ArticleDown-syndrome-induced senescence disrupts the
nuclear architecture of neural progenitorsGraphical abstractHighlightsd Trisomy 21 disrupts nuclear architecture and transcriptome
of neural progenitors
d Trisomy 21 harboring neural progenitors display signatures of
cellular senescence
d Senolytic drugs ameliorate trisomy-21-associated molecular
and cellular dysfunctionsMeharena et al., 2022, Cell Stem Cell 29, 116–130
January 6, 2022 ª 2021 The Authors. Published by Elsevier Inc.
https://doi.org/10.1016/j.stem.2021.12.002Authors
Hiruy S. Meharena, Asaf Marco,
Vishnu Dileep, ...,
Malvina Papanastasiou,
Jacob D. Jaffe, Li-Huei Tsai
Correspondence
hiruym@mit.edu (H.S.M.),
lhtsai@mit.edu (L.-H.T.)
In brief
In this issue of Cell Stem Cell, Meharena
et al. (2021) show that Down syndrome or
trisomy 21 disrupts the 3D-genome,
epigenome, and transcriptome of neural
progenitors similar to that observed in
senescent cells and find that the senolytic
drug combination of dasatinib and
quercetin ameliorates these disruptions.ll
OPEN ACCESSllArticle
Down-syndrome-induced senescence disrupts
the nuclear architecture of neural progenitors
Hiruy S. Meharena,1,2,* Asaf Marco,1,2 Vishnu Dileep,1,2 Elana R. Lockshin,1,2 Grace Y. Akatsu,1,2 James Mullahoo,3
L. Ashley Watson,1,2 Tak Ko,1 Lindsey N. Guerin,1,2 Fatema Abdurrob,1,2 Shruthi Rengarajan,1 Malvina Papanastasiou,3
Jacob D. Jaffe,3 and Li-Huei Tsai1,2,3,4,*
1Picower Institute for Learning and Memory, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
2Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
3Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA
4Lead contact
*Correspondence: hiruym@mit.edu (H.S.M.), lhtsai@mit.edu (L.-H.T.)
https://doi.org/10.1016/j.stem.2021.12.002SUMMARYDown syndrome (DS) is a genetic disorder driven by the triplication of chromosome 21 (T21) and characterized
by a wide range of neurodevelopmental and physical disabilities. Transcriptomic analysis of tissue samples
from individuals with DS has revealed that T21 induces a genome-wide transcriptional disruption. However,
the consequences of T21 on the nuclear architecture and its interplaywith the transcriptome remain unknown.
In this study, we find that unlike human induced pluripotent stem cells (iPSCs), iPSC-derived neural progenitor
cells (NPCs) exhibit genome-wide ‘‘chromosomal introversion,’’ disruption of lamina-associated domains, and
global chromatin accessibility changes in response to T21, consistent with the transcriptional and nuclear ar-
chitecture changes characteristic of senescent cells. Treatment of T21-harboring NPCs with senolytic drugs
alleviates the transcriptional, molecular, and cellular dysfunctions associated with DS. Our findings provide a
mechanistic link between T21 and global transcriptional disruption and indicate that senescence-associated
phenotypes may play a key role in the neurodevelopmental pathogenesis of DS.INTRODUCTION
Down syndrome (DS) is a genetic disorder caused by the tripli-
cation of chromosome 21 (T21) and is the leading cause of in-
tellectual disability associated with abnormal brain morphogen-
esis (Benda, 1940; Down, 1995; Pinter et al., 2001; Weijerman
and de Winter, 2010). These brain abnormalities have been
associated with dysfunctional neural progenitor cells (NPCs),
the multipotent stem cells of the developing brain that differen-
tiate into neurons, astrocytes, or oligodendrocytes (Stagni
et al., 2018; Tyler and Haydar, 2013). Recent whole-genome
transcriptional profiling of various tissue samples from individ-
uals with DS and DS mouse models revealed that T21 induces
genome-wide transcriptional disruption in addition to the
dosage-dependent upregulation of a subset of expressed
genes on chromosome 21, similar to those observed in
other whole-chromosome aneuploidies (Bianco et al., 2016;
Du€rrbaum et al., 2014; FitzPatrick et al., 2002; Gonzales
et al., 2018; Lejeune et al., 1959; Letourneau et al., 2014; Ol-
mos-Serrano et al., 2016; Sheltzer et al., 2012; Tyler and Hay-
dar, 2013; Walus et al., 2016; Wangsa et al., 2019). However,
the mechanisms by which T21 induces genome-wide transcrip-
tional disruption remain unknown.
Epigenetic modifications of chromatin determine TF-binding
site accessibility at gene promoters and regulatory elements,116 Cell Stem Cell 29, 116–130, January 6, 2022 ª 2021 The Authors
This is an open access article under the CC BY-NC-ND license (http://and the folding of chromatin within the interphase nucleus brings
these two regions into spatial proximity (McLaughlin et al., 2019;
Pombo and Dillon, 2015). This three-dimensional (3D) genome
organization of gene promoters and regulatory elements deter-
mines the transcriptional levels of specific genes within a cell
(Schoenfelder and Fraser, 2019). However, the influence of T21
on the 3D-genome organization and the role of the epigenome
in the global transcriptional disruption observed in DS remain
unexplored.
To explore the consequences of T21 on the molecular and
cellular changes observed in DS, we established human-derived
induced pluripotent stem cells (iPSCs) and iPSC-derived fore-
brain NPCs frommultiple individuals and investigated the conse-
quences of T21 on the 3D-genome organization, epigenome,
and transcriptome (Buenrostro et al., 2013; Chambers et al.,
2009; Rao et al., 2014; Topol et al., 2015; Weick et al., 2013).
Our findings show that T21 induces chromosomal introversion,
disrupts lamina-associated domains (LADs), and alters the
genome-wide chromatin accessibility of NPCs but not iPSCs.
While the overall organization of A/B compartments is conserved
in NPCs harboring T21, we observed global loss of chromatin
accessibility within the A-compartment that is associated with
transcriptional downregulation and increased long-range chro-
matin interactions in the B-compartment associated with
transcriptional upregulation. We find that these architectural. Published by Elsevier Inc.
creativecommons.org/licenses/by-nc-nd/4.0/).
ll
Article OPEN ACCESSchanges are similar to those observed in senescent cells, and
our transcriptional analysis confirms that differentially expressed
genes (DEGs) identified in NPCs harboring T21 are highly corre-
lated with DEGs identified in oxidative stress-induced senescent
cells. Finally, we demonstrate that the senolytic drug combina-
tion of dasatinib and quercetin (DQ) alleviates the genome-
wide transcriptional disruption as well as deficits in cellular pro-
liferation and migration observed in NPCs harboring T21.
RESULTS
T21 induces chromosomal introversion in NPCs
To determine the consequences of T21 on the global 3D-
genome organization, we performed chromosome conformation
capture (Hi-C) on an isogenic pair (euploid: Iso-E and trisomic:
Iso-T) of iPSCs and NPCs (Maclean et al., 2012) and identified

2.5 billion unique interactions per cell type (Figures 1A–1D; Ta-
ble S1A). We first sought to interrogate the consequences of T21
on the 3D-genome reorganization that occurs during differentia-
tion of iPSCs to NPCs. We observed significant gain and loss
of both cis- and trans-chromosomal interactions in Iso-E
NPCs compared with iPSCs (Figure 1E); however, in Iso-T, we
observed significant genome-wide reduction of trans-chromo-
somal interactions in NPCs compared with iPSCs (Figure 1F).
While T21 minimally impacted global trans-chromosomal inter-
actions in iPSCs (Iso-E = 19.82% ± 1.0% and Iso-T =
21.72% ± 2.8%, p value = 3.4 3 101) (Figures S1A and S1B),
we observed a significant decrease in the fraction of trans-chro-
mosomal interactions in NPCs harboring T21 (Iso-E = 21.55% ±
0.8% and Iso-T = 15.85% ± 1.2%, p value = 2.23 103) (Figures
1G and 1H; Table S1B). To evaluate whether the altered distribu-
tion of chromosomal interactions induced by T21 in NPCs is
conserved across individuals with DS, we derived NPCs from
an additional pair of male individuals with euploid (Ma-E) and
trisomic (Ma-T) karyotypes (Figure S1C). Consistent with our
findings from the isogenic pair, we observed that NPCs
harboring T21 exhibit reduced trans-chromosomal interactions
compared with euploid NPCs (Figures S1D and S1E). As ex-
pected, we observed increased signal from chromosome 21
(HSA21) in both iPSCs and NPCs because of its triplication; how-
ever, when normalized for the number of chromosomes, we
observed that loss of trans-chromosomal interactions in NPCs
is distributed across all chromosomes, including HSA21 (Fig-
ure S1F; Table S1C).
T21 NPCs exhibited reduced longer-range interactions (>
1
Mb) and increased shorter-range interactions (<
1 Mb),
whereas in the isogenic pair of iPSCs, the distribution of cis-
chromosomal interactions remained unchanged (Figure S1I).
This reduction of longer-range cis-chromosomal interactions
and an increase in shorter-range interactions was observed
across all chromosomes, including HSA21 in T21 NPCs but not
iPSCs (Figures 1I–1L and S1J). On average, we observed an

20% increase in shorter-range (<1 Mb) interactions in T21
NPCs (Table S1D). To ensure that these results were not
confounded because of the normalization of T21, we repeated
the analysis by removing chromosome 21 reads and found
similar genome-wide chromosomal introversion in T21-NPCs
(Figures S1K–S1N). In conclusion, we observed that T21 induces
loss of trans-chromosomal interactions and increases shorter-range (<1 Mb) cis-chromosomal interactions in NPCs but not
in iPSCs.
T21 induces reorganization of intra-TAD interactions
and chromosomal looping
The genome is organized into two compartments within the nu-
cleus: the active (A) compartment, which is localized toward the
nuclear core, and the inactive (B) compartment, which is local-
ized toward the periphery on the nuclear lamina (Lieberman-Ai-
den et al., 2009). We found that T21 minimally impacts the clas-
sification of A/B compartments in both iPSCs and NPCs (Figures
S2A and S2B). A/B compartments are composed of multiple
self-interacting structural units of the genome known as topolog-
ically associating domains (TADs) (Dixon et al., 2012; Nora et al.,
2012; Rao et al., 2014), and we identified 5,559 and 4,731 TADs
in iPSCs and NPCs, respectively (Table S2A). While T21 did not
impact the intra-TAD interaction density (ID) of iPSCs, the intra-
TAD ID of NPCs was significantly reorganized in trisomic NPCs
compared with euploid NPCs, where 1,800 (
38%) TADs had
significantly higher ID and 389 (
8%) TADs had significantly
reduced ID (Figures 2A–2D). Next, we performed an enrichment
analysis (10 million randomized permutations) to assess whether
the differentially interacting TADs preferentially localized to a
specific compartment. We found that TADs with increased ID
were enriched in the B-compartment, and unchanged TADs
were predominantly localized in the A-compartment; however,
the reduced-ID TADs were not enriched in either compartment
(Figure 2E; Table S2B). While the intra-TAD ID of 
46% of
TADs was significantly altered, the directionality index and insu-
lation score remained unchanged as a consequence of T21 in
both iPSCs and NPCs, indicating a general maintenance of
global chromatin architecture and TAD boundaries (Figure S2C).
Next, we identified 12,053 loops in iPSCs and 11,373 loops in
NPCs: of these, 2,653 and 1,973 loops are unique to iPSCs and
NPCs, respectively (Figures 2F and S2D; Table S2C). Differential
chromosomal looping analysis revealed that 
700 loops were
significantly reduced or lost in T21 NPCs, where 27 loops were
only identified in euploid NPCs (z score Iso-E = 17.75 and Iso-
T = 0.84), and 672 were significantly reduced in T21 NPCs
(z score Iso-E = 64.46 and Iso-T = 32.46). Conversely, we identi-
fied 
3,800 loops significantly enhanced in T21 NPCs, with 264
de novo loops (z score Iso-E = 2.19 and Iso-T = 33.97), and 3,494
significantly increased loops in T21 NPCs (z score Iso-E = 30.75
and Iso-T = 62.56). Additionally, enrichment analysis revealed
that the chromosomal loops that were lost/reduced were
predominantly localized in TADs with reduced ID (p value =
0.00596, 10 million randomized permutations), and the de
novo/increased loops were enriched in TADs with increased ID
(p value = 0.00539) (Figure 2E; Table S2B). In summary, we
find that T21-induced introversion is predominantly localized
within the inactive B-compartment.
Disruption of nuclear lamina is associated with
increased intra-TAD interactions
The B-compartment is localized toward the nuclear periphery on
the nuclear lamina and is enriched for the histone modification
H3K9me3 (Shevelyov andUlianov, 2019). To interrogate the con-
sequences of T21 on genome organization around the nuclear
lamina, we utilized immunofluorescence to label the nuclearCell Stem Cell 29, 116–130, January 6, 2022 117
ll
OPEN ACCESS Article
Figure 1. T21 induces genome-wide chromosomal introversion in NPCs, but not iPSCs
(A–C) Immunofluorescence staining of (A) iPSCs (OCT4 [green], NANOG [red]) and (B) NPCs (SOX1 [green], Nestin [red]) from the isogenic pair (Iso-E [top] and Iso-
T [bottom]) as well as (C) respective G-band karyotyping of NPCs.
(D) Schematic of experimental design to assess the 3D-genome of the isogenic pair iPSCs and NPCs as well as the euploid (Ma-E) and trisomic (Ma-T)
male NPCs.
(E–H) Comparative Hi-C analysis of Iso-E and Iso-T iPSCs and NPCs. Genome-wide differential (NPC/iPSC) Hi-C interaction maps for Iso-E (E) and Iso-T (F). (G)
Genome-wide differential (Iso-T/Iso-E) Hi-C interactionmaps for NPCs. (H) Percent distribution of trans- and cis-chromosomal interactions for the isogenic pair of
NPCs (t test).
(I–L) Representative images of chromosome 21 (chr21) differential cis-chromosomal interaction maps for the isogenic pair of iPSCs (I) and NPCs (K). Dot plots
representing the ratio of short-range (<1 Mb) to long-range (>1 Mb) interactions for the isogenic pair of iPSCs (J) and NPCs (L). Each dot represents a chro-
mosome, and chr21 is represented as a red dot. Wilcoxon rank-sum test shows that euploid and trisomic iPSCs were not statistically different (p value = 0.536),
whereas NPCs show significantly increased short-range interactions in the trisomic NPCs (p value = 0.004).lamina and H3K9me3. We examined the localization of
H3K9me3 and the lamina-associated protein LMNB1 to interro-
gate the consequences of T21 on lamina organization. While the
overall architecture of the nuclear envelope and heterochromatin
organization in T21 iPSCs remains unchanged (Figures S3A–
S3C; Table S3A), we observed decreased levels of LMNB1 in
T21 NPCs relative to the euploid NPCs (Figures 3A and 3B).118 Cell Stem Cell 29, 116–130, January 6, 2022The global intensity of H3K9me3 was unchanged in both T21
iPSCs and NPCs (Figures S3C and S3D); however, we observed
an increased number of H3K9me3 aggregates specifically within
the nuclei of T21 NPCs and not iPSCs (Figures S3E and S3F).
To interrogate whether B-compartment TAD introversion is
associated with disruption of the nuclear lamina, we performed
ChIP-seq for LMNB1 in the isogenic pair of NPCs (Figure S3G).
ll
Article OPEN ACCESS
Figure 2. T21 induced reorganization of TADs and loops are segregated into A/B compartments
(A and C) Representative images of differential Hi-C interaction maps (Iso-T/Iso-E) of a region on chromosome 6 (chr6) in the isogenic pair of iPSCs (A) and
NPCs (C).
(B and D) Volcano plot of differential interaction density (ID) within TADs in the isogenic pair of iPSCs (B) and NPCs (D).
(E) Randomized permutation test analysis between the differential interaction density (ID) TADs with A/B compartments and differential chromosomal loops in
NPCs. Color intensity (positive correlations are displayed in green, and negative correlations in black) and the size of the circle are proportional to the enrichment
coefficients of A/B compartments with differentially interacting TADs identified as a consequence of T21 in NPCs.
(F) Aggregate Peak Analysis (APA) of loops uniquely identified in iPSCs (top rows) and NPCs (bottom rows) in both Iso-E and Iso-T.We identified 949 LADs, which are associated with 
53% of the
genome, and from these, we observed a loss of
7%of the LADs
(113Mb) and a gain of <1% (10Mb) in T21 NPCs (Figures 3C and
S3H). As expected, LAD enrichment analysis demonstrated that
the A-compartment was depleted for LMNB1 association in both
Iso-E and Iso-T NPCs (r = 0.4134) (Figure S3I), whereas the
B-compartment was highly enriched for LMNB1-associated
genomic regions (r = 0.4267) (Figure 3D; Table S3B). Additionally,
LMNB1 occupancy was significantly (p value < 104) reduced in
the B-compartment of Iso-T NPCs compared with that in Iso-E
NPCs. Next, we sought to investigate the relationship between
TAD interaction and LAD enrichment and found that TADs with
increased ID have significantly lower LMNB1 enrichment than
TADs with reduced ID or unchanged ID (Figures 3E and 3F and
S3J–S3O; Table S3C). In conclusion, we observed reduced
LMNB1 association in TADs within the B-compartment that
have increased intrachromosomal interactions in Iso-T NPCs
compared with Iso-E NPCs.
T21-induced transcriptional upregulation is associated
with chromosomal introversion
To examine the relationship between T21-associated nuclear ar-
chitecture reorganization and the transcriptome, we performedRNA sequencing of iPSCs and NPCs derived from the isogenic
pair as well as NPCs derived from individuals with DS (female
[Fe-T] and male [Ma-T]) and euploid controls (female [Fe-E]
and male [Ma-E]) (Figure 4A and Figure S1C). Principal compo-
nent analysis showed that transcriptionally, NPCs derived from
the different individuals segregate by karyotype on the first prin-
cipal component (Figure S4A). We observed that T21 had a
greater impact on the NPC transcriptome than the transcriptome
of iPSCs: NPCs exhibited 1,869 downregulated and 2,724 upre-
gulated DEGs (false discovery rate [FDR] < 0.01, log2 fold change
R 0.5 or%0.5), while iPSCs exhibited 102 downregulated and
423 upregulated DEGs (Figures 4B and 4C). Comparison of the
DEGs identified in NPCs derived from different individuals (pair-
wise comparisons of Iso-T versus Iso-E, Fe-T versus Fe-E, and
Ma-T versus Ma-E) showed a significant concordance of the
upregulated and downregulated genes after running 10 million
randomized permutations of the expression data (Figure S4B;
Tables S4A and S4B). To ensure that the transcriptional observa-
tions were not due to normalization of T21, we repeated our
differential analysis of the isogenic pair after removing all reads
originating from chromosome 21 and observed a similar
genome-wide transcriptional disruption (Figure S4C). While
both iPSCs and NPCs exhibited upregulation of 
25% of theCell Stem Cell 29, 116–130, January 6, 2022 119
ll
OPEN ACCESS Article
Figure 3. T21-induced disruption of lamina-associated domains (LADs) is associated with chromosomal introversion
(A) Immunofluorescence of LMNB1 (red) and H3K9me3 (green) in the isogenic pair of NPCs.
(B) Quantification of LMNB1 staining intensity in the isogenic pair of NPCs. Each dot on the histogram represents a replicate experiment of 
440 total nuclei
analyzed for the isogenic pair of NPCs (t test).
(C) Venn diagram of LAD overlap between the euploid (Iso-E) and T21 (Iso-T) NPCs (intersect of replicates); bottom, schematic representation of genome
coverage (base pairs) by LADs gained or lost as a consequence of T21 relative to euploid.
(D) Scatter plot of LMNB1 enrichment over input in Iso-E (x axis) and Iso-T (y axis) of the genomic region in the B-compartment.
(E) Violin plot of fold change of LMNB1 (Iso-T/Iso-E) of the differential (increased-ID [red], reduced ID [blue], and unchanged ID [green]) interaction density TADs.
(F) IGV plot of A-compartment (Iso-E [yellow] and Iso-T [orange]) and B-compartment (Iso-E [blue] and Iso-T [light-blue]), LMNB1 ChIP-seq of Iso-E (blue) and Iso-
T (light-blue), as well as the location of LADs (black) and the differentially interacting TADs (reduced ID [blue], increased ID [red], and unchanged ID [gray]), an 87
Mb region of chromosome 1.genes expressed on HSA21, NPCs showed disruption of 
20%
(8.5% downregulation and 11.6% upregulation) of expressed
genes on all the other chromosomes. In contrast, iPSCs only ex-
hibited alterations of 
3% (0.6% downregulation and 2.6% up-
regulation) of the expressed genes on all the other chromosomes
(Figures S4D and S4E).
In both iPSCs and NPCs, we found that the A and B compart-
ments are composed of similar genome sizes, and, as expected,
the majority of expressed genes are located in the A-compart-
ment (
80% and 
76% of expressed genes in iPSCs and
NPCs, respectively; Table S4C). However, we found that the
DEGs identified in NPCs were distinctly segregated into the
two compartments, with downregulated genes predominantly
located in the A-compartment and upregulated genes enriched
in the B-compartment (Figure 4D; Tables S4D and S4E). We
found that TADs with increased interaction densities and
increased looping events are enriched for upregulated genes,
exhibit reduced LMNB1 association, and are predominantly
located in the B-compartment (p value < 1 3 1007; Figures
4E, S4F, and S4G; Tables S4F–S4H). Conversely, TADs with
reduced ID in T21 NPCs are enriched for downregulated genes
and are not enriched in either compartment (p value < 1 3
1007). Although the A-compartment is predominantly enriched
for downregulated genes, we observed that this compartment
is predominantly enriched for TADs with unchanged ID (p value <120 Cell Stem Cell 29, 116–130, January 6, 20221 3 1007), suggesting that the downregulation of genes
observed in the A-compartment is independent of altered long-
range promoter-enhancer interactions.
We examined gene ontology by focusing on those DEGs local-
ized within differential ID TADs and found that downregulated
genes within reduced-ID TADs were enriched for brain develop-
ment-associated processes, whereas upregulated genes local-
ized within increased-ID TADs were predominantly enriched for
biological processes involving cell migration, cell adhesion,
and ECM organization (Figures 4F and 4G). The downregulated
genes within unchanged-ID TADs were associated with tran-
scriptional regulation and brain development (Figure 4H). In sum-
mary, upregulated genes were enriched in B-compartment TADs
with increased ID, and downregulated genes were predomi-
nantly enriched in reduced-ID TADs and in the A-compartment.
However, the mechanism underlying the downregulation of
genes localized within unchanged-ID TADs remains unknown.
T21-induced reduction of chromatin accessibility is
associated with dysfunction of nuclear processes
Our transcriptional analysis showed a significant number of the
downregulated genes in T21 NPCs impact chromatin modifica-
tions on histone H3, including histone deacetylases (HDAC1
and HDAC2), polycomb-group proteins (EZH2 and EED), histone
acetyltransferase (KAT2A), as well as the eukaryotic protein
ll
Article OPEN ACCESS
Figure 4. T21 DEGs are distinctly separated into the A/B compartments in NPCs
(A) Schematic of RNA-seq experiment on the isogenic pair (Iso-E and Iso-T) of iPSCs and NPCs as well as NPCs derived from euploid (female [Fe-E] and male
[Ma-E]) and trisomic (female [Fe-T] and male [Ma-T]) individuals.
(B and C) Volcano plots of the differentially expressed genes (DEGs) identified in the isogenic pair of iPSCs (B) and NPCs.
(C and D) Randomized permutation test T21 NPC DEGs with A/B compartments. Histogram represents the range of expected overlap percentage, and the red
line represents the observed percentage overlap.
(E) Corrplot of the randomized permutation test between TADs with varying interaction densities (IDs) with DEGs and A/B compartments. Color intensity (positive
correlations are displayed in green, and negative correlations in black) and the size of the circle are proportional to the enrichment coefficients.
(F–H) Gene ontology visualized in semantic similarity-based scatterplots for downregulated genes associated with reduced-ID TADs (F), upregulated genes
associated with increased-ID TADs (G), and downregulated genes localized within unchanged-ID TADs (H).kinases AURKB and VRK1 (Figure S5A). To assess the post-
translational modification landscape of histone H3 residues,
we performed global chromatin profiling (GCP), a targeted
mass-spectrometry-based assay that quantifies histone post-
translational modifications in bulk chromatin (Figure S5B; Table
S5A) (Jaffe et al., 2013). Comparison of the isogenic pair of
NPCs showed a significant reduction of the mitotic marker H3
serine-10 phosphorylation (H3S10p), consistent with previous
studies indicating reduced cellular proliferation in DS (Contesta-
bile et al., 2009; Contestabile et al., 2007; Guidi et al., 2011; Mol-drich et al., 2009). While we did not identify any significant
changes to chromatin states that typically demarcate promoters
or enhancers (H3 lysine-4 mono-methylation [H3K4me] and tri-
methylation [H3K4me3], respectively), we observed a reduction
in H3 lysine-23 acetylation (H3K23ac) and an increase in H3
lysine-27 acetylation (H3K27ac), marks associated with chro-
matin activity, potentially indicating a relationship between the
observed genome-wide transcriptional changes and the overall
chromatin state (Allis and Jenuwein, 2016; Klemm et al., 2019;
Lane et al., 2014; Liu et al., 2016; Lu et al., 2015).Cell Stem Cell 29, 116–130, January 6, 2022 121
ll
OPEN ACCESS Article
Figure 5. T21-induced genome-wide disruption of chromatin state in NPCs is associated with transcriptional downregulation
(A and B) Volcano plots of differentially accessible regions (DARs) identified by ATAC-sequencing in the isogenic pair of iPSCs (A) and NPCs (B).
(C) Enrichment heatmap of the randomized permutation test between the DARs on promoters, exons, introns, and intergenic regions with accessible regions
genome wide. Color intensity (positive correlations are displayed in red, and negative correlations in black) and the size of the circle are proportional to the
enrichment coefficients.
(D) Enrichment heatmap of the randomized permutation test between the A/B compartments and DEGs and DARs. Color intensity (positive correlations are
displayed in green, and negative correlations in black) and the size of the circle are proportional to the enrichment coefficients of the specific genetic features with
the A/B compartments.
(E and F) Aggregate plots of less accessible regions harboring SP1 (E) and NFY (F) motifs.
(G and H) Gene ontology visualized in semantic similarity-based scatterplots for downregulated genes with less accessible promoters harboring SP1 (G) and NFY
(H) motifs.To assess the consequences of T21 on the genome-wide
chromatin state, we utilized the assay for transposase-acces-
sible chromatin using sequencing (ATAC-seq) on the isogenic
pair of iPSCs and NPCs (Buenrostro et al., 2013). We found
that while T21 minimally impacts the chromatin state of iPSCs
(
1.6% of the accessible chromatin), 
20% of the accessible
chromatin is altered as a consequence of T21 in NPCs with
an FDR < 0.01 (Figures 5A and 5B), a level of chromatin state
alteration that has recently been observed in astroglia derived
from individuals with DS (Bally et al., 2020). We identified
18,304 differentially accessible regions (DARs) in T21 NPCs
(14,814 more accessible and 3,490 less accessible). As ex-
pected, due to the chromosomal triplication, HSA21 is more
permissive to transposase in both T21 iPSCs and NPCs (Fig-122 Cell Stem Cell 29, 116–130, January 6, 2022ures S5C and S5D). Additionally, in T21 NPCs, we observed
genome-wide chromatin accessibility changes that were com-
parable to the distribution of DEGs across the chromosomes
(Figures S4D and S4E). We verified that these changes were
not an artifact of T21 by repeating the normalization and anal-
ysis in the absence of reads from chromosome 21 and
observed 3,419 regions with reduced accessibility and 14,605
with increased accessibility, FDR < 0.01 (Figure S5E). As both
the differential gene expression and differential accessibility
analysis of NPCs indicate increased transcriptional activity in
T21 NPCs (Figures 4C and 5B), we assessed whether this cor-
responded to an increase in nascent RNA expression by label-
ing nascent transcripts in live NPCs via a 5-min Ethylene uridine
(EU) pulse and quantified EU-RNA fluorescence. We observed
ll
Article OPEN ACCESS
Figure 6. T21 induces senescence in NPCs
(A) Senescence-associated b-galactosidase staining (SA-b-gal, green) in euploid (left) and T21 (right) NPCs.
(B) Quantification of SA-b-gal staining intensity in the isogenic pair of NPCs. Each dot on the histogram represents a replicate experiment of
200 cells analyzed
for the isogenic pair of NPCs (t test).
(C) Correlation heatmap of the randomized permutation test of T21 iPSCs and NPC DEGs, compared with DEGs identified in various modes of senescence
induction (oxidative stress-induced senescence [OSIS], replicative senescence [RS], stress-induced senescence [SIPS], early and late oncogene-induced
senescence [OIS] and HMGB2 knockdown (HMGB2-KD). Color intensity (positive correlations are displayed in yellow, and negative correlations in black) and the
size of the circle are proportional to the enrichment coefficients.
(D) Gene expression heatmap of senescence marker genes as fold change of T21/E21 or senescent/control.
(legend continued on next page)
Cell Stem Cell 29, 116–130, January 6, 2022 123
ll
OPEN ACCESS Articlean 
20% increase in nascent RNA levels in T21 NPCs (Figures
S5F and S5G; Table S5B).
To gain an understanding of the histone modifications associ-
ated with the T21 NPC DARs, we utilized published ChIP-seq
data from H1-derived NPCs (Hawkins et al., 2010). While less
accessible regions are equally distributed on H3K4me3,
H3K4me, H3K27ac, H3K23ac, and regions not marked by any
of thesemodifications, themore accessible regions are predom-
inantly localized on regions with H3K4me and regions that are
not marked (Figure S5H; Table S5C). However, global enrich-
ment analysis of DARs on promoters, exons, introns, and inter-
genic regions (GENCODE GRCh37) revealed that less acces-
sible regions are significantly enriched on promoters and
exons, whereas the more accessible regions are significantly
depleted on promoters and exons, and neither the less nor the
more accessible DARs are significantly enriched with introns
and intergenic regions (Figure 5C). Consistent with these find-
ings, we observed that the less accessible regions are enriched
for the promoter mark H3K4me3 as well as active marks such as
H3K27ac and H3K23ac (Figures S5H and S5I). We found a sig-
nificant concordance between the less accessible promoters
and downregulated genes and globally observed that less
accessible regions are predominantly localized to the A-
compartment (Figures 5D and S5J; Tables S5D and S5E). The
biological processes associated with these downregulated
genes with reduced promoter accessibility includes cell cycle
regulation, transcriptional regulation, mRNA splicing, and chro-
matin organization, the biological processes not associated
with the 3D-genome reorganization (Figure S5K). While we
observed a significantly higher number of more accessible re-
gions genome-wide, this increased chromatin accessibility is
not specific to any annotated genomic regions.
Motif analysis of the less accessible regions in T21 NPCs re-
vealed significant enrichment of TF-binding motifs for SP1
(p value = 1e44) and NFY (p value = 1e41) (Figures 5E and
5F). NFY and SP1 are transcription factors involved in the regu-
lation of cell cycle and proliferation (Benatti et al., 2011; Grinstein
et al., 2002; Oh et al., 2007; Zhang et al., 2014b). While the
accessibility of regions harboring the SP1 and NFY motifs is
reduced, the transcript levels of SP1 and the different isoforms
of NFY (NFYA, NFYB, and NFYC) remain unchanged in T21
NPCs (Figure S5L). The downregulated gene promoters with
less accessible SP1 motifs are enriched for biological processes
involved in cell cycle, gene expression regulation, RNA splicing,
nucleosome assembly, and DNA damage response, whereas
the downregulated genes with less accessible promoters
harboring the NFY motifs are associated with transcriptional
regulation, mRNAmetabolic processes, and cellular senescence
(Figures 5G and 5H). Overall, our findings indicate that the
biological processes disrupted by T21 are driven by three dis-
ruptions in chromatin organization: (1) brain development-asso-(E) Venn diagram of overlapping downregulated (top) and upregulated (bottom) g
OSIS-induced senescent cells.
(F) Enrichment heatmap of the randomized permutation test between DEGs ide
differential ID TADs as well as reduced promoter accessibility. Color intensity (pos
the size of the circle are proportional to the enrichment coefficients.
(G and H) Gene ontology visualized in semantic similarity-based scatterplots for d
NPCs and OSIS-induced senescent cells.
124 Cell Stem Cell 29, 116–130, January 6, 2022ciated downregulated genes are predominantly localized within
reduced-ID TADs, (2) cell migration-associated upregulated
genes are localized within increased-ID TADs, and (3) cell prolif-
eration-, transcriptional regulation-, and chromosomal organiza-
tion-related downregulated genes are associated with reduced
promoter accessibility (Figures 4F, 4G, and S5K).
T21 induces NPC senescence
The global changes we observed in T21 NPCs including chromo-
somal introversion, disruption of LADs, and transcriptional and
chromatin state changes have all been identified as key features
of senescent cells (Criscione et al., 2016; Hernandez-Segura
et al., 2017; Lenain et al., 2017; Parry et al., 2018; Sati et al.,
2020; Zirkel et al., 2018). Senescence is a stress response that
quenches the proliferative capacity of cells and is predominantly
associated with cellular aging (Hayflick and Moorhead, 1961).
To assess whether T21 NPCs undergo senescence, we per-
formed senescence-associated (SA) b-galactosidase staining
and observed an
15-fold increase (p value = 0.001) in the num-
ber of senescent cells in T21 NPCs (Figures 6A and 6B; Table
S6A). Additionally, we observed increased HP1-alpha and
decreased HMGB1 and H3K27me3 levels, similar to what has
been observed in senescent cells (Figures S6A and S6B; Table
S6B) (Zirkel et al., 2018). To confirm that T21-NPCs do not un-
dergo premature differentiation, we compared our RNA-seq
data from iPSCs and NPCs with previously published data sets
from iPSC-derived neurons and NPC-derived astrocytes and ol-
igodendrocytes. We observed that euploid and T21 NPCs tran-
scriptionally clustered together and were distinct from all other
cell types (Figure S6C) (Garcı́a-León et al., 2018; Lin et al.,
2018; Tcw et al., 2017). Additionally, we stained for the astrocytic
marker GFAP, a senescencemarker p16INK4a, and the intermedi-
ate filament marker vimentin. Vimentin staining indicated that
euploid and T21 NPCs were morphologically distinct from astro-
cytes, andwhile NPC-derived astrocytes expressGFAP, they are
absent from both euploid and T21 NPCs, indicating that despite
the elevated level of p16INK4a in T21 NPCs, they do not undergo
premature differentiation to astrocytes (Figures S6D and S6E).
To assess the transcriptional similarities between T21 and se-
nescent cells, we utilized previously published transcriptomic
data where senescence was induced through oxidative stress
(oxidative stress-induced senescence [OSIS]), replicative stress
(replicative senescence [RS]), radiation ionization (stress-
induced premature senescence [SIPS]), overexpression of RAS
(oncogene-induced senescence [OIS]), and HMGB2 knockdown
(HMGB2-KD) (Alspach et al., 2014; Borràs-Fresneda et al., 2016;
Casella et al., 2019; Crowe et al., 2016; Herranz et al., 2015; Mar-
thandan et al., 2014; Nelson et al., 2016; Sati et al., 2020; Zirkel
et al., 2018). Permutation test of DEGs identified in T21 iPSCs
from our study (Iso-E and Iso-T) as well as from previously pub-
lished data set (Gonzales et al., 2018) showed a lack ofenes identified in NPCs harboring T21 and compared with DEGs identified in
ntified in both T21-NPCs and OSIS-induced senescent cells (SA-DEGs) with
itive correlations are displayed in green, and negative correlations in black) and
ownregulated (G) and upregulated (H) overlapping genes identified in both T21-
ll
Article OPEN ACCESScorrelation between DEGs identified in iPSCs and the various
forms of senescence induction (Figure 6C; Tables S6C and
S6D). In contrast, the DEGs identified in T21 NPCs (isogenic
pair, female pair, and male pair) had the highest correlation
compared with the DEGs identified in OSIS (p value < 1 3
1007) and, to a lesser extent, with RS and SIPS. We found
that T21 induced a marked expression increase of the SA cy-
clin-dependent kinase inhibitors (p16, p15, and p21), upregula-
tion of SA secretory phenotype (SASP) genes (CXCL12,
MMP24,MMP25, and LAMB1), and decreased expression of nu-
clear markers of senescence (LMNB1, P53, HMGB1, and
HMGB2) similar to senescent cells induced through various
mechanisms, particularly OSIS and RS (Figure 6D).
To identify the DEGs associated with senescence in T21
NPCs, we focused our analysis on the DEGs identified through
OSIS.We observed a significant overlap (p value <13 1007) be-
tween DEGs identified in T21 NPCs and those in OSIS, where
426 genes were downregulated and 584 genes were upregu-
lated in both (hereafter referred to as SA-DEGs; Figure 6E).
Next, we performed an enrichment analysis of the SA-DEGs
identified in T21 NPCs and identified a significant enrichment
of SA-downregulated genes in reduced-ID TADs (log2(obs/
exp) = 1.17, p value < 1 3 107), SA-upregulated genes in
increased-ID TADs (log2(obs/exp) = 0.62, p value < 1 3 10
7),
and a modest but significant enrichment (log2(obs/exp) = 0.32,
p value = 0.00083) of SA-downregulated genes with less acces-
sible promoters (Figure 6F; Table S6E). The SA-downregulated
genes are predominantly involved in cell differentiation, prolifer-
ation, and chromosome organization, whereas the SA-upregu-
lated genes are associated with cell migration, vesicle-mediated
transport, and inflammatory response (Figures 6G and 6H).While
the DEGs unique to T21 NPCs andOSIS seem to encode for cell-
type-specific processes, there are also shared biological pro-
cesses (Figures S6G–S6J). In summary, our findings indicate
that T21 NPCs exhibit features of senescent cells, such as
expression of senescent markers and a global transcriptional
signature similar to OSIS.
Senolytics ameliorate the T21-induced phenotype
in NPCs
Early clinical trials utilizing the senolytic drug DQ have shown
positive results in ameliorating SA features in individuals with
diabetic kidney disease (Hickson et al., 2019). Furthermore,
studies indicate that senolytic drugs not only selectively remove
senescent cells but also alleviate the transcriptional and hetero-
chromatin architecture changes induced during senescence
(Geng et al., 2019); however, the beneficial impact of senolytic
drugs on T21-induced senescence was unexplored. The
isogenic NPCs were treated with either DQ (400 nM of dasatinib
and 15 mMof quercetin) or vehicle (DMSO) for 5 days (Figure 7A).
We observed that DQ treatment significantly reduced the
elevated level of p16INK4a observed in the vehicle-treated T21
NPCs to that observed in vehicle-treated euploid NPCs (ANOVA
p value = 1.2 3 102) (Figures 7B and 7C; Table S7A). Next, we
interrogated the consequences of senolytic drugs on hetero-
chromatin architecture in T21 NPCs (Figure 7D). We found that
5 days of DQ treatment restored the heterochromatin distribu-
tion, as measured by the co-localization of H3K9me3 and
LMNB1 (ANOVA p value = 4.7 3 103) and LMNB1 levels(ANOVA p value = 2.4 3 104) to similar levels observed in the
euploid NPCs (Figures 7E and 7F; Tables S7B and S7C).
Subsequently, we sought to explore the impact of senolytic
treatment on the transcriptome of T21 NPCs. RNA sequencing
revealed that 
54% of the DEGs induced by T21 in NPCs
were ameliorated after 5 days of DQ treatment (Figure S7A;
Table S7D). Clustering of the DEGs identified 8 clusters: 2 clus-
ters where gene expression was rescued to the same transcrip-
tional levels as the euploid NPCs (clusters I and II), 2 clusters
where gene expression changes were mitigated (clusters III
and IV), 2 clusters with de novo transcriptional changes (clusters
V and VI), and 2 clusters that remained unchanged after DQ
treatment (clusters VII and VIII) (Figure 5G). The de novo upregu-
lated genes (cluster VI) represent a small percentage of the tran-
scriptional changes induced by DQ treatment (<9%) and include
a transcriptional response to the antioxidant activity of quercetin
aswell as neuronal apoptosis in response to the senolytic activity
of DQ-T21 NPCs exhibited a significant increase in apoxin-pos-
itive apoptotic cells after 6 h of DQ treatment (p value = 0.0004)
(Figure S7B; Table S7E).
We also observed amelioration of a significant number of
genes associated with cell cycle regulation upon DQ treatment
of T21 NPCs (cluster III). Notably, cell cycle dysfunction during
neurodevelopment is a hallmark of DS (Contestabile et al.,
2009; Contestabile et al., 2007). To examine cell cycle dynamics
in T21 NPCs, we utilized a 2-h BrdU incorporation to label newly
replicated/synthesized DNA (a marker of S-phase cells) coupled
with Ki-67 staining, amarker of all cycling cells irrespective of cell
cycle phase (Figures S7C and S7D). After DQ treatment of
trisomic NPCs, the proportion of proliferative (Ki-67+ and
SOX1+) NPCs was rescued to the same levels as euploid
NPCs (Iso-E(veh) = 62.1%, Iso-T(veh) = 44.9% and Iso-T(DQ) =
73.1%, ANOVA p value = 2.9 3 103) (Figure S7E; Table S41).
Furthermore, analysis of the NPCs that have traversed through
S-phase (BrdU+ and SOX1+) in the 2-h BrdU incorporation
window showed that cycling T21 NPCs led to significant reduc-
tion in S-phase cells (Iso-E(veh) = 25.3%and Iso-T(veh) = 13.4%,
Tukey’s padj < 1.0 3 104), similar to what has previously been
observed in the developing brain of individuals with DS (Fig-
ure S7F; Table S7F), which was rescued to euploid levels upon
DQ treatment (Iso-T(DQ) = 26.7%, Tukey’s padj = 1.0 3 104).
Additionally, senolytic treatment rescued the levels of upregu-
lated genes and resulted in de novo downregulation of genes
involved in ECM organization, cell adhesion and cell migration
in T21 NPCs (clusters II and V, respectively; Figure 7G). Thus,
we explored the impact of senolytic treatment on the migratory
abilities of T21 NPCs, a process disrupted in DS as well as in se-
nescent cells (Huo et al., 2018; Romanov et al., 2010) by treating
the isogenic pair of NPCs with either vehicle or DQ for 5 days in a
monolayer culture system (Figure S7G). Next, NPCs were al-
lowed to self-aggregate into neurospheres for an additional
5 days without vehicle or DQ (day 5–10), before embedding
them into 3D-Matrigel scaffold (Figure S7H). After 3 days (day
13), the embedded neurospheres were stained for the NPC
marker SOX1, and the number of migratory cells as well as the
distance traveled by each cell from the edge of the neurosphere
were assessed (Figure S7I). We found that DQ treatment of T21
NPCs rescued both the average number of migratory cells per
neurosphere (ANOVA p value < 1.0 3 104) as well as theCell Stem Cell 29, 116–130, January 6, 2022 125
ll
OPEN ACCESS Article
Figure 7. Senolytic drug alleviates the transcriptional and cellular dysfunctions induced by T21 in NPCs
(A) Schematic of DQ treatment experimental design in isogenic T21 NPCs.
(B and C) Immunofluorescence (B) and quantification (C) of p16INK4a (green) in euploid (Iso-E) and T21 (Iso-T, vehicle or DQ) NPCs (ANOVA).
(D) Immunofluorescence of H3K9me3 (green) and LMNB1 (red) in euploid (Iso-E) and T21 (Iso-T, vehicle or DQ) NPCs.
(E and F) Quantification of co-localization between H3K9me3 and LMNB1 (E) and LMNB1 staining intensity (F) in euploid (Iso-E) and T21 (Iso-T, vehicle or DQ)
NPCs. Each dot on the histogram represents a replicate, and 
200 nuclei were analyzed per condition (ANOVA).
(G) K-means clustering of the transcriptomic analysis of euploid and T21 NPCs (vehicle or DQ) identified in 8 distinct clusters. Values in heatmap are plotted as the
log2-fold change from the median.
(H) Gene ontology of biological process associated with the 8 transcriptional clusters identified in response to senolytic drug treatment.
(I) Enrichment plot of the randomized permutation test between the genes identified in the 8 different clusters with differential ID TADs as well as less accessible
promoters. Genes from cluster I and III as well clusters II and IV were merged for this analysis. Color intensity (positive correlations are displayed in green, and
negative correlations in black) and the size of the circle are proportional to the enrichment coefficients.
(J) Volcano plots of the differentially expressed genes (DEGs) identified in euploid NPCs treated with vehicle (DMSO) or DQ.
(K) Enrichment plot of the randomized permutation test betweenDEGs identified in DEGs identified after DQ treatment with differential ID TADs as well as reduced
promoter accessibility. Color intensity (positive correlations are displayed in green, and negative correlations in black) and the size of the circle are proportional to
the enrichment coefficients.average distance migrated by NPCs away from the edge of the
neurosphere (ANOVA p value < 1.0 3 104) (Figures S7J and
S7K; Table S7G).
Next, we sought to characterize the genes impacted by
senolytic treatment in terms of chromatin accessibility and126 Cell Stem Cell 29, 116–130, January 6, 2022TAD organization. We identified that the genes downregulated
in T21 NPCs and upregulated by DQ (clusters I and III) are
enriched for the geneswith reduced promoter accessibility (log2(-
obs/exp) = 1.28, p value < 1 3 107), whereas T21-upregulated
genes that were downregulated after DQ treatment (clusters II
ll
Article OPEN ACCESSand IV) are predominantly localized in increased-ID TADs (log2(-
obs/exp) = 0.37, p value < 1 3 107) (Figure 7I; Table S7H). The
de novo downregulated (cluster V) and upregulated (cluster VI)
genes after DQ treatment were both enriched for genes with
reduced promoter accessibility in T21 NPCs. Upregulated genes
identified in T21 NPCs that remained upregulated after DQ treat-
ment (cluster VIII) were localized in increased-ID TADs (log2(obs/
exp) = 0.42, p value < 1 3 107), whereas the downregulated
genes that were not impacted by DQ (cluster VII) were enriched
in reduced-ID TADs (log2(obs/exp) = 0.71, p value < 1 3 10
7)
and genes with less accessible promoters (log2(obs/exp) = 0.8,
p value < 13 107). Moreover, we observed that
20%of down-
regulated (clusters I/III) and
30%of upregulated genes (clusters
II/IV) ameliorated by senolytic drug treatment are SA-DEGs (Fig-
ure S7L). The ameliorated SA-DEGs are involved in biological
processes predominantly associated with cell cycle regulation
(clusters I/III) and cell migration (clusters II/IV).
To interrogate whether senolytics directly impact the transcrip-
tome of nonsenescent cells, we treated euploid NPCswith DQ for
5 days and performed RNA-seq. We identified 338 upregulated
and509downregulatedgenes in Iso-ENPCs treatedwithDQ (Fig-
ure 7J). We observed that the upregulated genes were predomi-
nantly enriched forbiological processesassociatedwith response
to oxidative stress, which is expected because of the antioxidant
activity of quercetin (Figure S7M). Conversely, we observed that
many of the keybiological processes associatedwith genes upre-
gulated in T21 NPCs were downregulated in DQ-treated euploid
NPCs, such as ECM organization, cell migration, cell differentia-
tion, and chemotaxis (Figure S7N). Accordingly, DQ treatment of
Iso-E NPCs significantly increased the average number of migra-
tory cells per neurosphere as well as the average distance
migrated by Iso-E NPCs away from the edge of the neurosphere
(Figures S7J and S7K; Table S7G). Finally, we observed that DQ
treatment of euploid NPCs induced downregulation of genes
localized to increased-ID TADs in T21 NPCs (log2(obs/exp) =
0.34, p value = 8.33 105), the region of the genome that harbors
manyof thegenes found tobeupregulated inT21NPCs (Figure7J;
Table S7I). In contrast, the genes upregulated in DQ-treated
euploid NPCs were predominantly enriched in unchanged-ID
TADs (log2(obs/exp) = 0.4, p value < 1 3 10
7) and in genes that
exhibit reduced promoter accessibility in T21 NPCs (log2(obs/
exp) = 0.95, p value < 13 107). In summary, senolytic drug treat-
ment of T21 NPCs reduced the number of senescent cells, allevi-
ated the heterochromatin architecture changes, ameliorated the
genome-wide transcriptional disruption, mitigated the cell migra-
tion deficits, and increased the population of proliferative cells.
DISCUSSION
T21 impacts the brain development of nearly every individual
with DS, and studies utilizing human-derived tissue samples
and DS mouse models have established that T21 induces
genome-wide transcriptional disruption (Bianco et al., 2016;
Gonzales et al., 2018; Letourneau et al., 2014; Olmos-Serrano
et al., 2016; Tyler and Haydar, 2013; Walus et al., 2016). How-
ever, the underlying molecular mechanisms contributing to an
altered global transcriptional profile as a consequence of T21
were unknown. Here, we utilized human iPSCs and iPSC-derived
NPCs to decode the molecular mechanisms underlying thesegenome-wide transcriptional changes. While both iPSCs and
NPCs exhibit a similar level of dosage-dependent transcriptional
upregulation of a subset of expressed genes on HSA21, unex-
pectedly, we found that T21 has a distinct genome-wide impact
on the transcriptome, epigenome, and 3D-genome of NPCs that
is absent in iPSCs. Even though a cell-type-specific response to
T21 is expected from the organ-specific etiologies observed in
DS (Weijerman and deWinter, 2010), the mechanisms governing
this cell-type-specific response remain unclear and require
further studies to fully elucidate this distinct response to T21.
Globally, we observed chromosomal introversion in NPCs
harboring T21,where every chromosomehas reduced interactions
with other chromosomes and increased self-interactions
compared with euploid NPCs. Furthermore, this increase in chro-
mosomal self-interactions occurs within TADs, where promoter-
enhancer interactions are typically observed (Robson et al.,
2019). In T21 NPCs, the chromosomal introversion is predomi-
nantly observed in regions of the genome associated with the nu-
clear lamina (B-compartment), and this feature is associated with
disruption of LADs, which leads to intranuclear B-compartment
bundles. In T21 NPCs, we find that the genome-wide transcrip-
tional upregulation is enriched within TADs in the B-compartment
that have increased chromosomal interaction densities. We find
that the genes associated with specific biological processes dis-
rupted in T21 NPCs are localized to distinct spatial genomic
regions.
Recent studies have established the promising therapeutic
potential of targeting senescent cells (Kirkland et al., 2017). Se-
nolytic drugs not only eliminate senescent cells but also restore
key cellular and molecular hallmarks associated with senes-
cence (Geng et al., 2019; Head et al., 2012; Zhang et al.,
2019). Chromosomal introversion and disruption of the nuclear
lamina are key features of senescence (Criscione et al., 2016; Zir-
kel et al., 2018), and our analysis of three different individuals
with DS revealed that T21 NPCs exhibit a transcriptional profile
that resembles senescence and display several key SA pheno-
types. We find that treating T21 NPCs with the senolytic drug
DQ ameliorated the SA genome-wide transcriptional and hetero-
chromatin architecture disruptions. Moreover, treatment with
senolytics also restored the migratory and proliferative capacity
in T21 NPCs, which are hallmark dysfunctions of the developing
brain in individuals with DS (Chakrabarti et al., 2007; Contesta-
bile et al., 2007, 2009).
In summary, our findings indicate that senescence may play a
key role in the neurodevelopmental pathogenesis of DS, and we
show that senolytics provide an exciting therapeutic avenue for
treating individuals with DS by restoring NPC dysfunctions
induced by T21.
Limitations of study
The iPSCs andNPCs generated for this study originated from indi-
viduals who passed away at a very young age, indicating the
severity of trisomy 21 in these individuals, and thus, samples
derived from older or live individuals may yield amilder phenotype
than what has been described in this study. There are currently
several approaches for generating iPSCs and NPCs, as well as
identifying chromatin accessibility, LADs, and spatial chromatin
organization, and thus,utilizing thesedifferent approachesorsam-
pling at different time points may yield varying results to thoseCell Stem Cell 29, 116–130, January 6, 2022 127
ll
OPEN ACCESS Articledescribed in this article. While we compared NPCs derived from
both male and female patients, the relatively small number of pa-
tients sampled also represents a limitation of our study. Further-
more, patient iPSCs and iPSC-derived NPCs are a valuable tool
tomodel the neurodevelopmental aspects of DS in vitro; however,
they may not fully recapitulate in vivo disease pathogenesis.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
d KEY RESOURCES TABLE
d RESOURCE AVAILABILITYB Lead contact
B Materials availability
B Data and code availability
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Induced pluripotent stem cell (iPSC) generation and
neural progenitor cell (NPC) differentiation
d METHOD DETAILS
B Immunofluorescence imaging and analysis
B Global Chromatin Profiling (GCP)
B Library preparation and analysis
d QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.
stem.2021.12.002.
ACKNOWLEDGMENTS
This work was supported by grants from the Alana foundation (received
byL.-H.T.), the LuMind Foundation (received by H.S.M.), Burroughs Wellcome
Fund (received by H.S.M.), UNCF-Merck postdoctoral fellowship (received by
H.S.M.), and the National Institute of Health (3U54HG008097-05S1, received
by J.D.J. and L.-H.T.). We would like to thank Dr. Jay Penney, Dr. Scarlett
Barker, Dr. Matheus Victor, Dr. Ravikiran Raju, Dr. Priyanka Narayan, Dr.
Ping-Chieh Pao, Dr. Emily Niederst, and the members of Tsai Lab for all the
constructive discussions and feedback on the manuscript. We would also
like to thank Ying Zhou andDr. Jennie Young of the Tsai Lab, aswell as Dr. Stu-
art Levine and his entire team at theMIT BioMicro Center. Wewould also like to
thank Dr. Stuart Orkin for sharing the isogenic pair iPSCs utilized in this manu-
script. We would also like to acknowledge that cartoon illustrations were
created with BioRender.com.
AUTHOR CONTRIBUTIONS
H.S.M. and L.-H.T. conceptualized and designed the project. H.S.M. and T.K.
generated iPSCs and NPCs. H.S.M., E.R.L., G.Y.A., and L.N.G. performed cell
culture experiments. H.S.M., L.A.W., E.R.L., G.Y.A., L.N.G., and F.A. per-
formed immunostaining, imaging, and IMARIS analysis. H.S.M., E.R.L.,
G.Y.A., and S.R. performed RNA-seq experiments. H.S.M., A.M., E.R.L., and
S.R. performed ATAC-seq library preparation. H.S.M. and S.R. performed
Hi-C library preparation, and H.S.M. performed ChIP-seq library preparation
and data analysis. H.S.M., A.M., and V.D. performed computational analysis.
J.M. andM.P. performed global chromatin profiling experiments, and J.M. and
H.S.M. performed data analysis. H.S.M., L.A.W., A.M., V.D., E.R.L., G.Y.A.,
J.M., M.P., and L.-H.T. wrote the manuscript with the input from all authors.
L.-H.T. and J.D.J. provided the resources for the project.
DECLARATION OF INTERESTS
The authors declare no competing interests.128 Cell Stem Cell 29, 116–130, January 6, 2022Received: May 10, 2020
Revised: August 30, 2021
Accepted: December 9, 2021
Published: January 6, 2022
REFERENCES
Abdennur, N., and Mirny, L.A. (2020). Cooler: scalable storage for Hi-C data
and other genomically labeled arrays. Bioinformatics 36, 311–316.
Allis, C.D., and Jenuwein, T. (2016). The molecular hallmarks of epigenetic
control. Nat. Rev. Genet. 17, 487–500.
Alspach, E., Flanagan, K.C., Luo, X., Ruhland, M.K., Huang, H., Pazolli, E.,
Donlin, M.J., Marsh, T., Piwnica-Worms, D., Monahan, J., et al. (2014).
P38MAPK plays a crucial role in stromal-mediated tumorigenesis. Cancer
Discov 4, 716–729.
Anders, S., Pyl, P.T., and Huber, W. (2015). HTSeq–a Python framework to
work with high-throughput sequencing data. Bioinformatics 31, 166–169.
Bally, B.P., Farmer, W.T., Jones, E.V., Jessa, S., Kacerovsky, J.B., Mayran, A.,
Peng, H., Lefebvre, J.L., Drouin, J., Hayer, A., et al. (2020). Human iPSC-
derived Down syndrome astrocytes display genome-wide perturbations in
gene expression, an altered adhesion profile, and increased cellular dynamics.
Hum. Mol. Genet. 29, 785–802.
Benatti, P., Dolfini, D., Viganò, A., Ravo, M., Weisz, A., and Imbriano, C. (2011).
Specific inhibition of NF-Y subunits triggers different cell proliferation defects.
Nucleic Acids Res 39, 5356–5368.
Benda, C.E. (1940). Growth disorder of the skull in mongolism. Am. J. Pathol.
16, 71–86.3.
Bianco, K., Gormley,M., Farrell, J., Zhou, Y., Oliverio, O., Tilden, H.,McMaster,
M., and Fisher, S.J. (2016). Placental transcriptomes in the common aneu-
ploidies reveal critical regions on the trisomic chromosomes and genome-
wide effects. Prenat. Diagn. 36, 812–822.
Borràs-Fresneda, M., Barquinero, J.F., Gomolka, M., Hornhardt, S., Rössler,
U., Armengol, G., and Barrios, L. (2016). Differences in DNA repair capacity,
cell death and transcriptional response after irradiation between a radiosensi-
tive and a radioresistant cell line. Sci. Rep. 6, 27043.
Buenrostro, J.D., Giresi, P.G., Zaba, L.C., Chang, H.Y., and Greenleaf, W.J.
(2013). Transposition of native chromatin for fast and sensitive epigenomic
profiling of open chromatin, DNA-binding proteins and nucleosome position.
Nat. Methods 10, 1213–1218.
Buenrostro, J.D., Wu, B., Chang, H.Y., and Greenleaf, W.J. (2015). ATAC-seq:
A method for assaying chromatin accessibility genome-wide. Curr. Protoc.
Mol. Biol. 109, 21–29, 1–21.29.9.
Casella, G., Munk, R., Kim, K.M., Piao, Y., De, S., Abdelmohsen, K., and
Gorospe, M. (2019). Transcriptome signature of cellular senescence. Nucleic
Acids Res 47, 7294–7305.
Chakrabarti, L., Galdzicki, Z., and Haydar, T.F. (2007). Defects in embryonic
neurogenesis and initial synapse formation in the forebrain of the Ts65Dn
mouse model of Down syndrome. J. Neurosci. 27, 11483–11495.
Chambers, S.M., Fasano, C.A., Papapetrou, E.P., Tomishima, M., Sadelain,
M., and Studer, L. (2009). Highly efficient neural conversion of human ES
and iPS cells by dual inhibition of SMAD signaling. Nat. Biotechnol. 27,
275–280.
Contestabile, A., Fila, T., Bartesaghi, R., and Ciani, E. (2009). Cell cycle elon-
gation impairs proliferation of cerebellar granule cell precursors in the
Ts65Dn mouse, an animal model for Down syndrome. Brain Pathol 19,
224–237.
Contestabile, A., Fila, T., Ceccarelli, C., Bonasoni, P., Bonapace, L., Santini,
D., Bartesaghi, R., and Ciani, E. (2007). Cell cycle alteration and decreased
cell proliferation in the hippocampal dentate gyrus and in the neocortical
germinal matrix of fetuses with Down syndrome and in Ts65Dn mice.
Hippocampus 17, 665–678.
Creech, A.L., Taylor, J.E., Maier, V.K., Wu, X., Feeney, C.M., Udeshi, N.D.,
Peach, S.E., Boehm, J.S., Lee, J.T., Carr, S.A., and Jaffe, J.D. (2015).
Building the Connectivity Map of epigenetics: chromatin profiling by quantita-
tive targeted mass spectrometry. Methods 72, 57–64.
ll
Article OPEN ACCESSCriscione, S.W., DeCecco,M., Siranosian, B., Zhang, Y., Kreiling, J.A., Sedivy,
J.M., and Neretti, N. (2016). Reorganization of chromosome architecture in
replicative cellular senescence. Sci. Adv. 2, e1500882.
Crowe, E.P., Tuzer, F., Gregory, B.D., Donahue, G., Gosai, S.J., Cohen, J.,
Leung, Y.Y., Yetkin, E., Nativio, R., Wang, L.S., et al. (2016). Changes in the
transcriptome of human astrocytes accompanying oxidative stress-induced
senescence. Front. Aging Neurosci. 8, 208.
Dixon, J.R., Selvaraj, S., Yue, F., Kim, A., Li, Y., Shen, Y., Hu, M., Liu, J.S., and
Ren, B. (2012). Topological domains in mammalian genomes identified by
analysis of chromatin interactions. Nature 485, 376–380.
Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut,
P., Chaisson,M., andGingeras, T.R. (2013). STAR: ultrafast universal RNA-seq
aligner. Bioinformatics 29, 15–21.
Down, J.L. (1995). Observations on an ethnic classification of idiots. 1866.
Ment. Retard. 33, 54–56.
Du€rrbaum, M., Kuznetsova, A.Y., Passerini, V., Stingele, S., Stoehr, G., and
Storchová, Z. (2014). Unique features of the transcriptional response to model
aneuploidy in human cells. BMC Genomics 15, 139.
Eden, E., Navon, R., Steinfeld, I., Lipson, D., and Yakhini, Z. (2009). GOrilla: a
tool for discovery and visualization of enriched GO terms in ranked gene lists.
BMC Bioinformatics 10, 48.
Egan, B., Yuan, C.C., Craske, M.L., Labhart, P., Guler, G.D., Arnott, D., Maile,
T.M., Busby, J., Henry, C., Kelly, T.K., et al. (2016). An alternative approach to
ChIP-seq normalization enables detection of genome-wide changes in histone
H3 lysine 27 trimethylation upon EZH2 inhibition. PLoS One 11, e0166438.
FitzPatrick, D.R., Ramsay, J., McGill, N.I., Shade, M., Carothers, A.D., and
Hastie, N.D. (2002). Transcriptome analysis of human autosomal trisomy.
Hum. Mol. Genet. 11, 3249–3256.
Garcı́a-León, J.A., Kumar, M., Boon, R., Chau, D., One, J., Wolfs, E.,
Eggermont, K., Berckmans, P., Gunhanlar, N., de Vrij, F., et al. (2018).
SOX10 single transcription factor-based fast and efficient generation of oligo-
dendrocytes from human pluripotent stem cells. Stem Cell Rep 10, 655–672.
Geng, L., Liu, Z., Zhang,W., Li,W.,Wu, Z.,Wang,W., Ren, R., Su, Y., Wang, P.,
Sun, L., et al. (2019). Chemical screen identifies a geroprotective role of quer-
cetin in premature aging. Protein Cell 10, 417–435.
Gonzales, P.K., Roberts, C.M., Fonte, V., Jacobsen, C., Stein, G.H., and Link,
C.D. (2018). Transcriptome analysis of genetically matched human induced
pluripotent stem cells disomic or trisomic for chromosome 21. PLoS One 13,
e0194581.
Grinstein, E., Jundt, F., Weinert, I., Wernet, P., and Royer, H.D. (2002). Sp1 as
G1 cell cycle phase specific transcription factor in epithelial cells. Oncogene
21, 1485–1492.
Guidi, S., Ciani, E., Bonasoni, P., Santini, D., and Bartesaghi, R. (2011).
Widespread proliferation impairment and hypocellularity in the cerebellum of
fetuses with down syndrome. Brain Pathol 21, 361–373.
Hawkins, R.D., Hon, G.C., Lee, L.K., Ngo, Q., Lister, R., Pelizzola, M., Edsall,
L.E., Kuan, S., Luu, Y., Klugman, S., et al. (2010). Distinct epigenomic land-
scapes of pluripotent and lineage-committed human cells. Cell Stem Cell 6,
479–491.
Hayflick, L., and Moorhead, P.S. (1961). The serial cultivation of human diploid
cell strains. Exp. Cell Res. 25, 585–621.
Head, E., Powell, D., Gold, B.T., and Schmitt, F.A. (2012). Alzheimer’s disease
in Down syndrome. Eur. J. Neurodegener. Dis. 1, 353–364.
Heinz, S., Texari, L., Hayes, M.G.B., Urbanowski, M., Chang, M.W., Givarkes,
N., Rialdi, A., White, K.M., Albrecht, R.A., Pache, L., et al. (2018). Transcription
elongation can affect genome 3D structure. Cell 174, 1522–1536.e22.
Hernandez-Segura, A., de Jong, T.V., Melov, S., Guryev, V., Campisi, J., and
Demaria, M. (2017). Unmasking transcriptional heterogeneity in senescent
cells. Curr. Biol. 27, 2652–2660.e4.
Herranz, N., Gallage, S., Mellone, M., Wuestefeld, T., Klotz, S., Hanley, C.J.,
Raguz, S., Acosta, J.C., Innes, A.J., Banito, A., et al. (2015). mTOR regulates
MAPKAPK2 translation to control the senescence-associated secretory
phenotype. Nat. Cell Biol. 17, 1205–1217.Hickson, L.J., Langhi Prata, L.G.P., Bobart, S.A., Evans, T.K., Giorgadze, N.,
Hashmi, S.K., Herrmann, S.M., Jensen, M.D., Jia, Q., Jordan, K.L., et al.
(2019). Senolytics decrease senescent cells in humans: preliminary report
from a clinical trial of dasatinib plus quercetin in individuals with diabetic kidney
disease. EBioMedicine 47, 446–456.
Huo, H.Q., Qu, Z.Y., Yuan, F., Ma, L., Yao, L., Xu, M., Hu, Y., Ji, J.,
Bhattacharyya, A., Zhang, S.C., and Liu, Y. (2018). Modeling Down syndrome
with patient iPSCs reveals cellular and migration deficits of GABAergic neu-
rons. Stem Cell Rep 10, 1251–1266.
Jaffe, J.D., Wang, Y., Chan, H.M., Zhang, J., Huether, R., Kryukov, G.V.,
Bhang, H.E., Taylor, J.E., Hu, M., Englund, N.P., et al. (2013). Global chromatin
profiling reveals NSD2 mutations in pediatric acute lymphoblastic leukemia.
Nat. Genet. 45, 1386–1391.
Kirkland, J.L., Tchkonia, T., Zhu, Y., Niedernhofer, L.J., and Robbins, P.D.
(2017). The clinical potential of senolytic drugs. J. Am. Geriatr. Soc. 65,
2297–2301.
Klemm, S.L., Shipony, Z., and Greenleaf, W.J. (2019). Chromatin accessibility
and the regulatory epigenome. Nat. Rev. Genet. 20, 207–220.
Lane, A.A., Chapuy, B., Lin, C.Y., Tivey, T., Li, H., Townsend, E.C., van
Bodegom, D., Day, T.A., Wu, S.C., Liu, H., et al. (2014). Triplication of a
21q22 region contributes to B cell transformation through HMGN1 overex-
pression and loss of histone H3 Lys27 trimethylation. Nat. Genet. 46, 618–623.
Langmead, B., and Salzberg, S.L. (2012). Fast gapped-read alignment with
Bowtie 2. Nat. Methods 9, 357–359.
Lejeune, J., Gauthier, M., and Turpin, R. (1959). Human chromosomes in tissue
cultures. C. R. Hebd. Seances Acad. Sci. 248, 602–603.
Lenain, C., de Graaf, C.A., Pagie, L., Visser, N.L., de Haas, M., de Vries, S.S.,
Peric-Hupkes, D., van Steensel, B., and Peeper, D.S. (2017). Massive reshap-
ing of genome-nuclear lamina interactions during oncogene-induced senes-
cence. Genome Res 27, 1634–1644.
Letourneau, A., Santoni, F.A., Bonilla, X., Sailani, M.R., Gonzalez, D., Kind, J.,
Chevalier, C., Thurman, R., Sandstrom, R.S., Hibaoui, Y., et al. (2014).
Domains of genome-wide gene expression dysregulation in Down’s syn-
drome. Nature 508, 345–350.
Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with
Burrows-Wheeler transform. Bioinformatics 25, 1754–1760.
Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G.,
Abecasis, G., and Durbin, R.; 1000 Genome Project Data Processing
Subgroup (2009). The Sequence Alignment/Map format and SAMtools.
Bioinformatics 25, 2078–2079.
Lieberman-Aiden, E., van Berkum, N.L., Williams, L., Imakaev, M., Ragoczy,
T., Telling, A., Amit, I., Lajoie, B.R., Sabo, P.J., Dorschner, M.O., et al.
(2009). Comprehensive mapping of long-range interactions reveals folding
principles of the human genome. Science 326, 289–293.
Lin, Y.T., Seo, J., Gao, F., Feldman, H.M., Wen, H.L., Penney, J., Cam, H.P.,
Gjoneska, E., Raja, W.K., Cheng, J., et al. (2018). APOE4 causes widespread
molecular and cellular alterations associated with Alzheimer’s disease pheno-
types in human iPSC-derived brain cell types. Neuron 98, 1141–1154.e7.
Litichevskiy, L., Peckner, R., Abelin, J.G., Asiedu, J.K., Creech, A.L., Davis,
J.F., Davison, D., Dunning, C.M., Egertson, J.D., Egri, S., et al. (2018). A library
of phosphoproteomic and chromatin signatures for characterizing cellular re-
sponses to drug perturbations. Cell Syst 6, 424–443.e7.
Liu, F., Wang, L., Perna, F., and Nimer, S.D. (2016). Beyond transcription fac-
tors: how oncogenic signalling reshapes the epigenetic landscape. Nat. Rev.
Cancer 16, 359–372.
Love, M.I., Huber, W., and Anders, S. (2014). Moderated estimation of fold
change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550.
Lu, L., Chen, X., Sanders, D., Qian, S., and Zhong, X. (2015). High-resolution
mapping of H4K16 and H3K23 acetylation reveals conserved and unique dis-
tribution patterns in Arabidopsis and rice. Epigenetics 10, 1044–1053.
Maclean, G.A., Menne, T.F., Guo, G., Sanchez, D.J., Park, I.H., Daley, G.Q.,
and Orkin, S.H. (2012). Altered hematopoiesis in trisomy 21 as revealed
through in vitro differentiation of isogenic human pluripotent cells. Proc. Natl.
Acad. Sci. U. S. A. 109, 17567–17572.Cell Stem Cell 29, 116–130, January 6, 2022 129
ll
OPEN ACCESS ArticleMarthandan, S., Priebe, S., Hemmerich, P., Klement, K., and Diekmann, S.
(2014). Long-term quiescent fibroblast cells transit into senescence. PLoS
One 9, e115597.
McLaughlin, K., Flyamer, I.M., Thomson, J.P., Mjoseng, H.K., Shukla, R.,
Williamson, I., Grimes, G.R., Illingworth, R.S., Adams, I.R., Pennings, S.,
et al. (2019). DNAmethylation directs polycomb-dependent 3D genome re-or-
ganization in naive pluripotency. Cell Rep 29, 1974–1985.e6.
Moldrich, R.X., Dauphinot, L., Laffaire, J., Vitalis, T., Hérault, Y., Beart, P.M.,
Rossier, J., Vivien, D., Gehrig, C., Antonarakis, S.E., et al. (2009).
Proliferation deficits and gene expression dysregulation in Down’s syndrome
(Ts1Cje) neural progenitor cells cultured from neurospheres. J. Neurosci.
Res. 87, 3143–3152.
Nelson, D.M., Jaber-Hijazi, F., Cole, J.J., Robertson, N.A., Pawlikowski, J.S.,
Norris, K.T., Criscione, S.W., Pchelintsev, N.A., Piscitello, D., Stong, N.,
et al. (2016). Mapping H4K20me3 onto the chromatin landscape of senescent
cells indicates a function in control of cell senescence and tumor suppression
through preservation of genetic and epigenetic stability. Genome Biol 17, 158.
Nora, E.P., Lajoie, B.R., Schulz, E.G., Giorgetti, L., Okamoto, I., Servant, N.,
Piolot, T., van Berkum, N.L., Meisig, J., Sedat, J., et al. (2012). Spatial partition-
ing of the regulatory landscape of the X-inactivation centre. Nature 485,
381–385.
Oh, J.E., Han, J.A., and Hwang, E.S. (2007). Downregulation of transcription
factor, Sp1, during cellular senescence. Biochem. Biophys. Res. Commun.
353, 86–91.
Olmos-Serrano, J.L., Kang, H.J., Tyler, W.A., Silbereis, J.C., Cheng, F., Zhu,
Y., Pletikos, M., Jankovic-Rapan, L., Cramer, N.P., Galdzicki, Z., et al.
(2016). Down syndrome developmental brain transcriptome reveals defective
oligodendrocyte differentiation and myelination. Neuron 89, 1208–1222.
Parry, A.J., Hoare, M., Bihary, D., Ha€nsel-Hertsch, R., Smith, S., Tomimatsu,
K., Mannion, E., Smith, A., D’Santos, P., Russell, I.A., et al. (2018). NOTCH-
mediated non-cell autonomous regulation of chromatin structure during
senescence. Nat. Commun. 9, 1840.
Pinter, J.D., Eliez, S., Schmitt, J.E., Capone, G.T., and Reiss, A.L. (2001).
Neuroanatomy of Down’s syndrome: a high-resolution MRI study. Am. J.
Psychiatry 158, 1659–1665.
Pombo, A., and Dillon, N. (2015). Three-dimensional genome architecture:
players and mechanisms. Nat. Rev. Mol. Cell Biol. 16, 245–257.
Raja, W.K., Mungenast, A.E., Lin, Y.T., Ko, T., Abdurrob, F., Seo, J., and Tsai,
L.H. (2016). Self-organizing 3D human neural tissue derived from induced
pluripotent stem cells recapitulate Alzheimer’s disease phenotypes. PLoS
One 11, e0161969.
Ramı́rez, F., Bhardwaj, V., Arrigoni, L., Lam, K.C., Gru€ning, B.A., Villaveces, J.,
Habermann, B., Akhtar, A., and Manke, T. (2018). High-resolution TADs reveal
DNA sequences underlying genome organization in flies. Nat. Commun.
9, 189.
Ramı́rez, F., Ryan, D.P., Gru€ning, B., Bhardwaj, V., Kilpert, F., Richter, A.S.,
Heyne, S., Du€ndar, F., and Manke, T. (2016). deepTools2: a next generation
web server for deep-sequencing data analysis. Nucleic Acids Res 44,
W160–W165.
Rao, S.S., Huntley, M.H., Durand, N.C., Stamenova, E.K., Bochkov, I.D.,
Robinson, J.T., Sanborn, A.L., Machol, I., Omer, A.D., Lander, E.S., and
Aiden, E.L. (2014). A 3D map of the human genome at kilobase resolution re-
veals principles of chromatin looping. Cell 159, 1665–1680.
Risso, D., Ngai, J., Speed, T.P., and Dudoit, S. (2014). Normalization of RNA-
seq data using factor analysis of control genes or samples. Nat. Biotechnol.
32, 896–902.
Robson, M.I., Ringel, A.R., and Mundlos, S. (2019). Regulatory landscaping:
how enhancer-promoter communication is sculpted in 3D. Mol. Cell 74,
1110–1122.
Romanov, V.S., Abramova, M.V., Svetlikova, S.B., Bykova, T.V., Zubova, S.G.,
Aksenov, N.D., Fornace, A.J., Jr., Pospelova, T.V., and Pospelov, V.A. (2010).130 Cell Stem Cell 29, 116–130, January 6, 2022p21(Waf1) is required for cellular senescence but not for cell cycle arrest
induced by the HDAC inhibitor sodium butyrate. Cell Cycle 9, 3945–3955.
Ross-Innes, C.S., Stark, R., Teschendorff, A.E., Holmes, K.A., Ali, H.R.,
Dunning, M.J., Brown, G.D., Gojis, O., Ellis, I.O., Green, A.R., et al. (2012).
Differential oestrogen receptor binding is associated with clinical outcome in
breast cancer. Nature 481, 389–393.
Sati, S., Bonev, B., Szabo, Q., Jost, D., Bensadoun, P., Serra, F., Loubiere, V.,
Papadopoulos, G.L., Rivera-Mulia, J.C., Fritsch, L., et al. (2020). 4D genome
rewiring during oncogene-induced and replicative senescence. Mol. Cell 78,
522–538.e9.
Schoenfelder, S., and Fraser, P. (2019). Long-range enhancer-promoter con-
tacts in gene expression control. Nat. Rev. Genet. 20, 437–455.
Sheltzer, J.M., Torres, E.M., Dunham, M.J., and Amon, A. (2012).
Transcriptional consequences of aneuploidy. Proc. Natl. Acad. Sci. U. S. A.
109, 12644–12649.
Shevelyov, Y.Y., and Ulianov, S.V. (2019). The nuclear lamina as an organizer
of chromosome architecture. Cells 8, 136d.
Stagni, F., Giacomini, A., Emili, M., Guidi, S., and Bartesaghi, R. (2018).
Neurogenesis impairment: an early developmental defect in Down syndrome.
Free Radic. Biol. Med. 114, 15–32.
Supek, F., Bosnjak, M., Skunca, N., and Smuc, T. (2011). REVIGO summarizes
and visualizes long lists of gene ontology terms. PLoS One 6, e21800.
Tcw, J., Wang, M., Pimenova, A.A., Bowles, K.R., Hartley, B.J., Lacin, E.,
Machlovi, S.I., Abdelaal, R., Karch, C.M., Phatnani, H., et al. (2017). An efficient
platform for astrocyte differentiation from human induced pluripotent stem
cells. Stem Cell Rep 9, 600–614.
Topol, A., Tran, N.N., and Brennand, K.J. (2015). A guide to generating and us-
ing hiPSC derived NPCs for the study of neurological diseases. J. Vis. Exp. 96,
e52495.
Tyler, W.A., and Haydar, T.F. (2013). Multiplex genetic fate mapping reveals a
novel route of neocortical neurogenesis, which is altered in the Ts65Dn mouse
model of Down syndrome. J. Neurosci. 33, 5106–5119.
Walus, M., Kida, E., Rabe, A., Albertini, G., and Golabek, A.A. (2016).
Widespread cerebellar transcriptome changes in Ts65Dn Down syndrome
mouse model after lifelong running. Behav. Brain Res. 296, 35–46.
Wangsa, D., Braun, R., Stuelten, C.H., Brown, M., Bauer, K.M., Emons, G.,
Weston, L.A., Hu, Y., Yang, H.H., Vila-Casadesús, M., et al. (2019). Induced
chromosomal aneuploidy results in global and consistent deregulation of the
transcriptome of cancer cells. Neoplasia 21, 721–729.
Weick, J.P., Held, D.L., Bonadurer, G.F., 3rd, Doers, M.E., Liu, Y., Maguire, C.,
Clark, A., Knackert, J.A., Molinarolo, K., Musser, M., et al. (2013). Deficits in hu-
man trisomy 21 iPSCs and neurons. Proc. Natl. Acad. Sci. U. S. A. 110,
9962–9967.
Weijerman, M.E., and de Winter, J.P. (2010). Clinical practice. The care of chil-
dren with Down syndrome. Eur. J. Pediatr. 169, 1445–1452.
Zhang, J.P., Zhang, H., Wang, H.B., Li, Y.X., Liu, G.H., Xing, S., Li, M.Z., and
Zeng, M.S. (2014b). Down-regulation of Sp1 suppresses cell proliferation, clo-
nogenicity and the expressions of stem cell markers in nasopharyngeal carci-
noma. J. Transl. Med. 12, 222.
Zhang, P., Kishimoto, Y., Grammatikakis, I., Gottimukkala, K., Cutler, R.G.,
Zhang, S., Abdelmohsen, K., Bohr, V.A., Misra Sen, J., Gorospe, M., and
Mattson, M.P. (2019). Senolytic therapy alleviates Abeta-associated oligoden-
drocyte progenitor cell senescence and cognitive deficits in an Alzheimer’s
disease model. Nat. Neurosci. 22, 719–728.
Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E.,
Nusbaum, C., Myers, R.M., Brown, M., Li, W., and Liu, X.S. (2008). Model-
based analysis of ChIP-Seq (MACS). Genome Biol 9, R137.
Zirkel, A., Nikolic, M., Sofiadis, K., Mallm, J.P., Brackley, C.A., Gothe, H.,
Drechsel, O., Becker, C., Altmu€ller, J., Josipovic, N., et al. (2018). HMGB2
loss upon senescence entry disrupts genomic organization and induces
CTCF clustering across cell types. Mol. Cell 70, 730–744.e6.
ll
Article OPEN ACCESSSTAR+METHODSKEY RESOURCES TABLEREAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
OCT-4 Cell Signaling Technology Cat# 2890, RRID:AB_2167725
Nanog Cell Signaling Technology Cat# 4893, RRID:AB_10548762
SOX1 R&D Systems Cat# AF3369, RRID:AB_2239879
Nestin Millipore Sigma Cat# MAB5326, RRID:AB_2251134
PAX6 R&D Systems Cat# AF8150, RRID:AB_2827378
H3K9me3 Abcam Cat# ab176916, RRID:AB_2797591
HP1a Cell Signaling Technology Cat# 2616, RRID:AB_2070987
H3K27me3 Abcam Cat# ab6002, RRID:AB_305237
HMGB1 Thermo Fisher Scientific Cat# PA1-16926, RRID:AB_2248274
LMNB1 Abcam Cat# ab16048, RRID:AB_10107828
p16INK4a Thermo Fisher Scientific Cat# MA5-32133, RRID:AB_2809424
KI-67 Abcam Cat# ab15580, RRID:AB_443209
Vimentin Abcam Cat# ab24525, RRID:AB_778824
GFAP Millipore Cat# AB5804, RRID:AB_2109645
BrdU Cell Signaling Technology Cat# 5292, RRID:AB_10548898
Alexa-Fluor 488 (Donkey anti-Goat) Thermo Fisher Scientific Cat# A-11015, RRID:AB_2534082
Alexa-Fluor 488 (Donkey anti-Rabbit) Thermo Fisher Scientific Cat# A-21206, RRID:AB_2535792
Alexa-Fluor 488 (Donkey anti-Mouse) Thermo Fisher Scientific Cat# A-21202, RRID:AB_141607
Alexa-Fluor 488 (Donkey anti-Sheep) Thermo Fisher Scientific Cat# A-11015, RRID:AB_2534082
Alexa-Fluor 555 (Donkey anti-Goat) Thermo Fisher Scientific Cat# A-21432, RRID:AB_2535853
Alexa-Fluor 555 (Donkey anti-Rabbit) Thermo Fisher Scientific Cat# A-31572, RRID:AB_162543
Alexa-Fluor 555 (Donkey anti-Mouse) Thermo Fisher Scientific Cat# A-31570, RRID:AB_2536180
Alexa-Fluor 555 (Donkey anti-Sheep) Thermo Fisher Scientific Cat# A-21099, RRID:AB_2535753
Alexa-Fluor 647 (Donkey anti-Goat) Thermo Fisher Scientific Cat# A-21447, RRID:AB_2535864
Alexa-Fluor 647 (Donkey anti-Rabbit) Thermo Fisher Scientific Cat# A-31573, RRID:AB_2536183
Alexa-Fluor 647 (Donkey anti-Mouse) Thermo Fisher Scientific Cat# A-31571, RRID:AB_162542
Alexa-Fluor 647 (Donkey anti-Sheep) Thermo Fisher Scientific Cat# A-21448, RRID:AB_2535865
DAPI Thermo Fisher Scientific Cat# D1306, RRID:AB_2629482
Biological samples
Iso-E and Iso-T iPSCs (AG05397) Coriell (Maclean et al., 2012) Cat# AG05397
Ma-E iPSCs Cohen, McLean, HMS (Raja et al., 2016) Cat# CS-0020-01
Ma-T Fibroblast Coriell Cat# AG06922
Fe-E iPSCs WiCell Cat# WA09
Fe-T Fibroblasts Coriell Cat# GM04616
CF1 Mouse embryonic fibroblasts, Thermo Fisher Scientific Cat# A34181
irradiated
Chemicals, peptides, and recombinant proteins
Fluoromount-G Electron Microscopy Sciences Cat# 17984-25
PBS, pH 7.4 Thermo Fisher Scientific Cat# 10010049
Triton X-100 Sigma-Aldrich Cat# T8787-100ML
Paraformaldehyde Electron Microscopy Sciences Cat# 15714-S
5-Bromo-20-deoxyuridine Sigma-Aldrich Cat# B5002-5G
Corning Matrigel hESC-Qualified Matrix VWR International Cat# 354277
(Continued on next page)
Cell Stem Cell 29, 116–130.e1–e7, January 6, 2022 e1
ll
OPEN ACCESS Article
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Corning Matrigel Basement Membrane VWR International Cat# 356254
Matrix
DMEM/F12 Glutamax Life Technologies Cat# 10565-042
Knockout Serum Replacement Life Technologies Cat# 10828028
MEM non-Essential amino acid solution Sigma-Aldrich Cat# M7145
Glutamax Thermo Fisher Scientific Cat# 35050-079
recombinant human FGF-basic PeproTech Cat# 100-18B
2-mercaptoethanol Sigma-Aldrich Cat# M6250
mTeSR1 medium Stemcell technologies Cat# 85850
ReLeSR Stemcell Technology Cat# 05872
AggreWell800 Stemcell Technology Cat# 34811
Neurobasal Thermo Fisher Scientific Cat# 21103049
N-2 Supplement Invitrogen Cat# 17502-048
B-27 Serum-Free Supplement Gibco Cat# 12587010
insulin Thermo Fisher Scientific Cat# 41400045
Pen/strep Gemini Bio-Products Cat# 400-109
Dorsomorphin PeproTech Cat# 8666430
SB431542 Tocris Cat# 1614
STEMdiff Neural Rosette Selection Stemcell technologies Cat# 05832
Reagent
Sulfuric acid Sigma-Aldrich Cat# 339741-500ML
Trichloroacetic acid BDH Cat# BDH0310500G
Trypsin Promega Cat# V5113
C18 Sep-Pak cartridges Waters Cat# 186002318
TrypLE Thermo Fisher Scientific Cat# 12605028
Igepal CA-630 (NP-40) Sigma-Aldrich Cat# I8896
AMPure beads Beckman Coulter Cat# A63880
Critical commercial assays
CytoTune-iPS Sendai Reprogramming kit Thermofisher scientific Cat# A16518
Kapa mRNA hyperprep kit for standard Roche Cat# KK8540
RNA libraries
Nextera DNA Library Prep Kit Illumina Cat# FC-121-1030
Qiagen MinElute PCR Purification Kit Qiagen Cat# 28006
Arima Hi-C kit Arima Genomics N/A
Kappa Hyperprep Kit Roche Cat# KK8502
Deposited data
Hi-C This paper GEO: GSE185192
LMNB1 ChIP-seq This paper GEO: GSE185192
RNA-seq This paper GEO: GSE185192
ATAC-seq This paper GEO: GSE185192
Senescence RNA-seq Casella et al., 2019; Crowe et al., 2016; GEO: GSE130727, GSE58910, GSE56293,
Alspach et al., 2014; Sati et al., 2020; Zirkel GSE130306, GSE98448, GSE80207
et al., 2018; Borras-Fresneda et al., 2016;
RNA-seq of iPSC derived from Down Gonzales et al., 2018 GEO: GSE101942
syndrome and control individuals
RNA-seq of iPSC derived neurons, Garcia-Leon et al., 2018; Lin et al., 2018; GEO: GSE106984, GSE102956, GSE97904
astrocytes and OPCs Tcw et al., 2017
ChIP-seq of Histone Modifications Hawkins et al., 2010 GEO: GSM956008
Experimental models: Cell lines
Iso-E and Iso-T (AG05397) Coriell (Maclean et al., 2012) Cat# AG05397
(Continued on next page)
e2 Cell Stem Cell 29, 116–130.e1–e7, January 6, 2022
ll
Article OPEN ACCESS
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Ma-E Cohen, McLean, HMS (Raja et al., 2016) Cat# CS-0020-01
Ma-T Coriell Cat# AG06922
Fe-E WiCell Cat# WA09
Fe-T Coriell Cat# GM04616
CF1 Mouse embryonic fibroblasts, Thermo Fisher Scientific Cat# A34181
irradiated
Software and algorithms
Imarisx64 9.2.1 Bitplane http://www.bitplane.com/Default.aspx
ZEN imaging software Carl Zeiss https://www.zeiss.com/microscopy/int/
products/microscope-software/zen.html
STAR (v.2.6.1a) Dobin et al., 2013 https://github.com/alexdobin/STAR
HTSeq Anders et al., 2015 https://htseq.readthedocs.io/en/master/
RUVSeq Risso et al., 2014 https://bioconductor.org/packages/
release/bioc/html/RUVSeq.html
DESeq2 Love et al., 2014 https://bioconductor.org/packages/
release/bioc/html/DESeq2.html
Bowtie2 Langmead and Salzberg, 2012 http://bowtie-bio.sourceforge.net/
bowtie2/index.shtml
Samtools Li et al., 2009 http://samtools.sourceforge.net/
Homer http://homer.ucsd.edu/homer/
interactions2/index.html
MACS2.1.0 Zhang et al., 2008 https://github.com/taoliu/MACS
R https://www.r-project.org/
Corrplot https://github.com/taiyun/corrplot
Diffbind v1.16.3 Ross-Innes et al., 2012 https://bioconductor.org/packages/
release/bioc/html/DiffBind.html
Burrows-Wheeler Aligner (BWA) Li and Durbin, 2009 http://bio-bwa.sourceforge.net/
Juicer Rao et al., 2014 https://github.com/aidenlab/juicer/wiki
Graphpad Prism 8 Graphpad https://www.graphpad.com/
Distiller Abdennur and Mirny, 2020 https://github.com/mirnylab/distiller-nf
Hi-C explorer Ramirez et al., 2018 https://hicexplorer.readthedocs.io/en/
latest/content/list-of-tools.html
BEDTools https://bedtools.readthedocs.io/en/latest/
deepTools Ramı́rez et al., 2016 https://deeptools.readthedocs.io/en/
develop/RESOURCE AVAILABILITY
Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Li-Huei
Tsai (lhtsai@mit.edu).
Materials availability
Unique resources and reagents generated in this study are available from the lead contact with a completed Material Transfer
Agreement.
Data and code availability
Genomic sequencing data have been deposited at Gene Expression Omnibus (GEO) and are publicly available as of the date of pub-
lication. Accession numbers are listed in the key resources table. This paper does not report original code. Any additional information
required to reanalyze the data reported in this paper is available from the lead contact upon request.Cell Stem Cell 29, 116–130.e1–e7, January 6, 2022 e3
ll
OPEN ACCESS ArticleEXPERIMENTAL MODEL AND SUBJECT DETAILS
Induced pluripotent stem cell (iPSC) generation and neural progenitor cell (NPC) differentiation
iPSCs
The isogenic pair (Iso-E and Iso-T) of induced pluripotent stem cells (iPSCs) were provided by Dr. Stuart Orkin, Dana-Farber Cancer
Institute, which have been described previously (Coriell, AG05397) (Maclean et al., 2012). The DS-iPSCs from the male individual
(Ma-T, AG06922) purchased from Coriell and the female control iPSC (Fe-E, H9 WA09) purchased from WiCell. Fibroblasts of the
derived from a female with DS, purchased from Coriell (GM04616) and fibroblasts from the male control (CS-0020-01) were utilized
to generate iPSCs using Sendai virus to overexpress OCT4, SOX2, KLF4, and c-MYC, utilizing the CytoTune-iPS Sendai Reprogram-
ming kit (Thermofisher scientific, A16518) as previously described (Raja et al., 2016). iPSCs were cultured on irradiated mouse em-
bryonic fibroblasts (MEFs, Thermo Fisher Scientific, A34181) in DMEM/F12 media supplemented with knockout serum replacement
(KSR, 20% v/v), MEM non-Essential amino acid solution (Sigma-Aldrich M7145), Glutamax (Thermo Fisher Scientific 35050-079),
8 ng/ml recombinant human FGF-basic (FGF2, PeproTech, 100-18B) and 71.5mM 2-mercaptoethanol (Sigma-Aldrich M6250). The
quality of cells was monitored daily and differentiated cells were mechanically removed under a light microscope in a biosafety
hood. iPSCs were cultured to 80% confluence and non-differentiated colonies were transferred to a feeder-free system and grown
on Matrigel hESC-Qualified Matrix coated plates (VWR, 354277) using mTeSR1 medium (Stemcell technologies, 85850). iPSCs for
RNA-seq, ATAC-seq, Hi-C and immunofluorescence imaging were grown under feeder-free conditions.
NPCs
NPCs were differentiated as previously described with minor adjustments (Chambers et al., 2009). iPSCs were grown to 80% conflu-
ence on feeder-free conditions in mTeSR1 medium. iPSCs were then dissociated into single cell suspension using ReLeSR (Stem-
cell Technology, 05872) following the manufacturers guidelines. 3 million cells were then transferred into a single well of
AggreWell800 (Stemcell Technology, 34811) to form 300 embryoid bodies of 10,000 iPSCs each in NPC-differentiation medium
[1:1 of DMEM/F12 Glutamax and Neurobasal (Life Technologies 10565-042 and Thermo Fisher Scientific 21103049), N-2 Supple-
ment (Invitrogen,17502-048), B-27 Serum-Free Supplement (Gibco, 12587010), 2.5mg/l insulin (Thermo Fisher Scientific,
41400045), Glutamax (Thermo Fisher Scientific, 35050-079), MEM non-Essential amino acid solution (Sigma-Aldrich, M7145),
71.5mM 2-mercaptoethanol, Pen/strep (Gemini Bio-Products, 400-109), Dorsomorphin (PeproTech, 8666430) and SB431542 (Toc-
ris,1614)]. Half NPC-induction medium replacement was performed daily for 5 days and on day 6 EBs were washed and transferred
onto Matrigel coated plates. Complete NPC-induction medium replacement was performed daily for 7 days, and on day 12 STEM-
diff Neural Rosette Selection Reagent (Stemcell technologies, 05832) was utilized to isolate NPC-rosettes and re-plated on onto
fresh Matrigel coated plates in NPC-induction medium. Complete NPC-induction medium replacement was performed for 7 days
and on day 20 NPCs were transferred onto fresh Matrigel coated plates in NPC-expansion medium [1:1 of DMEM/F12 Glutamax
and Neurobasal (Life Technologies 10565-042 and Thermo Fisher Scientific 21103049), N-2 Supplement (Invitrogen,17502-048),
B-27 Serum-Free Supplement (Gibco, 12587010), 2.5mg/l insulin (Thermo Fisher Scientific, 41400045), Glutamax (Thermo Fisher
Scientific, 35050-079), MEM non-Essential amino acid solution (Sigma-Aldrich, M7145), 71.5mM 2-mercaptoethanol, Pen/strep
(Gemini Bio-Products, 400-109) and 8 ng/ml recombinant human FGF-basic (FGF2, PeproTech, 100-18B)]. NPCs were passaged
once a week and NPCs for RNA-seq, ATAC-seq, Hi-C, senolytic drug treatment and immunofluorescence imaging were used on
passage 3-5.
METHOD DETAILS
Immunofluorescence imaging and analysis
iPSCs and NPCs grown on coverslips were fixed with 4% paraformaldehyde in 1X PBS for 15 minutes, then incubated overnight at
4C in primary antibody in 1X PBS with 0.3% Triton X-100. Primary antibodies used were OCT-4 (CST C52G3), Nanog (CST 1E6C4),
SOX1 (R&DSystems AF3369), Nestin (Millipore SigmaMAB5326), PAX6 (R&DSystems AF8150), H3K9me3 (Abcam 176916), LMNB1
(Abcam 16048), p16INK4a (Thermo Fisher Scientific MA5-32133), KI-67 (Abcam 15580) and BrdU (CST 5292S) 1:500 dilution over-
night at 4oC. Coverslips were washed three times with PBS and incubated with Alexa-Fluor 488, Alexa-Flour 555 and/or Alex-Fluor
647 secondary antibodies (Thermo Fisher Scientific) 1:1,000 dilution, and DAPI (Invitrogen D1306) 1:5,000 dilution for two hours at
room temperature. The coverslips were mounted on microscope slides with Fluoromount-G (Electron Microscopy Sciences 17984-
25). The slides were imaged using a confocal microscope (LSM 710 and 880, Zeiss) with a 5X, 20x or 63x objective with identical
settings for all matched images.
Heterochromatin architecture
iPSCs and NPCs grown on coverslips were fixed with 4% paraformaldehyde in 1X PBS for 12 minutes, then incubated overnight at
4C in primary antibody in 1X PBS with 0.3% Triton X-100. Primary antibodies used were anti-LMNB1 (1:500) and anti-H3K9Me3
(1:1,000). Primary antibody targets were visualized with Alexa-Fluor 488 and Alex-Fluor 647 secondary antibodies, and cell nuclei
were stained with DAPI. The coverslips were mounted on microscope slides with Fluoromount-G. The slides were imaged using a
confocal microscope (LSM 880, Zeiss) with a 63x objective and identical settings were used for all coverslips. Images were quantified
using Imarisx64 9.2.1 (Bitplane, Zurich, Switzerland), to determine the number of heterochromatin aggregates (as labelled with
H3K9Me3) and degree of co-localization with LMNB1 for the isogenic euploid, trisomic, and drug-treated trisomic cells. For each
image, first, a ‘‘generous’’ surface was created on the LMNB1 staining and used to mask the H3K9Me3 channel, to eliminate generale4 Cell Stem Cell 29, 116–130.e1–e7, January 6, 2022
ll
Article OPEN ACCESSextranuclear and background signal. The ‘‘generous’’ LMNB1 surface had surface detail of 0.6, the Imaris auto-generated surface
threshold, and a volume gate of greater than 40 mm3. Then, a ‘‘stringent’’ LMNB1 surface was created, to hone in on the areas of
the nuclei with the strongest LMNB1 signal. These ‘‘stringent’’ surfaces had surface detail of 0.2, surface threshold of 105, and no
volume gate. Finally, spots were created on the H3K9Me3 channel, with diameter 0.65 mm and quality threshold 15.6. To quantify
the co-localization, the ‘‘Find spots close to surface’’ command was used between the H3K9Me3 spots and the ‘‘stringent’’
LMNB1 surface, with a threshold of 0.35 mm. For each of the images, a total number was determined for both H3K9Me3 spots
co-localized with LMNB1 and intra-nuclear H3K9Me3 aggregates.
Proliferation
NPCs were treated with 10mM BrdU for 2hrs before fixing the cells with 4% paraformaldehyde in 1X PBS for 12 minutes. DNA hy-
drolysis was performed by treating NPCs with 1.5M HCl for 10mins at room temperature and washed 3x with 1X PBS. Cells were
then permeabilized and primary antibodies was used against BrdU (1:500), KI-67 (1:1,000) and SOX1 (1:100). Primary antibody tar-
gets were visualized with Alexa-Fluor 488, Alexa-Fluor 555 and Alex-Fluor 647 secondary antibodies, and cell nuclei were stained
with DAPI. The coverslips were mounted on microscope slides with Fluoromount-G. The slides were imaged using a confocal micro-
scope (LSM 880, Zeiss) with a 20x objective and identical settings for all coverslips. Images were quantified using Imarisx64 9.2.1
(Bitplane, Zurich, Switzerland), to determine the proportion of proliferative NPCs. SOX1 surface were used to identify all the NPCs
in the images with a setting of 0.75mm surface grain size. Then BrdU and KI-67 positive cells under the SOX1 surface were quantified
with a mean intensity of >18 and >10, respectively.
Cell Migration
Matrigel embedded neurospheres in 96-well glass bottom plates were fixed with 4% paraformaldehyde in 1X PBS for 15 minutes,
then incubated overnight at 4C in primary antibody in 1X PBS with 0.3% Triton X-100 with primary SOX1 (1:100) and Nestin
(1:500). SOX1 was visualized with Alexa-Fluor 488 and Nestin was visualized with Alexa-Fluor 555 secondary, and nuclei were
stained with DAPI. Fluoromount-G was then added onto Matrigel embedded neurospheres. The Matrigel embedded neurospheres
were imaged using a confocal microscope (LSM 710, Zeiss) with a 5x objective and identical settings were used for all images. Im-
ages were then quantified using Imarisx64 9.2.1 (Bitplane, Zurich, Switzerland), to determine how far individual NPCs had migrated
out of the neurospheres for the isogenic pair of euploid, trisomic and drug-treated trisomic neurospheres. For each image, first, a
DAPI sphere surface was created to demarcate the boundaries of the neurosphere core. This surface had surface detail of 2, surface
threshold of the Imaris auto-generated threshold divided by two, and a volume gate such that all but the largest surface is eliminated.
Next, the DAPI sphere surface was used to mask the SOX1 channel, to isolate the signal from NPCs that had migrated out of the
primary neurosphere. Finally, spots were created on the masked SOX1 channel, with diameter 15 mm and quality threshold 5.
Then, the distance between the SOX1 spots and the DAPI sphere surface was calculated and a list of individual distance values
for each migrated NPC was generated for each neurosphere.
Global Chromatin Profiling (GCP)
The Global Chromatin Profiling (GCP) assay was performed as previously described (Creech et al., 2015; Litichevskiy et al., 2018).
Briefly, three biological and three technical replicates of neural progenitor cells (NPCs), derived from isogenic-pair (AG-005397 (iso-E
and Iso-T)), were grown in 6-well plates, washed in ice cold PBS 2x, scraped, and collected by centrifugation. After nuclei were iso-
lated using mild lysis conditions (0.3% NP40), histones were extracted using sulfuric acid (Sigma-Aldrich, 339741-500ML), and
precipitated using trichloroacetic acid (BDH, BDH0310-500G). Samples (10mg each) were propionylated, desalted and digested
by trypsin (Promega, V5113) overnight following standard protocols. A second round of propionylation followed by desalting using
C18 Sep-Pak cartridges (Waters,186002318) was employed. A mix of isotopically labeled synthetic peptides for histone H3 was
added as a reference to each sample prior to mass-spectrometric (MS) analysis. Peptides were separated on a C18 column
(EASY-nLC 1000, Thermo Scientific) and analyzed by mass spectrometry (MS) in a parallel reaction monitoring mode (QExactiveTM
Plus, Thermo Scientific) as previously described(Creech et al., 2015). Detailed protocols of sample preparation steps can be found in
https://panoramaweb.org/labkey/wiki/LINCS/Overview%20Information/page.view?name=sops.
Library preparation and analysis
RNA-seq
Biological replicates (n=3) of the isogenic pair (Iso-E and Iso-T) of iPSCs and NPCs as well as the male (Ma-E and Ma-T) and female
(Fe-E and Fe-T) pairs were used for library preparation. RNA quality was checked using RNA Integrity number on an Agilent Bio-
analyzer 2100, and samples with a RIN score of >9 were used for library preparation. 150 ng of RNA was used to prepare libraries
using the Kapa mRNA hyperprep kit for standard RNA libraries (Roche, KK8540) on a TecanEvo 150s (Tecan). >50 million 75-bp
paired-end reads were acquired for each replicate using the NextSeq 500 platform at the MIT BioMicro Center. FASTQ paired-
end reads were aligned using STAR (v.2.6.1a) to GRCh37 reference genome (GENCODE 19) (Dobin et al., 2013). Transcripts
were quantified using HTSeq, data was normalized utilizing RUV-seq and differential gene expression analysis was performed
through DESeq2 (Anders et al., 2015; Love et al., 2014; Risso et al., 2014). Significant differentially expressed genes (DEGs)
were called with an FDR < 0.01, log2 fold change R 0.5 or % -0.5. Gene ontology analysis was performed using GORILLA with
two unranked lists of genes (target and background lists) followed by REViGO a program that can takes a long list of Gene Ontology
terms and removes redundant GO terms and generates a scatterplot with an unbiased reduction of the GO terms in semantic space
(x, y) (Eden et al., 2009; Supek et al., 2011). Correlation analysis of the pairwise comparisons was performed using corrplotCell Stem Cell 29, 116–130.e1–e7, January 6, 2022 e5
ll
OPEN ACCESS Article(https://github.com/taiyun/corrplot), utilizing the log2 fold overlap enrichment and significance p-value generated through random-
ized permutation test. Randomized permutation test was done in R using custom scripts. Briefly, a sample set of equal size as the
observed data was randomly selected from the background set and then the statistic test was calculated for each randomly sampled
data (s* = s1, s2,.sN) exactly as for the observed data (sObs). This was iterated N=107 times with replacement to create the null
distribution and the p-value was calculated using the formula.
p-Value = sum(s* >= sObs)/ N (For enrichment) p-Value = 1- sum(s* >= sObs)/ N (For depletion)
If sObs was greater than or less than all values in the null distribution, the p-value was set as < < 1E-07.
ChIP-seq
Libraries were generated from 2million iPSC derived NPCs using the ChIP-IT Express Kit and Sonication Shearing Kit (Active Motif)
with Spike-in Chromatin from Drosophila (Active Motif) following to the manufacturer’s instructions without any modifications (Egan
et al., 2016). Briefly, the ChIP-seq library preparation consists of 8 steps: (i) cells were crosslinked with 1% formaldehyde, (ii) chro-
matin was sheared (Covaris E220 Focused-ultrasonicator) after lysing cells, (iii) digested chromatin was incubated with 2ug of
LMNB1 antibody (Abcam, ab8982), samples were also incubated without any antibody for input and Spike-in chromatin was incu-
bated with Spike-in antibody (Active motif). (iv) same amount of Spike-in chromatin/antibody was added to each sample. (V) anti-
body-bound DNA/protein complexes and input samples were captured and washed using Protein Gmagnetic beads. (vi) cross-links
were reversed and protein was digested using proteinase K. (vii) DNA was purified utilizing QIAquick PCR Purification Kit (Qiagen,
28104) (viii) libraries were prepared utilizing the Kappa Hyperprep Kit (Roche, KK8502). Quality control was performed using the Agi-
lent Bioanalyzer 2100. 
30 million 40-bp single-end reads were acquired for each replicate on the NextSeq 500 platform at the MIT
BioMicro Center. FASTQ reads for each replicate were aligned using Bowtie 2.0 (bowtie2 –very-sensitive -k 30) to hg19 (human) and
dm6 (drosophila) (Langmead and Salzberg, 2012). Samtools was used to sort (samtools sort), remove duplicates (samtools
markdup), removemitochondrial reads (samtools view) and index BAM files (samtools index) (Li et al., 2009). Bigwigs were generated
from BAM files adjusted to the scale-factor for each sample using the percentage of mapped reads from the drosophila genome
(dm6) to total reads over input utilizing deepTools (3.5.0) bamCoverage (Egan et al., 2016; Ramirez et al., 2016). Peak regions
were identified using homer (4.11), with the function findPeaks utilizing the reads generated from input samples as background. Scat-
ter plots and violin plots were generated utilizing the R package ggplot2.
ATAC-seq
Libraries were prepared as previously described (Buenrostro et al., 2015). 50,000 cells from the biological replicates (n=3) of the
isogenic pair of iPSCs and NPCs were used for library preparation. Cells were harvested after treatment for 4 mins with ReLeSR
(Stemcell Technology, 05872) for iPSCs and TrypLE (Thermo Fisher Scientific, 12605028) for NPCs were then counted and 50,000
cells were spun down at 5003g for 5 min, 4C. Cells were with 50ml PBS and spun down at 5003g for 5 min, 4C. Cells were then
resuspended in fresh 50ul cold lysis buffer (0.15% igepal CA-630 (NP-40)) and left on ice for 10 min and spun down at 5003g for
5 min, 4C. Next, tagmentation (Nextera FC-121-1030, Illumina) was carried out in 50ul volume (25 ml TD, 2.5 ml TDE1, 22.5 ml
nuclease-free water), for 30 min at 37C with gentle rocking. Transposed DNA fragments were purified using the Qiagen MinElute
PCR Purification Kit (28006) and were barcoded, amplified (PCR reaction: 1 cycle; 5min at 72C, 30sec at 98C, 8 cycles; 15sec
at 98C, 30sec at 60C, 3min at 72C) and purified using AMPure beads (Beckman Coulter, A63880). ATAC-seq library fragment
size distribution was assessed on an Agilent Bioanalyzer 2100. >50 million 40-bp paired-end reads were acquired for each replicate
on the NextSeq 500 platform at the MIT BioMicro Center. FASTQ paired-end reads for each replicate were aligned using Bowtie 2.0
(bowtie2 –very-sensitive -k 30) to hg19 (Langmead and Salzberg, 2012). Samtools was used to sort (samtools sort), remove dupli-
cates (samtoolsmarkdup), removemitochondrial reads (samtools view), index BAMfiles (samtools index), and BAMfiles were shifted
by +4bp and -5bp using an in-house algorithm (Li et al., 2009). Accessible chromatin peak calling was performed using Model-based
Analysis of ChIP-Seq (macs2 callpeak -f BAMPE -g hs -q 0.05) (Zhang et al., 2008). Differential peaks analysis between the isogenic
pairs was performed using Diffbind v1.16.3 (Ross-Innes et al., 2012). Correlation between differential accessibility of gene promoters
(-2,000bp and +500bp of TSS) with differential expression was computed by first calculating the overlap percentage and significance
of enrichment/depletion using a randomized permutation test then data was plotted as a histogram.
In situ Hi-C
Hi-C libraries were generated from 2 million iPSCs and NPCs using the Arima Hi-C kit (Arima Genomics, San Diego) following to the
manufacturer’s instructions without any modifications. Briefly, the Arima Hi-C kit consists of 6 steps: (i) cells were crosslinked with
1% formaldehyde, (ii) DNAwas digested using the Arima restriction enzyme cocktail within intact permeabilized nuclei, (iii) ends were
filled and biotinylated, (iv) blunt ends were ligated, (v) ligated fragments were purified, sheared and a biotin pull down was performed
using streptavidin, and (vi) adapters were ligated, barcoded and amplified using the Kappa Hyperprep Kit (Roche, KK8502). Quality
control was performed after shearing and library-prep using the Agilent Bioanalyzer 2100. Finally, >1.5 billion 40-bp paired-end reads
were acquired for each replicate on the NextSeq 500 platform at the MIT BioMicro Center. FASTQ paired-end reads were processed
through Juicer as previously described (Rao et al., 2014). In Brief, paired-end read pairs were aligned separately to the human
genome (hg19) using BWA, then, duplicates/invalid pairs were removed, and the remaining reads were filtered based on mapping
quality score of 30 (Li and Durbin, 2009). Interaction matrices were generated at resolutions of 2.5Mb, 1Mb, 500kb, 250kb,
100kb, 50kb, 25kb, 10kb, and 5kb and normalized interactions matrices were produced at all resolutions using Knight and Ruiz
(
/juicer/scripts/CPU/juicer.sh -D 
/juicer/ -y 
/juicer/restriction_sites/hg19_Arima.txt -s Arima).e6 Cell Stem Cell 29, 116–130.e1–e7, January 6, 2022
ll
Article OPEN ACCESSTrans-chromosomal interactions stats were generated through Juicer (Table S1) and a t-test was run comparing euploid and
trsiomic iPSCs andNPCs. Differential trans-interactionmapswere generated through Hi-C Explorer utilizing the default setting (Ram-
irez et al., 2018). In brief, interaction matricies generated in juicer (.hic) were converted to cooler and then to h5 (hicConvertFormat -m
matrix.hic –inputFormat hic –outputFormat cool -o matrix.cool –resolutions 1000000 -> hicConvertFormat -mmatrix.cool –inputFor-
mat cool –outputFormat h5 -omatrix.h5). Next, a matrix containing the normalized log2ratio values of trisomy/euploid was generated
(hicCompareMatrices -m Trisomy.h5 Euploid.h5 –operation log2ratio -o Trisomy_Euploid.h5) and plotted (hicPlotMatrix -m Triso-
my_Euploid.h5 –clearMaskedBins –vMin -1 –vMax 1 -o Trisomy_Euploid.png). Distribution of normalized counts was generated us-
ing distiller pipeline on nextflow (nextflow run distiller.nf -params-file project.yml) as previously described (https://github.com/
mirnylab/distiller-nf) (Abdennur andMirny, 2020). Trans-chromosomal interaction statistics generated by distiller was used to identify
the distribution of the trans-chromosomal interaction for each chromosome and calculated as the percentage of total interactions per
sample. A statistical t-test comparison of euploid and trisomic cells was performed for each chromosome after normalizing for the
number of chromosomes, where all chromosomeswere divided by 2with the exception of chromosome 21whichwas divided by two
in euploid samples and divided by 3 in trisomic samples (Figure S1E; Table S3).
Cis-chromosomal interactions stats were generated through Juicer (Table S1) and a t-test was run comparing euploid and
trisomic iPSCs and NPCs. Differential trans-interaction maps were generated using juicebox, and represents the foldchange of
trisomic over euploid cells. Global decay plots were generated using an in-house algorithm. In brief, decay plots were made by plot-
ting the ‘‘Fraction of total paired-end tags’’ vs ‘‘Distance between the paired-end tags" generated by homer Hi-C analysis pipeline
(Heinz et al., 2018). Fold enrichment values for each distance intervals were calculated as log2(mean of trisomy replicates/ mean of
euploid replicates).The FDR p-values for the enrichment between the tree replicates for trisomy and disomy was calculated using
t-test (two-tailed) and corrected using p.adjust function in the R library qvalue using standard methods (Benjamini & Hochberg).
Shorth-range (<1Mb) to long-range (>1Mb) interaction ratios were generated using Hi-C explorer utilizing the default settings (hic-
PlotSVL -m Euploid_10kb.cool Trisomy_10kb.cool –distance 1000000) and a Wilcoxon rank-sum test was run between euploid
and trisomic samples and represented as dot plots for each chromosome.
A/B compartmentswere analyzed utilizing the HOMERHi-C analysis pipeline (http://homer.ucsd.edu/homer/interactions2/index.
html) utilizing the default settings (Heinz et al., 2018). In brief, FASTQ paired reads were aligned separately using BWA (bwa mem -t8
-A1 -B4 -E50 -L0 
/BWA_Index/hg19.fa), then tag directories were generated (makeTagDirectory sample_R1.sam,sample_R2.sam
-tbp 1 -removePEbg -removeSpikes 10000 5). Principal component analysis of the Hi-C interaction matrices was used to reveal
the active and inactive compartments of the genome (runHiCpca.pl auto sample_TagDir/ -res 25000 -window 50000 -genome
hg19). Regions were annotated (annotatePeaks.pl) and edgeR/limma was used to identify the differential compartments
(getDiffExpression.pl). Comparative analysis was performed to define the concordance between first principal component-based
compartments identified through HOMER and the first principal component of the Pearson’s matrix (eigenvector) identified through
Juicer using bedtools intersect. Enrichment analysis with differentially expressed genes (DEGs) and differentially accessible regions
(DARs) with A and B compartments was performed by first calculating the overlap percentage of DEGs and DARs with the different
compartments and the significance of enrichment/depletion was calculated using a randomized permutation test (p-Value). Data was
plotted as histogram or as corrplot on R.
Loops and Topologically associating domains (TADs)
Loopswere called using the default setting of HICCUPS on juicer at 5Kb or 10Kb resolution (java -Xmx125g -jar
/juicer/juicer_tools.-
jar hiccups). Differential looping between euploid and trisomic cells was identified using the default setting of HiCCUPS Diff on juicer
using a graphics processing unit (GPU) (java -Xmx125g -jar 
/juicer/juicer_tools.jar hiccupsdiff). Aggregate plot analysis was per-
formed using the APA command tool on juicer using the default settings (java -Xmx125g -jar 
/juicer/juicer_tools.jar apa) and
genome-wide aggregate plots were generated from the normalized APA matrices on R. TADs boundaries, insulation scores and
directionality index was computed using the default setting on HOMER (findTADsAndLoops.pl find TagDir/ -cpu 5 -res 3000 -window
15000 -genome hg19 -p badRegions.bed). For differential intra-tad interaction density analysis TAD files for the biological replicates
per cell-type were merged (merge2Dbed.pl) and then interaction densities were quantified across replicates (findTADsAndLoops.pl
score -tad). Finally, edgeR/limmawas utilized to identify the differentially interacting TADs (getDiffExpression.pl). Enrichment analysis
with differentially expressed genes was performed between down and upregulated genes with differentially interacting TADs and
unchanged TADs, by first calculating the overlap percentage and significance of enrichment/depletion (randomized permutation
test, p-Value). Data was plotted using corrplot on R.
QUANTIFICATION AND STATISTICAL ANALYSIS
Data was analyzed using GraphPad Prism version 8. Replicate sizes and error bars definitions are indicated in the figure legends.
Correct application of standard parametric procedures was confirmed and stated in the supplemental data. Two-tailed Student’s
t test were performed for datasets with two groups. ANOVA analyses were used for comparisons of data with more than two groups.
Post hoc group comparisons were performed and stated in the figure legends and supplemental tables.
Differences were considered statistically significant at * (p < 0.05), ** (p < 0.01), *** (p < 0.001), and **** (p < 0.0001).Cell Stem Cell 29, 116–130.e1–e7, January 6, 2022 e7