news & views
CANCER EPIGENETICS
A role for chromatin regulatory dynamics in
breast cancer evolution
Enhancer profiling of breast tumors reveals that chromatin regulatory elements contribute to the clonal fitness
landscape, treatment resistance and phenotypic divergence.
Christopher Probert and Christina Curtis
M
odern genomic data demonstrate
the central role of chromatin
regulatory elements in cellular
identity and behavior; however, little is
known about their heterogeneity and
clonal dynamics in cancer. Breast cancer
is the most commonly diagnosed cancer
among women worldwide, but the extent
of chromatin regulatory variation and its
impact on clinical outcomes are largely
elusive. Although large-scale studies
have found extensive mutational and
transcriptional heterogeneity between
and within breast cancer subtypes1,2, no
efforts to date have rigorously characterized
the diversity of genome-wide chromatin
regulatory activity in a meaningful number
of tumors using modern histone-based
sequencing techniques. In a new study in
this issue, Patten et al.3 carry out extensive
molecular profiling and functional analysis
of breast cancer regulatory elements
using modern histone-based sequencing
techniques to examine the role of chromatin
in breast cancer phenotypes and evolution.
Recent analyses of breast cancer–
associated genomic mutations
demonstrate pervasive patient-to-
patient heterogeneity1,2; however, several
integrative breast cancer subgroups2 lack
recurrent somatic alterations, suggesting
that additional sources of heterogeneity
contribute to clinical outcomes such as
drug resistance. Meanwhile, studies of
chromatin dynamics during differentiation
of blood (hematopoiesis)4 and neurons
(neurogenesis)5 reveal intricately regulated
patterns of chromatin element activity
that impact transcription, morphology
and differentiation. Additionally, a recent
pan-cancer study of only a small subset of
the genomic enhancer landscape showed
an association between enhancer activity
and clinical outcomes6. The genomic
heterogeneity within individual breast
cancers7 provides a rich substrate for
clonal evolution, resulting from mutation
acquisition, selection and genetic drift;
Nature Medicine | www.nature.com/naturemedicine
Active
enhancers
identified
with
H3K27Ac
ChIP-seq
TF-binding
sites
predicted
ER
YY1
Fig. 1 | Active enhancers are linked to breast cancer heterogeneity and drug resistance. Pattern et al.3
profile the levels of H3K27Ac—an epigenetic mark associated with active enhancers—in breast cancers.
They find that the activity of enhancers is linked to tumor heterogeneity and that hormone therapy
selects for subclones in which the transcription factors YY1 and ER bind to active enhancers.
TF, transcription factor.
however, the role of chromatin regulatory
elements and epigenetic variation in this
process is largely unknown.
Patten et al.3 use chromatin
immunoprecipitation sequencing
(ChIP-seq) to profile the chromatin
landscapes of 39 primary and 16 metastatic
breast tumors. They identify 350,695 unique
active chromatin enhancer elements marked
with H3K27ac, a post-translational histone
modification commonly used to define
active regulatory elements. They develop
new analyses based on the signal intensity of
this mark within each sample, allowing them
to infer the clonality or lineage relationships
and prevalence of each enhancer among
cells within the tumor—alterations present
in the founding tumor cell are present
in all descendent cells and are clonal by
definition, whereas mutations that arise
later are present in subclones in only a
fraction of tumor cells but can expand if the
environment selects for them. Further, they
Clone
Hormone
therapy–resistant
tumor
Hormone Subclone
therapy expansion
Heterogenous
tumor
Subclone
compared the identified chromatin states
across patients.
The authors found that the active
regulatory elements they identified in breast
cancers almost completely overlap with
known breast cancer risk variants identified
through genome-wide association studies
(GWAS). Furthermore, active enhancers
identified in samples from patients with
breast cancer correlate with known breast
cancer gene expression profiles and to
enhancers identified in the ENCODE/
Roadmap breast cell line. This provides
the first catalog of breast cancer functional
elements and surveys their inter- and
intrapatient heterogeneity. Future studies
that sample multiple tumor regions within
a patient may reveal more extensive within-
patient epigenetic heterogeneity.
Patten et al.3 then predicted transcription
factor–binding sites in their identified
active breast cancer enhancers using
binding-motif analysis. Their analysis
news & views
indicates that active enhancers in clones and
subclones are regulated by largely distinct
transcription factor–regulatory programs
and that the vast majority of predicted
active estrogen receptor α​ (ER)–binding
sites are in subclonal regions of tumors.
Further, they discover that the transcription
factor YY1—a known master regulator of
chromatin activity and organization that
has recently been shown to be essential
in developmental processes such as
neurogenesis8,9—is predicted to bind in
a high proportion of active regulatory
elements shared between clones. However,
the authors’ in vitro YY1 ChIP-seq analysis
reveals that in tumor subclones, ER
cooperates with YY1 at subclone-specific
active elements. This demonstrates a
bifunctional role of YY1 binding without ER
in dominant clones, but with ER at active
elements in subclones.
ER is a nuclear-localized protein
overexpressed in a large proportion of breast
cancers, which are commonly treated with
hormone (endocrine) therapy. Resistance
to hormone therapy is relatively common
with treated tumors exhibiting either
intrinsic or acquired resistance10; however,
the mechanisms and molecular basis of the
resistance are largely unknown. The current
study reveals that YY1–ER cooperation may
enable subclonal resistance to endocrine
therapy in luminal breast cancer. The
authors treated breast cancer cells with
increasing doses of tamoxifen (a small-
molecule ER-binding modulator widely used
to treat ER+ breast cancer) and showed that
tamoxifen selects for the subclones with
YY1–ER binding at active enhancers. The
observation that YY1 is a master regulator
of breast cancer–specific chromatin
architecture and its cooperative effect
on therapeutic resistance and subclonal
enhancer fitness is particularly interesting
in the context of the central role of YY1
in other developmental processes8,9. This
finding also highlights the importance of
perturbation-based functional investigation
of disease regulatory element activity—the
authors only elucidated the clinical relevance
of YY1–ER cooperation in treatment
resistance by carefully combining cohort-
based functional element discovery with in
vitro perturbation experimental data.
Although the findings reported by
Patten et al.3 provide fresh insights into
chromatin regulatory dynamics in breast
cancer, this study only scratches the surface
of potentially clinically relevant epigenomic
variation in cancer. As multimodal single-
cell genomic sequencing technologies
mature, they will likely improve the ability
to identify functional regulatory elements
that differ between clones as well as enable
a deeper understanding of the interactions
between regulatory elements and genomic
alterations in individual cancer cells that
may be of clinical importance11. Cell-
centric analyses will likely be important in
understanding, for example, the emergence
of treatment-resistant subclones because
they can rapidly emerge from small
populations under treatment selective
pressure. Additionally, perturbation-based
assays of primary tumor cells, including
genome-wide functional screens, will bore
out additional information about sufficiency,
dependence and synthetic lethality of
regulatory elements that will further shape
our understanding of the tumor epigenomic
fitness landscape and will ultimately enable
the development of evolutionarily informed
treatment strategies Fig. 1.
A better understanding of tumor genomic
regulatory element dynamics, clonality
and function may ultimately improve our
ability to detect and treat cancer. While
current tumor molecular profiling focuses
almost exclusively on mutations, genomic
copy-number aberrations, translocations
and expressed protein markers, these results
and others point to a role for functional
genomic elements in determining clinical
outcome, and tumor epigenetic profiling
could become part of workflows for tumor
molecular profiling. Additionally, the ability
to directly target epigenetic misregulation
in vivo with small-molecule inhibitors
of tumor epigenetic modifications draws
near, and in the future, precisely targeted
correction of individual tumor-specific
chromatin modifications may be possible
and could translate to personalized
epigenomic therapies. Although these
advances are largely out of reach today,
improved understanding of the state and
functional role of genomic regulatory
elements in cancer, such as those presented
by the current study3, may help guide their
development in the near future. ❐
Christopher Probert1 and Christina Curtis1,2*
1
Department of Genetics, Stanford University School
of Medicine, Stanford Cancer Institute, Stanford, CA,
USA. 2Department of Medicine, Stanford University
School of Medicine, Stanford Cancer Institute,
Stanford, CA, USA.
*e-mail: cncurtis@stanford.edu
Published: xx xx xxxx
https://doi.org/10.1038/s41591-018-0182-8
References
1. Cancer Genome Atlas Network. Nature 490, 61–70 (2012).
2. Curtis, C. et al. Nature 486, 346–352 (2012).
3. Patten, D. K. Nat. Med. https://doi.org/10.1038/s41591-018-
0091-x (2018).
4. Pogrebniak, K. L. & Curtis, C. Trends Genet. 34, 639–651 (2018).
5. Dulken, B. W., Leeman, D. S., Boutet, S. C., Hebestreit, K. &
Brunet, A. Cell Reports 18, 777–790 (2017).
6. Chen, H. et al. Cell 173, 386–399.e12 (2018).
7. Yates, L. R. et al. Nat. Med. 21, 751–759 (2015).
8. Weintraub, A. S. et al. Cell 171, 1573–1588 (2017).
9. He, Y. et al. Neuron 55, 217–230 (2007).
10. Osborne, C. K. & Schiff, R. Annu. Rev. Med. 62, 233–247 (2011).
11. Azizi, E. et al. Cell https://doi.org/10.1016/j.cell.2018.05.060 (2018).
Competing interests
The authors declare no competing interests.
Nature Medicine | www.nature.com/naturemedicine