REGULAR ARTICLE

Inflammation-driven activation of JAK/STAT signaling reversibly

accelerates acute myeloid leukemia in vitro

Jan Habbel,1,* Lucas Arnold,1,* Yiyang Chen,1,2,* Michael Möllmann,1 Kirsten Bruderek,3 Sven Brandau,3 Ulrich Dührsen,1 and
Maher Hanoun1
1
Department of Hematology, University Hospital Essen, Essen, Germany; 2Division of Hematology and Oncology, Department of Medicine, Chang Gung Memorial Hospital,

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

Chiayi, Taiwan; and 3Research Division, Department of Otorhinolaryngology, West German Cancer Center, University Hospital Essen, Essen, Germany

Acute myeloid leukemia (AML) is characterized by a high relapse rate and dismal long-term
Key Points
overall survival which is related to persistence of leukemia-initiating cells in their niche.
• Inflammatory response Different animal models of myeloid malignancies reveal how neoplastic cells alter the
induces activation of structural and functional characteristics of the hematopoietic stem cell niche to reinforce
JAK/STAT signaling in
malignancy. Understanding and disruption of the microenvironmental interactions with
AML that fosters leuke-
AML cells are a vital need. Malignant niches frequently go along with inflammatory
mia proliferation.
responses, but their impact on cancerogenesis often remains unexplored. Here, we
• Inhibition of JAK/STAT uncovered an aberrant production of inflammatory cytokines in untreated AML bone
signaling leads to sig- marrow that was proved to promote the proliferation of leukemia cells. This inflammatory
nificant antileukemic
response induced an activation of the Janus kinase/signal transducer and activator of
activity in vitro but lacks
transcription (JAK/STAT) signaling pathway in AML blasts as well as bone marrow stromal
substantial effects as
cells that also fostered leukemia proliferation. Inhibition of JAK/STAT signaling using the
monotherapy in vivo.
selective JAK1/2 inhibitor ruxolitinib resulted in significant antileukemic activity in AML
in vitro which is mediated through both cell-autonomous and microenvironment-mediated
mechanisms. However, in a xenograft transplantation model, monotherapy with ruxolitinib
did not achieve substantial antileukemic activity, possibly suggesting a complementary
function of JAK1/2 inhibition in AML.

Introduction

Patients with acute myeloid leukemia (AML) face a high failure rate of achieving complete remission as
well as high relapse rates, which result in poor long-term overall survival. Resistance to therapy is largely
a result of the persistence of leukemia-initiating cells.1 For decades, the mainstay of treatment has been
based on chemotherapeutic agents. In recent years, increasing evidence points to the significance of
the bone marrow microenvironment for the pathogenesis of AML.2,3 Meanwhile, the concept of the
hematopoietic stem cell niche where different cellular and noncellular constituents regulate hemato-
poietic stem cell self-renewal and differentiation is well established.4 In murine models of AML, it has
been shown that the microenvironmental regulation of hematopoietic stem cells is significantly perturbed
to the advantage of leukemia growth.2 Therefore, treatment should not only encompass cytotoxic agents
but should also aim to disrupt the crosstalk between leukemia-initiating cells and their microenvironment.
A common feature of different hematologic malignancies is an inflammatory response in the bone
marrow niche.5-7 There is evidence for the concept of inflammation as a chronic and self-perpetuating
stimulus that forces malignant clones to develop additional subclones by triggering additional mutations
in hematopoietic cells.7,8 Furthermore, in murine AML, increased inflammatory signals were linked with

Submitted 2 December 2019; accepted 26 May 2020; published online 2 July 2020. Requests for data should be addressed to Maher Hanoun at maher.hanoun@uk-
DOI 10.1182/bloodadvances.2019001292. essen.de.
The full-text version of this article contains a data supplement.
*J.H., L.A., and Y.C. contributed equally to this study. © 2020 by The American Society of Hematology

3000 14 JULY 2020 x VOLUME 4, NUMBER 13

 A IL-1E IL-6 IL-8 IL-10 TNFD
20 75 ** 100 **** 85 * 40

70 80 80
15 20 15 30
60
pg/ml

pg/ml

pg/ml

pg/ml
pg/ml
15
10 10 20
10 40
5 5 10
5 20

0 0 0 0 0
control AML control AML control AML control AML control AML

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

B C

2000 *** 2500 *

5ng/ml IL-6 5ng/ml IL-6
+Ruxo 2PM +Ruxo 2PM
2000
1500
MFI (pSTAT3)

MFI (pSTAT5)

50ng/ml IL-6 50ng/ml IL-6

1500
1000
5ng/ml IL-6 5ng/ml IL-6
1000
500 DMSO DMSO
500
FMO FMO
0 0
3 3 4 5 3 3 4
control AML control AML -10 0 10 10 10 -10 0 10 10 105

pSTAT3 pSTAT5

D AML control AML control

* *
12 ** 5.0 * 2.0 2.0
11 *
10 4.0
pSTAT3 (fold change)

pSTAT5 (fold change)

pSTAT5 (fold change)

pSTAT3 (fold change)

9 3.0 1.5 1.5

2.0 2.0
1.5 1.5 1.0 1.0
1.0 1.0
0.5 0.5
0.5 0.5
0 0 0 0
SO

5
5+ 50
xo

SO

5
50

xo

SO

5
5+ 50
xo

SO

5
5+ 50
xo
Ru

Ru

Ru

Ru
M

M
5+
D

ng/ml IL-6 ng/ml IL-6 ng/ml IL-6 ng/ml IL-6

E
FMO mono +HS-5 AML control AML control
2.5 * 2.5 1.5 1.5
* *
pSTAT3 (fold change)

pSTAT5 (fold change)

2.0 2.0
1.0 1.0
1.5 1.5

1.0 1.0
0.5 0.5
0.5 0.5

0 0 0 0
-103 0 103 104 105
SO

D xo
SO

xo

SO

D xo
SO

xo

SO

D xo
SO

xo

SO

D xo
SO

xo
Ru

Ru

Ru

Ru

Ru

Ru

Ru

Ru
M

pSTAT3
D

+HS-5 +HS-5 +HS-5 +HS-5

Figure 1. AML induces an inflammatory response in the human bone marrow niche linked to an activation of JAK/STAT signaling. (A) Cytokine concentrations
in the extracellular bone marrow fluid of AML patients (n 5 19; mean age, 62.1 years 6 standard deviation [SD] 12.1) and controls (n 5 19; mean age, 58.2 years 6 15.7)
quantified by Magnetic Luminex Performance Assay. Corresponding P values are given for each cytokine tested. (B) Median fluorescence intensity (MFI) of pSTAT3 and

14 JULY 2020 x VOLUME 4, NUMBER 13 INFLAMMATION AND JAK/STAT INHIBITION IN AML 3001
severe remodeling of the bone marrow architecture.3 Therefore,
understanding the regulation of inflammatory pathways in the AML
bone marrow microenvironment may uncover novel options for
therapeutic intervention.
Here we report how AML infiltration in humans induces an inflam-
matory response in the bone marrow followed by an activation of the
Janus kinase/signal transducer and activator of transcription (JAK/
STAT) signaling pathway. Inhibition of JAK1/2 disrupts the micro-
environmental crosstalk with AML cells and results in significant
antileukemic activity in vitro.
Patients and methods
spleen, bones, and peripheral blood were harvested for further
processing.
In vitro experiments
The human cell lines THP-1, HL-60, and KG-1 were purchased from
the American Type Culture Collection (ATCC); Kasumi-1 cells were
purchased from the German Collection of Microorganisms and Cell
Cultures (DSMZ) and cultured according vendor´s recommenda-
tions. MS-5 cells and HS-5 cells were cultured in Iscove modified
Dulbecco medium supplemented with 20% and 10% fetal calf
serum (FCS), respectively. All cell lines were kept for a maximum of
5 passages in culture. Primary hematopoietic cells were cultured in

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

Patient samples
Primary patient samples were obtained from peripheral blood or
bone marrow of newly diagnosed AML patients and nonleukemic
donors after informed consent of all patients according to institu-
tional guidelines. The study was carried out in accordance with
the approved protocol of the University of Duisburg-Essen Ethics
Committee. Mononuclear cells (MNCs) were purified using Lympho-
prep (STEMCELL Technologies, Vancouver, BC, Canada). Patient
characteristics are summarized in supplemental Table 1. Primary
human bone marrow mesenchymal stem and progenitor cells were
isolated from newly diagnosed AML patients or patients with non-
Hodgkin lymphoma without bone marrow involvement as controls.
Briefly, bone marrow MNCs were digested with Gibco collagenase
IV (1 mg/mL) and dispase (2 mg/mL) (Thermo Fisher Scientific,
Waltham, MA) followed by magnetic CD45 depletion using the
EasySep Human CD45 Depletion Kit (STEMCELL Technologies)
and sorted for CD45–CD235a–CD31–CD2711CD1461 cells or
expanded in vitro.
In vivo treatments
Procedures were performed in accordance with the German Animal
Welfare Act and approved by the local animal ethics committees.
NOD-scid Il2Rg2/2 (NSG) mice were bred and used in the animal
care facility at the University Hospital Essen. Human MNCs, derived
from untreated patients with the initial diagnosis of AML, were
depleted of CD31 cells using 1 mL OKT3 per 1 3 106 cells
(BioLegend, San Diego, CA) and intravenously transplanted in
NSG mice. Bone marrow and spleen cells were harvested from
leukemic mice and used for secondary transplantation without
previous conditioning. Leukemic mice were randomly assigned (on
the basis of their peripheral blood leukemia burden) to begin
treatment with ruxolitinib (90 mg/kg twice per day) dissolved in
PEG200 or vehicle and administered by oral gavage. Ruxolitinib
was provided by Incyte/Novartis. Once mice started to show signs
of being moribund, treatment for the entire experiment was
discontinued. At the study end point, mice were euthanized, and
StemSpan serum-free medium (STEMCELL Technologies) supple-
mented with recombinant human stem cell factor (50 ng/mL),
thrombopoietin (50 ng/mL), interleukin-3 (IL-3) (5 ng/mL), IL-6
(5 ng/mL) (all from PeproTech, London, United Kingdom), human
high-density lipoprotein (40 mg/mL) (Millipore, Burlington, MA), and
Primocin (2 mL/mL) (InvivoGen, San Diego, CA). Human bone
marrow–derived mesenchymal stem and progenitor cells were
derived from bone marrow MNCs and cultured in MesenCult mes-
enchymal stem cell (MSC) basal medium (STEMCELL Technolo-
gies) containing 2 mM L-glutamine and MesenCult MSC stimulatory
supplement (STEMCELL Technologies). Cells were incubated at
37°C in 5% CO2 with a full medium change after 48 hours followed
by a half medium change once per week until reaching 80%
confluence. Purity was assessed by flow cytometry. For in vitro
experiments, the JAK1/2-selective inhibitor ruxolitinib (INCB18424)
was purchased from Abmole Bioscience (Houston, TX), dipheny-
leneiodonium chloride (DPI) was purchased from Sigma-Aldrich
(St. Louis, MO).
Statistical analyses
Shapiro-Wilk test was applied as a test of normality. In case of
normality, Student t test was applied for comparisons between 2
groups. Mann-Whitney U test and Wilcoxon signed-rank test were
used for nonparametric unpaired and paired analyses. Analyses
were performed with GraphPad Prism software (San Diego, CA).
Results
AML induces an inflammatory response in the
bone marrow niche linked to an activation of
JAK/STAT signaling
To evaluate the magnitude of a putative inflammatory response in
the human AML bone marrow niche, we measured the expression of
inflammatory cytokines in the extracellular bone marrow fluid of AML
patients at first diagnosis using multiplex bead-based Luminex
technology. In fact, AML infiltration paralleled an upregulation of
circulating inflammatory cytokines, in particular IL-8, IL-1b, IL-6, and

Figure 1. (continued) pSTAT5 in unstimulated peripheral blood mononuclear cells (PBMCs) from AML patients (n 5 8) and controls (n 5 7) assessed by BD Phosflow
technology. (C) Representative flow cytometry plots showing fluorescence intensity of pSTAT3 (left) and pSTAT5 (right) of 1 AML patient sample under indicated conditions.
(D) MFI of pSTAT3 (left) and pSTAT5 (right) in PBMCs from AML patients and controls after incubation with either IL-6 at the indicated concentrations or in combination with
2 mM ruxolitinib (Ruxo) for 15 minutes. MFI is given in relation to the unstimulated control condition; linked dots represent 1 patient sample under indicated conditions.
(E) Primary human AML cells and controls were cocultured with HS-5 cells and treated with 2 mM ruxolitinib. Left, representative fluorescence intensity of 1 AML patient
sample after mono- and coculture compared with fluorescence minus one (FMO) control. Right, MFI of pSTAT3 and pSTAT5 (normalized to monocultured control [n 5 4-6];
each primary sample was tested in duplicate). DMSO, dimethyl sulfoxide. Data are shown as mean 6 standard error of the mean (SEM). See also supplemental Figure 1.
*P , .05; **P , .01; ***P , .001; ****P , .0001 as determined by Mann-Whitney U test or Wilcoxon signed-rank test.

3002 HABBEL et al 14 JULY 2020 x VOLUME 4, NUMBER 13

A
THP-1 Kasumi-1 KG-1 HL-60
*** * **

CD45+ cells (relative to control)

2.0 2.0 2.0 2.0
**** **** **** **** ** **
-49.97 -70.72 -48.75 -52.00 -17.13 -42.43 +8.413 -5.665
1.5 1.5 1.5 1.5

1.0 1.0 1.0 1.0

0.5 0.5 0.5 0.5

0 0 0 0
SO

D o
SO

xo

SO

D o
SO

xo

SO

D o
SO

xo

SO

D o
SO

xo
x

x
Ru

Ru

Ru

Ru

Ru

Ru

Ru

Ru
M

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

D

D
+MS-5 +MS-5 +MS-5 +MS-5

B
THP-1 KG-1 THP-1 KG-1
120 120 10 15

Cell number x103 per colony

Cell number x103 per colony
*** ***
100 100 8
Colony number
Colony number

80 80 10
6
60 60
4
40 40 5

20 20 2

0 0 0 0
DMSO Ruxo DMSO Ruxo DMSO Ruxo DMSO Ruxo

C D
DMSO THP-1 primary AML control
120 * 2.5 **
CD45+ cells (relative to control)

*** 3.0 *
**** **** +2.08 -14.35
100 2.0 -48.37 -69.47 2.0
1.5
Colony number

80
1.5
60 1.0
Ruxo
1.0
40
0.5
20 0.5

0 0 0
DMSO Ruxo
SO

xo
SO

xo

SO

xo
SO

xo
Ru

Ru

Ru

Ru
M

M
D

+HS-5 +HS-5

E F
CD33+ + DMSO CD33+ + Ruxo
Primary AML 250K 44.2% 250K 66.9%

53.7% 30.0%
***
CD45+ cells (relative to control)

200K 200K 1.5

*** *
-19.05 150K 150K
-7.74 -53.05
CD45+ cells (relative to control)

2.5
100K 100K
1.0
2.0 50K 50K

1.5 ***
0 0
3
-10 0 103 104 105 -103 0 103 104 105 0.5
-21.54
1.0
SSC

- -
CD33 + DMSO CD33 + Ruxo
0.5 250K
42.7%
250K
46.8%
0
200K 55.7% 200K 51.2%
SO

xo
SO

xo

0
Ru

Ru
M

150K 150K
D

D
SO

D o
SO

xo
x
Ru

Ru
M

100K 100K CD33- CD33+

D

50K 50K
+HuMSPC +HS-5
0 0
-103 0 103 104 105 -103 0 103 104 105

CD45

Figure 2.

14 JULY 2020 x VOLUME 4, NUMBER 13 INFLAMMATION AND JAK/STAT INHIBITION IN AML 3003
IL-10 (Figure 1A). The expression of IL-6 and IL-8 positively
correlated with the degree of leukemic infiltration in the bone
marrow (supplemental Figure 1A). To assess the direct effect of
inflammatory cytokines on AML proliferation, we treated AML cell
lines and primary human AML samples with increasing concen-
trations of IL-6, IL-1b, and IL-8 (supplemental Figure 1B). Depend-
ing on the added cytokines, their concentration and the AML
subtype leukemia growth increased up to 1.5-fold. Cytokine
receptor–derived signals are predominantly mediated through the
JAK/STAT pathway, which plays a major role in maintaining hema-
topoietic balance.9 Chronic inflammation is associated with JAK/
STAT activation and JAK2 overexpression.10 To test whether the
reduction in cell growth in different human AML cell lines
(Figure 2A; supplemental Figure 2A). Applying a series of
concentrations of ruxolitinib resulted in values for the concentra-
tion that inhibits 50% (IC50) between 650 nM and 21.17 mM,
whereas human MLL-AF9–mutated THP-1 and AML1/ETO-
mutated Kasumi cell lines showed the highest sensitivity to
ruxolitinib (Figure 2A; supplemental Figure 2A-B). In contrast,
KG-1 and HL-60 cell lines showed only minor effects at higher
dosages (Figure 2A; supplemental Figure 2A-B). Of note, THP-1
cells showed comparatively high phosphorylation of STAT3
(supplemental Figure 2D). The cytotoxic effects on THP-1 cells
resulted in a dramatic reduction in the number of leukemia

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

increased expression of inflammatory cytokines in the AML bone
marrow niche was correlated with an activation of JAK/STAT sig-
naling, we measured phosphorylation of STAT3 and STAT5 by
phosphoflow cytometry. We found a 1.9-fold higher phosphoryla-
tion of STAT3 and a 2.4-fold higher phosphorylation of STAT5 in
primary AML blasts compared with MNCs from nonleukemic
controls (Figure 1B). Furthermore, the degree of activation of the
JAK/STAT pathway depended on inflammatory signals. With
increasing concentrations of IL-6, phosphorylation of STAT3 was
gradually increased in both AML and control samples, whereas
phosphorylation of STAT5 showed high variability to IL-6 (Figure
1C-D). Of note, JAK2 mutations are uncommon in de novo AML11
and the majority of AML cell lines.12 Given the role of the microen-
vironment in regulating inflammatory processes,13,14 we investi-
gated the influence of stromal cells on JAK/STAT activity in human
AML. Coculturing primary human AML cells with HS-5 stromal cells
for 4 days induced a 1.6-fold increase in phosphorylation of STAT3,
but not STAT5, in AML and importantly was not changed in
nonleukemic MNCs (Figure 1E). These data indicate that AML
induces an inflammatory response in the human bone marrow niche,
which leads to an increased activity of the JAK/STAT pathway.
Inhibition of JAK/STAT pathway has specific
antileukemic activity
We hypothesized that inhibition of the JAK/STAT pathway might
exert antileukemic activity in AML. JAK1/2 inhibition previously showed
efficient reduction of elevated levels of inflammatory cytokines, but with
modest activity on mutational load in myelofibrosis.15,16 In fact,
application of the JAK1/2 inhibitor ruxolitinib showed dramatic
colonies, with formation of strikingly smaller colonies indicating
reduced proliferation capacity (Figure 2B). Notably, plating equal
numbers of viable cells after treatment with ruxolitinib still showed
a deficient colony formation, suggesting loss of leukemogenic
potential upon treatment with ruxolitinib (Figure 2C). To better
differentiate cell-autonomous effects from microenvironment-
mediated effects, we used a coculture model with MS-5 stromal
cells, which proved to significantly support proliferation of most
AML cell lines (Figure 2A; supplemental Figure 2B)17 or HS-5
stromal cells and bone marrow–derived primary human mesen-
chymal stem and progenitor cells for coculture with primary human
AML or nonleukemic MNCs. Culturing primary human AML cells
with HS-5 stromal cells led to variable increases of inflammatory
cytokines in the supernatant compared with monocultured AML
cells (supplemental Figure 1C). In addition, coculture with bone
marrow–derived primary human mesenchymal stem and pro-
genitor cells led to significantly higher secretion of IL-6, but not
IL-1b or IL-8 (supplemental Figure 1D), suggesting that this in vitro
model replicates aspects of an inflammatory niche. Interestingly,
the antileukemic activity of ruxolitinib in coculture with stromal
cells was around 30% higher compared with its effects in
monoculture (Figure 2A). Considering the substantial heteroge-
neity in AML, we used a number of primary human AML samples,
including samples with favorable (;15%) and adverse (;40%)
genetic abnormalities according to the European LeukemiaNet
recommendations (supplemental Table 1), and we screened their
sensitivity to ruxolitinib treatment in a serum-free coculture model.1
Again, coculture for 4 days led to a significant growth advantage for
AML cells. We confirmed a significant cell-autonomous and more
pronounced microenvironment-mediated antileukemic efficacy of

Figure 2. Inhibition of JAK/STAT signaling has strong antileukemic efficacy in AML in vitro. (A) Human AML cell lines were either mono- or cocultured with MS-5
stromal cells and treated with 2 mM ruxolitinib or vehicle for 4 days. Absolute numbers of CD451 cells were normalized to monocultured control (n 5 6); for each pair, the
mean change in cell count is given in percentage points. (B) In all, 5 3 104 THP-1 or KG-1 cells were treated with 2 mM ruxolitinib or vehicle for 4 days, and then 200 cells
were transferred to methylcellulose medium (n 5 4). Number of colonies and cell numbers of single colonies were determined after 10 days. Representative colonies of THP-1
cells are shown below. (C) Primary colony formation of equal numbers of viable THP-1 cells after treatment with ruxolitinib or vehicle (n 5 4). (D) Primary human AML cells or
PBMCs as controls were either mono- or cocultured with HS-5 cells and treated with 2 mM ruxolitinib or vehicle for 4 days. Absolute numbers of CD451 cells were normalized
to monocultured control (n 5 18-19 different AML or control donors; each primary sample was tested in triplicate); for each pair the change in cell count is given in percentage
points. (E) Primary human AML cells were either mono- or cocultured with bone marrow–derived primary human mesenchymal stem and progenitor cells (HuMSPCs) and
treated with 2 mM ruxolitinib or vehicle. Absolute numbers of CD451 cells were normalized to monocultured control (n 5 10 different AML and 2 different HuMSPC donors;
each primary AML sample was tested in triplicate); for each pair, the change in cell count is given in percentage points. (F) CD331 and CD33– PBMCs from an individual AML
patient were separated by FACS and treated with 2 mM ruxolitinib or vehicle in coculture with HS-5 stromal cells for 4 days. Left, representative flow cytometry plots gated on
49,6-diamidino-2-phenylindole (DAPI)– single cells after treatment with 2 mM ruxolitinib or vehicle. Right, absolute cell numbers of isolated CD331 and CD33– cells after
ruxolitinib or vehicle treatment normalized to control (n 5 6 different AML donors; each primary sample was tested in triplicate); for each pair, the change in cell count is given
in percentage points. Data are shown as mean 6 SEM. See also supplemental Figure 2. *P , .05; **P , .01; ***P , .001; ****P , .0001 determined by Student t test or
Wilcoxon signed-rank test. SSC, side scatter.

3004 HABBEL et al 14 JULY 2020 x VOLUME 4, NUMBER 13

 A
DMSO THP-1 Kasumi-1 KG-1
105
S phase 100 100 100
44%
104
75 75 75

% (CD45+ cells)
103
50 50 50
0 G2+M
G0/G1
3.02%
48.5%
-103 25 25 25
BrdU

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

0

0K

0K

0K

0K
50

10

20
15

25

0 0 0
DMSO Ruxo DMSO Ruxo DMSO Ruxo
5 Ruxo 2PM
10
S phase S G2+M G0/G1 apoptotic
25.4%
104
S G1/G0 G2+M
THP-1 **** **** *
103
Kasumi-1 ** n.s. n.s.
0 G2+M
G0/G1 KG-1 n.s. * n.s.
3%
67%
-103
0

0K

0K

0K

0K
50

10

20
15

25

DAPI

B
THP-1 Kasumi-1 KG-1
0.8 15 2.5
***
2.0
0.6
Annexin+ (%)

10
1.5
0.4
1.0
5
0.2
0.5

0 0 0
DMSO Ruxo DMSO Ruxo DMSO Ruxo

Figure 3. Ruxolitinib inhibits proliferation of AML cells. (A) Cell proliferation analysis by FACS using BD bromodeoxyuridine (BrdU) flow assay. Left, representative plots;
right quantification for different AML cell lines treated with 2 mM ruxolitinib or vehicle for 2 days (n 5 3-6). (B) Quantification of apoptotic cells by Annexin V assay in different
AML cell lines after treatment with 2 mM ruxolitinib or vehicle for 3 days (n 5 3-6). Data are shown as mean 6 SEM. See also supplemental Figure 3. *P , .05; **P , .01;
***P , .001; ****P , .0001 as determined by Student t test. n.s., not significant.

treatment with ruxolitinib on primary human AML cells (Figure 2D;
supplemental Figure 2C). Depending on the drug concentration,
AML cell growth was reduced by approximately 70% in the
coculture model. In addition, these findings could be confirmed by
coculture of primary human AML cells with bone marrow–derived
primary human mesenchymal stem and progenitor cells (Figure 2E).
Of note, the inhibition of cell growth was AML specific. Ruxolitinib
had no substantial effects on nonleukemic MNCs in either the
monoculture or the coculture system, nor did we observe any
cytotoxic effects on stromal cells (Figure 2D; supplemental
Figure 2C,E). To prove the specificity to leukemia cells, AML blasts
from a single donor were separated by fluorescence-activated cell
sorting (FACS) from residual nonleukemic cells, whereas both
fractions were exposed to ruxolitinib in coculture with HS-5 stromal
cells. In fact, only leukemia cells showed cell growth that was
significantly decreased by about 50% (Figure 2F). This antileukemic
specificity was independent of genetic characteristics of AML cells
(supplemental Figure 2F). In summary, ruxolitinib showed selective
antileukemic activity in AML by interfering in the microenvironmental
crosstalk and sparing healthy hematopoiesis and stromal cells.
Dissecting the antileukemic effects of ruxolitinib revealed a signifi-
cant inhibition in AML proliferation. Bromodeoxyuridine analysis
showed a 40% to 60% reduction of proliferating cells in ruxolitinib-
responsive AML cell lines after 2 days (Figure 3A), whereas apo-
ptosis was induced only to a small extent (Figure 3B). Given the fact

14 JULY 2020 x VOLUME 4, NUMBER 13 INFLAMMATION AND JAK/STAT INHIBITION IN AML 3005
 A
JAK1 JAK2 STAT1 STAT3 STAT5a STAT5b
0.20 0.015 0.5 0.04 0.004 0.020
* ** * *

Relative to GAPDH

Relative to GAPDH

Relative to GAPDH

Relative to GAPDH

Relative to GAPDH
Relative to GAPDH

0.4
0.15 0.03 0.003 0.015
0.010
0.3
0.10 0.02 0.002 0.010
0.2
0.005
0.05 0.01 0.001 0.005
0.1

0.00 0.000 0.0 0.00 0.000 0.000

control AML control AML control AML control AML control AML control AML

B 4.0
C
DAPI vimentin pSTAT3
5

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

10 Relativte JAK2 mRNA *

3.0
(relative to control)
104
CD271

103 2.0

Isotype Isotype
1.0
0

0 103 104 105 0

control AML

CD146

D 1200 2500
control control

2000
MFI (pSTAT3)

MFI (pSTAT5)

800
1500

1000
400
500

0 0 AML AML
control AML control AML

E F 80
G 60 60
100 mono +HS-5 * 50 50
* 40 40
30 30
MitoSOX (MFI x103)
DCFDA (MFI x103)

80 60
10 10
60 8 8
40
40 6 6

20 4 4
20
2 2
0 0 0 0
-103 0 103 104 105 DMSO Ruxo DMSO Ruxo DMSO Ruxo DMSO Ruxo

DCFDA (MFI) +HS-5 +HS-5

H 30 *
I 2.0
CD45+ cells (normalized to ctrl.)

100 DPI DMSO

DCFDA (MFI x103)

80 1.5
20
60
1.0
40
10
0.5
20

0 0 0
-103 103 104 105
PI

0 DMSO DPI DMSO DPI

SO

SO
PI

D
D
M

M
D

DCFDA (MFI) +HS-5

+HS-5

Figure 4. Ruxolitinib targets inflammation-induced JAK/STAT signaling in leukemia microenvironment and ROS production. (A) JAK1/2, STAT1/3/5a/5b gene
expression analysis in sorted bone marrow-derived primary human mesenchymal stem and progenitor cells after coculture with primary human AML cells (n 5 7) or healthy
controls (n 5 2) for 4 days by real-time polymerase chain reaction (PCR) (each primary sample was tested in triplicate). (B) Left, representative FACS plot of CD2711CD1461
primary human bone marrow mesenchymal stem and progenitor cells, gated on CD45–CD235–CD31– cells. Right, JAK2 gene expression analysis in sorted bone marrow

3006 HABBEL et al 14 JULY 2020 x VOLUME 4, NUMBER 13

that ruxolitinib decreased the leukemogenic potential in certain AML
cell lines (Figure 2C), we tested for the capacity to induce differ-
entiation of myeloid blasts. Strikingly, MLL-AF9–mutated THP-1
cells showed an increased expression of CD11b upon treatment,
but that was not reflected by a more differentiated cytomorphologic
phenotype (supplemental Figure 3A-B). Furthermore, 1 of 6 tested
primary AML samples revealed an increased expression of CD11b,
CD14, CD16, and CD66b upon ruxolitinib treatment, again without
morphologic signs of differentiation (supplemental Figure 3C). Other
ruxolitinib-responsive AML cell lines also failed to show increased
differentiation by FACS or cytomorphology (supplemental Figure 3B).
In summary, the specific antileukemic activity of ruxolitinib is based on
results do not show an obvious difference in pSTAT3 (Figure 4C).
To validate this, we measured phosphorylation of STAT3 and
STAT5 by phosphoflow cytometry in bone marrow–derived primary
human mesenchymal stem and progenitor cells. Coculture with 4 of
6 primary human AML cells led to a 1.25-fold increase in pSTAT3
and an almost twofold increase in pSTAT5 in bone marrow–derived
primary human mesenchymal stem and progenitor cells compared
with coculture with nonleukemic MNCs (Figure 4D). This indicates
that the activation of JAK/STAT signaling in mesenchymal stem and
progenitor cells is primarily observed at the messenger RNA level
and to lesser degree at increased phosphorylation of STAT proteins.
It raises the question of how AML infiltration induces JAK/STAT

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

both microenvironment-mediated and cell-autonomous mechanisms,
whereas the latter is predominantly a result of significant decrease in
cell proliferation.
Ruxolitinib targets inflammation-induced JAK/STAT
signaling in leukemia microenvironment and
production of ROS
Previously, it was suggested that JAK1/2 inhibitors not only affect
the malignant clone in myeloproliferative neoplasia.18 It is still unclear
how ruxolitinib exerts its antileukemic activity in the AML microenvi-
ronment. To evaluate putative targets of ruxolitinib in the microenvi-
ronment, we assessed the expression of JAK/STAT signaling-related
genes in stromal cells by quantitative real-time polymerase chain
reaction. We found that purified bone marrow–derived primary
human mesenchymal stem and progenitor cells cocultured with
primary AML cells had an increased JAK1/2, STAT3, and STAT5a
expression compared with stromal cells cultured with nonleukemic
cells (Figure 4A). Similarly, HS-5 cells cultured with primary AML cells
showed increased JAK2 expression, whereas this effect could be
abolished by treatment with ruxolitinib (supplemental Figure 4A-B).
To assess the clinical relevance of this observation, bone marrow
mesenchymal stem and progenitor cells from untreated AML patients
or nonleukemic controls were isolated by FACS. In fact, CD1461
2711 mesenchymal stem and progenitor cells from AML patients
showed an almost threefold higher expression of JAK2 (Figure 4B).
To further substantiate the role of JAK/STAT signaling in
the microenvironment, we performed immunofluorescence stain-
ing for phosphorylated STAT3 (pSTAT3) on bone marrow–
derived primary human mesenchymal stem and progenitor cells
after coculture with primary AML or nonleukemic MNCs. The
signaling in the niche. To examine the role of leukemia-derived
inflammatory signals, we treated HS-5 stromal cells with increasing
concentrations of IL-6 and IL-1b. IL-6 did not induce any changes in
expression of JAK1, JAK2, or JAK3 (supplemental Figure 4C).
However, IL-1b significantly increased JAK2 expression in stromal
cells (supplemental Figure 4D). Of note, stromal cells cocultured with
primary AML cells did not show increased expression of inflammatory
cytokines (supplemental Figure 4E). This implies that AML-derived
inflammatory cytokines, in particular IL-1b, induce JAK/STAT signal-
ing in mesenchymal cells.
Generally, chronic inflammation and cancer are closely linked to
elevated levels of intracellular reactive oxygen species (ROS).19
Recently, it has been shown that genetic alterations in the mesen-
chymal stem cell niche can induce oxidative stress in hematopoietic
stem cells through inflammatory signals, ultimately contributing to
AML transformation.7 To elucidate whether ruxolitinib interferes with
inflammation-induced ROS production, we first measured ROS
levels in our coculture system. Four of 9 of the applied primary
human AML cells showed a two- to threefold increase in ROS
generation after coculture with stromal cells compared with mono-
culture after 4 days (Figure 4E-F). Treatment with ruxolitinib did not
lead to a significant change in ROS levels in monocultured AML
cells. However, 5 of 6 primary AML samples cocultured with stromal
cells showed an almost 50% reduction in intracellular ROS levels
after treatment (Figure 4F). Next, we evaluated the source of niche-
mediated ROS generation and assessed mitochondrial superoxide
production using the mitochondrial superoxide (MitoSOX) assay.
We did not observe any difference in mitochondrial superoxide in
mono- or cocultured primary AML cells, nor did the levels change
after treatment with ruxolitinib (Figure 4G). Next to the mitochondrial

Figure 4. (continued) mesenchymal stem and progenitor cells from AML patients or nonleukemic controls at first diagnosis by real-time PCR (n 5 3, data normalized to
control). (C) Confocal images from primary human mesenchymal stem and progenitor cells after coculture with nonleukemic mononuclear cells (control) or primary human AML
cells stained for vimentin and pSTAT3. The scale bars, 50 mm. (D) primary human AML cells (n 5 7) and controls (n 5 2) were cocultured with bone marrow–derived primary
human mesenchymal stem and progenitor cells. MFI of pSTAT3 and pSTAT5 of stromal cells assessed by BD Phosflow technology (each sample was tested in duplicate).
(E-F) Intracellular ROS production detected by measuring 29,79-dichlorofluorescin diacetate (DCFDA) in primary human AML cells in mono- or coculture with HS-5 stromal cells.
(E) Exemplary FACS plot showing the MFI of DCFDA in 1 DMSO-treated AML patient sample mono- or cocultured with HS-5 stromal cells. (F) Intracellular ROS production in
mono- or cocultured primary human AML cells after treatment with 2 mM ruxolitinib for 4 days (n 5 7-9 different primary AML samples; each primary sample was tested in triplicate;
connecting lines represent 1 patient sample). (G) Mitochondrial superoxide (MitoSOX) production in mono- or cocultured primary human AML cells after treatment with 2 mM
ruxolitinib or vehicle using MitoSOX Red reagent (n 5 6 different primary AML samples; each primary sample was tested in triplicate, connecting lines represent 1 patient sample).
(H) Intracellular ROS production was detected by measuring DCFDA in mono- or cocultured primary human AML cells treated with 0.5 mM diphenyleneiodonium (DPI) or vehicle for
2 days. Left, representative FACS plot showing the MFI of DCFDA in 1 AML patient sample treated with DPI or vehicle. Right, quantification of intracellular ROS production in
primary human AML cells in mono- or coculture with HS-5 stromal cells (n 5 4-6 different primary AML samples; each primary sample was tested in triplicate; connecting lines
represent 1 patient sample). (I) Primary human AML cells were either mono- or cocultured with HS5 cells and treated with 0.5 mM DPI or vehicle for 2 days. Absolute numbers of
CD451 cells were normalized to monocultured control (n 5 4-6 different AML donors; each primary sample was tested in triplicate). Data are shown as mean 6 SEM. See also
supplemental Figure 4. *P , .05; **P , .01 determined by Student t test, Wilcoxon signed-rank test, or Mann-Whitney U test.

14 JULY 2020 x VOLUME 4, NUMBER 13 INFLAMMATION AND JAK/STAT INHIBITION IN AML 3007
 A B
control Ruxo

2 1,2 **

mCD45
1
(relative to control)

1,5

(relative to control)
Cells per femur

Spleen weight
0,8 99.5% 99.5%

1 0,6 hCD45
0,4
0,5

hCD3
0,2

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

0 0
control Ruxo control Ruxo 100% 100%

hCD33

C D
bone marrow blood spleen control control

100
Human CD45+/CD33+ (%)

80

60
Ruxo Ruxo
40

20

0
control Ruxo control Ruxo control Ruxo

E
2 1,4 4
1,8 * 3,5
1,2
1,6 3
(relative to control)

(relative to control)

(relative to control)

1,4 1 2
Thrombocytes
Hemoglobin

1,2 0,8
WBC

1 1,5
0,8 0,6
1
0,6 0,4
0,4 0,5
0,2
0,2
0 0 0
control Ruxo control Ruxo control Ruxo

Figure 5. Ruxolitinib reduces spleen size but lacks antileukemic activity in vivo. (A) cells per femur and wet spleen weight (data normalized to control). (B) Flow
cytometry gating strategy for bone marrow analysis of human hematopoietic engraftment by gating on human hCD451 cells, detecting exclusively myeloid hCD331 cells, and
excluding hCD31 cells for each experimental condition. (C) Human AML engraftment in bone marrow, peripheral blood, and spleen (n 5 4 different primary AML samples; 2-
4 mice per group per sample. Data normalized to control). (D) Representative images of reticulum staining on femur sections. Arrowheads indicate reticular fibers. Scale bars,
50 mm. (E) Peripheral blood counts of ruxolitinib- or vehicle-treated leukemic NSG mice (data normalized to control). Data are shown as mean 6 SEM. See also supplemental
Figure 5. *P , .05; **P , .01 determined by Wilcoxon signed-rank test.

source, nicotinamide adenine dinucleotide phosphate oxidases
(NOX) are a known major producer of ROS. We therefore applied
DPI, a specific inhibitor of NOX and nitric oxide synthetase in mono-
and coculture. Treating primary AML cells with DPI led to significant
reduction of ROS production (Figure 4H) alongside a moderate
20% to 25% reduction in cell growth in mono- and cocultured AML
cells (Figure 4I). However, given the striking effects of DPI on ROS
production, in particular in monoculture, it is unlikely that ruxolitinib
interferes in ROS production via inhibition of NOX.
This suggests that AML cells secrete inflammatory cytokines that
activate the JAK/STAT pathway in the microenvironment which
again perpetuates JAK/STAT signaling in AML cells. Ruxolitinib
interferes in this crosstalk by downregulating JAK/STAT signaling in

3008 HABBEL et al 14 JULY 2020 x VOLUME 4, NUMBER 13

the microenvironment and inhibiting microenvironmentally induced
ROS production in AML cells which eventually contributes to its
antileukemic efficacy.
Ruxolitinib reduces spleen size but lacks
antileukemic activity in vivo
To test the in vivo antileukemic activity of ruxolitinib in a clinically relevant
model, we transplanted different primary human AML cells into immu-
nodeficient NSG mice. Secondary transplantation of primary human
AML cells allowed robust engraftment without any conditioning and
thus avoiding any interference with the microenvironment. With a mean
leukemia burden of ;30% (mean, 27.2%, 6 standard error of the
mean, 2.4) in the peripheral blood, mice were randomly assigned and
treated twice per day with 90 mg/kg ruxolitinib or vehicle for an average
of 20 days (supplemental Figure 5A). Strikingly, ruxolitinib-treated mice
showed a reduction of 30% in spleen size (Figure 5A; supplemental
Figure 5B). However, monotherapy with ruxolitinib did not show any
significant change in cellularity or reduction in leukemia burden in bone
marrow, spleen, or peripheral blood (Figure 5A-C; supplemental
Figure 5B-C). Of note, transplanted primary human AML samples
proved to be sensitive to ruxolitinib in vitro (median cytoreduction, 30%;
range, 12% to 67% in monoculture and 55%; range, 19% to 65% in
coculture). In myelofibrosis, ruxolitinib has been shown to reduce
fibrotic tissue in bone marrow.20 Because there are hints of an increase
in bone marrow fibers in AML, we also screened murine bones for the
abundance of reticulin fibers. Overall, leukemic bone marrow showed
only scattered reticulin fibers in the central marrow and endosteal
region, which did not differ between the groups (Figure 5D). We were
not able to detect significant levels of human inflammatory cytokines in
the extracellular bone marrow fluid of engrafted NSG mice using the
multiplex bead-based Luminex technology (data not shown). Notably,
ruxolitinib-treated mice showed lower hemoglobin levels, which
is a known adverse effect that occurs early in myelofibrosis
patients (Figure 5E; supplemental Figure 5C). In summary,
ruxolitinib significantly reduced spleen size, which is a well
described effect of the drug in symptomatic myelofibrosis, but it
lacked antileukemic efficacy as a monotherapy in AML in vivo.
Discussion
Increasing insight into the structural and functional organization of the
bone marrow microenvironment in AML supports that it has an impact
on AML initiation and progression. Likely, the high relapse rate in AML
is a result of an aberrant regulation of residual leukemia stem cells in
their niche. Therefore, understanding of and interference with this
crosstalk are of high priority for improving outcome. Inflammatory
response in the microenvironment is a characteristic feature of
myeloproliferative neoplasia and has been proved to predict clinical
outcome.21 Inflammatory responses have also been observed in
murine models of myelodysplastic syndromes and AML and were
found to be relevant for their pathobiology.6,7,22 Furthermore, in
human AML, aberrant IL-1 signaling promoted AML progenitor
cells.23 IL-1 receptor accessory protein (IL1RAP) was overexpressed
on hematopoietic stem cells of patients with AML and high-risk
myelodysplastic syndrome, and it was associated with poor overall
survival.24 However, the clinical and therapeutic significance of this
sterile inflammation in AML has not yet been thoroughly explored.
Here we presented how AML infiltration induces an inflammatory
response in the human bone marrow microenvironment, which is linked
to an aberrant activation of JAK/STAT signaling in AML blasts as well
14 JULY 2020 x VOLUME 4, NUMBER 13
as the microenvironment that again fosters proliferation of leukemia
cells. Aberrant activation of JAK/STAT signaling has been observed in
a number of hematologic malignancies, and JAK inhibition contributed
to antineoplastic effects in myeloproliferative neoplasia and murine
models of acute lymphoblastic leukemia.15,25-28 However, the
mechanisms by which these agents achieve clinical efficacy have not
been fully delineated. More recently, it has been shown that inhibition of
JAK-STAT3 pathway in both mutant and nonmutant cells is required to
reduce inflammatory signaling to achieve clinical benefit in murine
myeloproliferative neoplasia.18 Here we show that inhibition of JAK/
STAT signaling results in antileukemic activity in AML, which is medi-
ated through both cell-autonomous and microenvironment-mediated
Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020
mechanisms. AML infiltration goes along with increased expression of
IL-1b, which induces JAK/STAT signaling in stromal cells. This again
perpetuates the JAK/STAT pathway in AML cells and induces ROS
generation in leukemia cells. JAK1/2 inhibition disrupts this crosstalk
and thus enhances its antileukemic activity. In line with an in vitro drug
screening study, many AML samples were found to be sensitive to JAK
inhibitors after coculturing with stromal cells.29 Karjalainen et al29
showed that adding ruxolitinib to a BCL2 inhibitor overcame stroma-
mediated drug resistance in AML.
Modulating the microenvironment by attenuating the inflammatory
response is a feasible approach to directly inhibit AML progression
and also to increase sensitivity of other antileukemic compounds.
However, monotherapy did not achieve substantial antileukemic
activity in vivo. This could be a result of lower efficiency in diminishing
the autocrine effects of inflammatory cytokines on AML blasts in vivo
as well as the complexity of the microenvironmental regulation which
our in vitro system does not account for. Conversely, a xenograft
model may not accurately reproduce the influence of human AML-
derived inflammatory cytokines on murine stromal cells. More likely,
a combination of ruxolitinib with different antileukemic agents could
improve the efficacy of long-term therapy in AML.
Acknowledgments
The authors thank the staff at the Imaging Center Essen for their
assistance with expert cell sorting and imaging, as well as the Institute
of Pathology of the University Hospital Essen for hematoxylin and eosin
and reticulum stainings. Ruxolitinib was provided by Incyte/Novartis.
This work was supported by a grant from the Dr. Werner Jackstädt
Foundation. J.H. is supported by a doctoral study course of the
University Hospital Essen promoted by the Else Kröner-Fresenius
Foundation.
Authorship
Contribution: J.H., L.A., Y.C., and M.M. designed and performed
experiments, analyzed data, and provided valuable input on the
manuscript; K.B. and S.B. performed cytokine measurements; U.D.
discussed data and provided valuable input on the manuscript; and
M.H. designed and supervised the study, analyzed and discussed
data, and wrote the manuscript.
Conflict-of-interest disclosure: The authors declare no compet-
ing financial interests.
ORCID profiles: J.H., 0000-0003-3798-3947; Y.C., 0000-
0002-7675-0129; M.H., 0000-0002-1754-8940.
Correspondence: Maher Hanoun, Department of Hematology,
University Hospital Essen, Hufelandstrasse 55, 45147 Essen,
Germany; e-mail: maher.hanoun@uk-essen.de.
INFLAMMATION AND JAK/STAT INHIBITION IN AML 3009
References

1. Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel.
Blood. 2017;129(4):424-447.
2. Hanoun M, Zhang D, Mizoguchi T, et al. Acute myelogenous leukemia-induced sympathetic neuropathy promotes malignancy in an altered hematopoietic
stem cell niche. Cell Stem Cell. 2014;15(3):365-375.
3. Duarte D, Hawkins ED, Akinduro O, et al. Inhibition of endosteal vascular niche remodeling rescues hematopoietic stem cell loss in AML. Cell Stem Cell.
2018;22(1):64-77.e6.
4. Pinho S, Frenette PS. Haematopoietic stem cell activity and interactions with the niche. Nat Rev Mol Cell Biol. 2019;20(5):303-320.
5. Medyouf H, Mossner M, Jann JC, et al. Myelodysplastic cells in patients reprogram mesenchymal stromal cells to establish a transplantable stem cell niche
disease unit. Cell Stem Cell. 2014;14(6):824-837.

Downloaded from https://ashpublications.org/bloodadvances/article-pdf/4/13/3000/1747763/advancesadv2019001292.pdf by guest on 18 August 2020

6. Schepers K, Pietras EM, Reynaud D, et al. Myeloproliferative neoplasia remodels the endosteal bone marrow niche into a self-reinforcing leukemic niche.
Cell Stem Cell. 2013;13(3):285-299.
7. Zambetti NA, Ping Z, Chen S, et al. Mesenchymal inflammation drives genotoxic stress in hematopoietic stem cells and predicts disease evolution in
human pre-leukemia. Cell Stem Cell. 2016;19(5):613-627.
8. Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002;420(6917):860-867.
9. Springuel L, Renauld JC, Knoops L. JAK kinase targeting in hematologic malignancies: a sinuous pathway from identification of genetic alterations
towards clinical indications. Haematologica. 2015;100(10):1240-1253.
10. Hermouet S, Vilaine M. The JAK2 46/1 haplotype: a marker of inappropriate myelomonocytic response to cytokine stimulation, leading to increased risk of
inflammation, myeloid neoplasm, and impaired defense against infection? Haematologica. 2011;96(11):1575-1579.
11. Steensma DP, McClure RF, Karp JE, et al. JAK2 V617F is a rare finding in de novo acute myeloid leukemia, but STAT3 activation is common and remains
unexplained. Leukemia. 2006;20(6):971-978.
12. Quentmeier H, MacLeod RA, Zaborski M, Drexler HG. JAK2 V617F tyrosine kinase mutation in cell lines derived from myeloproliferative disorders.
Leukemia. 2006;20(3):471-476.
13. Hanoun M, Maryanovich M, Arnal-Estapé A, Frenette PS. Neural regulation of hematopoiesis, inflammation, and cancer. Neuron. 2015;86(2):360-373.
14. Takizawa H, Manz MG. Impact of inflammation on early hematopoiesis and the microenvironment. Int J Hematol. 2017;106(1):27-33.
15. Verstovsek S, Kantarjian H, Mesa RA, et al. Safety and efficacy of INCB018424, a JAK1 and JAK2 inhibitor, in myelofibrosis. N Engl J Med. 2010;
363(12):1117-1127.
16. Deininger M, Radich J, Burn TC, Huber R, Paranagama D, Verstovsek S. The effect of long-term ruxolitinib treatment on JAK2p.V617F allele burden in
patients with myelofibrosis. Blood. 2015;126(13):1551-1554.
17. Auffray I, Dubart A, Izac B, Vainchenker W, Coulombel L. A murine stromal cell line promotes the proliferation of the human factor-dependent leukemic
cell line UT-7. Exp Hematol. 1994;22(5):417-424.
18. Kleppe M, Kwak M, Koppikar P, et al. JAK-STAT pathway activation in malignant and nonmalignant cells contributes to MPN pathogenesis and
therapeutic response. Cancer Discov. 2015;5(3):316-331.
19. Panieri E, Santoro MM. ROS homeostasis and metabolism: a dangerous liason in cancer cells. Cell Death Dis. 2016;7(6):e2253.
20. Harrison CN, Vannucchi AM, Kiladjian JJ, et al. Long-term findings from COMFORT-II, a phase 3 study of ruxolitinib vs best available therapy for
myelofibrosis. Leukemia. 2017;31(3):775.
21. Tefferi A, Vaidya R, Caramazza D, Finke C, Lasho T, Pardanani A. Circulating interleukin (IL)-8, IL-2R, IL-12, and IL-15 levels are independently prognostic
in primary myelofibrosis: a comprehensive cytokine profiling study. J Clin Oncol. 2011;29(10):1356-1363.
22. Ye H, Adane B, Khan N, et al. Leukemic stem cells evade chemotherapy by metabolic adaptation to an adipose tissue niche. Cell Stem Cell. 2016;19(1):23-37.
23. Carey A, Edwards DK 5th, Eide CA, et al. Identification of interleukin-1 by functional screening as a key mediator of cellular expansion and disease
progression in acute myeloid leukemia. Cell Rep. 2017;18(13):3204-3218.
24. Barreyro L, Will B, Bartholdy B, et al. Overexpression of IL-1 receptor accessory protein in stem and progenitor cells and outcome correlation in AML and
MDS. Blood. 2012;120(6):1290-1298.
25. Quintás-Cardama A, Kantarjian H, Cortes J, Verstovsek S. Janus kinase inhibitors for the treatment of myeloproliferative neoplasias and beyond. Nat Rev
Drug Discov. 2011;10(2):127-140.
26. Roberts KG, Morin RD, Zhang J, et al. Genetic alterations activating kinase and cytokine receptor signaling in high-risk acute lymphoblastic leukemia.
Cancer Cell. 2012;22(2):153-166.
27. Appelmann I, Rillahan CD, de Stanchina E, et al. Janus kinase inhibition by ruxolitinib extends dasatinib- and dexamethasone-induced remissions in
a mouse model of Ph1 ALL. Blood. 2015;125(9):1444-1451.
28. Maude SL, Dolai S, Delgado-Martin C, et al. Efficacy of JAK/STAT pathway inhibition in murine xenograft models of early T-cell precursor (ETP) acute
lymphoblastic leukemia. Blood. 2015;125(11):1759-1767.
29. Karjalainen R, Pemovska T, Popa M, et al. JAK1/2 and BCL2 inhibitors synergize to counteract bone marrow stromal cell-induced protection of AML.
Blood. 2017;130(6):789-802.

3010 HABBEL et al 14 JULY 2020 x VOLUME 4, NUMBER 13