professor Philippe Gaulard of Hôpital Henri Mondor, Université Paris-Est, Créteil, FR, delivered a presentation on the new insights into the molecular pathogenesis of T-cell lymphomas.³ 

Follicular helper T cell peripheral T-cell lymphoma (TFH-PTCL)

In 2017 WHO recognised nodal PTCL derived from follicular helper T cells (TFH) (TFH-PTCL) in its classification. Between 20% and 40% of PTCL-NOS share features with AITL, including expression of TFH markers, TFH gene signature, mutational landscape and genomic alterations.⁴

The mutational landscape of TFH-PTCL (summarized in Table 1) is unique, with aberrations in epigenetic regulators (>80%), signalling, and cell cycle alterations. Interestingly, DNA and histone hypermethylation are not mutually exclusive. Aberrations in T-cell receptor signalling pathways, mainly activating mutations and fusions are common in TFH-PTCL, but also in other PTCLs. Mutations in RhoA are relatively specific to TFH-PTCL.

Angioimmunoblastic T‐cell lymphoma

Based on molecular and animal model studies, a model of TFH lymphomagenesis has been established. The model indicates initial mutations in TET2 and DNMT3A genes of early progenitor cells, followed by RHOA^G17V expression which leads to the development of the AITL.

A role of metabolism has also been suggested since 20–40% of AITL tumors have a gain of function mutation in IDH2 which increases levels of D2-hydroxyglutarate, an oncometabolite.⁸ Additionally, preclinical studies have shown that GAPDH overexpression in T cells promotes AITL via an NF-κβ pathway and analysis of AITL tumors have confirmed GAPDH induction of the pathway.⁹ Professor Gaulard highlighted that this mouse model could be used in future efficacy tests of new drugs for patients with AITL.

Anaplastic large cell lymphoma (ALCL)

ALCL, which was initially described as a unique disease, encompasses several distinct entities with different molecular characteristics and clinical presentations:

• Systemic ALCL ALK-positive (sALCL AKT+)
• Systemic ALCL ALK-negative (sALCL AKT-)
• Cutaneous ALCL
• Breast ALCL (BI-ALCL)

Other forms of the disease are characterized by fusions in IRF4/DUSP22, TP63, VAV1, and tyrosine kinases, as well as overexpression of ERBB4 and mutations in JAK1/STAT3 pathways.

Systemic ALK+ ALCL is characterized by ALK rearrangement. In ALK+ ALCL, Wiskott-Aldrich syndrome protein (WASP) and WASP-interacting (WIP) proteins have been found to associate with STAT3, resulting in reduced expression.¹⁰ In NPM-ALK transgenic mice, WASP /WIP deficiencies accelerate the development of lymphoma and activate CDC42 and MAPK, which could provide therapeutic vulnerability to MEK inhibition.

A distinct subset of sALCL ALK- was also identified with DUSP22 rearrangement and a distinct gene signature, with DNA hypomethylation, expression of cancer-testis (CT) antigens and lack of STAT3 activation.¹¹ This immunogenic phenotype may contribute to more favorable patient prognosis. Out of this subset, 35% of cases have MSC^E116K mutation and were shown to have a dominant-negative bHLH transcription factor. As the mutation drives cell cycle progression via the CD30/IRF4 activation of MYC, it could potentially be targeted with BET inhibitors.¹² Moreover, genetic alterations in sALCL were also demonstrated to have potential use as prognostic factors^13,14 and could be used to guide the selection of optimal therapy. This demonstrates how molecular genetics determines the susceptibility of different subsets of disease to treatment, depending on the abnormalities which are present.

A rare subset of ALCL, with a long latency period, is associated with macro-textured breast implants, which caused a recall earlier this year by the U.S Food and Drug Agency (FDA). Breast implant associated-ALCL (BI-ALCL) is ALK- and has two forms: in situ seroma and infiltrative. The molecular analysis of 34 BI-ALCL tumors revealed mutations in epigenetic modifiers (74% of tumors), genes involved in JAK/STAT signalling (59%), cell cycle or apoptosis (26%) and PI3K/AKT/mTOR (6%).¹⁵ The current model of lymphomagenesis assumes that a local chronic inflammation caused by immunogenicity of the implant itself, bacteria, auto-antigens or host susceptibility leads to the polyclonal T cell proliferation. This is then followed by proliferation of oligoclonal T cells expressing PDL1, leading to immune escape of cells, followed by JAK/STAT activation (KMT2C, KMT2D, CHD2), and epigenetic deregulation (STAT3, JAK1, STAT5B, SOCS3, SOCS1) culminating in BI-ALCL.^16,17

Extranodal PTCL

Different cells of the innate immune system can be involved in the formation of PTCL. Those include NK cells, Tγδ, and Tδβ cells. The extranodal subsets of the disease include:

• Extranodal NK/T cell lymphoma, nasal type (ENKTCL)
• Enteropathy-associated T-cell lymphoma (EATL)
• Monomorphic epitheliotropic intestinal T-cell lymphoma (MEITL)
• Hepatosplenic T-cell lymphoma (HSTL)

They all share extranodal site presentation and are characterized by cytotoxic activity, chronic antigenic stimulation, and possibly a genetic susceptibility. Additionally, they are associated with constitutive activation of the JAK-STAT3 pathway via activating mutations or fusions, that could potentially be targeted.

In recent years inactivating mutations of SETD2 methyltransferase were also found to be implicated in around 90% of MEITL cases, 40% of HSTL, and less frequently in EATL. The inactivation is often linked with aggressive disease, Tγδ derivation, and could potentially be targeted by WEE1 inhibitors.^18,19

Cell counterpart

In Professor Gaulard’s opinion, the concept of cell counterpart should be expanded. He highlighted again that PTCL-NOS is more than one disease and recent gene expression profiling studies revealed distinct molecular signatures including some with activation of PI3K-AKT pathway, others with APMK-mTOR, CDKN2A-TP53, or PTEN-PI3K axis as oncogenic drivers.²⁰

Immune evasion

It was suggested that the role of immune evasion in PTCL should be further explored. PD-1 and PD-L1 are overexpressed by many subsets of PTCL and could potentially be a therapeutic target. In particular, PD-L1, which is associated with immune surveillance, is overexpressed in the majority of ENKTCLs and 20% of adult T-cell lymphoma (ATLL), as well as in many reactive cells including T cells, B cells and macrophages. The overexpression by the neoplastic cells could be caused by oncogenic activation via NPM-ALK pathway, STAT3 activation, be induced by viruses such as Epstein-Barr virus (EBV) or genetic alterations.^21-23

PD-1 is overexpressed on TFH-PTCL and other tumor cells and its role is less well understood. It can act as a tumor suppressor counteracting TCR activation. But on the other hand, it has been found to be inactivated in ATLL and CTCL. Moreover, PD-1 inhibitors have been shown to be able to accelerate PTCL. Therefore, more studies are needed to establish the role of immunotherapy in PTCL.^24–26

Conclusions

Various events and pathways are involved in PTCL oncogenesis including changes in epigenetic regulation, abnormal signalling and cell cycle regulation, as well as compromised immune surveillance. Some of those processes are shared among all PTCL entities, while others are more specific to a particular subset. This information could potentially lead to improved diagnosis and treatment. The final message from Professor Gaulard was that PTCL is a very heterogenous disease and that biology should play a key role in designing efficient targeted therapies.