GDF15 secreted by senescent endothelial cells improves vascular progenitor cell functions
GDF15 secreted by senescent endothelial cells improves vascular progenitor cell functions
Guillaume HaID 0 2
Fanny De Torres 0 2
Nassim Arouche 0 2
Nassima Benzoubir 1 2
S e?gol?ne Ferratge 2
Elie Hatem 0 2
Adrienne Anginot 0 2
Georges Uzan 0 2
? These authors contributed equally to this work. 2
0 INSERM U1197, Ho?pital Paul Brousse , Villejuif , France , 2 Universite ? Paris-Diderot , Paris , France
1 UMRS-1193, Ho?pital Paul Brousse , Villejuif, France, 4 Cellenion, Ho?pital Paul Brousse, Villejuif , France
2 Editor: Rudolf Kirchmair, Medical University Innsbruck , AUSTRIA
Endothelial dysfunction (ED) is part of the first steps in the development of cardiovascular diseases (CVD). Growth Differentiation Factor 15 (GDF15) is a cytokine belonging to the Transforming Growth Factor ? superfamily and its expression is increased both during ED and in CVD. Because high blood levels of GDF15 have been reported during ED, we hypothesized that GDF15 could be produced by endothelial cells in response to a vascular stress, possibly to attenuate endothelial function loss. Since senescence is mainly involved in both vascular stress and endothelial function loss, we used Endothelial Colony Forming Cells generated from adult blood (AB-ECFCs) as a model of endothelial cells to investigate GDF15 expression during cellular senescence. Then, we analyzed the potential role of GDF15 in AB-ECFC functions and senescence. When AB-ECFCs become senescent, they secrete increased levels of GDF15. We investigated GDF15 paracrine effects on nonsenescent AB-ECFCs and showed that GDF15 enhanced proliferation, migration, NO production and activated several signaling pathways including AKT, ERK1/2 and SMAD2 without triggering any oxidative stress. Taken together, our results suggest that GDF15 production by senescent AB-ECFCs could act in a paracrine manner on non-senescent ABECFCs, and that this interaction could be beneficial to its model cells. Therefore, GDF15 could play a beneficial role in a dysfunctional vascular system as previously reported in patients with CVD, by limiting ED related to vascular stress occurring in these diseases.
Data Availability Statement: All relevant data are
within the manuscript and its Supporting
Funding: The authors received no specific funding
for this work.
Competing interests: The authors have declared
that no competing interests exist.
According to the World Health Organization, cardiovascular diseases (CVD) such as ischemic
heart, cerebrovascular and peripheral arterial diseases have been the worldwide leading cause
of death over the past 15 years. In vessels, the endothelium is composed of a monolayer of
endothelial cells (ECs) involved in several physiological processes, including the control of
vasomotor tone and the regulation of coagulation and inflammation [
]. However, with
aging, the endothelium becomes progressively dysfunctional and its ability to maintain its full
functions declines while emergence of CVD is observed .
Decrease nitric oxide (NO) bioavailability due to an increase in reactive oxygen species
] and a decrease in Nitric Oxide Synthase 3 (NOS3) activity [
] leads to impaired
endothelial functions, including reduced vasomotor responses and increased levels of
inflammatory and adhesion molecules [
] involved in CVD development.
Growth Differentiation Factor 15 (GDF15) is a stress response cytokine belonging to the
Transforming Growth Factor ? superfamily [
]. Physiologically, GDF15 is expressed at basal
level in healthy human tissues except in the placenta, and is upregulated in several pathologies
such as cancers [
], mitochondrial disorders [
] and CVD [
]. Among CVD, high
GDF15 levels have been found in peripheral artery diseases [
]. In addition, GDF15 has been
described as a protective cytokine, preventing cold ischemia-reperfusion injury in heart
transplantation  and limiting atherosclerosis development [
]. Furthermore, GDF15 is
induced in several EC types in culture following a cellular stress such as a high glucose stimulus
], oxygen concentration changes [
] or after chemotherapy treatment . Taken
together, these data suggest that GDF15 could play a role in endothelial damage and
dysfunction. However, the functional relationship between GDF15 and EC physiology remains elusive
because its biological effects vary according to EC type, environment and concentration
Endothelial progenitor cells (EPCs) were discovered in 1997 [
] and are involved in
endothelium homeostasis, through their physiological ability to regenerate the vascular system
]. Endothelial Colony Forming Cells (ECFCs) are derived from EPCs in culture. They
can be either isolated from peripheral adult blood (AB-ECFCs) or from umbilical cord blood
(CB-ECFCs) . CB-ECFCs exhibit high clonal, proliferative and functional capacities
whereas AB-ECFCs rapidly undergo premature replicative senescence and display poor
angiogenic potential [
]. Therefore, AB-ECFCs provide a convenient model to study EC
In this study, we hypothesized that GDF15 could be produced by ECs during a vascular
stress to attenuate EC loss of function. Because cellular senescence is involved in vascular
damage and stress, we used AB-ECFCs as a model of ECs to investigate GDF15 expression during
senescence. Then, we investigated GDF15 biological effects in AB-ECFCs.
We first assessed the expression of GDF15 in ECs with different functional capacities.
GDF15 was expressed both in AB-ECFCs and CB-ECFCs in vitro and a higher level of GDF15
was found in AB-ECFCs as compared to CB-ECFCs. This higher level of GDF15 was related to
AB-ECFC senescence, while these cells became dysfunctional (decrease in their proliferation
rate, angiogenic properties and NO bioavailability). Since circulating GDF15 was elevated
during EC senescence, we then analyzed whether the addition of recombinant GDF15 to
AB-ECFCs impacted their function. We showed that GDF15 improves AB-ECFC
proliferation, migration, NO production and activated the AKT, ERK1/2 and SMAD2 signaling
pathways without triggering oxidative stress. Taken together, our data suggest that in an impaired
vascular system, dysfunctional ECs contribute to GDF15 production that could be beneficial
to still functional endothelial cells as a tentative to delay or reduce endothelium damage.
Materials and methods
Recombinant human GDF15 protein was purchased from Peprotech. GDF15 optimal
concentration was determined by proliferation assay (S1 Fig) according to data from the literature
(tested concentration range: 10?100 ng/mL). In this study, a GDF15 concentration of 50 ng/
mL was found optimal. TGF-?, IL-6 and IL-8 ELISA kits were purchased from Abcam.
Because it was described that recombinant GDF15 protein may contain traces of TGF-?
], we performed a TGF-? ELISA assay in EGM2 containing GDF15
2 / 17
recombinant protein at 50 ng/ml, the medium we used for our experiments and we did not
revealed any detectable TGF-? contamination (S2 Fig).
Human sample collection and handling met the tenets of the declaration of Helsinki.
Umbilical cord blood samples from healthy full-term newborns were obtained through a partnership
with the Cord Blood Bank of St Louis Hospital (Paris, France), which is authorized by the
French Regulatory Authority (authorization N? PPC51). Peripheral adult blood was obtained
from human healthy males aged 18?70 years, through a partnership with the French
Establishment of Blood (EFS, Ile de France, authorization 14/5/011). This activity was declared to and
authorized by the French Ministry of Research under number AC-2008-376, and the French
Organization for standardization under number 201/51848.1.
ECFCs were isolated as previously described [
]. Briefly, mononuclear cells (MNC) from
adult or cord blood were isolated using a Ficoll gradient. MNC were plated into collagen I
(50 ?g/mL) precoated 6- or 12-well plates in EGM2 medium (Lonza). The next day, cells were
washed with 1X PBS and non-adherent cells were discarded daily for 7 days when the medium
was changed. Thereafter, the medium was changed every other day. All cells were
characterized by FACS using CD31, CD144 and KDR markers (S3 Fig).
Rabbit anti-human antibodies against GDF15, peroxyredoxin-3 and peroxyredoxin-SO3 were
purchased from Abcam. Anti-GFRAL antibody was purchased from Thermo Fisher Scientific.
Anti-P44/42 anti-MAPK, anti-phospho-p44/42 MAPK, anti-Akt, anti-phospho-Akt, Smad2,
and anti-phospho-Smad2 antibodies were purchased from Cell Signaling Technology.
For classical western-blots, proteins were extracted using RIPA lysis buffer, then treated with
dithiothreitol (100 mM final) and Laemmli (1X final). For redox western-blots, cells were
treated with 20% trichloroacetic acid for 1h. Pellets were recovered and washed 3 times with
cold acetone then dried. Dried pellets were resuspended in 70 ?L of TES buffer (100 mM
TrisHCl, 1% SDS, 10 mM EDTA, pH 8.8) containing 100 mM N-Ethylmaleimide. Protein
concentration was determined using Pierce BCA Protein Assay Kit. For redox western-blots, the
lysates were subjected to SDS-PAGE under non-reducing conditions. Transfer was done onto
PVDF membranes. After blocking, membranes were incubated with primary antibodies, then
HRP-conjugated secondary antibody (Jackson Immunoresearch) and revealed with HRP
substrate (Immobilon Western Chemiluminescent HRP Substrate) using enhanced
chemiluminescence detection system (Fusion solo, Vilber lourmat).
Measuring GDF15 secretion by AB-ECFC under a laminar flow
100.000 cells were seeded into a micro slide (Ibidi) in EGM2 for 2 hours then a 10?l/min flow
was applied on these cells with a pressure pump (Fluigent, MFCS-EZ) for 48h. Then,
supernatant was recovered and a GDF15 ELISA (Abcam) assay was performed following the
3 / 17
Senescence-associated ?-galactosidase activity assay
Cells were seeded in triplicate into 12-well plates at a density of 5,000 cells/cm2. At 80% of
confluence, a senescence-associated ?-galactosidase activity assay was performed according to the
manufacturer?s instructions (Sigma). After an overnight incubation with X-gal, cells were
washed and five pictures (Nikon D5300) were taken per well at x10 magnification. Cells with
positive and negative ?-gal staining were counted using Image J cell counter software.
ROS assessment using fluorescent probes
Cells were washed with PBS at 37?C and 1 mL of DMEM without phenol red (Thermo Fisher
Scientific) containing 10 ?M carboxy-H2DCFDA (Molecular Probes) or 5 ?M MitoSOX
(Molecular Probes) was added to each well. After a 40-min incubation at 37?C with 5% CO2,
cells were collected, washed and resuspended in 500 ?L of DMEM without phenol red. Green
(530 nm) and red (585 nm) fluorescence was detected by flow cytometry using Accuri C6
cytometer (BD Biosciences) and mean fluorescence intensity (MFI) were quantified.
RNA extraction, reverse transcription and quantitative RT-PCR
RNA extraction, reverse transcription and quantitative RT-PCR were performed as described
]. Total RNA was extracted with RNeasy mini kit (Qiagen) and reverse
transcription was performed with High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher
Scientific) following the manufacturer?s instruction. Quantitative Taqman RT-PCR was
performed in triplicate with 7000 Real-time PCR system (Applied Biosystems, Themo Fisher
Scientific). GDF-15: Hs00171132_m1, NQO1: Hs01045993_g1, GCLM: Hs00978072_m1 and
HMOX1: Hs01110250_m1 probes were purchased from Applied Biosystems. The 2??Ct
method was used to quantify the relative transcriptional level of each gene. Endogenous
GAPDH was used as housekeeping gene.
Cells were seeded in triplicate into a 96-well plate (5,000 cells/well) and incubated with EBM2 +
0.5% FBS for 7h. Then, cells were treated with 50 ng/mL of GDF15 overnight. Cell viability was
assessed using the Cell titer gloR Luminescent cell viability assay (Promega) following the
manufacturer?s instructions. Luminescence was measured using a Microplate Luminometer Centro
LB 960 (Berthold Technologies).
Cell cycle analysis
Cells were starved in EBM2 containing 0.5% FBS for 8h and then treated overnight with
GDF15 (50 ng/mL). Cells were then detached and fixed with 70% cold ethanol for 2h. Then,
cells were recovered, washed and resuspended in 1X PBS and incubated with 0.5 mg/mL of
RNAse A for 1h at 37?C. Finally, cells were treated with 10 ?g/mL of propidium iodide and
analyzed by flow cytometry.
BrDU incorporation assay
BrdU incorporation experiment was performed using the BrdU Cell Proliferation Assay Kit
(Cell Signaling Technology) following the manufacturer?s instructions. Briefly, cells were
seeded in triplicate into 96-well plates. The day after, cells were treated with EBM2 + 0.5% FBS
for 8h. Then, cells were treated with GDF15 (50 ng/mL) or VEGF (50 ng/mL) in EBM2 + 0.5%
FBS for 16h. BrdU was added for 4h in EGM2 without VEGF and FGF. Proliferation was
4 / 17
quantified by measuring BrDU absorbance using a plate reader (Multiskan EX, Thermo Fisher
Cell migration assay
Cells were starved for 16h in EBM2 + 0.2% FBS and seeded into the upper compartment of
fibronectin (bovine plasma) precoated (20 ?g/mL, Sigma-Aldrich) 24-well, 8-?m culture
inserts with transparent polyethylene terephthalate track-etched membrane (BD falcon).
Then, GDF15 (50 ng/mL) and VEGF (50 ng/mL) were added into the lower compartment.
After 5 hours, inserts were recovered, fixed with 4% paraformaldehyde and stained with
MayGru?nwald Giemsa coloration variant (RAL 555kit, RAL diagnostics, France). After cleaning
the insert upper compartment with a cotton bud, we extracted and mounted the membrane
from the insert. Membranes were photographed using a camera associated with an inverted
microscope with a 10X objective in phase contrast mode. The number of cells was counted
manually using Image J cell counter software.
Assessment of nitric oxide production
DAF-FM diacetate (D23844, Invitrogen) was used to quantify NO production. At 80%
confluence, cells were treated for 16h with GDF15 (50 ng/mL) or LPS (100 ng/mL) then washed and
incubated with 1 ?M DAF-FM diacetate probe in EBM2 + 0.5% FBS for 1h at 37?C. After
removing the supernatant, cells were incubated in fresh EGM2 medium for 5 min at 37?C,
resuspended in 1X PBS + 0.5% BSA and fluorescence intensity at 530 nm was quantified using
Accuri C6 cytometer (BD Biosciences).
All data were analyzed using Graph Pad 7.0 software. A Mann-Whitney T-test or Wilcoxon
matched pairs signed rank t-test was used to compare two groups. Then, a Kruskal Wallis test
was used. A non-parametric One-Way Anova test was used to compare three groups or more.
Then, a Kruskal Wallis test and a Dunn?s multiple comparisons test was performed. A p value
<0.05 was considered significant.
GDF15 expression level is higher in AB-ECFCs than in CB-ECFCs
To evaluate GDF15 expression in and secretion AB-ECFCs and CB-ECFCs, we analyzed
GDF15 gene and protein expression levels by RT-qPCR, Western-Blot and ELISA in these
cells at same passages (between 4 and 5). We found that the levels of GDF15 mRNA (Fig 1A),
and intracellular (Fig 1B) and extracellular (Fig 1C) proteins were significantly higher in
AB-ECFCs than in CB-ECFCs (6.6, 2.5 and 8-fold higher, respectively). Using a Senescence
Associated ?-galactosidase assay (SA-?-gal), we showed a higher percentage of senescent cells
in AB-ECFCs (75.96 ? 4.84%) than in CB-ECFCs (1.97 ? 0.79%, p = 0.0006, n = 7) (Fig 1D).
These data suggested that ECs produce higher levels of GDF15 while they become senescent
GDF15 expression and secretion are associated with senescence
To confirm the relationship between GDF15 expression and AB-ECFC senescence, we
measured GDF15 level in non-senescent (NS) and senescent (S) AB-ECFCs (Fig 2A). We found
that GDF15 expression was strongly increased in S AB-ECFCs compared to NS AB-ECFCs by
RTq-PCR (~ 4 folds) (Fig 2B), Western-Blot (~ 3.2 folds) (Fig 2C) and ELISA (~ 7 folds) (Fig
5 / 17
Fig 1. GDF15 level is increased in AB-ECFCs and appears to be related to senescence. Analysis of GDF15 expression and production by AB-ECFCs and CB-ECFCs
between passages 4?5 by (a) RTqPCR (n = 7), (b) western-blot (n = 6) and (c) ELISA (n = 6). (d) Pictures and quantification of senescence by SA-?-gal assay in
AB-ECFCs and CB-ECFCs (n = 7). Data are presented as the mean ? SEM. P <0.01 and P <0.001 compared to CB-ECFCs.
2D). In addition, we found that GDF15 expression was gradually increased along the
successive passages (WB and ELISA) (S4 Fig) and we found that GDF15 was only over-secreted by
senescent AB-ECFCs under a laminar flow condition (S5 Fig). Moreover, GDF15
overexpression was also found in senescent CB-ECFCs, which senesce very lately beyond passage 20 (S6
Fig). These results confirm that GDF15 expression is related to AB-ECFCs senescence. Beside
GDF15 overexpression in senescent AB-ECFCs, we also found that IL-6 and IL-8
pro-inflammatory cytokines, hallmark of senescent cells [
], were also over-secreted by these cells (S7
Fig). To identify the role of GDF15 in senescence induction, we added a recombinant human
GDF15 protein at 50 ng/mL to AB-ECFC cultures at each passage from passages 3 to 6. After
these three passages, we observed that the number of ?-galactosidase positive cells was not
6 / 17
Fig 2. GDF15 expression and secretion is associated with senescence. (a) Pictures and quantification of NS AB-ECFCs and S AB-ECFCs
stained using the SA-?-gal assay (n = 13). Analysis of GDF15 expression and secretion by NS AB-ECFCs and S AB-ECFCs by (b) RTqPCR
(n = 13), (c) western-blot (n = 11) and (d) ELISA (n = 8). (e) GDF15 effects on cellular senescence from passages 3 to 6 (n = 3). Data are
presented as the mean ? SEM. ns = non-significant, P <0.01 and P <0.001 compared to NS AB-ECFCs.
increased in GDF15-treated cultures (Fig 2E). Thus, recombinant GDF15 had no significant
effect on senescence induction in AB-ECFCs.
These data confirmed that GDF15 upregulation and expression is significantly associated
with AB-ECFC senescence and that this expression does not increase the senescence
GDF15 improves AB-ECFC functions
Since S AB-ECFCs secreted high levels of GDF15, we then investigated the paracrine effect of
GDF15 on NS AB-ECFCs. We added recombinant GDF15 to NS AB-ECFC cultures and
analyzed its effects on cell proliferation, migration and NO production. We first investigated
whether GDF15 treatment modified cell cycle progression in AB-ECFC cultures.
GDF15treated and non-treated (NT) cells were labeled with propidium iodide and their cell cycle
profile was analyzed by flow cytometry. Here, VEGF at 50 ng/mL was used as a positive control.
In the presence of GDF15, the number of cells in G0-G1 (61.6 ? 3.4% in NT cells and 67.1 ?
3.6% in GDF15-treated cells) and in G2-M (18.8 ? 1.6% in NT cells and 21.9 ? 2% in
GDF15treated cells) phases was increased (Fig 3A and 3B). Then, we showed a 20% increase in BrDU
incorporation in GDF15-treated cells compared to NT cells (p <0.0001, n = 6) (Fig 3C). We
then investigated the effect of GDF15 on AB-ECFC migration capabilities using a transwell
migration assay. GDF15 significantly improved AB-ECFC migration compared to NT cells
(p = 0.0012, n = 4), reaching a migration level similar to that obtained with VEGF (Fig 3D and
3E). Then, we used a specific probe (DAF-FM-2A) to quantify intracellular NO in AB-ECFCs.
Here, LPS was used as positive controls. GDF15- and LPS-treated cells contained a higher level
of NO (~21% and ~37%, respectively) compared to NT cells (p = 0.0021, n = 8) (Fig 3F, data
were normalized to MFI of NT cells). Simultaneously, we analyzed NOS3 expression by
RT-PCR in AB-ECFCs treated or not with GDF15. NOS3 is the endothelium-specific nitric
oxide synthase [
]. NOS3 mRNA was upregulated in GDF15-treated cells (~26%) compared
to NT cells (Fig 3G). Taken together, these results demonstrated that GDF15 improves
AB-ECFC proliferation, migration and induction of NO production.
GDF15 induces the production of non-deleterious superoxide ions
To compare basal ROS levels between AB-ECFCs and CB-ECFCs, H2-DCFDA, a specific
intracellular probe whose oxidization is proportional to the amount of ROS present in the
cells, was used. Quantification of H2-DCFDA MFI by flow cytometry showed that the basal
ROS level was ~2 times higher in AB-ECFCs than in CB-ECFCs (p = 0.0021, n = 5) (Fig 4A).
As ROS are byproducts of cellular proliferation and migration [
] which are rapidly
generated by mitochondria, we then investigated after only 1h of GDF15 treatment on AB-ECFCs
whether superoxide ions, the main mitochondrial ROS byproduct [
], were produced. The
analysis of Mitosox, a mitochondrial superoxide probe, revealed that GDF15 treatment led to a
21% increase in superoxide production compared to NT cells (Fig 4B). However, the basal
ROS level was already high in AB-ECFCs. We then investigated whether superoxide ion
production could be toxic to AB-ECFCs. NRF2 is the main transcription factor responsible for the
induction of key antioxidant enzymes under oxidative stress conditions [
]. We did not
detect any significant increase in mRNA levels of 3 genes regulated by NRF2 (GCLM, HMOX1
8 / 17
Fig 3. GDF15 improves AB-ECFC functions. AB-ECFCs were starved for 8h and treated overnight with GDF15 (50 ng/mL) or VEGF (50 ng/mL), then
(a) the cell cycle was analyzed by FACS focusing on the G0-G1, S and G2-M phases (n = 6). Quantifications are presented in (b). BrdU proliferation assay
was performed on AB ECFCs in the presence of GDF15 or VEGF (c). Migration assays were performed after GDF15 (50 ng/mL) or VEGF (50 ng/mL)
treatment. Representative pictures and quantification of AB-ECFC migration assays are shown in (d) and (e), respectively (n = 4). (f) Intracellular nitric
oxide (NO) MFI analyzed by flow cytometry using DAF-FM-2A probes in NT, LPS- (100 ng/mL) and GDF15- (50 ng/mL) treated AB-ECFCs (n = 8). (g)
NOS3 RTqPCR analysis in NT and GDF15-treated AB-ECFCs (n = 8). Data are presented as the mean ? SEM. ns = non-significant, P <0.05, P <0.01
and P <0.001 compared to NT AB-ECFCs.
and NQO1) in AB-ECFCs after GDF15 treatment (Fig 4C). This showed that the
GDF15-induced increase in superoxide ion production in AB-ECFCs did not trigger any oxidative stress.
Peroxiredoxin 3 (PRX3) is the main mitochondrial peroxiredoxin to be readily oxidized by
hydrogen peroxide (H2O2), a ROS produced after superoxide dismutation. The evaluation of
the PRX3 redox status after treatment of AB-ECFCs with GDF15 or H2O2 (400 ?M) revealed
that, unlike H2O2, GDF15 did not lead to the accumulation of the hyper-oxidized form of
PRX3 (Fig 4D). This suggested that the superoxide ions produced following GDF15 treatment
were properly scavenged by the antioxidant system. Furthermore, under these conditions,
GDF15 was not toxic to AB-ECFCs as the cell viability remained unchanged after GDF15
treatment (Fig 4E), confirming the absence of oxidative stress despite superoxide ion
Taken together, these results suggested that treatment of AB-ECFCs with GDF15 generates
superoxide ions but at a higher level than the threshold required for PRX3 oxidation and for
NRF2 activation, in response to oxidative stress.
GDF15 activates the ERK1/2, SMAD2 and AKT signaling pathways
We then analyzed whether the GDF15 receptor, GFRAL, the presence of which has recently
been identified in the nervous system [
], was present on CB-ECFCs and AB-ECFCs and
thus could also be the GDF15 receptor on ECs. The Western-Blot analysis did not reveal any
expression of GFRAL in AB-ECFCs and CB-ECFCs, unlike in MDA MB231 cells that were
used as a positive control (Fig 5A).
Then, we focused on several previously described signaling pathways known to be activated
by GDF15 [
]. Western-Blots showed that GDF15 activated the ERK1/2, SMAD2
and AKT signaling pathways (Fig 5B).
Here, we showed that GFRAL is not the receptor of GDF15 on AB-ECFCs and CB-ECFCs
and that GDF15 activates the ERK1/2, SMAD2 and AKT signaling pathways in AB-ECFCs.
Endothelial dysfunction plays an important role in CVD progression. GDF15, a circulating
cytokine normally expressed at basal levels in most human tissues in steady-state situations, is
overexpressed in various pathologies, including CVD [
], or in response to cellular
stresses 12?25, in particular by ECs.
Based on these observations, we assumed that circulating GDF15 could be produced by
ECs during vascular stress to attenuate EC loss of function and damage. In this study, we used
AB-ECFCs and CB-ECFCs, which are the EPCs progeny obtained in culture, as models of ECs
displaying different functional properties. Indeed, CB-ECFCs are highly active ECs and are
resistant to replicative senescence whereas AB-ECFCs are less active and rapidly undergo
We determined whether GDF15 was differentially expressed in these ECs with contrasted
functionalities. GDF15 was specifically overexpressed in AB-ECFCs between passages 4 and 5
or when enforcing a laminar flux, while it was expressed in CB-ECFCs at lower levels. We also
10 / 17
Fig 4. GDF15 induces a non-deleterious superoxide ion production. (a) Intracellular ROS level measurement using H2-DCFDA
probes in CB-ECFCs (n = 8) and AB-ECFCs (n = 6). (b) Mitochondrial superoxide ion quantification using Mitosox probes in NT or
GDF15- (50 ng/mL) treated AB-ECFCs (n = 5). MFI detected for each probe were measured by FACS. (c) Analysis of HMOX1, GCLM
and NQO1 expression by RT-qPCR (n = 3). (d) Redox western-blot of peroxiredoxin 3 (Prx3) protein in NT, GDF15- (50 ng/mL) or
H2O2- (400 ?M) treated AB-ECFCs (n = 3). (Prx3 red: Peroxireoxin 3 reduced form, Prx3 ox: Peroxiredoxin 3 oxidized form, Prx
hyper-ox: Peroxiredoxin hyper-oxidized form). ?-actin was used as the loading control. (e) Cell viability analysis in NT and
GDF15-treated AB-ECFCs (n = 5). Data are presented as the mean ? SEM. ns = non-significant and P <0.05, P <0.01 and
P <0.001 compared to CB-ECFCs or NT cells.
showed that AB-ECFCs included a higher proportion of senescent cells than CB-ECFCs at the
same passage, thus suggesting a relationship between GDF15 expression and cellular
senescence. These findings are consistent with our previous study [
] showing that the number of
?-galactosidase positive cells (senescent) was higher in AB-ECFCs compared to CB-ECFCs.
Then, to determine whether GDF15 expression was associated to cellular senescence, it
expression was quantified in AB-ECFCs and CB-ECFCs. We found that AB-ECFCs expressed
basal levels of GDF15 at early passages when they are not yet senescent, like in CB-ECFCs.
GDF15 level then raised progressively while the number of ?-galactosidase positive cells
increased. GDF15 expression was maximal when most AB-ECFCs were senescent. In other
studies, GDF15 has also been shown to be overexpressed in radiation-induced senescent
], supporting the relationship between senescence and GDF15 over expression.
Taken together, all these findings suggest that senescent ECs could contribute to the release of
high GDF15 levels as previously found in patients with CVD [
]. Therefore, GDF15 could
be referred to as a ?senescence-associated secreted protein? (SASP) like IL-6 or IL-8 which are
also secreted by senescent AB-ECFCs.
Fig 5. GDF15 activates ERK1/2, SMAD2 and AKT signaling pathways. (a) Western-blot of GFRAL in MDA-MB
231 cells (positive control), CB-ECFCs and AB-ECFCs. (b) Western-blot of pERK1/2, ERK1/2, pSMAD2, SMAD2,
pAKT and AKT in AB-ECFCs treated with GDF15 (50 ng/mL) at different timepoints: 30 min, 1h, 1h30 and 2h.
12 / 17
Senescence may be induced by multiple cellular stresses such as telomere attrition, genomic
damage and oncogenic stress [
]. One of the hallmarks of senescent cells is the increased
expression and production of SASP, that include several types of factors such as
proinflammatory cytokines, interleukins, chemokines, growth factors and proteases [
]. SASP contribute
to create a proinflammatory environment that influences neighboring cell behavior [
]. In order to analyze the paracrine effect of GDF15 on ECs, we treated non senescent
AB-ECFCs with recombinant GDF15 to determine whether this could affect their
functionality. We first found that the long-term treatment of non senescent AB-ECFCs with
recombinant GDF15 neither contributed nor delayed their senescence. Recently, GDF15 has been
shown to improve HUVEC proliferation and angiogenic capacities [
]. These data are
consistent with our current findings because we demonstrate that GDF15 improves AB-ECFC
proliferation, migration and NO production. This suggests thus that high levels of GDF15 is
beneficial for a dysfunctional endothelial system.
Mitochondria play a major role in cellular metabolism and are the main source of ROS
]. ROS may trigger various cellular responses depending on their concentrations [
high concentrations, ROS may trigger oxidative stress resulting in a stress adaptive response,
cellular damages or apoptosis [
]. Beside GDF15 positive effects on ECs, GDF15 could also
have toxic effects, since it induces the production of ROS, as shown in previous studies [
Since AB-ECFCs express higher ROS levels than CB-ECFCs, we showed that GDF15 treatment
led to an additional mitochondrial superoxide ion (O??) production, a main ROS byproduct
that could trigger a toxic oxidative stress . In order to evaluate the possible deleterious
effects of this O?? production, we analyzed different markers of cellular response to oxidative
stress. First, we focused on NRF2, the main transcription factor regulating the anti-oxidant
response. We showed that NRF2-dependent gene (GCLM, NQO1 and HMOX1) expression
was not induced by GDF15 treatment, indicating that the cellular response to oxidative stress
was not activated. Second, we did not detect any peroxiredoxin hyper-oxidation, a
phenomenon due to a mitochondrial anti-oxidant enzyme, indicating that the protein oxidation
pathway was not induced. Third, GDF15 treatment did not affect AB-ECFC viability and
senescence status. Taken together, in a context of high ROS levels, our results demonstrated
that, despite an increase in O?? production, GDF15 treatment of NS AB-ECFCs has beneficial
Finally, we investigated GDF15 signaling pathways. GFRAL, previously described as an
orphan receptor expressed on area postrema neurons and solitary tract nuclei in mice and
humans, has been recently identified as a specific receptor of GDF15 [
]. We assessed
whether AB-ECFCs, that are sensitive to GDF15 actions, expressed GFRAL. We did not detect
any GFRAL expression in AB-ECFCs and CB-ECFCs, suggesting that GFRAL is not the
GDF15 receptor on ECFCs. Therefore, GDF15 might use another receptor that remains to be
identified or a receptor shared with other TGF-? molecules, as previously suggested [
Moreover, we demonstrated that GDF15 activated several signaling pathways such as the
ERK1/2, AKT and SMAD2 pathways. Several studies have described that the activation of
these pathways induces EC proliferation, migration, NO production and angiogenesis
], which is confirmed and supported by our findings. These results thus confirmed
that GDF15 is able to activate various signaling pathways to achieve its functions.
In summary, we showed that GDF15, produced by senescent ECs, improves AB-ECFC
proliferation, migration and NO production through the activation of various signaling pathways.
We also confirmed that GDF15 induces ROS production and showed for the first time that
GDF15 is not deleterious to ECs because it does not trigger any oxidative stress. Therefore,
GDF15 could have a therapeutic effect on a damaged vascular system. However, further studies
in animal models are needed to confirm GDF15 beneficial effects on the vascular system.
13 / 17
S1 Fig. GDF15 optimal concentration determined using a proliferation assay. BrdU
incorporation assay in AB-ECFCs in the presence of 10, 50 and 100 ng/mL of GDF15.
S2 Fig. TGF-? is not detected in GDF15 prep. ELISA of TGF-? protein in EGM2 or in EGM2
supplemented with 50ng/ml of GDF15. The point TGF-? at 32pg/ml of the ELISA curve is
used here as positive control. ND: non detected
S3 Fig. AB-ECFC and CB-ECFC phenotypic characterization. (a) Pictures and (b)
phenotypic characterization by FACS of AB-ECFC and CB-ECFC colonies. AB-ECFCs and
CB-ECFCs are both positive for the CD31, CD144 and KDR endothelial markers and negative
for the CD45 hematopoietic marker.
S4 Fig. GDF15 expression gradually increases with passages. GDF15 expression by (a)
western-blot and (b) ELISA between passages 3 to 5.
S5 Fig. GDF15 is over-secreted by senescent AB-ECFC under a laminar flow. ELISA of
GDF15 in non senescent (NS) and senescent (S) AB-ECFC under a laminar flow (n = 6). Data
are presented as the mean ? SEM. P <0.05, compared to NS AB-ECFC.
S6 Fig. GDF15 mRNA is also over-expressed in senescent CB-ECFC. RT-PCR analysis
(n = 5) in senescent AB-ECFC (around passage 5), in CB-ECFC at passage 5, 10 and in
senescent CB-ECFC (beyond passage 20). Data are presented as the mean ? SEM. ns =
non-significant and P <0.05, compared to AB-ECFCs.
S7 Fig. IL-6 and IL-8 are over-secreted in senescent AB-ECFC. Analysis of (a) IL-6 (n = 8)
and (b) IL-8 (n = 7) secretion in non senescent (NS) and senescent (S) AB-ECFC in ELISA.
Data are presented as the mean ? SEM. P <0.01 compared to NS AB-ECFCs.
We acknowledge the French Blood institution (EFS) of Rungis and Cord Blood Bank of
SaintLouis Hospital for providing us with peripheral and umbilical cord blood. We also
acknowledge Evelyne Mutabazi for her participation in this study, Julien Giron-Michel for his technical
support and advice, and Se?bastien Banzet for his helpful advice and review of the manuscript.
Conceptualization: Guillaume Ha, Adrienne Anginot.
Data curation: Guillaume Ha, Adrienne Anginot.
Formal analysis: Guillaume Ha, Adrienne Anginot.
Funding acquisition: Guillaume Ha, Georges Uzan.
Investigation: Guillaume Ha, Fanny De Torres, Nassim Arouche, Se?gol?ne Ferratge, Adrienne
14 / 17
Methodology: Guillaume Ha, Elie Hatem, Adrienne Anginot.
Project administration: Guillaume Ha, Georges Uzan.
Resources: Guillaume Ha, Georges Uzan.
Software: Guillaume Ha, Adrienne Anginot.
Supervision: Guillaume Ha, Adrienne Anginot, Georges Uzan.
Validation: Guillaume Ha, Adrienne Anginot, Georges Uzan.
Visualization: Guillaume Ha, Nassima Benzoubir, Adrienne Anginot.
Writing ? original draft: Guillaume Ha, Adrienne Anginot, Georges Uzan.
Writing ? review & editing: Guillaume Ha, Adrienne Anginot, Georges Uzan.
15 / 17
16 / 17
1. Sandoo A , Veldhuijzen van Zanten JJCS , Metsios GS , Carroll D , Kitas GD . The Endothelium and Its Role in Regulating Vascular Tone . The Open Cardiovascular Medicine Journal . 2010 ; 4 : 302 - 312 . https://doi.org/10.2174/1874192401004010302 PMID: 21339899
2. van Hinsbergh VWM. Endothelium-role in regulation of coagulation and inflammation . Seminars in Immunopathology. 2012 ; 34 : 93 - 106 . https://doi.org/10.1007/s00281-011 -0285-5 PMID: 21845431
3. Brandes R , Fleming I , Busse R . Endothelial aging . Cardiovascular Research . 2005 ; 66 : 286 - 294 . https://doi.org/10.1016/j.cardiores. 2004 . 12 .027 PMID: 15820197
4. van der Loo B , Labugger R , Skepper JN , Bachschmid M , Kilo J , Powell JM , et al. Enhanced Peroxynitrite Formation Is Associated with Vascular Aging . The Journal of Experimental Medicine . 2000 ; 192 : 1731 - 1744 . https://doi.org/10.1084/jem.192.12.1731 PMID: 11120770
5. Matsushita H , Chang E , Glassford AJ , Cooke JP , Chiu C-P , Tsao PS . eNOS Activity Is Reduced in Senescent Human Endothelial Cells: Preservation by hTERT Immortalization . Circulation Research . 2001 ; 89 : 793 - 798 . https://doi.org/10.1161/hh2101.098443 PMID: 11679409
6. Deanfield JE , Julian P. Halcox , Ton J. Rabelink. Endothelial Function and Dysfunction: Testing and Clinical Relevance . Circulation . 2007 ; 115 : 1285 - 1295 . https://doi.org/10.1161/CIRCULATIONAHA. 106.652859 PMID: 17353456
7. Bootcov MR , Bauskin AR , Valenzuela SM , Moore AG , Bansal M , He XY , et al. MIC-1, a novel macrophage inhibitory cytokine, is a divergent member of the TGF- superfamily . Proceedings of the National Academy of Sciences . 1997 ; 94 : 11514 - 11519 . https://doi.org/10.1073/pnas.94.21.11514 PMID: 9326641
8. Ago T. GDF15 , a Cardioprotective TGF- Superfamily Protein . Circulation Research . 2006 ; 98 : 294 - 297 . https://doi.org/10.1161/01.RES. 0000207919 .83894.9d PMID: 16484622
9. Altena R , Fehrmann RSN , Boer H , de Vries EGE , Meijer C , Gietema JA . Growth Differentiation Factor 15 (GDF-15) Plasma Levels Increase during Bleomycin- and Cisplatin-Based Treatment of Testicular Cancer Patients and Relate to Endothelial Damage . de Mello RA , editor. PLOS ONE . 2015 ; 10 . https:// doi.org/10.1371/journal.pone. 0115372 PMID: 25590623
10. Blanco-Calvo M , Tarr??o N , Reboredo M , Haz-Conde M , Garc??a J , Quindo?s M, et al. Circulating levels of GDF15, MMP7 and miR-200c as a poor prognostic signature in gastric cancer . Future Oncology . 2014 ; 10 : 1187 - 1202 . https://doi.org/10.2217/fon.13.263 PMID: 24947260
11. Shnaper S , Desbaillets I , Brown DA , Murat A , Migliavacca E , Schluep M , et al. Elevated levels of MIC-1/ GDF15 in the cerebrospinal fluid of patients are associated with glioblastoma and worse outcome . International Journal of Cancer . 2009 ; 125 : 2624 - 2630 . https://doi.org/10.1002/ijc.24639 PMID: 19521960
12. Staff AC , Bock AJ , Becker C , Kempf T , Wollert KC , Davidson B . Growth differentiation factor-15 as a prognostic biomarker in ovarian cancer . Gynecologic Oncology . 2010 ; 118 : 237 - 243 . https://doi.org/10. 1016/j.ygyno. 2010 . 05 .032 PMID: 20576287
13. Fujita Y , Taniguchi Y , Shinkai S , Tanaka M , Ito M. Secreted growth differentiation factor 15 as a potential biomarker for mitochondrial dysfunctions in aging and age-related disorders: GDF15 in aging and age-related disorders . Geriatrics & Gerontology International . 2016 ; 16 : 17 - 29 . https://doi.org/10.1111/ ggi.12724 PMID: 27018280
14. Ikeda Y , Kumagai H , Motozawa Y , Suzuki J . Growth Differentiation Factor 15 ( GDF15 ) as a Reliable Biomarker for Cardiovascular Risk Assessment . International Heart Journal . 2016 ; 57 : 1 - 2 . https://doi. org/10.1536/ihj.15-324 PMID: 26742705
15. Hsu L-A , Wu S , Juang J-MJ , Chiang F-T , Teng M-S , Lin J-F , et al. Growth Differentiation Factor 15 May Predict Mortality of Peripheral and Coronary Artery Diseases and Correlate with Their Risk Factors . Mediators of Inflammation . 2017 ; 2017 : 1 - 13 . https://doi.org/10.1155/ 2017 /9398401 PMID: 28798540
16. De Haan JJ , Haitjema S , den Ruijter HM , Pasterkamp G , de Borst GJ , Teraa M , et al. Growth Differentiation Factor 15 Is Associated With Major Amputation and Mortality in Patients With Peripheral Artery Disease . Journal of the American Heart Association . 2017 ; 6 . https://doi.org/10.1161/JAHA.117. 006225 PMID: 28855167
17. Zhang Y , Moszczynski LA , Liu Q , Jiang J , Zhao D , Quan D , et al. Over-expression of growth differentiation factor 15 (GDF15) preventing cold ischemia reperfusion (I/R) injury in heart transplantation through Foxo3a signaling . Oncotarget . 2017 ; 8 . https://doi.org/10.18632/oncotarget.16607 PMID: 28388574
18. Johnen Heiko , Kuffner Tamara, Brown David A. , Wu Ben J., Stocker Roland , Breit Samuel N. Increased expression of the TGF-b superfamily cytokine MIC-1/GDF15 protects ApoE ?/? mice from the development of atherosclerosis . Cardiovascular Pathology . 2012 ; 499 - 505 . https://doi.org/10.1016/j.carpath. 2012 . 02 .003 PMID: 22386250
19. Li J , Yang L , Qin W , Zhang G , Yuan J , Wang F . Adaptive Induction of Growth Differentiation Factor 15 Attenuates Endothelial Cell Apoptosis in Response to High Glucose Stimulus . Tsilibary EC, editor. PLoS ONE . 2013 ; 8 . https://doi.org/10.1371/journal.pone. 0065549 PMID: 23799024
20. Song Hejian , Yin Delu, Liu Zhihua. GDF-15 promotes angiogenesis through modulating p53/HIF-1a signaling pathway in hypoxic human umbilical vein endothelial cells . Mol Biol Rep . 2012 ; 39 : 4017 - 4022 . https://doi.org/10.1007/s11033-011 -1182-7 PMID: 21773947
21. Tiwari KK , Moorthy B , Lingappan K. Role of GDF15 (growth and differentiation factor 15) in pulmonary oxygen toxicity . Toxicology in Vitro . 2015 ; 29 : 1369 - 1376 . https://doi.org/10.1016/j.tiv. 2015 . 05 .008 PMID: 26004619
22. Nickel N , Jonigk D , Kempf T , Bockmeyer CL , Maegel L , Rische J , et al. GDF -15 is abundantly expressed in plexiform lesions in patients with pulmonary arterial hypertension and affects proliferation and apoptosis of pulmonary endothelial cells . Respiratory Research . 2011 ; 12 . https://doi.org/10.1186/ 1465 -9921-12-12
23. Whitson RJ , Lucia MS , Lambert JR . Growth differentiation factor-15 (GDF-15) suppresses in vitro angiogenesis through a novel interaction with connective tissue growth factor (CCN2) . Journal of Cellular Biochemistry . 2013 ; 114 : 1424 - 1433 . https://doi.org/10.1002/jcb.24484 PMID: 23280549
24. Asahara T , Murohara T , Sullivan A , Silver M , van der Zee R , Li T , et al. Isolation of putative progenitor endothelial cells for angiogenesis . Science . 1997 ; 275 : 964 - 967 . PMID: 9020076
25. Kalka C , Masuda H , Takahashi T , Kalka-Moll WM , Silver M , Kearney M , et al. Transplantation of ex vivo expanded endothelial progenitor cells for therapeutic neovascularization . MEDICAL SCIENCES.: 6.
26. Kawamoto A , Gwon H-C , Iwaguro H , Yamaguchi J-I , Uchida S , Masuda H , et al. Therapeutic Potential of Ex Vivo Expanded Endothelial Progenitor Cells for Myocardial Ischemia . Circulation . 2001 ; 103 : 634 - 637 . https://doi.org/10.1161/01.CIR. 103 .5.634 PMID: 11156872
27. Ingram DA . Identification of a novel hierarchy of endothelial progenitor cells using human peripheral and umbilical cord blood . Blood . 2004 ; 104 : 2752 - 2760 . https://doi.org/10.1182/blood-2004 -04-1396 PMID: 15226175
28. Au P , Daheron LM , Duda DG , Cohen KS , Tyrrell JA , Lanning RM , et al. Differential in vivo potential of endothelial progenitor cells from human umbilical cord blood and adult peripheral blood to form functional long-lasting vessels . Blood . 2007 ; 111 : 1302 - 1305 . https://doi.org/10.1182/blood-2007 -06- 094318 PMID: 17993613
29. Ferratge S , Ha G , Carpentier G , Arouche N , Bascetin R , Muller L , et al. Initial clonogenic potential of human endothelial progenitor cells is predictive of their further properties and establishes a functional hierarchy related to immaturity . Stem Cell Research . 2017 ; 21 : 148 - 159 . https://doi.org/10.1016/j.scr. 2017 . 04 .009 PMID: 28499264
30. Olsen OE , Skjaervik A , St?rdal BF , Sundan A , Holien T. TGF-? contamination of purified recombinant GDF15. Goumans MJ, editor . PLOS ONE . 2017 ; 12 : e0187349. https://doi.org/10.1371/journal.pone. 0187349 PMID: 29161287
31. Coppe? J-P , Desprez P-Y, Krtolica A , Campisi J . The Senescence-Associated Secretory Phenotype: The Dark Side of Tumor Suppression . Annual Review of Pathology: Mechanisms of Disease . 2010 ; 5 : 99 - 118 . https://doi.org/10.1146/annurev-pathol- 121808 -102144 PMID: 20078217
32. Forstermann U , Closs EI , Pollock JS , Nakane M , Schwarz P , Gath I , et al. Nitric oxide synthase isozymes. Characterization, purification, molecular cloning, and functions . Hypertension . 1994 ; 23 : 1121 - 1131 . https://doi.org/10.1161/01.HYP. 23 .6.1121 PMID: 7515853
33. Schieber M , Chandel NS. ROS Function in Redox Signaling and Oxidative Stress . Current Biology . 2014 ; 24 : R453 - R462 . https://doi.org/10.1016/j.cub. 2014 . 03 .034 PMID: 24845678
34. Turrens JF . Mitochondrial formation of reactive oxygen species . The Journal of Physiology . 2003 ; 552 : 335 - 344 . https://doi.org/10.1113/jphysiol. 2003 .049478 PMID: 14561818
35. Ma Q. Role of Nrf2 in Oxidative Stress and Toxicity . Annual Review of Pharmacology and Toxicology . 2013 ; 53 : 401 - 426 . https://doi.org/10.1146/annurev-pharmtox- 011112 -140320 PMID: 23294312
36. Emmerson PJ , Wang F , Du Y , Liu Q , Pickard RT , Gonciarz MD , et al. The metabolic effects of GDF15 are mediated by the orphan receptor GFRAL . Nature Medicine . 2017 ; 23 : 1215 - 1219 . https://doi.org/ 10.1038/nm.4393 PMID: 28846098
37. Mullican SE , Lin-Schmidt X , Chin C-N , Chavez JA , Furman JL , Armstrong AA , et al. GFRAL is the receptor for GDF15 and the ligand promotes weight loss in mice and nonhuman primates . Nature Medicine . 2017 ; 23 : 1150 - 1157 . https://doi.org/10.1038/nm.4392 PMID: 28846097
38. Yang L , Chang C-C , Sun Z , Madsen D , Zhu H , Padkjaer SB , et al. GFRAL is the receptor for GDF15 and is required for the anti-obesity effects of the ligand . Nature Medicine . 2017 ; 23 : 1158 - 1166 . https://doi. org/10.1038/nm.4394 PMID: 28846099
39. Hsu J-Y, Crawley S , Chen M , Ayupova DA , Lindhout DA , Higbee J , et al. Non-homeostatic body weight regulation through a brainstem-restricted receptor for GDF15 . Nature . 2017 ; 550 : 255 - 259 . https://doi. org/10.1038/nature24042 PMID: 28953886
40. Park H , Kim C-H , Jeong J-H , Park M , Kim KS . GDF15 contributes to radiation-induced senescence through the ROS-mediated p16 pathway in human endothelial cells . Oncotarget . 2016 ; 7 . https://doi. org/10.18632/oncotarget.7457 PMID: 26909594
41. Wollert KC . Growth-differentiation factor-15 in cardiovascular disease: From bench to bedside, and back . Basic Research in Cardiology. 2007 ; 102 : 412 - 415 . https://doi.org/10.1007/s00395-007 -0662-3 PMID: 17546529
42. Lind L , Wallentin L , Kempf T , Tapken H , Quint A , Lindahl B , et al. Growth-differentiation factor-15 is an independent marker of cardiovascular dysfunction and disease in the elderly: results from the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) Study . European Heart Journal . 2009 ; 30 : 2346 - 2353 . https://doi.org/10.1093/eurheartj/ehp261 PMID: 19561023
43. Burton DGA , Krizhanovsky V . Physiological and pathological consequences of cellular senescence . Cellular and Molecular Life Sciences . 2014 ; 71 : 4373 - 4386 . https://doi.org/10.1007/s00018-014 -1691- 3 PMID: 25080110
44. Davalos AR , Coppe J-P , Campisi J , Desprez P -Y. Senescent cells as a source of inflammatory factors for tumor progression . Cancer and Metastasis Reviews . 2010 ; 29 : 273 - 283 . https://doi.org/10.1007/ s10555-010 -9220-9 PMID: 20390322
45. Rodier F , Campisi J . Four faces of cellular senescence . The Journal of Cell Biology . 2011 ; 192 : 547 - 556 . https://doi.org/10.1083/jcb.201009094 PMID: 21321098
46. Lecot P , Alimirah F , Desprez P-Y , Campisi J , Wiley C. Context-dependent effects of cellular senescence in cancer development . British Journal of Cancer . 2016 ; 114 : 1180 - 1184 . https://doi.org/10. 1038/bjc. 2016 .115 PMID: 27140310
47. Wang S , Li M , Zhang W , Hua H , Wang N , Zhao J , et al. Growth differentiation factor 15 promotes blood vessel growth by stimulating cell cycle progression in repair of critical-sized calvarial defect . Scientific Reports . 2017 ; 7 . https://doi.org/10.1038/s41598-017 -09210-4 PMID: 28831101
48. Murphy MP . How mitochondria produce reactive oxygen species . Biochemical Journal . 2009 ; 417 : 1 - 13 . https://doi.org/10.1042/BJ20081386 PMID: 19061483
49. Liu D-D , Lu J-M , Zhao Q-R , Hu C , Mei Y -A. Growth differentiation factor-15 promotes glutamate release in medial prefrontal cortex of mice through upregulation of T-type calcium channels . Scientific Reports . 2016 ; 6 . https://doi.org/10.1038/srep28653 PMID: 27353765
50. Johnen H , Lin S , Kuffner T , Brown DA , Tsai VW -W, Bauskin AR , et al. Tumor-induced anorexia and weight loss are mediated by the TGF-? superfamily cytokine MIC-1 . Nature Medicine . 2007 ; 13 : 1333 - 1340 . https://doi.org/10.1038/nm1677 PMID: 17982462