sciencemag.org SCIENCEINSIGHTS | PERSPECTIVESGRAPHIC: V. ALTOUNIAN/SCIENCEWagenblast et al. use CRISPR-Cas9 ed-
iting of human fetal hematopoietic stem
and progenitor cells (HSPCs) trisomic for
Hsa21 to induce DS-specific leukemogenic
deletions in GATA1 and stromal antigen
2 (STAG2). These edited cells were then
transplanted into immunocompromised
mice, creating xenograft mouse models
that faithfully recapitulate DS leukemia. In
contrast to previous models, it seems likely
that targeting primary human fetal HSPCs
to induce the pathognomonic GATA1 muta-
tions provided the permissive cellular sub-
strate for transformation, as predicted by
clinical and biological studies in DS ( 2 , 4 ,
5 ). Whereas in disomic fetal cells, expres-
sion of GATA1s caused severe impairment
of erythropoiesis (corresponding to anemia
in non-DS individuals with germline GATA1
mutations), in T21 fetal cells, GATA1s ex-pression increased megakaryocytes and
leukemia-like blast cells similar to those in
the preleukemia in DS babies ( 2 , 4 ).
Studies to decipher the molecular basis for
the effects of T21 have generated a wealth of
data from a range of cell types, most focus-
ing on gene dosage of Hsa21-located genes.
Hsa21, the smallest human chromosome, has
~230 protein-coding genes and almost twice
as many non–protein coding genes, includ-
ing five miRNAs ( 1 , 6 ). Although many Hsa21
genes are expressed at 1.5-fold higher levels
than in matching disomic populations, this
is highly tissue and cell population specific,
and the overall effects of the supernumer-
ary Hsa21 extend far beyond this. Almost all
datasets show genome-wide perturbation of
gene expression by T21. This likely explains,
at least in part, why causative links betweenaltered gene expression and phenotypes in
DS have been so difficult to establish and
why none of these phenotypes has so far
been explained by a single gene acting alone.
Studies aimed at narrowing the region(s) of
Hsa21 responsible for specific phenotypes
are often inconsistent and hampered by the
rarity of individuals with DS owing to partial
T21 (<1%) and by cellular systems that im-
perfectly capture the genetic and epigenetic
complexity of DS.
Given the importance of cell context,
the development of a biologically accurate
model, recreating the precise initiating step
necessary for all DS myeloid leukemias
(GATA1s-encoding mutation) in the precise
cell type where this event occurs (fetal liver
HSPCs), should allow more specific and
tractable questions about the role of T21 to
be addressed. Focusing on the mechanismof cooperation between T21 and GATA1s,
Wagenblast et al. asked how binding of
mutant GATA1s protein differed from full-
length, wild-type GATA1 in fetal HSPCs.
They found that GATA1s specifically bound
to thousands of promoters, including those
that regulate genes controlling miRNA
production. They focused on three Hsa21
miRNA genes—MIR-125b-2, MIR-155, and
LET-7C—whose expression was up-regulated
specifically in T21 long-term hematopoietic
stem cells (LT-HSCs). Enforced overexpres-
sion of these miRNAs in normal fetal LT-
HSCs mimicked some features of T21 cells,
whereas ablating them partially reduced
the preleukemic phenotype of GATA1s,
suggesting that GATA1s-mediated up-regu-
lation of these Hsa21 miRNAs’ expression
could partially explain the mechanism ofGATA1s-induced preleukemia in DS. This
adds DS myeloid leukemia to the list of hu-
man diseases, including other phenotypic
abnormalities in DS, that may be caused by
altered miRNA expression ( 1 ). The precise
mechanisms underlying the distinct coop-
eration between T21 and GATA1s in human
fetal cells remain enigmatic but are likely
to include several Hsa21 genes with known
roles in embryonic and fetal hematopoiesis,
acting together in a cell context–dependent
manner (see the figure).
The reasons for the high frequency of
somatic GATA1 mutations in DS fetal blood
cells (most likely owing to selection) and
why most GATA1-mutant cells are rapidly
cleared after birth—a feature not captured
in Wagenblast et al.’s model, which may in-
volve the T21 hematopoietic cell niche—are
not yet known. Evidence of a mutagenic
phenotype or generalized defect in DNA re-
pair in DS tissues, including blood cells, is
sparse. Children with DS have a lower than
expected incidence of solid tumors, point-
ing to specific properties of T21 in hemato-
poietic cells. This might also explain why
T21 appeared to be less relevant for post-
natal progression of myeloid leukemia in
Wagenblast et al.’s model.
T21 is the most frequent constitutional
aneuploidy compatible with survival into
adulthood, although fetal loss is high ( 1 ).
Yet, despite the importance of genome or-
ganization on function ( 7 ), the effects of an-
euploidy on nuclear architecture of human
fetal cells and subsequent consequences for
gene expression and protein production
are unknown. Similarly, the development
of new model systems is needed to better
understand the function of each element
of Hsa21, only 48% of which has currently
been mapped in detail ( 1 ). Moreover, it can
be argued that the value of such models
lies with the opportunities for improving
outcomes for individuals with DS and their
families. This will likely be realized through
small steps in specific complications such
as leukemia but could include bolder ini-
tiatives such as harnessing X chromosome
inactivation in female cells by perhaps in-
tegrating an X inactive specific transcript
(XIST) transgene to reduce Hsa21 transcrip-
tional output back to disomic levels ( 8 ). jREFERENCES AND NOTES- S. E. Antonarakis et al., Nat. Rev. Dis. Primers 6 , 9 (2020).
- I. Roberts et al., Blood 122 , 3908 (2013).
- E. Wagenblast et al., Science 373 , eabf6202 (2021).
- A. Roy et al., Proc. Natl. Acad. Sci. U.S.A. 109 , 17579
(2012). - M. Labuhn et al., Cancer Cell 36 , 123 (2019).
- M. Hattori et al., Nature 405 , 311 (2000).
- C. Do, Z. Xing, Y. E. Yu, B. Tycko, Epigenomics 9 , 189
(2017). - J. C. Chiang, J. Jiang, P. E. Newburger, J. B. Lawrence, Nat.
Commun. 9 , 5180 (2018).
10.1126/science.abj3957
Disomy 21Trisomy 21GATA1GATA1sGATA1sGATA1sErythropoiesisTransplantation
into mice
No leukemia+ Myeloid leukemiaGATA1GATA1s Megakaryopoiesis and blast cells Preleukemia+ Myeloid leukemiaSTAG2 deletionSTAG2 deletionCooperating changes to initiate leukemia in fetal cells
The supernumerary copy of human chromosome 21 in Down syndrome [trisomy 21 (T21)] perturbs fetal gene
expression and hematopoietic stem and progenitor cell (HSPC) development. Acquisition of exon 2 mutations
in the GATA binding protein 1 (G ATA 1) gene in fetal T21 but not disomic HSPCs causes preleukemia in mice
because of expression of a short GATA1 protein (GATA1s). Secondary mutations in GATA1s-expressing cells,
such as in stromal antigen 2 (STAG 2), cause myeloid leukemia in disomic and T21 mouse fetal HSPCs.156 9 JULY 2021 • VOL 373 ISSUE 65510709Perspectives.indd 156 7/1/21 6:34 PM