Complementary Roles of DNMT3A and TET2 KO in Sculpting CD4-T Cells for Sustained Proliferation without Exhaustion to Promote Lymphomagenesis

Angioimmunoblastic T-cell lymphoma (AITL) is a type of peripheral T-cell lymphoma characterized by frequent genetic mutations affecting TET2 and often together with DNMT3A. We have derived mice with conditional KO of both Tet2 and Dnmt3a mutations in CD4-T cells. These mice develop T-cell lymphoma with a faster kinetics than mice with Tet2 KO only supporting the cooperativity of these mutations. We therefore generated TET2 KO, DNMT3A KO and TET2/DNMT3A KO (DKO) human CD4+ T cells using the CRISPR/Cas9 system to uncover the molecular mechanisms by which TET2 and DNMT3A loss of function may cooperate in the pathogenesis of AITL. We used in vitro CD3/CD28 bead stimulation to activate these cells in 10-day cycles where cells underwent activation, expansion, effector differentiation, and contraction similar with in vivo stimulation. We further performed transcriptomic, DNA methylation, and histone H3K27 acetylation analysis to gain insights on the mechanism underlying observed differences.

Both TET2 KO and DKO cells exhibited higher proliferation rates than wild-type (WT) cells during the acute expansion phase, a difference that became more pronounced with increased stimulation cycles. WT cells developed an exhausted phenotype and lost memory markers, which were blocked in TET2 KO cells. Transcriptomic analysis revealed two sets of “TET2 regulons”: TET2-upregulated exhaustion genes and TET2-suppressed memory genes. TET2 may directly contribute to the exhaustion phenotype through DNA demethylation (eg IKZF4, TOX). Interestingly, despite TET2 and DNMT3A having opposing roles in DNA methylation, DNMT3A KO did not negate TET2 KO's consequences in sustaining memory and preventing exhaustion.

However, TET2 KO cells quickly reverted to quiescence after activation, whereas combined DNMT3A KO showed prolonged activation, with increased spontaneous division even at the end of the stimulation cycle. This phenotype was reflected at the transcriptional level and was directly linked to DNMT3A's role in DNA methylation. In the resting state, DKO cells exhibited enriched effector and proliferation gene signatures compared to TET2 KO cells. Specifically, effector modulators like TBX21, PRDM1, and JUN were hypomethylated and upregulated in DKO cells. Additionally, RNA-seq identified upregulated pathways such as 'IL2_STAT5', 'TNFA_signaling via NFKB', and 'DNA damage response pathway,' all of which were hypomethylated in DKO cells.

In addition to its role in DNA demethylation, TET2 performs non-enzymatic functions, such as recruiting histone deacetylases (HDAC1/2) to repress gene expression. Loss of TET2 may, therefore be associated with increase in gene expression of these loci without alteration in 5mC. Interestingly, 'T-memory genes' upregulated by TET2 KO, including CCR4, CCR7 and IRF4 did not exhibit significant methylation changes; however, they showed increased H3K27 acetylation in both TET2 KO and DKO cells. Furthermore, HDAC inhibition with Trichostatin A in DNMT3A KO cells promoted similar genes upregulated as in TET2 KO. We are performing ChIP-seq experiments to validate directly TET2 binding sites.

ICOS, a critical costimulatory molecule, is highly expressed in AITL and thought to play a role in its pathogenesis. ICOS was highly expressed with increased H3K27 acetylation in TET2 KO and DKO cells. Surprisingly, upon CD3/ICOS Ab stimulation, DKO cells showed sustained PLCγ1 activation. Notably, PLCG1 was also highly expressed with prominent-H3K27 acetylation in both TET2 KO and DKO cells. These findings suggest that the enzymatic and non-enzymatic functions of TET2 and DNMT3A cooperated to shape the ICOS stimulatory pathway in CD4-T cells.

Chronic T cell activation with antigen stimulation typically leads to T cell anergy and exhaustion. This study provides insight into how the combined knockout of TET2 and DNMT3A sustained long-term T cell proliferation and activation without inducing exhaustion, thus increasing tumorigenic potential of the altered cells. TET2 and DNMT3A appeared to have divergent and yet often complementary roles in balancing T cell exhaustion and activation despite their opposite functions in DNA methylation. Importantly, some of the phenotypes observed with TET2 KO may be related to its normal function in histone deacetylation whose loss could lead to aberrant gene expression and the amplified expression of some genes with hypomethylation due to DNMT3A KO.