TET2 Deficiency Sets the Stage for B-cell Lymphoma

The TET family of enzymes (TET1, TET2, and TET3) plays a critical role in DNA demethylation by converting 5-methylcytosine (5mC) into 5-hydroxymethylcytosine (5hmC), a precursor step. Loss of TET2 has been shown to lead to aberrant gene expression that can disrupt normal differentiation processes. Work in knockout mouse models has shown that TET2 loss enhances hematopoietic stem cell (HSC) self-renewal, expanding the HSC pool (reviewed in refs. 1 and 2). An enlarged HSC pool increases the potential for successive genetic events that drive development of diverse hematopoietic malignancies. The type of malignancy may depend on the identity of these later genetic events. TET2-knockout mice predominantly develop chronic myelomonocytic leukemia (CMML), but they can also develop other myeloproliferative as well as lymphoproliferative diseases, reflecting the prevalence of TET2 mutations in these varied hematologic malignancies.

Significant work has been performed investigating the effects of TET2 loss in myeloid diseases. Frameshift and nonsense mutations occur frequently in CMML, acute myeloid leukemia (3), and myeloproliferative neoplasm (MPN) cases (4). TET2 mutations have been associated with reduced patient survival in acute myeloid leukemia. Mutations have also been found in the blood of elderly patients without hematopoietic cancers, indicating that TET2 loss alone is not sufficient to trigger leukemic transformation (reviewed in ref. 5). Nevertheless, these patients manifest a condition termed clonal hematopoiesis of indeterminate potential (CHIP), referring to the clonally expanded HSC pool, and are at increased risk for developing a hematologic malignancy.

The role of TET2 and cytosine hydroxymethylation in B-cell lymphomagenesis is less known. Genomic studies have uncovered silencing TET2 mutations in B-cell and T-cell lymphomas (6–8), indicating a tumor suppressor role for TET2 in lymphoid malignancies. In this issue of Cancer Discovery, Dominguez and colleagues report their findings on the mechanistic link between TET2 mutation and germinal center (GC) B-cell transformation (3).

The authors used a conditional TET2 knockout model to demonstrate that loss of TET2 in either the HSC or early B-cell stage led to enrichment of GC B cells. Furthermore, this increase in the GC percentage was due to larger but not greater numbers of GCs. However, this phenotype was not observed when TET2 loss was restricted to the GC B-cell stage; therefore, the phenotype was dependent on TET2 loss occurring earlier in the hematopoietic lineage. This stage-specific phenotype has been observed in other diffuse large B-cell lymphoma (DLBCL) models involving epigenetic tumor suppressors such as KMT2D and CREBBP, suggesting that loss of these epigenetic factors is an early tumorigenic event in lymphomas. Over time, mice with TET2-deficient HSCs displayed splenomegaly and disrupted splenic architecture. A more severe neoplastic phenotype was seen when GC-specific Bcl6 overexpression was combined with TET2 deficiency. This combination led to morbidity, splenomegaly, and mature B-cell lymphoma–like disease with complete penetrance. Bcl6 overexpression alone led to an incompletely penetrant neoplastic phenotype, whereas TET2 loss alone did not lead to lymphoma or splenic effacement.

Further investigation showed that TET2 loss in all three conditional knockout models led to a decrease in IgG levels and a block in plasma cell differentiation. This effect was reproduced in an in vitro system that recapitulates the GC environment, demonstrating that this effect is B-cell autonomous. Examination of gene expression changes in TET2-null GC B cells revealed downregulation of genes known to be involved in the exit of GC B cells from the GC reaction, as well as genes whose enhancers are normally repressed during the GC reaction but are induced upon GC exit. These genes are normally repressed by BCL6 and reactivated by CREBBP. In the in vitro system, the investigators demonstrated aberrant expression of Bcl6 and Prdm1, transcription factors that must be turned off and on, respectively, for GC exit and plasma cell differentiation. The authors concluded that TET2 loss impairs exit from the GC as GC B cells are no longer able to activate the genes necessary for terminal differentiation.

To determine how TET2 alters gene expression, the authors examined the changes in DNA hydroxymethylcytosine (5hmC) marks through hMeDIP-Seq of purified TET2-null or wild-type murine GC B cells. Approximately 24,000 differentially hydroxymethylated regions were found, and gene set enrichment analysis revealed enrichment of gene sets involved in GC exit and plasma cell differentiation including NF-κB pathway genes. Interestingly, there was significant correlation between the genes that lost 5hmC marks in the absence of TET2 and those that lost H3K27ac marks upon CREBBP loss. This correlation is consistent with the authors’ observation that TET2 mutations are mutually exclusive with CREBBP mutations in a cohort of 128 patients with DLBCL and suggests that TET2 and CREBBP have at least partially overlapping mechanisms of transformation. Because the authors had previously shown that CREBBP-deficient lymphomas are addicted to HDAC3 activity, they hypothesized that HDAC3 addiction may apply to TET2-deficient lymphomas as well, and showed that cells were more sensitive to an HDAC3 inhibitor when TET2 expression was diminished by RNAi.

This report significantly advances our knowledge of the mechanism of TET2-driven B-cell transformation, as well as the genetic lesions that can cooperate with TET2 deficiency to drive B-cell lymphoma. In the absence of TET2, GC B cells are blocked from terminal differentiation because they are unable to regulate expression of key sets of genes (Fig. 1A). This model predicts that TET2 mutations would be found more frequently in DLBCL of the GCB-like subtype than the ABC-like subtype, which may be derived from a more mature activated B cell. Consistent with this prediction, we observed more TET2 mutations in GCB-like than in ABC-like DLBCL cases in a cohort of 1,001 patient samples that we sequenced (8).

The data from the series of conditional knockout models in this study demonstrate that TET2 must be lost before cells enter the GC reaction in order to encounter the terminal differentiation block. This conclusion implies that in patients with TET2-deficient B-cell lymphoma, the initial TET2 mutation likely occurs in a cell between the HSC and immature B-cell stages. Prior studies have found TET2-mutated patients who present with identical mutations in both myeloid and lymphoid cells, or who present with more than one hematologic malignancy, indicating that TET2 loss occurred very early in the hematopoietic lineage (Fig. 1B; refs. 6, 9). Presumably, the identity and/or cell state of later cooperating mutations determines the type of malignancy that arises. For example, Dominguez and colleagues showed that GC-specific overexpression of the transcription factor Bcl6 cooperates with TET2 loss to generate mature B-cell lymphomas. In addition, our DLBCL sequencing data revealed that mutations in members of B-cell signaling pathways such as PTEN, PIK3R1, and BCL10 significantly co-occurred with TET2 mutations (8). Further investigation is needed to determine how these genes and others combine with TET2 to give rise to a diverse array of hematologic malignancies.

Many questions surrounding the role of TET2 in hematopoietic malignancies remain outstanding: Why do TET2-deficient HSCs skew toward the myeloid differentiation pathway? Following TET2 loss-of-function in HSCs, are there more “pathways” via cooperating mutations to myeloid malignancies than to lymphoid?

Here, Dominguez and colleagues have presented a compelling story that illustrates the critical role of DNA demethylation in GC biology and B-cell lymphomagenesis. The study adds significantly to our understanding of the diverse roles of TET2 in the hematopoietic lineage.

S.S. Dave is founder of and has ownership interest (including stock, patents, etc.) in Data Driven Bioscience. No potential conflicts of interest were disclosed by the other author.