ArticlePDF Available

The Aurora kinase/β-catenin axis contributes to dexamethasone resistance in leukemia

Authors:

Abstract

Glucocorticoids, such as dexamethasone and prednisolone, are widely used in cancer treatment. Different hematological malignancies respond differently to this treatment which, as could be expected, correlates with treatment outcome. In this study, we have used a glucocorticoid-induced gene signature to develop a deep learning model that can predict dexamethasone sensitivity. By combining gene expression data from cell lines and patients with acute lymphoblastic leukemia, we observed that the model is useful for the classification of patients. Predicted samples have been used to detect deregulated pathways that lead to dexamethasone resistance. Gene set enrichment analysis, peptide substrate-based kinase profiling assay, and western blot analysis identified Aurora kinase, S6K, p38, and β-catenin as key signaling proteins involved in dexamethasone resistance. Deep learning-enabled drug synergy prediction followed by in vitro drug synergy analysis identified kinase inhibitors against Aurora kinase, JAK, S6K, and mTOR that displayed synergy with dexamethasone. Combining pathway enrichment, kinase regulation, and kinase inhibition data, we propose that Aurora kinase or its several direct or indirect downstream kinase effectors such as mTOR, S6K, p38, and JAK may be involved in β-catenin stabilization through phosphorylation-dependent inactivation of GSK-3β. Collectively, our data suggest that activation of the Aurora kinase/β-catenin axis during dexamethasone treatment may contribute to cell survival signaling which is possibly maintained in patients who are resistant to dexamethasone.
ARTICLE OPEN
The Aurora kinase/β-catenin axis contributes
to dexamethasone resistance in leukemia
Kinjal Shah
1,2
, Mehreen Ahmed
1,2
and Julhash U. Kazi
1,2
Glucocorticoids, such as dexamethasone and prednisolone, are widely used in cancer treatment. Different hematological
malignancies respond differently to this treatment which, as could be expected, correlates with treatment outcome. In this study,
we have used a glucocorticoid-induced gene signature to develop a deep learning model that can predict dexamethasone
sensitivity. By combining gene expression data from cell lines and patients with acute lymphoblastic leukemia, we observed that
the model is useful for the classication of patients. Predicted samples have been used to detect deregulated pathways that lead to
dexamethasone resistance. Gene set enrichment analysis, peptide substrate-based kinase proling assay, and western blot analysis
identied Aurora kinase, S6K, p38, and β-catenin as key signaling proteins involved in dexamethasone resistance. Deep learning-
enabled drug synergy prediction followed by in vitro drug synergy analysis identied kinase inhibitors against Aurora kinase, JAK,
S6K, and mTOR that displayed synergy with dexamethasone. Combining pathway enrichment, kinase regulation, and kinase
inhibition data, we propose that Aurora kinase or its several direct or indirect downstream kinase effectors such as mTOR, S6K, p38,
and JAK may be involved in β-catenin stabilization through phosphorylation-dependent inactivation of GSK-3β. Collectively, our
data suggest that activation of the Aurora kinase/β-catenin axis during dexamethasone treatment may contribute to cell survival
signaling which is possibly maintained in patients who are resistant to dexamethasone.
npj Precision Oncology (2021) 5:13 ; https://doi.org/10.1038/s41698-021-00148-5
INTRODUCTION
Acute lymphoblastic leukemia (ALL) is the most common pediatric
malignancy of lymphoid progenitor cells, with ~80% of cases
occurring in children and the remainder in adults
1
. Glucocorticoids
such as dexamethasone and prednisolone are important drugs in
the chemotherapeutic regimen for the treatment of ALL
1
.Apart
from their strong anti-inammatory and immune-suppressive
actions, glucocorticoids induce growth arrest and apoptosis in ALL.
Their action is exerted by binding to the glucocorticoid receptor
(GR), a member of the nuclear hormone receptor superfamily that
also acts as a transcription factor
2
. Sensitivity to glucocorticoids
serves as a positive prognostic indicator, and patients unresponsive
to glucocorticoids often relapse, which leads to poor prognosis.
Multiple mechanisms of glucocorticoid resistance have been
identied such as alterations in expression and function of GR or
of GR-associated proteins like chaperones and co-chaperones,
thereby affecting their function, and causing defects in the target
genes, leading to the inhibition of apoptosis and defective
metabolism, cross-talk with other cell signaling pathways and
transcription factors, as well as changes in chromatin accessibility
25
.
The current chemotherapeutic regimen used for the treatment
of ALL, which consists of corticosteroids in conjunction with
chemotherapeutic drugs, has resulted in the long-term survival of
8090% in children and of 40% in adults, though the complete
remission (CR) rates are similar in both groups
6
. Despite these
positive statistics, the highest cancer-related mortality in children
is associated with ALL due to relapse or treatment-related toxic
effects. This shows a great need for improving the existing
treatments by not only identifying new molecular targets to
develop better and less toxic therapeutic agents but also by
identifying patients who require less intensive therapy
7
.
The heterogeneity in the patients responses to different
treatments has resulted in great advancements in the bioinfor-
matics area, where different machine learning approaches are
used to better match patients to drugs. Several studies have
compared machine learning approaches for monotherapy predic-
tion. The ridge regression model built using baseline gene
expression data from cell lines was shown to outperform several
machine learning algorithms including random forests, nearest
shrunken centroids, principal component regression, lasso, and
elastic net to predict in vivo chemotherapy response
8
. A landmark
study compared 44 drug sensitivity prediction algorithms in which
genomic, epigenomic, and proteomic proling data for human
breast cancer cell lines were used
9
. In that study, models
developed using Bayesian multitask multiple kernel learning and
random forests were found to perform better than other
algorithms. Another large-scale study compared multiple machine
learning algorithms using multi-omics data and reported elastic
net and ridge regression as the top-performing algorithms
10
.
These studies used several thousand features for a limited number
of samples to develop the model; therefore, the selection of
features might have affected the performance of the algorithm.
Besides those low complexity models, deep learning models have
been implicated in monotherapy prediction
11
. Because deep
learning models are highly sensitive to the feature-to-sample ratio,
the feature selection approach was found to be an effective
method in which only features highly conserved with the
phenotype were used to build the model
1214
. Such methods
are also useful in identifying a potential molecular marker and
driver of sensitivity to the drugs
12
.
In this study, we employed transcriptomics, peptide substrate-
based kinase proling, deep learning models, and drug synergy
studies to identify the deregulated signaling pathways in
1
Division of Translational Cancer Research, Department of Laboratory Medicine, Lund University, Lund, Sweden.
2
Lund Stem Cell Center, Department of Laboratory Medicine,
Lund University, Lund, Sweden. email: kazi.uddin@med.lu.se
www.nature.com/npjprecisiononcology
Published in partnership with The Hormel Institute, University of Minnesota
1234567890():,;
dexamethasone-resistant ALL. We show that the short-term
dexamethasone treatment induces the Aurora kinase/β-catenin
signaling axis which is also found to be enriched in ALL patients
that are predicted to be dexamethasone resistant.
RESULTS
Glucocorticoids induce a distinct gene signature in ALL
Long-term treatment with glucocorticoids induces resistance that is
mediated by several mechanisms. Even with the drug that can
induce a complete response, some cells always survive the
treatment. This surviving proportion of cells is referred to as drug-
tolerant cells, which eventually provoke drug resistance. Drug-
tolerant cells display dynamic uctuations in gene expression,
where survival is thought to be mediated by a group of genes that
are upregulated during the drug treatment
15
. To understand how
genes are regulated in ALL during glucocorticoid treatment, we
used the SUP-B15 cell line as a model system. SUP-B15 cells are
highly sensitive to glucocorticoids such as dexamethasone and
prednisolone (Supplementary Fig. 1a). We observed that these
drugs at micromolar concentrations inhibited more than 90% of cell
viability after 48 h of incubation (Supplementary Fig. 1b) and
induced a signicant level of apoptosis after 24 h of incubation
(Supplementary Fig. 1c). However, apoptosis remained almost equal
to the DMSO control after treatment of the drugs for 4 h or 6 h
(Supplementary Fig. 1c). Therefore, to identify glucocorticoid-
induced gene regulations at an earlier time point, we treated
SUP-B15 cells with dexamethasone (1 µM), prednisolone (2 µM), and
DMSO for 6 h before collecting total RNA. The experiment was
repeated three times and all nine samples were analyzed by
Affymetrix Human Gene 2.0 ST array. Deregulated genes were
identied by individually comparing expression between dexa-
methasone- and DMSO-treated samples, and prednisolone- and
DMSO-treated samples. Signicance Analysis of Microarrays (SAM)
16
was used to detect the deregulated genes. We used a 5% false
discovery rate (FDR) as the cut-off. Most of the deregulated genes
were common between dexamethasone- and prednisolone-treated
cells (Supplementary Table 1ad), where 366 common genes were
downregulated (Fig. 1a) and 217 common genes were upregulated
in drug-treated cells (Fig. 1b). A comparison between dexametha-
sone- and prednisolone-treated samples did not result in any
signicantly deregulated genes (data not shown). These data
suggest that both dexamethasone and prednisolone act on cells by
a similar mechanism. Upregulated (Supplementary Fig. 2a) or
downregulated (Supplementary Fig. 2b) genes were analyzed by
hierarchical clustering showing several different clusters. However,
clusters for dexamethasone- and prednisolone-treated cells remain
similar, as observed by SAM. Further analysis of the top-listed
commonly upregulated genes also showed similar clusters (Fig. 1c).
Several core cellular signaling regulators such as DUSP1, FBXW7,
Sprouty family proteins, etc. were upregulated in both dexametha-
sone- and prednisolone-treated cells in a similar fashion. Further-
more, gene set enrichment analysis (GSEA) identied enrichment of
similar signaling pathways in dexamethasone- and prednisolone-
treated cells (Fig. 1d, e). Although SUP-B15 displayed potential
variations in gene expression in response to glucocorticoids, several
glucocorticoid-resistant ALL cell lines such as JURKAT, MOLT-4, and
CCRF-CEM did not show any signicant difference in gene
expression upon treatment with dexamethasone (1µM) or
prednisolone (2 µM, data not shown).
Dexamethasone treatment induces activation of serine/
threonine protein kinases
Protein kinases regulate core cellular signaling by phosphorylating
the substrate proteins. Because glucocorticoid treatment potentially
deregulated gene expression leading to specic pathway enrich-
ment, we hypothesized that glucocorticoid treatment will also
modulate the core cellular signaling by altering the activation of
protein kinases. To measure global kinase activity, we used peptide
substrate-based kinase proling. Pamgene peptide substrate-based
kinase proling has been demonstrated to be a powerful technique
for determining kinase activity
17,18
. SUP-B15 cells were treated with
dexamethasone (1 µM) or DMSO for 6 h before lysis. Lysates were
applied to the Pamgene protein tyrosine and protein serine/
threonine chips following the manufacturers guidelines. While
comparing kinase activity enrichment between DMSO and
dexamethasone-treated cells, we observed that the activity of
protein tyrosine kinases was completely downregulated in
dexamethasone-treated cells (Fig. 2a) and the activity of several
protein serine/threonine kinases was upregulated (Fig. 2b). To
understand the possible mechanisms of how kinase activities were
regulated, we used the negative regulation of protein kinase
activity (GO:0006469)related gene signature to compare gene
expression data in dexamethasone- and prednisolone-treated SUP-
B15 cells. We observed that SPRY1 expression was more than
vefold upregulated in both dexamethasone- and prednisolone-
treated cells (Fig. 2c). Sprouty family proteins include four members
(SPRY1-4) that can control receptor tyrosine kinase activation
19,20
.
The upregulation of serine/threonine kinase activity, for example,
Aurora family kinases, Polo-like kinases (PLKs), p70S6K, and PIM2,
seems to be interesting. Further, using a panel of 378 kinase
inhibitors, we observed that ALL cell lines that were either sensitive
(NALM6 and 697) or resistant (TANOUE and JURKAT) to dexametha-
sone were highly responsive to inhibitors targeting Aurora kinases,
PLKs, and PI3K/mTOR pathway components (Supplementary Fig. 3).
Aurora kinase activity can inuence several signaling proteins
including β-catenin
21
. Because we observed that cells treated with
dexamethasone and prednisolone display enrichment in the
β-catenin responsive genes (Fig. 1d, e), we reasoned that Aurora
kinase may regulate the β-catenin pathway in dexamethasone-
treated cells. β-catenin is a component of the canonical WNT
signaling pathway where its protein level is maintained by the AXIN/
APC/CK1/GSK-3βcomplex. This complex phosphorylates β-catenin
resulting in β-TrCP mediated ubiquitination and, thereby, degrada-
tion in proteasome
22
. Aurora kinase directly phosphorylates GSK-3β
on serine 9 residue, which reduces the catalytic activity of GSK-3β,
resulting in the accumulation of β-catenin
23
. We observed that the
β-catenin level was upregulated in dexamethasone-treated cells (Fig.
2d). GSK-3βinactivation can also be mediated by several serine/
threonine kinases such as ERK
24
,p38
25
, and AKT
26
. GSEA shows two
opposite signatures in the MAP kinase pathway: KRAS down-
regulation and MEK upregulation (Fig. 1d, e). We observed that
while ERK phosphorylation was inhibited upon dexamethasone
treatment (Fig. 2e), p38 phosphorylation was upregulated (Fig. 2f).
The KRAS downregulation signature is in agreement with the
downregulation of ERK phosphorylation in dexamethasone-treated
cells. Nevertheless, MEK signaling can activate p38 without
affecting ERK phosphorylation at a certain condition
27
. Additionally,
dexamethasone-induced activation of p38 was previously described
in dexamethasone-sensitive RS4;11 and SUP-B15 cells
28
. Finally, cells
treated with dexamethasone displayed an upregulation of S6K
phosphorylation (Fig. 2g), similar to the kinase proling data (Fig.
2b), while it showed a downregulation of AKT phosphorylation (Fig.
2h). S6K is a downstream effector of the PI3K/AKT pathway and can
be activated by mTORC1 complex
26
. Likewise, S6K can directly
phosphorylate and inactivate GSK-3β
29
. However, dexamethasone-
induced S6K phosphorylation is likely to be independent of AKT
activation. Because we observed that ERK and AKT phosphorylation
was inhibited, while p38 and S6K phosphorylation was upregulated
by dexamethasone, we then checked the phosphorylation of their
substrate GSK-3βSerine 9 residue. We observed that GSK-3βSerine
9 phosphorylation was initially reduced and then increased over
time (Fig. 2i). Furthermore, cells treated with the Aurora kinase
inhibitor Tozasertib in the presence of dexamethasone displayed a
reduction of β-catenin accumulation (Fig. 2j), suggesting that Aurora
K. Shah et al.
2
npj Precision Oncology (2021) 13 Published in partnership with The Hormel Institute, University of Minnesota
1234567890():,;
kinase activity may be required for β-catenin accumulation in
dexamethasone-treated cells. SUP-B15 cells display constitutive
activation of ERK (Fig. 2e) and AKT (Fig. 2h), which probably
maintains initial GSK-3βSerine 9 phosphorylation. However, later,
this might be mediated by Aurora kinase, p38, and S6K. Several
other kinases such as protein kinase C (PKC) family proteins, protein
kinase A (PKA), ribosomal protein S6 kinase (RSK), etc. can also
phosphorylate GSK-3βat Serine 9 residue
30
. We have not tested
those kinases in this context and, therefore, we cannot exclude the
possibility of their involvement. Collectively, these data suggest that
β-catenin protein levels are stabilized by dexamethasone-mediated
regulation of multiple signaling pathways.
106 17217
Upregulated genes
Downregulated genes
196 54366
DMSO Dexamethasone Prednisolone
OLAH
STYK1
BAALC
EDNRB
TMEM236
TMEM217
ISG20
BMP2
STAG3
OR7C1
P2RY14
PLEKHG1
LDB3
EPHA4
CHRNA1
GLIS3
RCAN1
TSC22D3
SLC22A23
OR7A5
LILRA2
CALCRL
SMIM3
YBX3
NFIL3
BMP3
ELL2
TMEM204
FKBP5
ZC3H12D
TGFB1I1
NFKBIZ
GP9
NGFR
LMO2
SPTLC3
KLF9
ATMIN
CYTH3
DDR1
TLR5
KLF2
TP53INP1
SMAP2
SLC44A1
DDIT4
LONRF1
DENND3
RASD1
BTG2
ITGA6
EPC1
CPM
TNFSF4
FAM43A
RASAL2
SAP30L
SPRY1
SYNE2
DPEP1
MYO10
DKK1
EPAS1
FBXW7
CTHRC1
PON2
RAPGEF5
VPS37B
KLF7
LDLRAD4
TXNIP
TLE1
RNASET2
GADD45A
DUSP1
4681012
Expression level
a
b0.35 0.36 0.37 0.38 0.39 0.40 0.41 0.42
ES
HALLMARK_KRAS_
SIGNALING_DN
HALLMARK_TNFA_
SIGNALING_VIA_NFKB
HALLMARK_COAGULATION
HALLMARK_APOPTOSIS
HALLMARK_EPITHELIAL_
MESENCHYMAL_TRANSITION
HALLMARK_HYPOXIA
0.06
0.07
0.08
0.09
0.10
0.34
Dexamethasone vs DMSO
NES
1.44
1.46
1.47
1.54
1.61
1.63
0.45 0.5 0.55 0.6
ES
MEK_UP.V1_UP
PTEN_DN.V1_ UP
VEGF_A_UP.V1_UP
MYC_UP.V1_DN
IL15_UP.V1_DN
TGFB_UP.V1_UP
KRAS.50_UP.V1_DN
BCAT.100_UP.V1_UP
NES
1.59
1.59
1.61
1.62
1.65
1.67
1.68
1.94 0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.045
0.4
0.325 0.35 0.375 0.4 0.425 0.45 0.475
HALLMARK_IL2_
STAT5_SIGNALING
HALLMARK_COAGULATION
HALLMARK_APOPTOSIS
HALLMARK_HYPOXIA
HALLMARK_HEDGEHOG
_SIGNALING
0.10
0.12
0.14
0.16
0.18
0.20
0.3
NES
1.30
1.32
1.32
1.43
1.46
1.52
1.53
1.53
HALLMARK_EPITHELIAL_
MESENCHYMAL_TRANSITION
HALLMARK_KRAS_
SIGNALING_DN
HALLMARK_TNFA_
SIGNALING_VIA_NFKB
Prednisolone vs DMSO
ES
0.35 0.45 0.55 0.6 0.65
ES
MEK_UP.V1_UP
TGFB_UP.V1_UP
IL15_UP.V1_DN
MYC_UP.V1_DN
JNK_DN.V1_DN
PTEN_DN.V1_UP
VEGF_A_UP.V1_UP
BCAT.100_UP.V1_UP
NES
1.35
1.48
1.50
1.52
1.57
1.60
1.78
2.07 0.025
0.050
0.075
0.100
0.125
0.150
0.175
0.200
0.3 0.4 0.5
c
d
e
Fig. 1 Glucocorticoid-induced regulation in gene expression. SUP-B15 cells were treated with 1 µM dexamethasone, 2 µM prednisolone, or
DMSO for 6 h. Global gene expression was measured by Affymetrix Human Gene 2.0 ST Array. aThe number of downregulated genes in
dexamethasone- and prednisolone-treated cells. bThe number of upregulated genes in dexamethasone and prednisolone-treated cells. cThe
heatmap displaying clusters of upregulated genes in dexamethasone- and prednisolone-treated cells was generated by the heatmap.2
function of Gplots library in R. dGene set enrichments in dexamethasone-treated cells were measured by GSEA (Hallmarks and Oncogenic
signatures) and visualized by the pyplot.scatter function of Matplotlib. eGene set enrichments in prednisolone-treated cells were analyzed
and visualized as described above.
K. Shah et al.
3
Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2021) 13
Prediction of dexamethasone sensitivity in ALL
Because we observed that dexamethasone treatment mediates
the deregulation of certain survival signaling pathways, we next
aimed to verify the ndings in ALL patient materials. The TARGET
ALL dataset provides more than 200 annotated samples with gene
expression data but is missing dexamethasone sensitivity
information. Therefore, we aimed to predict dexamethasone
sensitivity in those samples. Various methodologies have been
Dexamethasone
(1000 nM)
Tozasertib 0 0 100 1000 (nM)
55 kDa
100 kDa
70 kDa
-actin
-catenin
55 kDa
100 kDa
70 kDa
-actin
Dexamethasone
0 8 16 24 (h)
-catenin
Aurora Kinase
GSK-3
-catenin
AKT
GSK-3
-catenin
S6K
mTO RC 1
pS6K
-actin
Dexamethasone
081624(h)
p-p38
p38
Dexamethasone
081624(h)
p38
GSK-3
-catenin
ERK2
Dexamethasone
081624(h)
ERK
GSK-3
-catenin
0-2-4-6 Normalized kinase statistic
2
1.3
Normalized kinase statistic
01 23-1
2
1.3
0
5
10
SPRY1 SPRY2 SPRY3 SPRY4
DMSO Dexamethasone Prednisolone
c
d
ef g
AKT
GSK-3
-catenin
70 kDa
55 kDa
pAKT
AKT
Dexamethasone
081624(h)
70 kDa
55 kDa
hij
55 kDa
55 kDa
-actin
pS9-GSK-3
Dexamethas one
081624(h)
Fig. 2 Glucocorticoid-induced regulation of kinase activation. SUP-B15 cells were treated with 1 µM dexamethasone or DMSO for 6 h before
lysis. aTo measure protein tyrosine kinase activity in dexamethasone-treated cells, lysates were applied in Pamgene peptide substrate-based
tyrosine kinase array and analyzed for tyrosine kinase activity using Pamgene software. bTo measure protein serine/threonine kinase activity
in dexamethasone-treated cells, lysates were applied in Pamgene peptide substrate-based serine/threonine kinase array and analyzed for
serine/threonine kinase activity using Pamgene software. cChanges in the expression of Sprouty family genes in dexamethasone and
prednisolone-treated SUP-B15 cells compared to DMSO-treated cells. Error bars show standard deviation. diSUP-B15 cells were treated with
1 µM dexamethasone for the indicated time period before lysis. Lysates were analyzed by SDS-PAGE and western blotting using specic
antibodies as labeled. jSUP-B15 cells were treated with 1 µM dexamethasone and with different concentrations of Tozasertib for 24 h before
lysis. Lysates were analyzed by SDS-PAGE and western blotting using specic antibodies as labeled. Blots shown in each panel were from the
same experiment and processed similarly.
K. Shah et al.
4
npj Precision Oncology (2021) 13 Published in partnership with The Hormel Institute, University of Minnesota
proposed to predict monotherapy response using different
modalities including single nucleotide polymorphisms (SNPs),
copy number variations (CNVs), mutations, DNA methylation,
protein expression, and RNA expression (Reviewed in ref.
11
). Both
the Cancer Cell Line Encyclopedia (CCLE)
31
and the Genomics of
Drug Sensitivity in Cancer (GDSC)
32
provide drug sensitivity scores
and various genetic features for >1000 cell lines. The GDSC dataset
lacks drug sensitivity scores for dexamethasone; therefore, we
used the CCLE dataset to predict dexamethasone sensitivity. The
CCLE dataset contains dexamethasone sensitivity scores for 708
cell lines from different cancers, where 138 cell lines are from
hematological malignancies. The use of RNA expression remains
the most popular modality for drug sensitivity prediction
9,10
, and
therefore, we used RNA expression data to serve the purpose. The
CCLE RNA expression dataset contains >18,000 genes for each
sample. The use of >18,000 genes with drug sensitivity data for
138 cell lines can perhaps be used to build a prediction model but
the model will suffer from poor generalization performance when
tested on new data. A high-dimensional dataset with too many
features can lead to overtting such that the model captures both
the real and random effects
11
. Low complexity models such as
logistic regression, principal component regression, partial least
square regression, or support vector machine can handle higher
dimensional data but, due to the linear nature of the algorithm,
predictions introduce signicant modeling bias
11,33,34
. Deep
learning models provide superior prediction accuracy over linear
models where the number of samples is comparable to the
number of features
11
. To increase the ratio between sample and
feature, we attempted to reduce the number of genes. We used a
combined gene signature (Supplementary Fig. 4) of 500 genes to
develop a deep learning model. Dexamethasone sensitivity data
for 138 cell lines relating to hematological malignancies were
collected from the PharmacoDB database
35
and gene expression
data for those cell lines were collected from the CCLE database.
Cell lines with IC
50
< 700 nM were considered sensitive and cell
lines with IC
50
> 1000 nM were marked as resistant. We used the
Keras
36
sequential model to build the drug sensitivity prediction
model. The model was tested using 708 CCLE cell lines data and
two small ArrayExpress datasets: E-MTAB-7781 and E-MTAB-9250
(Fig. 3a). When combining all three datasets, the model predicted
684 samples as resistant (out of 709) and 43 samples as sensitive
(out of 43; Fig. 3b). The model displayed more than 95%
prediction accuracy, while the negative predictive value was
63% (Fig. 3c). This is because, to test the model, we used a larger
number of resistant samples as compared to sensitive samples.
Overall although the model performed well with these datasets, it
has not been thoroughly tested with a large dataset, and
therefore, we cannot exclude the possibility that the model might
perform poorly with a large unknown dataset. Furthermore, we
observed that in a small set of RNAseq data where three samples
were experimentally dened as sensitive to dexamethasone and
ve samples as resistant, the model predicted two samples from
the resistant group as to be sensitive (Fig. 3a). To understand why
the model failed to detect two samples as resistant, we analyzed
the gene expression data (FPKM) from those samples. In this
dataset, three sensitive cell lines (697, NALM6, and RS4;11) were
treated with dexamethasone to generate dexamethasone-
resistant cell lines
5
. The resistant cell line TANOUE was used as a
control and was also treated with dexamethasone for the same
period. Comparisons between dexamethasone-treated cells and
naïve cell lines demonstrated an extremely high correlation of
gene expression for 697 (R
2
=0.9767), NALM6 (R
2
=0.9581), and
TANOUE (R
2
=0.9929), while RS4;11 displayed a comparatively
lower correlation (R
2
=0.5977) (Supplementary Fig. 5a). This
suggests that the model expects some differences in gene
expression between sensitive and resistant cell lines. This is also
evident by the fact that TANOUE and other sensitive cell lines
displayed a moderate correlation (Supplementary Fig. 5b).
However, the model could efciently detect sensitive cells even
though there was a moderate correlation between them
(Supplementary Fig. 5c).
Dexamethasone sensitivity prediction in ALL patients using
the deep learning model
We used the ALL patient dataset from the TARGET database to
predict dexamethasone sensitivity. The gene signature (500
genes) from 205 ALL patients was used to predict dexamethasone
sensitivity. The model predicted 96 patient samples as dex-
amethasone sensitive and 109 samples as dexamethasone-
resistant (Fig. 4a). Because dexamethasone sensitivity can predict
patient survival
3739
, we then compared event-free survival
between the dexamethasone-sensitive and -resistant groups. We
observed that the dexamethasone resistant group had signi-
cantly (p=0.0098) reduced event-free survival (Fig. 4b). The
TARGET ALL dataset contains samples that have been classied as
those obtained from bone marrow and peripheral blood. Thus, we
further determined the difference in event-free survival by
dividing the patients into bone marrow and peripheral blood
groups. Interestingly, a signicant survival difference was
observed in the bone marrow group (Fig. 4c), but not in the
peripheral blood group (Fig. 4d). However, this observation must
be interpreted carefully, as the peripheral blood group contains a
smaller number of patients.
The dexamethasone-resistant ALL patient group displays
similar pathway enrichment as the glucocorticoid-treated cells
Next, we compared gene signatures between the predicted
dexamethasone-sensitive and -resistant patient groups. We used
ALL samples from bone marrow to identify pathway enrichment.
Epithelialmesenchymal transition (EMT) was found to be highly
enriched in samples from dexamethasone-resistant patients (Fig.
5a), which was similar to that of dexamethasone-treated SUP-B15
cells (Fig. 1d). Several other interesting pathways that were
enriched included PTEN-DN, β-catenin, TGF-β, and WNT upregula-
tion signatures (Fig. 5b). Similar pathways were found to be
enriched in dexamethasone-resistant ALL cell lines (Supplemen-
tary Fig. 6a, b). Further, we observed the downregulation of
several genes including FLT3, BCL2, SOCS2, etc. in samples from
dexamethasone-resistant ALL patients (Fig. 5c). Additionally,
enrichment of the KEGG Cytokine-Cytokine Receptor Interaction
gene signature was observed in samples from dexamethasone-
resistant ALL patients (Fig. 5d), which is in line with the
observation that SOCS2 expression is downregulated as SOCS
family members act as negative regulators of cytokine receptor
signaling
40
. To verify the observation that several genes were
downregulated in samples from dexamethasone-resistant ALL
patients, we compared the expression of FLT3 and SOCS2 in
dexamethasone-sensitive SUP-B15 and -resistant TANOUE cells.
Data are in line with the ndings from patient samples, in that
FLT3 (Fig. 5e) and SOCS2 (Fig. 5f) were downregulated in TANOUE
cells. Although we observed that FLT3 expression was down-
regulated in predicted dexamethasone-resistant ALL patient
samples, in vitro development of dexamethasone resistance in
cell lines using long-term culture retained FLT3 expression in
resistant cells and even selected for oncogenic FLT3 mutations
5
.
Therefore, it is likely that the loss of FLT3 expression occurs in a
portion of ALL patients and might not be a common feature for all
dexamethasone-resistant ALL samples.
Several kinase inhibitors display synergy with dexamethasone
in dexamethasone-resistant cells
Next, we sought to identify drugs that can overcome dexametha-
sone resistance. We attempted to develop a method for predicting
synergy between dexamethasone and other drugs. For this, we
K. Shah et al.
5
Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2021) 13
used drug synergy data from the DrugComb database. We used
488 chemical descriptors to describe each drug and the 500-gene
signature to describe each cell line. Data were normalized before
being introduced to the deep learning model. Because we had a
limited number of cell lines and drug combinations, we used
binary classication such that, for drug combinations, a BLISS
score >3 was considered synergy (dened as 1), and a BLISS score
<2 was considered no synergy (dened as 0; Fig. 6a). Therefore,
the model was trained only for the absence or presence of
synergy. We were able to build a model with 82.4% test accuracy
(correlation coefcient =0.64; Fig. 6b). The deep learning model
achieved a moderate level of prediction accuracy (Fig. 6c). Further,
to predict synergy, we used 1454 kinase inhibitors and 40 ALL
patient samples from the TARGET dataset. The model predicted
synergy between dexamethasone and 226 inhibitors that inhibit
kinases involved in regulating apoptosis, the cell cycle, JAK/STAT,
MAPK, NFκ-B, PI3K/mTOR, Stem cells/WNT, and PKC/TGFβ/SMAD
pathways (Fig. 6d). Because we used a relatively small amount of
aSUP-B15
9 samples
7.1045 6.8036 7.9267 7.2883 8.3804 6.81564.7937 7.4253 7.0229 7.5627.4826 4.6459 7. 6792 5.89625.7176 7.8646 5.0615 4.5422 7.0957 5.3132 7. 4 558 7.40664.7239
4.3118 4.2407 4.1242 3.9557 4.924 4.79434.4007 4.2998 4.2722 4.4619 4.51 5.1424 4.4471 4.15314.3016 4.3911 4.7555 4.3484 4.4547 4.7317 4.5259 3.91654.5491
11.623 10.689 10.759 10.599 10.869 8.945 10.26 11.534 11.495 10.762 9.2136 11.187 4.3937 10. 98411.047 7.2427 10.351 11.297 11.163 11.383 12.387 12.5669.8118
4.5613 4.4014 4.6288 4.6252 6.4908 4.85194.4848 4.5594 4.4798 7.6185 7.6583 7.8632 4. 3198 6.17274.3173 4.3232 8. 0747.1978.1939 8.5051 4.4448 8.23957.9582
5.807 5.
0637 6.9797 6.0856 5.5515 4.99225.5143 4.6204 7.9521 6.0786 6.3109 7.1382 6.5664 6.66844.6969 4.3737 4.6067 6.2828 7.3251 6.6163 6.1815 7.01675.1127
6.3887 5.7243 5.2928 5.9684 5.3542 5.31735.4663 5.3846 5.4235 5.5014 5.9954 5.7772 5. 7943 5.36055.6869 5.1204 6.6104 6.4554 5.1465 6.7153 5. 4647 5.71416.0932
9.6645 9.6344 9.733 8.6469 9.7521 8. 96119.4817 9.614 9.5294 9.5423 9.806 8.6033 10.174 8.73839.1517 9.451 8.6378 8.933 9.9126 8.6381 9.4842 9.92078.8756
4.1167 3.7791 4.0528 3.9851 4.136 4.18143.7499 3.7692 3.8993.9649 4.0339 3. 9025 3.701 4.04483.7684 4.0432 3.9979 3.9973 3.7153 3.7599 3.7747 3.76123.7499
3.7567 3.6589 3.8643 3.7
151 3.7823 3. 88773.9512 3.6862 4.4559 3.6184.274 3.8548 3.6346 7.63693.6765 3.8488 3.7513.5977 3.7626 4.0977 3.6427 5.06953.9686
4.8141 4.3307 4.7705 4.9953 5.0123 4.80185.8157 4.4458 5.2542 4.5171 4.5239 4.7492 5. 2895 4.39444.7186 4.7079 4.8508 4.9744 4.3163 4.7252 4. 7365 4.59884.6951
3.9183 4.2529 4.1707 4.1291 4.329 4.13134.2429 4.1777 4.3106 4.4079 4.1143 4.5044 4.2397 4.20384.7171 3.9245 4.1333.9147 4.2759 4.6319 5.2884 4.25124.6519
5.7173 4.8425 5.197 4.9501 4.6135 4. 73086.7148 5.3781 5.4206 4.9635 5.8687 4.4022 4.4483 5.612 5.8892 5.0155 4.3496 4.3403 4.894 5.4211 4.91 5.09324.3789
7.9126 7.8728 7.8423 8.1105 8.4639 7.262
77.8777 7.505 8.0953 7.2326 7.9645 8.1973 8.4645 8.26118.1163 8.799 8.1091 8.1118 7.1725 8.069 8.7641 8.08347.894
9.8423 10.495 10.281 10.043 9.6925 10.47210.32 10.225 10.169.9143 9.8212 9.4768 8.9426 9.769210.051 10.263 10.2510.096 10.185 9.5889 8.6963 9.833410.506
7.059 4.4031 4.3454 4.5937 9.2153 9.95169.3088 6.2552 4.0192 6.3459 5.3738 10.466 10.06 9.77797.5939 9.3933 9.8278 11.052 7.548 9.2133 8.5284 8.27899.0541
4.173 4.0634.1168 7.0563 4. 3 651 6.27794.0234 4.0916 4.0321 4.1669 4.1912 3.8519 3. 9942 6.53544.0208 4.2589 4.2529 3.8862 3.8997 4.1883 4. 1121 3.99494.0416
8.4714 9.2141 8.9915 9.0409 8.7699 8.82118
.3096 9.7631 9.5627 9.3338 9.0293 8.8979 8.5995 8.68799.0326 8.6965 9.2088 8.292 9.6742 9.0274 8.932 9.09848.8506
12.263 11.225 12.015 11.632 10.607 12.26211.295 11.75 12.464 11.579 11.927 10.936 9.8074 11.69211.667 11.591 10.579 11.365 11.35 11.205 10.912 11.90312.412
6.1626 5.6007 5.9418 5.6744 5.7918 6.89015.6966 5.531 5.9302 5.4833 6.0609 5.9678 5.5034 5.56175.5993 5.6466 6.8944 5.4291 6.3824 5.8088 6. 4026 5.48846.0964
8.124 8.0296 8.2951 8.3889 10.086 8.793310.211 8.0178 7.9147.9548 8.6494 8.2849 8.8535 9.38539.3438 10.284 8.3403 9.0506 7.9444 9.072 8.3775 8.724 9.1207
6.3518 6.9057 6.4795 6.532 6. 1923 5.97546.3298 6.8348 6.3625 6.
7434 6.1879 6.6916 7.3053 6.55567.136 7.4934 6.7028 5.7244 6.4609 6.9765 7.4205 7.07117.0339
5.4571 5.2851 4.8804 4.6307 9.8167 5.89155.273 8.6811 5.7692 7.8559 5.022 5.4514 4.8946 5.78035.7145 5.6376 5.2386 5.2986 5.1181 5. 2226 5.2388 5.20715.2493
9.0468 8.3839 8.2668 9.234 8. 4926 8.397110.23 7.4263 8.5992 8.3389 9.3065 8.8827 9.213 9.11468.5678 8.7106 9.0062 9.2547 8.0555 9.3576 8.915 8.56928.9946
7.6801 6.3977 8.1442 5.8795 7.8105 9.91027.6879 8.1504 4.6886 4.8136 6.0052 6.1455 7.5706 9.83225.9674 7.6905 4.2372 5.3972 6.2066 6.8502 9. 6133 9.07667.2245
3.4996 3.4067 4.2367 4.5054 3.6764 3.52093.492 3.5593 3.6286 3.4055 3.
5282 3.4322 3. 3887 3.51083.2642 3.6526 3.3334 3.4929 3.3598 3.4734 3. 4908 3.47453.6136
5.164 4.7654.8686 5.1863 5.2374 4.82734.8763 4.8349 5.0723 4.7611 5.1678 4.5433 5. 3809 6.13145.3922 5.5644 4.9473 4.7678 4.5481 5.1934 5. 3321 4.65214.7433
3.7419 3.9231 3.6238 3.8327 4.0348 3.75343.7474 3.8584 3.8745 3.6546 3.6242 3.9612 3.663.57393.681 4.0388 3.9275 3.9903 3.7094 3.8425 3.8709 3.57093.7994
3.8754 3.9891 4.0796 4.0227 3.8495 3.87175.6198 3.8451 3.8269 4.0848 3.9579 4.1308 3.9487 4.19414.2053 3.7591 4.0037 3.8701 3.881 4.0986 3.99 3.87934.1665
5.498 4.8267 4.5961 5.495 5. 0184 5.16894.8256 4.728 4.5253 4.9554.8482 4.9261 4. 90
24 4.66765.1693 5.1076 5.0377 5.1337 4.5241 4.8961 4.9858 4.873 4.8178
4.3048 4.4314 4.4182 4.3242 4.6898 4.48954.3084 4.2988 4.6071 4.1839 4.1719 4.1927 4.5295 4.21814.3159 4.2526 4.6922 4.3501 4.0829 4.1364 4. 5777 4.21244.3147
6.0535 5.4089 5.1947 5.5984 5.4666 5.82415.1053 5.1644 5.2612 5.8716 5.6075 5.9706 5.6756 5.55435.1829 5.7797 5.5607 6.0473 5.3105 6.5851 5. 4207 4.99925.3624
3.5826 3.8118 3.9069 3.9532 3.4141 3.50143.5505 3.5989 3.6784 3.5376 3.9698 3.6909 3.6754 3.56433.6801 3.6528 3.4806 3.8003 3.6308 3.7229 3. 9119 3.62843.7369
3.4463 3.4776 3.5905 3.4012 3.4435 3.51693.6041 3.4686 3.5254 3.7119 3.6068 3.7775 3.5019 3.55243
.3916 3.6809 3.5952 3.6536 3.3619 3.6925 3. 6434 3.25563.419
7.7508 4.5096.2958 4.7473 4.3748 4.42789.2526 4.2847 4.2588.9176 8.8999 9.2227 9.5884 9.80654.8811 4.3151 4.4745 4.8093 7.3204 9.1476 9.0941 10.2785.8737
3.8401 4.0086 4.3607 3.9105 4.1445 4.66273.8117 3.8872 3.9793.74 3.9026 3.8088 3.9649 3.60823.8479 4.4016 3.9283 3.7369 3.5956 4.0576 4. 1475 4.15593.9287
5.4216 5.5034 4.8354 5.3348 5.4256 4.28595.2586 5.7599 5.5823 5.1276 4.9888 5.3376 5.1457 4.90095.1946 4.6942 5.8422 5.3795 4.569 5.6435 5.7852 5.59625.439
4.2639 4.0096 4.6137 4.5068 5.0071 4.88534.3939 4.2539 4.1996 4.4674 4.6286 4.6236 4.3727 4.17214.4387 5.
1159 9.3094.3376 4.1881 4.353 4.1343 4.38538.828
4.8246 3.6944 3.7296 3.9303 4.1082 4.49653.8942 3.6799 3.9421 3.7224 3.841 3.9969 3.7693 3.86913.6516 4.1935 4.0112 4.0796 4.4423 3.8724 4.147 3.63223.5096
4.1253 4.1609 4.0017 4.1214 4.4042 4.19514.2438 4.0953 4.2753 4.6966 4.2085 4.4783 3.7575 3.952 4.2671 4.095 4.3375 4.3415 3.9342 4. 3963 4.252 4.17744.5243
3.3054 3.2365 3.4322 3.7926 3.5602 4.03483.2453 3.2413 3.4958 3.2913 3.4422 3.2971 3.2098 3.55863.2773 3.5942 3.3749 3.5488 3.3598 3.4436 3. 4111 3.29993.3462
4.681 4.3303 4.4918 4.3448 4.7126 4.35873.98 4.2455 4.8262 4.3369 4.3872 4.1996 4.3769 4.451 4.2542 4.875 4.
8876 4.5606 4.1805 4.6515 4.2573 4.38664.4506
3.5187 3.7066 3.6408 3.6856 3.4991 3.64473.6252 3.6243 3.5271 3.6119 3.4945 3.6936 3.4466 3.42513.4203 3.7271 3.5798 3.6306 3.388 3.5582 3.4518 3.47713.3919
7.7826 7.2537 6.1952 7.5724 6.3999 5.80265.3993 7.6802 4.9809 7.2572 7.9672 6.6352 5.9492 6.73116.11 6.8033 6.1987 6.7743 7.6954 6.4984 7.2287 7.035 6.8849
3.8083 3.7869 4.1706 3.7839 4.1781 3.98823.8373 3.835 3.7501 3.8063.9424 4.0009 4. 0302 3.72273.9311 4.668 3.83 4.0049 3.6593 4.0018 3.889 3.76233.8723
4.3317 4.7534.2944 4.4227 5.1144 4.722 4.6533 6.1354 4.6605 3.9563 4.7107 9.7078 10.921 4.96774.3585 5.2411 8.6315 4.6882 4.
4144 8.2461 3. 7651 3.99435.0543
4.837 5.6547 3.7789 3.84 3. 6143 3.80913.6569 5.3303 3.6295 4.6435 4.3811 3.5915 3. 9367 3.65734.1498 3.8003 3.4935 3. 704 4.9137 3.187 6.1577 5.75314.0046
4.6329 4.7603 4.5209 4.6405 4.759 4.40914.537 4.6602 4.6079 4.4606 4.6936 4. 4629 4.2626 4.85674.6771 4.7095 4.7291 4.9058 4.1381 4.6525 4.579 4.34794.4012
6.937 6.9116.7684 6.6693 6.722 5.96916.7597 6.6158 7.3267 6.8826 6.5481 6.9214 6.5583 6.48796.0239 6.9032 6.0709 6.6191 5.9466 6. 5382 6.5179 6.418 6.7608
6.35 6.6199 6.3792 7.0788 5.5971 6.91784.6325 6.9523 4.8167 6.4866.2443 7.0409 7. 0831 5.56125.9729 7.0519 5.6681 5.5244 6.5467 5.7252 5.913 6
.25174.9263
4.0239 4.0949 4.2084 3.9145 4.3172 4.39294.2612 4.1765 4.7523 4.4873 4.2088 4.1044 4.0749 4.21974.0021 3.7744 4.3206 4.0168 4.1665 4.4348 4.068 4.31714.3921
TARGET
205 samples
4.962266 5.4960497.390905 7.3849955.842099 6.4713135.892537 6.444248 7.0625385.704079 5.2173537.053355 6.4757317.062844 5.9056836.1829296.776262 6.6644836.490469 6.4273396.777066 5.4491785.969378 6.724003 6.3604495.76378 6.9412666.530368
6.366588 4.6016555.0118
71 5.2055854.78089 3.9768764.129963 5.034856 5.3817075.236166 5.0313936.110261 4.7109595.142083 5.57029 5.0699935.011887 5.0400035.515817 6.1596654.378878 5.5733565.175855 5.663197 5.3450164.596672 4.8036414.22253
3.04495 3.523 3.125356 3.2061713.030879 3.5588823.684911 2.
912247 2.9152424.202212 2.8125663.141418 3.3797043.313725 2.7961063.1011363.286911 3.054312.766369 3.099372.862117 3.9204483.266078 3.42324 3.3385413.194318 3.0220453.174083
5.140967 5.3036215.82934 5.5507575.202031 6.9217366.598661 5.377669 5.5238795.762384 5.1
576055.171587 5.6256075.716035.3234925.7186945.988764 5.019654.933329 5.193375.811981 5.4034445.712669 5.80128 5.80885.888597 5.3419015.824795
2.554642 2.8464792.593315 2.6365692.661861 2.7788943.143265 2.627435 2.6122262.780144 2.675122.67322.6731262.702776 2.6
6356 2.6936592.702862 2.5813272.622279 2.4704232.822701 2.6780882.523595 2.781962 2.9805292.749078 2.6770312.828663
7.45377 6.5797687.642406 6.555847.100583 5.7321625.444057 6.419515 6.736246.411252 7.1021738.412145 6.8130357.580988 6.7759017.0396826.951032 7.1
093286.23991 7.7469856.906531 7.4040466.755989 7.612718 6.4305146.430788 7.4057146.425846
3.327709 3.491473.440642 3.7040013.356744 3.9341754.026879 3.19314 3.4999084.113081 3.5007973.299175 3.5690053.596159 3.5362643.9026973.717741 3.5280613.147004 3.4933693.5415
12 3.4420573.670137 3.562302 3.5872513.616655 3.4638933.944659
3.274916 3.1543573.10298 3.0162023.371606 3.5854583.347684 3.111011 3.0476933.755288 2.9984273.167709 3.2142663.216792 3.0968853.0485773.190392 3.0322513.078544 3.2300242.841486 3.0514343.18523.219178 3.2
268743.300607 3.1211093.266039
6.648789 3.4734544.781245 2.593193.371825 2.9359713.408514 4.23789 4.8743522.788488 3.4248023.16835 3.1019443.409173.6509853.1039222.636173 2.8477743.124163 4.1566912.929181 3.4426272.766817 4.389604 2.7046052.625716 3.1349783
.094915
2.717853 2.7962192.933841 2.778352.926863 2.7983982.780444 2.776618 2.8777592.858696 2.6453082.734573 2.8259552.917922.9526812.7870482.855292 2.8404352.803073 2.742632.728347 2.6538612.862062 2.6571 2.9024752.915989 3.1928782.751014
3.294846 3.11011
53.393367 3.3646683.266038 3.2454273.331404 3.217426 3.2621733.609811 3.2754643.163101 3.3372833.247562 3.0965813.2480573.304066 3.0759723.234429 3.273583.158696 3.3624093.499173.248824 3.5550563.263762 3.448653.184718
4.443083 4.3842874.493744 4.3447954.211759 4.3
919374.260103 4.295547 4.2776214.484783 4.7353894.417066 4.3747934.499134 4.1063244.4098164.168375 4.2262664.109906 4.2489764.164953 4.6204584.030147 4.318091 4.3905174.068924 4.5025154.188835
3.690213 3.6521833.647928 3.6640443.758452 3.9722854.252696 3.578092 3.6
0973.859571 3.7412243.570512 3.7251773.716238 3.6078013.77873.911196 3.5560183.424829 3.5393293.733154 3.4941923.639869 3.450127 3.7745934.270598 3.6175233.91894
6.560036 6.3044786.796744 6.6839735.417156 5.6346735.788019 4.689195 6.1340026.093253 5.0282997.352489 5.9
677286.886746 6.5494185.8265455.993345 6.2373126.161873 6.8886925.984032 5.3920626.699239 7.227452 5.8040997.423059 6.6745996.72251
5.645838 3.305644.277017 3.7909333.478091 3.4950294.759184 2.991302 3.5581243.51553 3.3197724.061367 4.1748324.421335 3.6154834.07747
23.271558 3.6170344.08911 4.4155254.006516 4.0802794.221434 4.112956 4.1275423.249222 3.7256453.579089
3.071705 3.2629913.052413 3.0612913.387486 3.3749013.633597 3.269655 3.4983523.21824 3.3137383.013373 3.4502333.242665 3.24504 3.6256033.206576 3.0561612.816182 3.0
343783.370386 2.9623083.158909 3.187486 3.4175573.553714 3.5286813.69926
3.445777 3.0483193.435181 3.4922423.337782 3.9933923.478695 3.29374 3.3041744.907913 3.0903383.313205 3.460033.379774 3.2529383.2918783.48395 3.2227233.187753 3.465983.409312 3.4769153.51
7294 3.464745 3.7042093.570093 3.2705263.429753
4.480082 3.1543273.790449 3.7136564.050282 3.1918562.801503 3.410098 3.2887672.809947 3.4925934.646204 3.2554413.262421 3.9731294.2039423.194362 3.399273.08993 4.8803783.709592 3.9978454.098811 3.693619 3.1629213.1699
04 3.4893813.019332
6.723762 5.4372296.089011 5.696635.678913 4.8091735.249387 5.308392 5.431325.032548 5.7601246.507622 4.9980595.231431 5.3472875.5417525.229084 6.049815.678303 6.6381035.426147 6.1591326.175904 5.635375 5.1503345.023016 5.2837444.771617
2.6341
37 2.7358972.80098 2.7568852.788657 3.6312173.157315 2.836105 2.8883882.976543 2.5533322.873346 3.1110482.713864 2.84434 2.8853982.970892 2.5832722.695384 2.6801562.544062 2.6139842.837223.092796 2.7531823.060874 2.7300443.032397
2.501462 2.6902432.788682 2.6084
72.689867 2.8210823.059595 2.845191 2.876723.014369 2.6910932.775124 2.752962.552314 2.50648 2.6615812.726832 2.7575962.651827 2.6137492.808493 2.56309 2.659285 2.66246 2.6793472.827496 2.8061352.827071
3.610495 3.7750183.727513 3.8090093.622038 4.0838594.258317 3.
506126 3.7147064.55836 3.6151673.711259 4.0280753.814476 3.3380554.1109643.801166 3.8303093.50908 3.5560523.563217 3.7043493.784059 3.579897 3.8041913.844813 3.8449893.838303
3.567535 3.4830453.487935 3.3510593.487444 3.4482614.186926 3.272183 3.5389534.401544 3.4
146443.461146 3.3305083.487493 3.3979113.4886423.431391 3.3911773.532873 3.4025333.413859 3.41374 3.650988 3.510673 3.4534583.430218 3.8027513.566991
6.597425 6.3317667.160222 6.3059935.324202 4.9297975.050385 5.434705 6.9720463.55883 7.0195026.617454 5.256915.37
5893 7.1415855.9695726.16392 5.7340036.958791 7.3267895.034802 5.9942066.775095 5.897751 5.8952044.881444 5.8108554.683116
6.433156 5.5030326.838887 4.8472265.636753 3.9214693.842158 5.206763 7.0888573.630466 6.7556696.388309 5.6544895.165848 7.1470255.4956916.
524445 4.9953916.282305 7.4215484.333613 5.4335816.402626 6.048034 5.3953883.590141 5.7076113.944967
3.759121 3.8219683.844206 3.9172113.723271 4.1137614.406469 3.886263 3.5692744.349499 3.9463073.688818 3.8485284.083833.8052383.6300544.101499 3.8984323.6346
94 4.0527983.644653 3.7156623.993271 3.886731 3.8882954.292526 3.7809894.070416
3.502001 3.7550284.272296 3.7689594.224998 3.4568763.638938 4.309554 4.1641453.552212 3.4133993.368959 3.777214.005571 3.89132 3.5571613.26779 3.9221245.26181 3.891324.352842 3.49291 4
.183854 3.935388 3.7378993.584429 4.4992763.943918
2.995695 2.9312972.802379 2.7439112.928455 3.23734 3.018822 2.758575 2.662883.467107 2.8120672.674209 2.9834362.883756 2.5362343.3016263.031775 2.7822932.844523 2.7846122.799708 2.8010162.953272 2.915428 2.9860052.
961855 2.9091313.106779
4.036089 4.6574444.481005 3.9707624.099623 4.9484275.393816 4.5055 4.4004554.852551 4.3963124.158611 4.4571014.703618 4.2938474.3856714.770094 4.3326623.987669 4.270864.549124 4.1940714.480277 4.219378 4.2507585.288807 4.3640514.954185
2.907431 3.0043372.905765 2.9560943.37157 3.8426353.516669 2.916557 2.9453583.39649 2.8888963.03416 3.2768693.036226 3.0001062.9192512.799091 2.9033573.048795 2.7780653.108809 2.9158523.06093.003643 3.1781393.323084 2.7963323.259679
4.368358 4.6336184.916895 4.46311
54.36057 5.06636 5.359789 4.126754 4.4968555.555518 4.5129214.33874.5957584.846743 4.2419054.8273944.775707 4.4211774.427625 4.4406854.543112 4.2086874.570944 4.574359 4.7946195.061302 4.5799494.844379
3.428197 3.107073.163586 3.111063.319702 3.0623453.272893.05
1831 3.5373032.996426 3.0498673.072494 2.9534432.972453 4.5709383.086563.082138 3.2534773.598244 2.982023.245876 3.0106462.855874 3.104927 3.1334153.225259 3.1137613.264391
3.455539 3.1602193.41782 3.4480023.235255 3.2633853.368249 3.426668 3.1755043.704806 3.250
9643.554156 3.1641333.359736 3.3934253.3516833.292578 3.1440733.159925 3.2757063.273243 3.22221 3.338864 3.296351 3.1836243.3264 3.2327543.129452
3.02468 4.746055.055789 6.6524135.244407 4.3599754.207372 4.854264 5.1810534.219902 3.7550196.542328 5.2018794.987088 5.3
128865.209544.710925 5.3066026.379941 5.955285.178231 3.74427 5.376162 5.473344 4.4232525.101427 4.708274.833483
3.017669 3.1776382.933674 2.9110733.202466 2.9183153.452999 2.733505 2.7295033.625902 2.8636873.069698 3.1763732.892606 2.9386092.8918643.053587 2.8
931982.873251 2.9183152.802214 2.7966542.882924 2.797549 2.920352.81189 3.016023.107859
3.257046 3.2692293.256966 3.5186263.180773 3.6530453.425843 3.254413 3.4452414.128544 3.1047393.127216 3.5732223.210637 3.5752743.3626583.612506 3.5390063.352903 3.3259563.256613 3.2
053623.389224 3.179034 3.5444583.402761 3.6816753.683555
3.546906 2.5937152.730734 3.2425892.538282 2.6677462.619898 2.708701 3.3302922.569194 3.2948512.928011 2.7305652.906557 3.1211172.9322552.631627 2.8970433.646343 3.4516552.759963 2.9191742.990683 2.945715 2.6
9172.439403 2.8035552.603105
3.326727 3.1437153.200542 3.270833.241086 3.3769383.680857 3.280915 3.0705623.549671 3.2909023.332717 3.2226423.028387 3.52272 3.3632083.688635 3.0660493.118305 3.1230583.447249 3.1809563.197917 3.144869 3.3700183.358186 3.4145763
.432168
3.884485 3.8744144.181056 3.9588763.755874 3.8120594.184297 4.249894 3.6027794.291664 4.1535643.840046 4.0262154.108535 3.9139184.0510593.689286 3.9554673.840404 3.8825043.906392 4.0371374.356667 4.099824 3.6844723.93656 4.1734743.738387
2.855263 3.37648
12.72792 2.6829312.740445 3.0343792.804137 2.930765 2.89813.392192 2.9435682.990618 3.0688563.126503 2.9127572.9202992.828001 2.923212.989278 2.8865822.915855 2.8892192.786963 2.674673 3.0609682.981215 3.284243.022441
2.694683 2.8470082.94066 2.6807152.867365 2.8165962.9235
65 2.826339 2.8875713.086244 2.7899632.89345 3.0189412.907913 2.6969182.7716813.254592 2.6331322.599783 2.7071852.723135 3.0800712.993043 2.904679 2.919133.166265 2.5918042.679049
3.530154 3.4255343.723635 3.871023.527304 4.0146944.446836 4.433846 3.9537833.648594 3
.4516713.589213 3.8331493.677154 3.9393863.7157143.487766 3.5273043.486256 3.5189763.601679 3.4885233.881363 3.744045 3.710663.508877 3.871023.605685
3.887156 3.9329974.340394 5.672644.597549 3.3960163.056864.865908 3.3724523.040542 5.7422986.026737 3.5912424.62
8339 3.5112314.9677343.332006 4.1954213.624325 5.3350824.579294 6.5513425.532634 3.878219 2.968293.214659 3.9515213.499144
4.045153 4.0845663.90068 5.2883174.056699 3.4463163.347874.343519 3.3976833.7432 5.4061735.666103 3.7574814.260087 3.82435 4.8939523.753135 4
.2317073.840951 4.8233294.485851 5.8273245.191436 4.264478 3.5915753.478786 4.4029734.0936
2.73306 2.6035392.62647 2.581492.869977 3.40703 2.771598 2.92346 2.7341242.948984 2.9838182.83887 2.7127222.773879 2.4888662.8420662.94001 2.7095292.601977 2.6715572.83
5568 2.5947732.766482.860245 2.8104352.828575 2.762962.862547
2.865628 2.7905942.733143 2.435922.529552 3.0177172.692024 2.713386 2.7173162.886413 2.6882872.627029 2.7827432.652377 2.5362342.7878772.833709 2.7980492.664065 2.7637372.783178 2.7189942.655843 2.8115
17 2.8864132.593698 2.6374552.769216
4.584054 3.647815.082536 5.0264533.694441 3.7345363.415534 3.414175 4.6084193.30229 3.5619684.946912 3.3297014.149469 4.2479574.1969023.480603 3.1420113.35233 5.0817323.737899 3.9292564.387206 3.528849 3.1891773.650475 4.2582573
.267109
6.265464 3.2170073.961873 3.6509623.070226 3.4107493.502886 3.559826 3.311742.81691 3.7354713.140566 3.238553.437054 3.8319953.0896973.822384 4.3619992.817104 4.965654.495529 2.97304 4.085184 4.05188 3.3966123.328936 3.8763783.354454
3.362448 2.8532382.
986624 3.1377482.774013 2.7120792.646102 3.25113 2.5735272.671935 2.9046783.645783 2.5214172.628032.9318552.9417642.696465 2.6707393.067895 3.9148 2.933398 3.3282873.138363 3.314829 2.6086392.780297 2.5687642.785635
4.729431 4.6934264.681956 4.6615524.567877 5.525
8986.033706 4.389911 4.3615854.816371 4.8419374.715657 4.9422574.551882 4.6566445.1920974.732853 4.6129114.428877 4.5369854.658873 4.80385 4.698201 4.656379 5.0879024.737953 4.2783495.096647
CCLE
917 samples
loss
accuracy
val_loss
val_accuracy
Binary classification
Sensitive = 0
Resistant =1
Hidden layers
677 Resistant
31 Sensitive
27 Resistant
9 Sensitive
5Resistant
3 Sensitive
True Resistant (654)
Resistant (Predicted) Sensitive (Predicted)
Resistant (Experimental)
Sensitive (Experimental) False Resistant (0)
False Sensitive (23)
True Sensitive (31)
True Resistant (27)
Resistant (Predicted) Sensitive (Predicted)
Resistant (Experimental)
Sensitive (Experimental) False Resistant (0)
False Sensitive (0)
True Sensitive (9)
True Resistant (3)
Resistant (Predicted) Sensitive (Predicted)
Resistant (Experimental)
Sensitive (Experimental) False Resistant (0)
False Sensitive (2)
True Sensitive (3)
Min-Max normalization
True Resistant (684)
Resistant (Predicted)
(684)
Sensitive (Predicted)
(68)
Resistant (Experimental)
(709)
Sensitive (Experimental)
(43) False Resistant (0)
False Sensitive (25)
True Sensitive (43)
Confusion Matrix
b%
%
%
%
%
%
%
Sensitivity
Specificity
Accuracy
Positive predictive value (PPV)
Negative predictive value (NPV)
Detection rate
Detection prevalence
96.5
100
96.7
100
63.2
91.0
91.0
:
:
:
:
:
:
:
c
Fig. 3 Dexamethasone sensitivity prediction model. a Deregulated gene signatures from dexamethasone- and prednisolone-treated SUP-
B15 cells were combined with genes displaying the highest level of variation in CCLE and TARGET (ALL) datasets and 500 genes were selected.
The 500 genes from 138 cell lines of hematological malignancies were used to build a deep learning model. The model was tested using
three sets of samples. bConfusion matrix for all three groups of samples. cThe performance of the model was calculated using three test
sample groups.
K. Shah et al.
6
npj Precision Oncology (2021) 13 Published in partnership with The Hormel Institute, University of Minnesota
data to build and test the model, it may display poor prediction
performance. To verify the synergy between dexamethasone and
these kinase inhibitors, we used 38 kinase inhibitors from those
groups in combination with dexamethasone to analyze synergy in
the dexamethasone-resistant TANOUE cell line. We observed that
several, but not all, kinase inhibitors displayed synergy with
dexamethasone in TANOUE cells (Supplementary Fig. 7 and Fig.
6e). Inhibitors displaying higher synergy scores include Aurora
kinase, JAK, mTOR, and S6K inhibitors (Fig. 6e). Several of these
kinases are known to be involved in β-catenin stabiliza-
tion
23,25,29,4145
, therefore, we tested whether β-catenin inhibition
synergizes with dexamethasone. We observed that several
inhibitors targeting β-catenin activity display synergy with
dexamethasone in TANOUE cells (Fig. 6f).
DISCUSSION
Glucocorticoids form the cornerstone of drugs in the chemother-
apeutic regimen cocktail for the treatment of ALL and serve as one
of the strongest predictors of relapse in ALL
46
. However, the
molecular mechanisms that contribute to glucocorticoid resis-
tance and relapse in ALL are quite divergent, and also remain
poorly understood
46,47
. In this study, we used transcriptomic and
peptide substrate-based kinase proling, which was supplemen-
ted with the deep learning method to dene deregulated cellular
signaling pathways in dexamethasone-resistant ALL. We observed
that dexamethasone resistance is mediated through the activation
of multiple parallel signaling pathways.
Six hours of dexamethasone treatment induced a gene
signature mimicking the loss of PTEN function (Fig. 1d), which
suggests that dexamethasone treatment might release a negative
feedback loop on PI3K signaling. Additionally, TGF-βpathway
enrichment (Fig. 1d) is linked to the PTEN suppression pheno-
type
48
. Enrichment of those pathways were also present in
prednisolone-treated cells (Fig. 1e) and in patient samples
predicted to be dexamethasone-resistant (Fig. 5b). Furthermore,
peptide substrate-based kinase proling (Fig. 2b) and western blot
using a phospho-specic antibody (Fig. 2g) showed upregulation
of S6K activity probably related to the activation of the PI3K/mTOR
pathway. The PI3K pathway plays a role in dexamethasone
resistance in acute lymphoblastic leukemia and inhibition of AKT
or mTOR reverses dexamethasone resistance
49,50
. However,
dexamethasone treatment inhibited AKT phosphorylation (Fig.
2h) and we have not seen an enrichment of the PI3K/mTOR gene
signature in GSEA raising the possibility that S6K activation may
be mediated by an alternative pathway. For example, Aurora
kinase activation is linked to the activation of S6K
43
.
Another interesting pathway that was enriched is involved in
epithelial-mesenchymal transition (Figs. 1d, e and 5a). In this
process, epithelial markers are downregulated while mesenchymal
markers get upregulated, a typical feature found in embryogen-
esis, organ development, and predominantly epithelial cancer
51,52
.
Several EMT-related transcription factors are also known to be
expressed in hematological cancers and thereby modulate cell
proliferation, viability, stemness, and drug resistance
53
. It has been
shown that dexamethasone favored EMT and cancer progression
in pancreatic cancer
54
. Aurora kinase activity has been linked to
EMT in several epithelial cancers
5557
. Our kinase activity proling
data suggest a strong activation of Aurora kinase activity in
dexamethasone-treated cells (Fig. 2b), linking to the enrichment of
EMT markers. The role of other cell cycle regulatory kinases such
as PLK family proteins in EMT has been poorly studied. Studies
suggest a role of PLK1 in EMT while the role of other family
members in EMT, such as PLK2 and PLK3, remains relatively
ab
ALL patient expression data
from the TARGET database
Dexamethasone Resistant (Predicted)
Dexamethasone Sensitive (Predicted)
Samples
96
109
Dexamethasone Sensitive (Predicted)
Dexamethasone Resistant (Predicted)
n=63
n=67
P = 0.0115
0 2000 4000 6000
0
50
100
Event-free survival (Days)
Bone marrow
Dexamethasone Sensitive (Predicted)
Dexamethasone Resistant (Predicted)
n=96
n=109
P = 0.0098
0200040006000
0
50
100
Event-free survival (Days)
All
Dexamethasone Sensitive (Predicted)
Dexamethasone Resistant (Predicted)
n=33
n=42
P = 0.3868
0200040006000
0
50
100
Event-free survival (Days)
Peripheral blood
cd
Fig. 4 Dexamethasone sensitivity prediction in ALL patient samples. The TARGET dataset for ALL was used to predict dexamethasone
sensitivity. aIn total 205 patient samples were used to predict dexamethasone sensitivity by the deep learning model. bEvent-free survival
between dexamethasone-sensitive and -resistant groups using 205 ALL patient samples was determined by GraphPad Prism. cEvent-free
survival for ALL patient samples collected from bone marrow. dEvent-free survival for ALL patient samples collected from peripheral blood.
K. Shah et al.
7
Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2021) 13
unknown
58
. However, a link has been established between PLK2
expression and Cyclin E stabilization where PLK2 induces FBXW7
degradation
59
. FBXW7 binds to cyclin E and directs it for
proteolytic degradation
60
. Finally, enrichment of WNT and
its canonical effector, the β-catenin pathway (Fig. 5b), also
links to EMT
61
.
Enrichment of cytokine and cytokine receptor signaling
pathways was observed in patient materials predicted to be
dexamethasone-resistant (Fig. 5d). This is probably because
SOCS2 expression was downregulated in dexamethasone-
resistant ALL patient samples (Fig. 5c), which was also evident
in a dexamethasone-resistant cell line (Fig. 5f). SOCS2 belongs to
the suppressor of the cytokine signaling family of proteins and
acts as a negative regulator of cytokine signaling
40
.Insilico
identication of synergy between dexamethasone and kinase
inhibitors provides additional support to the activation of distinct
signaling pathways (Fig. 6d). Furthermore, the in vitro drug
synergy experiment to follow up in silico data is in line
with the fact that inhibition of JAK, mTOR, S6K, and cell cycle
regulatory kinases including Aurora kinases, displays a synergistic
effect with dexamethasone in a dexamethasone-resistant cell
line (Fig. 6e).
While combining all the data, we observed that enrichment of
the β-catenin upregulation signature is common within all four
GSEAs (Figs. 1d, e and 5b, and Supplementary Fig. 6b), suggesting
a link between dexamethasone resistance and transcriptional
activation of β-catenin target genes. β-catenin activity is regulated
by stabilization of its protein levels through inactivation of the
destruction complex (AXIN/APC/CK1/GSK-3β) that phosphorylates
β-catenin, resulting in ubiquitin-mediated degradation
22
. Aurora
kinases play important roles in the inactivation of the destruction
complex by inducing phosphorylation-dependent inactivation of
GSK-3β. This process can take place via several different
mechanisms. GSK-3βis a substrate of Aurora kinase
23
. It can also
directly or indirectly activate JAK2
41
, p38
42
, and S6K
43,44
. While
inhibition of JAK2 activity suppressed β-catenin accumulation
45
,
S6K
29
, and p38
25
directly phosphorylate GSK-3βand negatively
regulate its activity. We also observed that inhibitors targeting
Aurora kinases, JAK, mTOR, and S6K display synergy with
dexamethasone (Fig. 6e), further suggesting that Aurora kinase
activity may be involved in dexamethasone resistance. A link
between Aurora kinase activity and dexamethasone sensitivity
was described where the pharmacological inhibition of Aurora
kinase enhanced dexamethasone-induced expression of cell
death genes
62
.
Collectively, our data suggest that stabilization of β-catenin
through phosphorylation-dependent inactivation of GSK-3βlikely
by Aurora kinases and their direct or indirect downstream
effectors such as JAK, mTOR, p38, and S6K may contribute to
dexamethasone resistance (Supplementary Fig. 8). These kinases
are possibly activated during the initial exposure to dexametha-
sone that may be contributed to dexamethasone-resistant
leukemia patients. Finally, drugs targeting these kinases and their
possible downstream effector β-catenin can partially restore
dexamethasone sensitivity.
55 kDa
25 kDa
SOCS2
-actin
SUP-B15 TANOUE
55 kDa
250 kDa
130 kDa
FLT3
SUP-B15 TANOUE
-actin
e
ab
c
Resistant vs Sensitive
NES
1.37
1.47
1.53
1.76
2.16
2.22
0.225 0.25 0.275 0.3 0.325 0.35 0.375
ES
HALLMARK_APICAL_JUNCTION
HALLMARK_COAGULATION
HALLMARK_MYOGENESIS
HALLMARK_APICAL_SURFACE
HALLMARK_KRAS_SIGNALING_DN
HALLMARK_EPITHELIAL_
MESENCHYMAL_TRANSITION
0.01
0.02
0.03
0.04
0.05
0.06
0.2 0.2
Resistant vs Sensitive
0.225 0.25 0.275 0.3 0.3250.35 0.375 0.4
ES
VEGF_A_UP.V1_UP
TGFB_UP.V1_UP
WNT_UP.V1_UP
PTEN_DN.V1_UP
P53_DN.V2_UP
SRC_UP.V1_UP
BCAT.100_UP.V1_UP
IL21_UP.V1_UP
NES
1.37
1.45
1.53
1.72
1.85
1.89
2.21
2.37 0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
CDC5L
JCHAIN
PBX1
ARL4C
PNO1
EPOR
SLC4A4
ID3
NID2
DDX58
PTPRM
IKZF1
ZAK
CD44
XAF1
CASP1
NR3C1
SOCS2
BCL2
FLT3
Fold change
d
ES: 0.3
NES: 1.97
FDR: 0.0068
f
Fig. 5 Pathway enrichment in dexamethasone-resistant ALL patient samples. Pathway enrichment in dexamethasone-resistant ALL patient
samples was analyzed using GSEA. aHallmarks and bOncogenic signatures gene sets were used for pathway enrichment analysis. c
Upregulated and downregulated genes in dexamethasone-resistant ALL patient samples were determined by SAM. The bar graph shows
selected top-listed genes. All the upregulated and downregulated genes are included in supplementary table T2. dThe enrichment of KEGG
cytokine and cytokine receptor interaction pathway in dexamethasone-resistant ALL patient samples was determined by GSEA. e,fSUP-B15
and TANOUE cells were lysed. Lysates were analyzed by SDS-PAGE and western blotting using specic antibodies as labeled. Blots shown in
each panel were from the same experiment and processed similarly.
K. Shah et al.
8
npj Precision Oncology (2021) 13 Published in partnership with The Hormel Institute, University of Minnesota
METHOD
Gene expression analysis
ALL cell lines SUP-B15, JURKAT, MOLT-4, and CCRF-CEM were treated with
1 µM dexamethasone or 2 µM prednisolone for 6 h before lysis. All cell lines
were collected from DSMZ (Braunschweig, Germany). Cells were main-
tained in the recommended medium by DSMZ and were regularly tested
for mycoplasma contamination. Total RNA was extracted using the RNeasy
mini kit (Qiagen) and the quality of total RNA was checked by Bioanalyzer.
Affymetrix Human Gene 2.0 ST Array was used for expression analysis. Raw
data were processed for robust multi-array average (RMA) normalization.
Signicance Analysis of Microarrays (SAM) was used to identify the
deregulated genes
16
, with the SAMR version developed by Michael Seo
being used
63
. A maximum false discovery rate (FDR) cut-off of 5% was used
to dene signicantly deregulated genes. Gene Set Enrichment Analysis
(GSEA) software
64
(Broad Institute) with the gene sets database MSigDB
v7.0 (Hallmarks and Oncogenic signatures) was used for pathway
enrichment analysis. ALL patient data were downloaded from the TARGET
database
63
. CCLE cell line data were collected from GSE36133
65
.
Peptide substrate-based kinase proling
Kinase activity in dexamethasone-treated SUP-B15 cells was measured
using the peptide substrate-based kinase proling (Pamgene, s-Herto-
genbosch, the Netherlands) method. SUP-B15 cells were treated with 1 µM
dexamethasone for 6 h before lysis. Tyrosine kinase proling and serine/
threonine kinase proling were performed separately using standard
protocols provided by the manufacturer.
a
Model summary:
Train samples: 421
Validation / test samples: 108
Test Accuracy: 82.4 %
Correlation coefficient: 0.64
True Synergy
(44)
Synergy (Predi cted)
(53)
No synergy (Predicted)
55
Synergy (Experimental)
(54)
No synergy (Experimental)
(54)
False Synergy
(9)
FalseNosynergy
(10)
True No synergy
(45)
Confusion matrix
(test samples)
Drugs: Chemical Descriptors
(488 features)
Cell lines: mRNA expression
(500 genes)
tanh norm
Min-Max norm Hidden layers
Binary classification
1=BLISSScore>3
0=BLISSScore<2
bc
d
1454 kinase inhibitors
(488 features)
40 ALL patient materials
(500 genes)
tanh norm
Min-Max norm Prediction 226 Kinase inhibitors
Prediction
Apoptosis
Cell cycle
JAK/STAT
MAPK
NF-kB
PI3K/mTOR
Stem cells / WNT
PKC/TGFbeta/SMAD
Major pathways:
Danusertib Tozasertib
Aurora Kinase in hibitors
Temsirolimus
mTOR inhibitor
PF-4708671
p70S6 kinase inhibitor
ZM 39923Ruxolitinib
Janus kinase (JAK) inhibitors
e
ICG-001 JW 67 CCT 031374 PNU 74654
f
Fig. 6 The synergy between dexamethasone and kinase inhibitors. a Using the same 500-gene signature and drug synergy data from the
DrugComb database, a deep learning binary classication model was developed to predict synergy between dexamethasone and other
inhibitors. bModel performance was tested using 108 test samples. cConfusion matrix showing the models performance using 108 test
samples. dIn silico synergy prediction between dexamethasone and 1454 kinase inhibitors using the deep learning model. eIn vitro synergy
measurement between dexamethasone and kinase inhibitors in TANOUE cells using Cell Titer Glo after 48 h of incubation with drug
combinations. The gure shows representative kinase inhibitors. fIn vitro synergy measurement between dexamethasone and β-catenin
inhibitors in TANOUE cells using Cell Titer Glo after 48 h of incubation with drug combinations.
K. Shah et al.
9
Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2021) 13
Western blot
Cells were lysed in 1% Triton X-100 lysis buffer, supplemented with
protease/phosphatase inhibitors (Na
3
VO
4
, Trasylol, and PMSF). The protein
concentration of the total cell lysates was determined by the bicinchoninic
acid (BCA) assay method (ThermoFisher Scientic, USA). The lysates were
then mixed with a 1:1 volume of SDS sample loading buffer, and 10 µg
proteins from each lysate were separated on SDS-PAGE gels; they were
then transferred to polyvinylidene diuoride (PVDF) membranes. The
membranes were immunoblotted with different primary antibodies. The
anti-phospho-ERK1/2 (1:2000), anti-phospho-S6K (1:1000), anti-phospho-
AKT (1:1000), and anti-phospho-p38 (1:2500) antibodies were obtained
from Cell Signaling Technologies, USA. The anti-ERK (1:200), anti-AKT
(1:1000), and anti-β-actin-HRP (1:2000) antibodies were from Santa Cruz
Biotechnology, USA. The anti-β-catenin (1:1000) and anti-p38 (1 µg/ml)
antibodies were from BD Biosciences, USA. The anti-FLT3 (1 µg/ml)
66
and
anti-SOCS2 (1 µg/ml)
67
antibodies were described previously. The anti-
phospho-serine 9-GSK-3β(1:1000) antibody was from ThermoFisher
Scientic, USA. For immunodetection, the blots were incubated with the
respective horseradish peroxidase-conjugated secondary antibodies and
developed with the Luminata Forte Western HRP Substrate (Millipore) and
Amersham Imager 600 (GE Healthcare, Sweden). Where applicable, all blots
were derived from the same experiment and were processed in parallel.
Deep learning models
To develop the binary classication model, 138 cell lines related to
hematological malignancies were selected from the CCLE database
65
. The
IC
50
value for dexamethasone was used to dene sensitivity. Cells with IC
50
values <700 nM were classied as sensitive, while cells with IC
50
values
>1000 nM were classied as resistant. With these criteria, 108 cell lines
were classied as resistant (dened as 1) and 30 cell lines were classied as
sensitive (dened as 0). Because the number of resistant cell lines was
higher than the number of sensitive cell lines, we used linear combinations
of gene expression for sensitive cell lines to increase the number of
samples in the sensitive cell line group. The sequential model from Keras
36
in TensorFlow
68
backend was used to develop the deep learning model.
The model was trained using a 500-gene signature after Min-Max
normalization. For prediction, a 500-gene signature from unknown
samples was collected and normalized using Min-Max normalization.
Further, to develop a model that can predict synergy between
dexamethasone and other drugs, we retrieved synergy data from the
DrugComb database
69
. Considering the amount of data, we used binary
classication to build the prediction model. We applied BLISS score >3 as
synergy (dened as 1) and BLISS score <2 as no synergy (dened as 0). We
used data from 529 combinations (29 cell lines and 23 drugs) following the
classication (263 samples with BLISS score >3 and 266 samples with BLISS
score <2). PyChem
70
and DeepChem
71
were used to dene features for
drugs (total 488 features) and 500-gene signature for cell lines. Data were
normalized twice: rst using tanh normalization and then by Min-Max
normalization. The sequential model from Keras in the TensorFlow
backend was used to develop the model. Due to the limited amount of
data available, we used 421 samples as training samples and 108 samples
as validation/test samples. Then we applied this model to predict synergy
between dexamethasone and 1454 kinase inhibitors (information collected
from Selleckchem) for 40 ALL patient data from the TARGET dataset.
Similarly, we used 488 features to describe each kinase inhibitor and a 500-
gene signature for each patient sample for the prediction.
In vitro drug synergy
For in vitro synergy analysis, 38 kinase inhibitors and a negative control
(dinaciclib, predicted as no synergy) were used. A vefold dilution (10000-
0.128 nM) of the single drugs was used. We used a 1:1 combination of 8
doses for synergy and then used Decrease
72
to predict the full
combination (8 × 8). Synergy was calculated using the SynergyFinder
web application
73
. We applied a similar method (6 doses) to measure
synergy between dexamethasone and β-catenin. Cells were seeded in a
384-well plate and treated with the drug(s) for 48 h. Cell Titer-Glo
(Promega, USA) was used to measure cell viability following the
manufacturers protocol.
Cell viability and apoptosis
Cells were seeded in a 96-well plate and treated with the drug or DMSO for
48 h. PrestoBlue (ThermoScientic, USA) was used to measure cell viability
following the manufacturers protocol. For apoptosis assay, cells were
seeded in a 24-well plate and treated with the drug or DMSO for different
time points. Apoptotic cells were quantied using the annexin-V/7-AAD kit
(BD Biosciences, USA) following the manufacturers protocol.
Reporting summary
Further information on research design is available in the Nature Research
Reporting Summary linked to this article.
DATA AVAILABILITY
The data generated and analysed during this study are described in the following
data record: https://doi.org/10.6084/m9.gshare.13475916
74
. The gene expression
data are openly available in ArrayExpress at https://identiers.org/arrayexpress:E-
MTAB-9250
75
. The following additional data les underlying Figs. 2,5, and 6, and
Supplementary Figs. 1 and 7 are openly available in gshare at https://doi.org/
10.6084/m9.gshare.13414706
74
: Uncropped_WB_Figure_2d_j.pdf, Uncropped_WB_-
Figure_5e_f.pdf, Synergy_data.xlsx, Upstream Kinase Score_PTK_Figure_2a.png,
Upstream Kinase Score_STK_Figure_2b.png, SUP-B15_EC50.pzf.
CODE AVAILABILITY
Python 3.7 was used to analyze data by using Tensorow (version 2.1.0), Rdkit
(version 2020.03.2), DeepChem (version 2.1.1), Scikit-learn (0.21.3) and Matplotlib
(version 3.1.1). Jupyter Notebook codes are available upon request.
Received: 15 July 2020; Accepted: 12 January 2021;
REFERENCES
1. Terwilliger, T. & Abdul-Hay, M. Acute lymphoblast ic leukemia: a comprehensive
review and 2017 update. Blood Cancer J. 7, e577 (2017).
2. Scheijen, B. Molecular mechanisms contributing to glucocor ticoid resistance in
lymphoid malignancies. Cancer Drug Resist. 2, 647664 (2019).
3. Ploner, C. et al. Glucocorticoid-ind uced apoptosis and glucocorticoid resistance in
acute lymphoblastic leukemia. J. Steroid Biochem. Mol. Biol. 93, 153160 (2005).
4. Norman, M. & Hearing, S. D. Glucocorticoid resistance - what is known? Curr. Opin.
Pharmacol. 2, 723729 (2002).
5. Chougule, R. A., Shah, K., Moharram, S. A., Vallon-Christersson, J. & Kazi, J. U.
Glucocorticoid-resistant B cell acute lymphoblastic leukemia displays receptor
tyrosine kinase activation. NPJ Genom. Med. 4, 7 (2019).
6. Hefazi, M. & Litzow, M. R. Recent advances in the biology and treatment of B-cell
acute lymphoblastic leukemia. Blood Lymphat. Cancer 8,4761 (2018).
7. Schwab, C. & Harrison, C. J. Advances in B-cell precursor acute lymphoblastic
leukemia genomics. Hemasphere 2, e53 (2018).
8. Geeleher, P., Cox, N. J. & Huang, R. S. Clinical drug response can be predicted
using baseline gene expression levels and in vitro drug sensitivity in cell lines.
Genome Biol. 15, R47 (2014).
9. Costello, J. C. et al. A community effort to assess and improve drug sensitivity
prediction algorithms. Nat. Biotechnol. 32, 12021212 (2014).
10. Jang, I. S., Neto, E. C., Guinney, J., Friend, S. H. & Margolin, A. A. Systematic
assessment of analytical methods for drug sensitivity prediction from cancer cell
line data. Pac. Symp. Biocomput.2014,6374 (2013).
11. Adam, G. et al. Machine learning approaches to drug response prediction:
challenges and recent progress. NPJ Precis. Oncol. 4, 19 (2020).
12. Lee, S. I. et al. A machine learning approach to integrate big data for precision
medicine in acute myeloid leukemia. Nat. Commun. 9, 42 (2018).
13. Dincer, A. V., Celik, S., Hiranuma, N. & Lee, S. I. DeepProle: Deep learning of
cancer molecular proles for precision medicine. bioRxiv https://doi.org/10.1101/
278739 (2018).
14. Rampasek, L., Hidru, D., Smirnov, P., Haibe-Kains, B. & Goldenberg, A. Dr.VAE:
improving drug response prediction via modeling of drug perturbation effects.
Bioinformatics 35, 37433751 (2019).
15. Shaffer, S. M. et al. Rare cell variability and drug-induced reprogramming as a
mode of cancer drug resistance. Nature 546, 431435 (2017).
16. Tusher, V. G., Tibshirani, R. & Chu, G. Signicance analysis of microarrays applied
to the ionizing radiation response. Proc. Natl Acad. Sci. USA 98, 51165121
(2001).
17. Gazi, M., Moharram, S. A., Marhall, A. & Kazi, J. U. The dual specicity PI3K/mTOR
inhibitor PKI-587 displays efcacy against T-cell acute lymphoblastic leukemia (T-
ALL). Cancer Lett. 392,916 (2017).
K. Shah et al.
10
npj Precision Oncology (2021) 13 Published in partnership with The Hormel Institute, University of Minnesota
18. Moharram, S. A., Shah, K., Khanum, F., Rönnstrand, L. & Kazi, J. U. The ALK inhibitor
AZD3463 effectively inhibits growth of sorafenib-resistant acute myeloid leuke-
mia. Blood Cancer J. 9, 5 (2019).
19. Mason, J. M., Morrison, D. J., Basson, M. A. & Licht, J. D. Sprouty proteins: multi-
faceted negative-feedback regulators of receptor tyrosine kinase signaling.
Trends Cell Biol. 16,4554 (2006).
20. Fely, H. & Klein, O. D. Sprouty genes regulate proliferation and survival of human
embryonic stem cells. Sci. Rep. 3, 2277 (2013).
21. Katsha, A., Belkhiri, A., Goff, L. & El-Rifai, W. Aurora kinase A in gastrointestinal
cancers: time to target. Mol. Cancer 14, 106 (2015).
22. Zhan, T., Rindtorff, N. & Boutros, M. Wnt signaling in cancer. Oncogene 36,
14611473 (2017).
23. Dar, A. A., Belkhiri, A. & El-Rifai, W. The aurora kinase A regulates GSK-3beta in
gastric cancer cells. Oncogene 28, 866875 (2009).
24. Ding, Q. et al. Erk associates with and primes GSK-3beta for its inactivation
resulting in upregulation of beta-catenin. Mol. Cell 19, 159170 (2005).
25. Thornton, T. M. et al. Phosphorylation by p38 MAPK as an alternative pathway for
GSK3beta inactivation. Science 320, 667670 (2008).
26. Manning, B. D. & Toker, A. AKT/PKB signaling: navigating the network. Cell 169,
381405 (2017).
27. Korneeva, N. L. et al. Mnk mediates integrin alpha6beta4-dependent eIF4E
phosphorylation and translation of VEGF mRNA. Mol. Cancer Res. 8, 15711578
(2010).
28. Consolaro, F. et al. FOXO3a and posttranslational modications mediate gluco-
corticoid sensitivity in B-ALL. Mol. Cancer Res. 13, 15781590 (2015).
29. Zhang, H. H., Lipovsky, A. I., Dibble, C. C., Sahin, M. & Manning, B. D. S6K1
regulates GSK3 under conditions of mTOR-dependent feedback inhibition of Akt.
Mol. Cell 24, 185197 (2006).
30. Eldar-Finkelman, H. Glycogen synthase kinase 3: an emerging therapeutic target.
Trends Mol. Med. 8, 126132 (2002).
31. Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modelling
of anticancer drug sensitivity. Nature 483, 603607 (2012).
32. Garnett, M. J. et al. Systematic identication of genomic markers of drug sensi-
tivity in cancer cells. Nature 483, 570575 (2012).
33. Niz, C. D., Rahman, R., Zhao, X. & Pal, R. Algorithms for drug sensitivity prediction.
Algorithms 9, 77 (2016).
34. Azuaje, F. Comp utational models for predicting drug responses in cancer
research. Brief. Bioinform. 18, 820829 (2017).
35. Smirnov, P. et al. PharmacoDB: an integrative database for mining in vitro
anticancer drug screening studies. Nucleic Acids Res. 46, D994D1002 (2018).
36. KERAS. https://keras.io/ (2020).
37. Vrooman, L. M. et al. Dexamethasone and individualized asparaginase dosing are
each associated with superior event-free survival in childhood acute lymphoblastic
leukemia: results from DFCI-ALL Consortium Protocol 00-01. Blood 114, 321 (2009).
38. Jackson, R. K., Irving, J. A. & Veal, G. J. Personalization of dexamethasone therapy
in childhood acute lymphoblastic leukaemia. Br. J. Haematol. 173,1324 (2016).
39. Inaba, H. & Pui, C. H. Glucocorticoid use in acute lymphoblastic leukaemia. Lancet
Oncol. 11, 10961106 (2010).
40. Krebs, D. L. & Hilton, D. J. SOCS proteins: negative regulators of cytokine sig-
naling. Stem Cells 19, 378387 (2001).
41. Katsha, A., Arras, J., Soutto, M., Belkhiri, A. & El-Rifai, W. AURKA regulates JAK2-
STAT3 activity in human gastric and esophageal cancers. Mol. Oncol. 8,
14191428 (2014).
42. Shang, Y. Y. et al. Alisertib promotes apoptosis and autophagy in melanoma through
p38 MAPK-mediated aurora a signaling. Oncotarget 8, 10707 6107088 (2017).
43. Pal, S. K. et al. RNA-seq reveals aurora kinase-driven mTOR pathway activation in
patients with sarcomatoid metastatic renal cell carcinoma. Mol. Cancer Res. 13,
130137 (2015).
44. Wu, J. et al. Aurora-A induces chemoresistance through activation of the AKT/
mTOR pathway in endometrial cancer. Front. Oncol. 9, 422 (2019).
45. Liu, Y. C. et al. Blockade of JAK2 activity suppressed accumulation of beta-catenin
in leukemic cells. J. Cell. Biochem. 111, 402411 (2010).
46. Wandler, A. M. et al. Loss of glucocorticoid receptor expression mediates in vivo
dexamethasone resistance in T-cell acute lymphoblastic leukemia. Leukemia
https://doi.org/10.1038/s41375-020-0748-6 (2020).
47. Shah, D. S. & Kumar, R. Steroid resistance in leukemia. World J. Exp. Med. 3,2125
(2013).
48. Chow, J. Y. et al. TGF-beta mediates PTEN suppression and cell motility through
calcium-dependent PKC-alpha activation in pancreatic cancer cells. Am. J. Physiol.
Gastrointest. Liver Physiol. 294, G899G905 (2008).
49. Piovan, E. et al. Direct reversal of glucocorticoid resistance by AKT inhibition in
acute lymphoblastic leukemia. Cancer Cell 24, 766776 (2013).
50. Zhang, C. et al. Synergistic activity of rapamycin and dexamethasone in vitro and
in vivo in acute lymphoblastic leukemia via cell-cycle arrest and apoptosis. Leuk.
Res. 36, 342349 (2012).
51. Kalluri, R. & Weinberg, R. A. The basics of epithelial-mesenchymal transition. J.
Clin. Invest. 119, 14201428 (2009).
52. Kazi, J. U. Mechanisms of anticancer therapy resistance: the role of cancer stem
cells. Int. J. Mol. Sci. https://doi.org/10.3390/ijms21239006 (2020).
53. Chen, S. C., Liao, T. T. & Yang, M. H. Emerging roles of epithelial-mesenchymal
transition in hematological malignancies. J. Biomed. Sci. 25, 37 (2018).
54. Liu, L. et al. Dexamethasone mediates pancreatic cancer progression by gluco-
corticoid receptor, TGFbeta and JNK/AP-1. Cell Death Dis. 8, e3064 (2017).
55. DAssoro, A. B. et al. The mitotic kinase Auroraa promotes distant metastases by
inducing epithelial-to-mesenchymal transition in ERalpha(+) breast cancer cells.
Oncogene 33, 599610 (2014).
56. Zhang, J. et al. Aurora B induces epithelial-mesenchymal transition by stabilizing
Snail1 to promote basal-like breast cancer metastasis. Oncogene 39, 25502567
(2020).
57. Tang, A. et al. Aurora kinases: novel therapy targets in cancers. Oncotarget 8,
2393723954 (2017).
58. Fu, Z. & Wen, D. The Emerging Role of Polo-Like Kinase 1 in Epithelial-
Mesenchymal Transition and Tumor Metastasis. Cancers https://doi.org/10.3390/
cancers9100131 (2017).
59. Ou, B. et al. Plk2 promotes tumor growth and inhibits apoptosis by targeting
Fbxw7/Cyclin E in colorectal cancer. Cancer Lett. 380, 457466 (2016).
60. Zhang, W. & Koepp, D. M. Fbw7 isoform interaction contributes to cyclin E pro-
teolysis. Mol. Cancer Res. 4, 935943 (2006).
61. Basu, S., Cheriyamundath, S. & Ben-Zeev, A. Cell-cell adhesion: linking Wnt/beta-
catenin signaling with partial EMT and stemness traits in tumorigenesis.
F1000Res. https://doi.org/10.12688/f1000research.15782.1 (2018).
62. Poulard, C. et al. Relapse-associated AURKB blunts the glucocorticoid sensitivity of B
cell acute lymphoblastic leukemia. Proc. Natl Acad. Sci. USA 116, 30523061 (2019).
63. Seo, M. SAMR. https://github.com/MikeJSeo/SAM (2016).
64. Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based
approach for interpreting genome-wide expression proles. Proc. Natl Acad. Sci.
USA 102, 1554515550 (2005).
65. GSE36133. https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE36133 (2012).
66. Kazi, J. U. et al. Tyrosine 842 in the activation loop is required for full transformation
by the oncogenic mutant FLT3-ITD. Cell Mol. Life Sci. 74,26792688 (2017).
67. Kazi, J. U. & Rönnstrand, L. Suppressor of cytokine signaling 2 (SOCS2) associates with
FLT3 and negatively regulates downstream signaling. Mol. Oncol.7,693703 (2013).
68. TensorFlow. https://www.tensorow.org (2020).
69. Zagidullin, B. et al. DrugComb: an integrative cancer drug combination data
portal. Nucleic Acids Res. 47, W43W51 (2019).
70. Cao, D. S., Xu, Q. S., Hu, Q. N. & Liang, Y. Z. ChemoPy: freely available python
package for computational biology and chemoinformatics. Bioinformatics 29,
10921094 (2013).
71. Ramsundar, B., Leswing, K., Eastman, P. & Pande, V. Deep Learning For The Life
Sciences: Applying Deep Learning To Genomics, Microscopy, Drug Discovery , And
More. ISBN13: 9781492039839 (Sebastopol, United States, 2019).
72. Ianevski, A. et al. Prediction of drug combination effects with a minimal set of
experiments. Nat. Mach. Intell. 1, 568577 (2019).
73. Ianevski, A., Giri, A. K. & Aittokallio, T. SynergyFinder 2.0: visual analytics of multi-
drug combination synergies. Nucleic Acids Res. 48, W488W493 (2020).
74. Shah, K., Ahmed, M. & Kazi, J. U. Metadata record for the manuscript: the Aurora
kinase/β-catenin axis contributes to dexamethasone resistance in leukemia. g-
share https://doi.org/10.6084/m9.gshare.13475916 (2020).
75. ArrayExpress. https://identiers.org/arrayexpress:E-MTAB-9250 (2020).
ACKNOWLEDGEMENTS
We would like to thank Professor Lars Rönnstrand for comments on the manuscript.
This research was supported by the Kungliga Fysiograska Sällskapet i Lund (KS and
MA), the Crafoord Foundation (JUK), Magnus Bergvalls Stiftelse (JUK), the Swedish
Cancer Society (JUK), and the Swedish Childhood Cancer Foundation (JUK).
AUTHOR CONTRIBUTIONS
K.S.,M.A.,andJ.U.K.designedandperformedtheexperiments,andcollectedandanalyzed
the data. J.U.K. supervised the research. K.S., M.A., and J.U.K. wrote the manuscript.
FUNDING
Open Access funding provided by Lund University.
COMPETING INTERESTS
The authors declare no competing interests.
K. Shah et al.
11
Published in partnership with The Hormel Institute, University of Minnesota npj Precision Oncology (2021) 13
ADDITIONAL INFORMATION
Supplementary information The online version contains supplementary material
available at https://doi.org/10.1038/s41698-021-00148-5.
Correspondence and requests for materials should be addressed to J.U.K.
Reprints and permission information is available at http://www.nature.com/
reprints
Publishers note Springer Nature remains neutral with regard to jurisdictional claims
in published maps and institutional afliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the articles Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the articles
Creative Commons license and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright
holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2021
K. Shah et al.
12
npj Precision Oncology (2021) 13 Published in partnership with The Hormel Institute, University of Minnesota
... This study used a semi-supervised VAE to predict monotherapy responses using cell line data, and the model was shown to perform better than several linear or nonlinear algorithms. The use of drug-induced gene expression perturbation seems to be useful in determining pathways that regulate drug response and therapy resistance [100]. Nevertheless, anomaly detection with density estimation can improve the prediction accuracy through false positive detection, but this still needs to be implemented [101]. ...
... This study provided mechanistic insights for designing a combination therapy with an understandable predictive model. Taken together, these studies suggest that drug perturbation data provide important information about the regulation of biological features that can be used to develop efficient ML models [100]. ...
Article
Full-text available
Resistance to therapy remains a major cause of cancer treatment failures, resulting in many cancer-related deaths. Resistance can occur at any time during the treatment, even at the beginning. The current treatment plan is dependent mainly on cancer subtypes and the presence of genetic mutations. Evidently, the presence of a genetic mutation does not always predict the therapeutic response and can vary for different cancer subtypes. Therefore, there is an unmet need for predictive models to match a cancer patient with a specific drug or drug combination. Recent advancements in predictive models using artificial intelligence have shown great promise in preclinical settings. However, despite massive improvements in computational power, building clinically useable models remains challenging due to a lack of clinically meaningful pharmacogenomic data. In this review, we provide an overview of recent advancements in therapeutic response prediction using machine learning, which is the most widely used branch of artificial intelligence. We describe the basics of machine learning algorithms, illustrate their use, and highlight the current challenges in therapy response prediction for clinical practice.
... However, there are still several possible missing links that need to be studied in the regulation of WNT/b-catenin. For instance, aurora family kinases seem to be involved in the stabilization of b-catenin (191). However, the exact mechanism of how b-catenin stabilization is mediated by aurora family kinases remains to be determined. ...
Article
Full-text available
WNT/β-catenin signaling is a highly complex pathway that plays diverse roles in various cellular processes. While WNT ligands usually signal through their dedicated Frizzled receptors, the decision to signal in a β-catenin-dependent or -independent manner rests upon the type of co-receptors used. Canonical WNT signaling is β-catenin-dependent, whereas non-canonical WNT signaling is β-catenin-independent according to the classical definition. This still holds true, albeit with some added complexity, as both the pathways seem to cross-talk with intertwined networks that involve the use of different ligands, receptors, and co-receptors. β-catenin can be directly phosphorylated by various kinases governing its participation in either canonical or non-canonical pathways. Moreover, the co-activators that associate with β-catenin determine the output of the pathway in terms of induction of genes promoting proliferation or differentiation. In this review, we provide an overview of how protein phosphorylation controls WNT/β-catenin signaling, particularly in human cancer.
... The website has grown to approximately 500 monthly users and has served as the source of curated and easily accessible data for numerous published studies. This includes researchers using the PharmacoDB database to understand specific compounds or molecular processes and pathways (7)(8)(9), as well as for more global analyses of drug sensitivity (10,11). PharmacoDB has also empowered machine learning researchers to develop and publish novel approaches to drug-response prediction (12,13). ...
Article
Cancer pharmacogenomics studies provide valuable insights into disease progression and associations between genomic features and drug response. PharmacoDB integrates multiple cancer pharmacogenomics datasets profiling approved and investigational drugs across cell lines from diverse tissue types. The web-application enables users to efficiently navigate across datasets, view and compare drug dose-response data for a specific drug-cell line pair. In the new version of PharmacoDB (version 2.0, https://pharmacodb.ca/), we present (i) new datasets such as NCI-60, the Profiling Relative Inhibition Simultaneously in Mixtures (PRISM) dataset, as well as updated data from the Genomics of Drug Sensitivity in Cancer (GDSC) and the Genentech Cell Line Screening Initiative (gCSI); (ii) implementation of FAIR data pipelines using ORCESTRA and PharmacoDI; (iii) enhancements to drug-response analysis such as tissue distribution of dose-response metrics and biomarker analysis; and (iv) improved connectivity to drug and cell line databases in the community. The web interface has been rewritten using a modern technology stack to ensure scalability and standardization to accommodate growing pharmacogenomics datasets. PharmacoDB 2.0 is a valuable tool for mining pharmacogenomics datasets, comparing and assessing drug-response phenotypes of cancer models.
... When comparing the transcriptomic profiles by Fusobacteriales RA in the TCGA-COAD-READ cohort, we identified dysregulation affecting cell architecture involving apical surface dynamics and Aurora A kinase signalling, which regulate cMYC, DNA repair, cell motility/migration and induce EMT transition via β-catenin and TGF-β, leading to metastasis and resistance to treatment in multiple cancer types. 36 Small molecule inhibitors against Aurora A have shown encouraging results in preclinical studies and clinical trials in CRC 37 and other cancers. 38 Cytoskeleton shape, filopodium protrusions and alterations in cell adhesion and structure are hallmarks of extracellular matrix invasion. ...
Article
Full-text available
Objectives Transcriptomic-based subtyping, consensus molecular subtyping (CMS) and colorectal cancer intrinsic subtyping (CRIS) identify a patient subpopulation with mesenchymal traits (CMS4/CRIS-B) and poorer outcome. Here, we investigated the relationship between prevalence of Fusobacterium nucleatum ( Fn ) and Fusobacteriales , CMS/CRIS subtyping, cell type composition, immune infiltrates and host contexture to refine patient stratification and to identify druggable context-specific vulnerabilities. Design We coupled cell culture experiments with characterisation of Fn / Fusobacteriales prevalence and host biology/microenviroment in tumours from two independent colorectal cancer patient cohorts (Taxonomy: n=140, colon and rectal cases of The Cancer Genome Atlas (TCGA-COAD-READ) cohort: n=605). Results In vitro, Fn infection induced inflammation via nuclear factor kappa-light-chain-enhancer of activated B cells/tumour necrosis factor alpha in HCT116 and HT29 cancer cell lines. In patients, high Fn / Fusobacteriales were found in CMS1, microsatellite unstable () tumours, with infiltration of M1 macrophages, reduced M2 macrophages, and high interleukin (IL)-6/IL-8/IL-1β signalling. Analysis of the Taxonomy cohort suggested that Fn was