Summary:

Aberrant DNA cytosine methylation is a critical contributor to compromised tissue regeneration and malignant transformation, particularly during aging. In this issue of Cancer Discovery, Huang and colleagues define a new class of disease-associated DNA (cytosine-5-)-methyltransferase 3 alpha (DNMT3A) variants with decreased de novo DNA methylation activity due increased proteasomal degradation that are able to drive clonal expansion of hematopoietic stem cells.

See related article by Huang et al., p. 220.

DNA cytosine methylation (5mC) is the addition of a methyl group to the C5 position of the pyrimidine ring of cytosines and a key epigenetic modification facilitating X-chromosome inactivation and gene regulation essential for cellular differentiation and organismic development and growth. De novo DNA methylation is governed by DNA methyl-transferase 3A (DNMT3A) 3A and DNMT3B, as opposed to DNMT1, which mainly ensures the maintenance and propagation of already established 5mC marks during DNA replication. Although DNMT3B is known to drive DNA cytosine methylation within transcriptionally active gene bodies as well as at repetitive elements, DNMT3A appears to mostly target hypomethylated enhancers and developmentally poised promoters. Both de novo methyltransferases are essential for healthy cell differentiation and tissue maintenance, but DNMT3A has been specifically implicated in age-associated clonal hematopoiesis, as well as in hematologic cancers, underscoring its likely unique role in healthy blood cell production (for review, see ref. 1). Nevertheless, the exact mechanistic underpinnings of how DNMT3A establishes differentiation-instructing gene regulation and how DNMT3A-dependent DNA methylation aberrations contribute to pathology are still incompletely resolved.

Lacking a consensus DNA-binding motif, DNMT3A is recruited to genomic target loci by specific transcription factors (2) or through recognition of histone H3 lysine 36 (H3K36) methylation marks (3), facilitated through its protein-interacting ATRX–DNMT3–DNMT3L [ADD, or plant homeodomain (PHD)-like] and Pro-Trp-Trp-Pro domains. DNMT3A mutations were first reported in acute myeloid leukemia (AML) and subsequently discovered across most hematologic malignancies (1). These mutations spread throughout the gene but are particularly found within catalytic methyltransferase (MT) and protein-interacting domain-encoding regions (1). Comprehensive clinical and preclinical work in the most recurrent mutation in AML, heterozygous R882, has shown R882 (R878 in mouse) to exert a dominant negative function on the wild-type protein, critically impairing MT activity and DNA binding, to result in aberrant DNA hypomethylation at CpG islands, shores, and promoters (1), as well as driving abnormal hematopoietic stem and progenitor cell expansion and AML development in mice (4). Accumulating evidence indicates that despite their compatibility with multilineage blood formation in humans, DNMT3A mutations likely aid in the malignant transformation process of hematopoietic stem cells (HSC) during aging (1). Albeit the precise molecular mechanism of action had remained largely elusive for variants other than R882, at least some DNMT3A variants had been suspected to facilitate the malignant transformation process through the expansion of differentiation-impaired HSC populations, as demonstrated in mouse models with Dnmt3A-deficient hematopoiesis (1).

In this issue of Cancer Discovery, Huang and colleagues demonstrate that most DNMT3A variants confer loss of de novo DNA methylation function and indeed increase intrinsic cellular fitness likely to drive the expansion of HSC clones in age-associated clonal hematopoiesis and AML, with a defined subset showing increased susceptibility to protein degradation (Fig. 1; ref. 5). Huang and colleagues took on the seemingly Herculean task of comprehensively characterizing MT activity of 253 disease-associated DNMT3A variants. The authors developed a fluorescent reporter assay to quantify DNMT3A variant MT activity at large scale through measuring highly DNMT3A-dependent promoter (HOXA5) activity driving a Snrpn–blue fluorescent protein reporter in human embryonic kidney cells (293T). Using this assay along with lentivirus-mediated ectopic expression of DNMT3A variants, the group uncovered that more than 70% of the DNMT3A mutations showed severe loss of function, which was subsequently validated by using whole-genome bisulfite sequencing upon ectopic expression of a selected group of variants in embryonic stem cells with near-complete absence of DNA methylation due to the lack of Dnmt3a and Dnmt3b. Following the observation that some of the selected variants also showed remarkably reduced protein levels, the investigators continued to determine the consequences of DNMT3A lesions on protein stability. They uncovered that indeed more than one third of the tested DNMT3A variants showed decreased protein stability, which they attributed to aberrant interdomain interactions of and bivalent metal coordination at protein–protein interacting domains. In contrast, MT domain–targeted mutations appeared to be less potent to decrease protein stability.

Figure 1.

Fluorescent DNA methylation reporter platform to assess DNA cytosine methylation activity of 253 disease-associated DNMT3A variants uncovers that most DNMT3A lesions confer severe impairment of DNA de novo methylation. Seventy percent of the DNMT3A variants tested by Huang and colleagues show severely compromised methyltransferase activity (bottom). Aside from the AML-prevalent MT domain impairing the heterozygous DNMT3AR882 lesion (bottom right), another third of the variants tested display decreased protein degradation, mediated by the UPS and selectively targeted by E3 ubiquitin ligase complex CUL4–DDB1–DCAF8 (bottom left). Functionally, such unstable DNMT3A mutants may likely increase HSC fitness and drive clonal expansion, as well as facilitate malignant transformation. BFP, blue fluorescent protein; WT, wild-type. This figure was created with Biorender.

Figure 1.

Fluorescent DNA methylation reporter platform to assess DNA cytosine methylation activity of 253 disease-associated DNMT3A variants uncovers that most DNMT3A lesions confer severe impairment of DNA de novo methylation. Seventy percent of the DNMT3A variants tested by Huang and colleagues show severely compromised methyltransferase activity (bottom). Aside from the AML-prevalent MT domain impairing the heterozygous DNMT3AR882 lesion (bottom right), another third of the variants tested display decreased protein degradation, mediated by the UPS and selectively targeted by E3 ubiquitin ligase complex CUL4–DDB1–DCAF8 (bottom left). Functionally, such unstable DNMT3A mutants may likely increase HSC fitness and drive clonal expansion, as well as facilitate malignant transformation. BFP, blue fluorescent protein; WT, wild-type. This figure was created with Biorender.

Close modal

Expression control via protein turnover is likely of particular relevance for HSCs, which tightly restrict protein translation rates (6), providing an effective mechanism to rapidly increase DNMT3A abundance upon stem cell activation to initiate differentiation commitment (7). In support, Huang and colleagues uncovered that unstable DNMT3A variants showed increased cell-intrinsic fitness advantages when modeled in silico using published human patient-derived data sets assessing for the dynamics of clonal hematopoiesis; they also found in silico evidence supporting that instable DNMT3A variant–dependent fitness alterations may increase the risk of AML development. Whether and how AML-associated gene lesions frequently co-occurring with DNMT3A variants, such as in FLT3 or NPM1, may facilitate malignant transformation through shared or distinct mechanisms along with stable and unstable mutant DNMT3A remains to be tested.

Wild-type DNMT3A protein abundance can be controlled by the ubiquitin–proteasome system (UPS), polyubiquitylated, and marked for UPS-mediated protein degradation by ubiquitin-like protein containing PHD and RING domains (UHRF) 1 and 2 acting as substrate-specific E3 ligases (8). Huang and colleagues found that pharmacologic inhibition of proteasome activity also partially restored global DNA methylation patterning and gene expression in DNMT3A unstable variant–expressing cells, demonstrating a functional role for UPS-mediated turnover (Fig. 1). To study the mechanistic underpinnings in depth, the authors used lymphoblastoid cell lines derived from patients with AML, as well as generated a new mouse line bearing one of the DNMT3A lesions created by an in-frame deletion of W297 (W293 in mice). DNMT3AW293Del resulted in severely compromised protein stability via enhanced degradation through the UPS, which appeared partially dependent on Cullin–RING E3 ligase activity. In line, they found higher levels of ubiquitination of DNMT3AW297Del than wild-type DNMT3A, suggesting that aberrantly increased polyubiquitination may cause enhanced proteasomal degradation of unstable variants. Huang and colleagues subsequently validated 20 additional unstable DNMT3A variants to be degraded in a similar manner. Moreover, ubiquitin ligase–focused CRISPR screens and coimmunoprecipitation on DNMT3AW297Del-expressing 293T cells revealed substrate adapter DCAF8, which interacts with the CUL4–DDB1 E3 ligase macromolecular complex, to facilitate ubiquitination and degradation of both wild-type and DNMT3A unstable mutants. Whether UHRF1/2 ubiquitinate DNMT3A wild-type and unstable mutants, and whether these or any other of the cell's more than 600 E3 ligases may work in concert with or compete for CUL4–DDB1–DCAF8 ubiquitination of DNMT3A, remains to be determined.

Overall, this exciting study provides crucial insights into common molecular and functional dominators of the remarkable array of diverse disease-associated DNMT3A mutations; it presents potentially new biomarkers for identifying aging individuals with an increased risk for AML development and suggests the rational design of DNMT3A-specific proteasome impairment as a new therapeutic avenue in higher risk individuals; it also raises the question of whether malignant transformation–associated increases in proteasome expression (9) may decrease even wild-type DNMT3A levels enough to support aberrant epigenetic reprogramming and leukemic stem cell formation. Finally, this work reveals new fundamental biology of DNMT3A, which posits further existential questions for this key epigenetic regulator: with the nonrandom distribution of DNMT3A variants increasing protein degradation, do posttranslational modifications within protein–protein interacting domains fine-tune DNMT3A degradation during healthy cell differentiation, perhaps allowing for a cell type– and/or context-specific regulation of DNMT3A? As DNMT3A-mediated epigenetic memory of autophagy (the second pillar of cellular protein degradation governed by lysosomes) increases susceptibility to apoptosis and senescence in cells upon repeated autophagy activation (10), are longer-lived (stem and multipotent progenitor) cells with increased DNMT3A degradation selected for during aging?

B. Will reports personal fees from Novartis Pharmaceuticals and Aileron; grants from Novartis Pharmaceuticals, and grants from LifeBiosciences outside the submitted work. No disclosures were reported by the other author.

1.
Yang
L
,
Rau
R
,
Goodell
MA
. 
DNMT3A in haematological malignancies
.
Nat Rev Cancer
2015
;
15
:
152
65
.
2.
Fuks
F
,
Burgers
WA
,
Godin
N
,
Kasai
M
,
Kouzarides
T
. 
Dnmt3a binds deacetylases and is recruited by a sequence-specific repressor to silence transcription
.
EMBO J
2001
;
20
:
2536
44
.
3.
Weinberg
DN
,
Papillon-Cavanagh
S
,
Chen
H
,
Yue
Y
,
Chen
X
,
Rajagopalan
KN
, et al
The histone mark H3K36me2 recruits DNMT3A and shapes the intergenic DNA methylation landscape
.
Nature
2019
;
573
:
281
6
.
4.
Dai
YJ
,
Wang
YY
,
Huang
JY
,
Xia
L
,
Shi
XD
,
Xu
J
, et al
Conditional knockin of Dnmt3a R878H initiates acute myeloid leukemia with mTOR pathway involvement
.
Proc Natl Acad Sci U S A
2017
;
114
:
5237
42
.
5.
Huang
Y-H
,
Chen
C-W
,
Sundaramurthy
V
,
Słabicki
M
,
Hao
D
,
Watson
CJ
, et al
Systematic profiling of DNMT3A variants reveals protein instability mediated by the DCAF8 E3 ubiquitin ligase adaptor
.
Cancer Discov
2022
;
12
:220–35.
6.
Signer
RA
,
Magee
JA
,
Salic
A
,
Morrison
SJ
. 
Haematopoietic stem cells require a highly regulated protein synthesis rate
.
Nature
2014
;
509
:
49
54
.
7.
Zaro
BW
,
Noh
JJ
,
Mascetti
VL
,
Demeter
J
,
George
B
,
Zukowska
M
, et al
Proteomic analysis of young and old mouse hematopoietic stem cells and their progenitors reveals post-transcriptional regulation in stem cells
.
Elife
2020
;
9
:
e62210
.
8.
Jia
Y
,
Li
P
,
Fang
L
,
Zhu
H
,
Xu
L
,
Cheng
H
, et al
Negative regulation of DNMT3A de novo DNA methylation by frequently overexpressed UHRF family proteins as a mechanism for widespread DNA hypomethylation in cancer
.
Cell Discov
2016
;
2
:
16007
.
9.
Kumatori
A
,
Tanaka
K
,
Inamura
N
,
Sone
S
,
Ogura
T
,
Matsumoto
T
, et al
Abnormally high expression of proteasomes in human leukemic cells
.
Proc Natl Acad Sci U S A
1990
;
87
:
7071
5
.
10.
Gonzalez-Rodriguez
P
,
Cheray
M
,
Fullgrabe
J
,
Salli
M
,
Engskog-Vlachos
P
,
Keane
L
, et al
The DNA methyltransferase DNMT3A contributes to autophagy long-term memory
.
Autophagy
2021
;
17
:
1259
77
.