CHAPTER 2: MATERIALS AND METHODS
2.11 Statistical analysis
Values were expressed as means ± SD. Statistical significance was determined using
a Student’s t-test. A P-value <0.05 was considered significant.
h
27
CHAPTER 3: RESULTS
28
3.1 TNF-D D treatment triggers MAPKs transient activation rather than cell apoptosis in endothelial EAhy926 cells
Before studying the signaling pathway between DUSPs and MAPKs in TNF-D
stimulated endothelium, we first eliminated cell apoptosis as the consequence of TNF-D
stimulation in endothelial EAhy926 cells, which were established by fusing primary
human umbilical vein endothelial cells (HUVECs) with a thioguanine-resistant clone of
A549 epithelia. We first checked caspase 8 (initiator caspase) and caspase3 (effector caspase) activation in EAhy926 cells exposed to TNF-D. The protein level of
pro-caspase 8 and 3 maintained intact and no cleaved form of caspases has been observed in 13 hours of TNF-D treatment (Fig. 1A). We also performed flow cytometry
analysis to check the cell cycle in TNF-D-treated EAhy926 cells and there was no
obvious apoptotic cells appeared in sub-G1 group (Fig. 1B). These preliminary tests removed the possibility that TNF-D triggers the cell death signaling in endothelial
EAhy926 cells.
Then we checked mitogen-activated protein kinases (MAPKs) activity changes by monitoring phosphorylation levels of MAPKs in TNF-D-treated EAhy926 cells.
Exposed to TNF-D, p38 MAPK and JNK were transiently activated whereas the activity
of ERK activity was not significantly changed in EAhy926 cells (Fig. 2). At resting
state, p38 MAPK and JNK were not activated whereas ERK has basal level activity.
o o o o
on n n n n n n n n n n n n n r r r r r r r r r r r r r r r r r ra a a a a a a a a a a a a a a at t t t th h h h he e e e er r r t t t t th h h h ha a a a an n n n n n n n n n n n n n n
MAPKs iiiiniiininininnnnn TTTTTTTTTTTTTTNNNNNNNNFNNNNNN D
29
Upon stimulation with TNF-D, p38 MAPK was activated at 5 minutes and then the
phosphorylation level of p38 decreased gradually with the time course. JNK activity
was peaked at 15 minutes and then diminished to basal level.
In order to know the mechanism by which TNF-D regulates MAPKs activity, we next
applied actinomycin D (transcription inhibitor) and cycloheximide (translation inhibitor)
to investigate the underling mechanism of MAPKs inactivation. As shown in Fig. 3,
actinomycin D and cycloheximide pretreatment sustained the phosphorylation level of all MAPKs in TNF-D-treated EAhy926 cells, indicating that TNF-D regulates the
activity of MAPKs through transcriptional and translational regulation mechanisms.
3.2 DUSPs are inducibly expressed in endothelial cells exposed to TNF-D D and function as MKPs
In order to know whether some DUSPs were inducibly expressed as negative
feedback regulators to down regulate MAPKs signaling in endothelial cell exposed to TNF-D, the mRNA levels of 12 DUSP genes, characterized as MKPs, were measured by
quantitative real-time PCR over a course of five hours after TNF-D stimulation. These
12 DUSPs were assigned to three groups based upon the magnitude and pattern of
inducible mRNA expression. Phosphatases in the first group, including DUSP6, DUSP8
and DUSP16, were rapidly induced in a transient manner and peaked at 1-2 hours of nu
30
TNF-D exposure (Fig. 4A). DUSP10 was the only phosphatase in the second group
whose mRNA levels gradually accumulated and remained at high levels for 5 hours
after stimulation (Fig. 4B). The DUSPs in these two groups showed at least a 5-fold
induction after treatment, and were, therefore, classified as genes with significant up-regulation in response to TNF-D. In contrast, the mRNA levels of the remaining
eight DUSPs, which were listed in the third group, did not change or were only
marginally increased (<3.5-fold) after treatment (Fig. 4C). Additional runs of real-time
PCR further confirmed the induction of four DUSPs (DUSP6, 8, 10 and 16) in cells exposed to TNF-D (Fig. 4D).
Next we investigated whether these four TNF-D-induced DUSPs function as MKP.
So we examined the phosphorylation levels of MAPKs in response to RNAi-mediated
knockdown of each DUSP separately. Sufficient knockdown effect of specific siRNA
targeting on DUSP6, 10 and 16 were confirmed by quantitative real-time PCR to check
the residual mRNA level of each DUSP (Fig. 5A). Although DUSP8 knockdown by
RNAi was not complete, the effect on regulating JNK and ERK dephosphorylation by
DUSP8 was confirmed (Fig. 5C). Through RNA interference technique, DUSP6 and
DUSP8 were identified as JNK and ERK phosphatases in EAhy926 cells stimulated with TNF-D, while DUSP16 showed a marginal effect and DUSP10 was not involved in
MAPKs regulation (Fig. 5B-5E).
31
3.3 DUSP6 is involved in TNF-D D-induced endothelial inflammation by
regulating intercellular adhesion molecules 1 (ICAM-1) expression
Inflammation is a dominant physiological consequence of endothelia exposed to TNF-D. TNF-D induces endothelial inflammatory response by regulating cell adhesion
molecules, such as ICAM-1, VCAM-1 and E-selectin, expression on cell surface through NF-NB pathway. Previous studies reported that ERK may function as a negative
regulator in cell adhesion molecules expression by inactivating NF-NB pathway. We
next examined whether TNF-D-induced DUSPs may also participate in regulation of
endothelial inflammation through targeting ERK activity. For this, ICAM-1 expression in TNF-D-treated EAhy926 cell has been checked by immunoblotting (Fig. 6A). In
EAhy926 cells, ICAM-1 expression was induced at 2 hours post TNF-D stimulation and
the protein was gradually accumulated with the time progression. In order to investigate
the role of inducible DUSPs in regulating inflammatory response, we examined the levels of TNF-D-induced ICAM-1 in response to RNAi-mediated knockdown of each
DUSP separately. Interestingly, ablation of DUSP6 led to a decrease of ICAM-1 (Figure
6B), whereas knockdowns of DUSP8, 10, or 16 did not (Figure 6C-6E). These results
suggest DUSP6 may regulate ICAM-1 expression through down-regulating ERK
activity. Although DUSP8 can function as ERK phosphatase (Figure 5C), knockdown of
DUSP8 did not affect ICAM-1 expression. The different regulation of ICAM-1 maybe
nf f f f f f f f f f f f f fl l l l l l l l l l la a a a a a a a a a a a a a a a a a a am m m m m m m m m m m m m m m m m m m m m mm m m m ma a a a at t ti i io o o o on n n n n b b b b b b b b b b b b b by y y y y y y y y y y y y y y y y y y y y y
32
due to the different subcellular localization of phosphatase, DUSP6 is a cytoplasmic
phosphatase and DUSP8 is found in both the cell nucleus and cytoplasm,37 therefore
DUSP8 may regulate different population of ERK which is not overlapped with DUSP6
regulated ERK population.
3.4 Inducible DUSP6 regulates TNF-D D-directed inflammatory responses in primary endothelial HUVECs
Focusing on the potential role of inducible DUSP6 in TNF-D-induced inflammation,
we used primary human umbilical vein endothelial cells (HUVECs) as a physiologically
relevant model for further investigations. The induction of DUSP6 was first confirmed
with quantitative real-time PCR by monitoring mRNA level of Dusp6 gene in TNF-D-stimulated HUVECs (Fig. 7A). We observed that the mRNA level of Dusp6 was
increased and peaked at 1 hour of TNF-D exposure. The protein levels of DUSP6 and
ICAM-1 were also checked in parallel by immunoblotting. Upon stimulation, DUSP6 was obviously increased in 1 hour after TNF-D treatment and then ICAM-1 started to
express at 2 hours of TNF-D exposure (Fig. 7B).
To evaluate the role of DUSP6 in regulating inflammatory response, we next examined the levels of TNF-D-induced ICAM-1 in HUVECs transfected with DUSP6
siRNA. Interestingly, ablation of DUSP6 led to significant suppression of ICAM-1, 6
6 6 6 6 6 6 6
6 isisisisisisiiisisisisisisisisis ssss aaaaaaaaaaaaaa cycycycycytototott plplplpllasasasasasmimimimimimimimimimmimimiiiiiicc c c ccc c c cc c
oplplplplplplplplplppplpplplllasaasasasasasasasasmsmmmmmmmmmmmmmmmmmm,,373737373737 ththththththerererererefefeefefe orororororrrrrrrrrrrrrrre eee eeeee e e
apped wiiiiiiiiiiiiiththththttthththththhhhDDDDDDDDDDDDDDUUUUUUUSUSUUUUUU P6
33
which was otherwise robustly expressed between 4 and 6 hours of TNF-D treatment
(Figure 8A). To validate the specific function of DUSP6 in TNF-D signaling, we
performed a rescue experiment by ectopically expressing a phosphatase-active, wild
type (WT) form of DUSP6 in HUVECs, in which endogenous DUSP6 had been
knocked down. As depicted in Figure 8B, when the WT form of DUSP6 was
re-expressed in RNAi-ablated HUVECs, there was a significant increase in the level of ICAM-1 in response to TNF-D stimulation. Previously, phosphatase activity of DUSP6
has been found to be required for inhibition of ERK.59 To further investigate the importance of DUSP6 enzymatic activity in regulating TNF-D-induced ICAM-1
expression, we re-expressed the phosphatase-dead mutant form of DUSP6 (catalytic
Cys293 replaced by Ser, the C/S mutant) in DUSP6 ablated HUVECs and checked ICAM-1 level. As shown in Fig. 9, in contrast to restoration of TNF-D signaling by the
WT form of DUSP6, the C/S mutant form of DUSP6, which was robustly accumulated
in cells leading to increased activity of endogenous ERK, did not promote the
expression of ICAM-1. These results suggest that DUSP6-promoted inflammatory
response in HUVECs depends on its catalytic activity.
We further examined whether endothelial leukocyte interaction, which is primarily
mediated by the accumulation of adhesion molecules on the surface of endothelium,60is regulated by DUSP6. Clearly, the binding of U937 monocytes to TNF-D-exposed f TTTTTTTTTTTTTTTTTTTNFNFNFNFNFNFNFNFNFNFNFNFNNFNNNFNNFNFNNF-F-DDDDD trtrtrtt eaeaeaaatmtmtmtmtmenenenenenenennnnnnnnnnnnnnntt t tt ttt t ttt
34
HUVECs was inhibited when DUSP6 in HUVECs was knocked down (Fig. 10A),
presumably due to the decreased levels of endothelial ICAM-1 (Fig. 8A).
Consistently, a loss of endothelial leukocyte interaction in response to DUSP6 ablation
was restored by re-expression of the WT form of DUSP6 (Fig. 10B). Together these findings suggested that inducible DUSP6 might play a key role in TNF-D-induced
endothelial inflammation.
3.5 DUSP6-mediated termination of ERK activity is essential for TNF-D D-induced inflammatory response in endothelium
Having demonstrated the involvement of DUSP6 in TNF-D-stimulated ICAM-1
expression, we next investigated whether down-regulation of ERK, the primary function
of DUSP6 thus far identified,36 is essential for this process. We first examined the detailed time-dependent regulation of ERK by DUSP6 in HUVECs exposed to TNF-D.
As shown in Fig. 11A, the immediate activation of ERK was terminated at 1 hour post
stimulation in control cells. However, ERK activation was sustained over the duration of 1-6 hours after TNF-Dtreatment in DUSP6-ablated cells (Fig. 11A). We further
evaluated whether DUSP6 also regulates JNK or p38 MAPKs in HUVECs exposed to TNF-DSurprisingly, unlike in TNF-D-treated EAhy926 that DUSP6 may also regulate
JNK activity. In contrast to the knockdown effect on ERK phosphorylation, ablation of do
35
DUSP6 did not cause a detectable change of TNF-D-dependent transient activation of
either JNK (Fig. 11B) or p38 (Fig. 11C), consistent with the current knowledge that
DUSP6 is a specific ERK phosphatase. These results suggested that termination of ERK
activity by inducible DUSP6 might promote the subsequent signaling essential for
ICAM-1 expression.
To test this hypothesis, we first elucidated the natural instincts of ERK to suppress the expression of ICAM-1 in HUVECs stimulated with TNF-DThe conventional ERK
inhibitors, PD184352, U0126 and PD98059 were used to pretreat cells before exposure to TNF-D As shown in Fig. 12, all three chemical inhibitor pretreatment increased
TNF-D-induced ICAM-1 expression. We also performed another approach by knocking
down ERK with siRNA to verify this result. Data shown in Fig. 13 demonstrated that ablation of ERK leads to enhanced ICAM-1 expression in HUVECs response to TNF-D
stimulation. Furthermore, forced inhibition of ERK via treatment with the chemical
inhibitors led to restoration of ICAM-1 levels in DUSP6 RNAi-ablated HUVECs (Fig.
14). Such findings supported the notion that ERK-caused negative constraint on
ICAM-1 expression may be lifted by inducible DUSP6. We tested this hypothesis by examining the knockdown effect of ERK on TNF-D-promoted ICAM-1 levels in
DUSP6 RNAi-ablated HUVECs. Clearly, DUSP6 deficiency-caused low levels of
ICAM-1 were partially restored when ERK expression was suppressed by the siRNA nsisisisisisissisisisisseneneeneneneneneneneneneneneneneneennt tttt t ttttt tttt acacacacactititititivavavatititititiononononon ooooooooooooooooooof f f f ff f ff f f f
36
specifically targeting ERK1 and ERK2 (Fig. 15). These results together suggest a
functional role in termination of ERK activity that DUSP6 plays during promotion of
endothelial inflammation.
3.6 Inhibition of ERK by DUSP6 promotes NF-NNB transcriptional activation in endothelium exposed to TNF-D
In order to figure out the underlining mechanism that DUSP6 increased ICAM-1
expression, mRNA level of ICAM-1 was checked by quantitative real-time PCR in
DUSP6 ablated HUVECs. As shown in Fig. 16, Dusp6 knockdown by siRNA decreased
ICAM-1 mRNA level, indicting Dudp6 regulate ICAM-1 in a transcriptional-dependent manner. It was suggested previously that activation of NF-NB is essential for expression
of adhesion molecules in endothelium under inflammatory response.49 Upon TNF-D
treatment, upstream kinase activation mediates inhibitor of NF-NB (INB)
phosphorylation and proteasome-dependent degradation, therefore releases free form of NF-NB to direct downstream inflammatory gene expression. We first examined NF-NB
signaling in HUVECs exposed to TNF-D stimulation. As shown in Fig. 17A,
degradation of INB-Dhappened in 5 minutes of TNF-D treatment and newly synthesized
INB-D appeared at 45 minutes of stimulation. Meanwhile the increased phosphorylation
level of NF-NB indicates TNF-D-induced activation. We further elucidate the role of o
37
NF-NB in regulating ICAM-1 expression in TNF-D-treated HUVECs. Using a specific
inhibitor BAY-117082 to pretreat cells, we showed intact INB-D level in TNF-D treated
sample which demonstrates sufficient inhibition of NF-NB signaling, and TNF-D-induced ICAM-1 expression was blocked completely (Fig. 17B). This result
suggests a critical role of NF-NB in TNF-D-mediated induction of ICAM-1 protein in
HUVECs. We wonder whether inducible DUSP6 would promote the expression of ICAM-1 through its up-regulation of NF-NB signaling. To test this hypothesis, we first
examined the dynamic change of INB-D levels over the duration of TNF-D stimulation
and investigated whether ablation of DUSP6 could affect this process. Interestingly, the immediate degradation of INB within the first 15 minutes post treatment occurred
independent of endogenous levels of DUSP6 (Fig. 17C), indicating that release of NF-NB from its inhibitor INB was not influenced by DUSP6. Consistently, NF-NB was
rapidly phosphorylated soon after TNF-D treatment even though DUSP6 was knocked
down (Fig.17C), further eliminating the possible involvement of DUSP6 at the initial phase of NF-NB pathway. In contrast, during the course of INB re-synthesis, a
NF-NB-dependent process61 that became evident between 1-6 hours after TNF-D
stimulation (Fig. 17C); we observed that DUSP6 played a clear role. Specifically, the accumulated levels of re-synthesized INB at 4-6 hours post treatment were significantly
suppressed when endogenous DUSP6 was ablated (Fig. 17C). Together, these results s
38
suggest that transcriptional activation of NF-NB, which is essential for re-synthesis of
INB and also inducible expression of ICAM-1, might be regulated by DUSP6 in
endothelium under TNF-Dstimulation.
This finding led us to wonder whether inducible DUSP6 would promote the expression of ICAM-1 through its up-regulation of NF-NB transcriptional activity; and
if this is the case, whether the underlying mechanism of such process is determined by
DUSP6-dependent inactivation of ERK. To test this hypothesis, we examined the role of DUSP6 in NF-NB-directed transcription of ICAM-1 gene. For this, luciferase reporter
assay was conducted in TNF-D-treated HUVECs. One unique NF-NB binding element
in the promoter region of ICAM-1 gene26 was selected for characterization (Fig. 18A).
The pilot test demonstrated that NF-NB-mediated transcription of ICAM-1 gradually
increased after TNF-D treatment, and that the robust enhancement of luciferase activity
occurred at 4 hours post stimulation (Fig. 18A). This led us to examine a role that DUSP6 plays in regulating NF-NB at this time. Clearly, upon the ablation of DUSP6,
NF-NB-mediated transcription of ICAM-1 promoter was significantly suppressed (Fig.
18B). We further investigated whether inactivation of ERK is essential for DUSP6-mediated activation of NF-NB. To do this, a mutant form of DUSP6 in which
the kinase interacting motif (KIM) was disrupted thereby losing its interaction with fo
39
ERK (termed KIM mutant thereafter),62 and the WT form of DUSP6, were ectopically
expressed in HUVECs treated with siRNA to ablate endogenous DUSP6. We found that, unlike the WT form of DUSP6 that restored the NF-NB activity significantly, the KIM
mutant form of DUSP6 was unable to do so (Fig. 18C). In addition, the KIM mutant form of DUSP6 was incapable of restoring TNF-D-induced expression of ICAM-1
protein (Fig. 19). We thus concluded that DUSP6 promotes canonical NF-NB signaling
through its inactivation of ERK for inducible expression of ICAM-1 during endothelial
inflammation.
3.7 TNF-D D-induced ICAM-1expression on the endothelial layer of
aorta and vein is attenuated in Dusp6
-/-mice
Having demonstrated the mechanism that illustrates how DUSP6 promotes inducible expression of ICAM-1 during initial six hours of TNF-D exposure, we further examined
the role of DUSP6 in endothelium with prolonged inflammatory response. As shown in Fig. 20, the induction of ICAM-1 in HUVECs peaked at 12-hour of TNF-D treatment,
and then sustained up to 24 hours. Importantly, RNAi ablation of DUSP6 led to a
significant decrease of ICAM-1 levels during the period of 12 to 24 hours after
stimulation (Fig. 20), suggesting that the maximal expression of ICAM-1 in the
inflamed endothelium is DUSP6-dependent. This hypothesis was subsequently tested in 6
40
vivo. For this, we used DUSP6 null mice (Dusp6-/- strain B6;129-Dusp6tm1Jmol/J)57 to
study whether DUSP6 regulates endothelial inflammation in aorta and inferior vena
cava (IVC). The degree of inflammatory response in animal tissues was measured by
the immunohistochemistry (IHC) staining with anti-ICAM-1 antibody. First, we
checked the mice knockout background by genotyping and the absence of DUSP6 expression in tissues also shown in Fig. 21. Then we challenged mice with TNF-D (5
Pg/kg) to examine ICAM-1 level on vascular endothelia of aorta and IVC. After
injection with TNF-D for 16 hours, there was significant increase in levels of ICAM-1
on the endothelial layer of both aorta and IVC isolated from the WT control mice
(B6129SF2/J) (Fig. 22A). Pairs of WT and Dusp6-/- mice were then examined under
various conditions. Basal expression of endothelial ICAM-1 was low regardless of the presence or absence of Dusp6 gene (Fig. 22B). Furthermore, TNF-D-induced ICAM-1
expression on the surface of aorta and IVC was visualized by the specific anti-ICAM-1
antibody but not by the isotype IgG (Fig. 22C). This data confirmed the reliability of
IHC staining. Interestingly, although endothelial ICAM-1 expression on the surface of aorta and IVC was robustly enhanced in the WT mice exposed to TNF-D, its level in the
Dusp6-/- mice remained low (Fig. 23). Quantitative results from multiple animals
revealed a significant difference in ICAM-1 levels on endothelial layer of aorta (Fig.
23A) and IVC (Fig. 23B) between the WT and Dusp6-/-mice under TNF-D stimulation.
Du
41
3.8 Deficiency of DUSP6 protects mice from acute lung injuries during experimental sepsis
We investigated whether DUSP6 regulates the pathological process of sepsis, which
is a severe medical condition characterized by a systemic inflammatory response to
infection.63 We particularly focused on the potential role of DUSP6 involved in
inflammatory consequence of sepsis within the pulmonary circulation, as the lung is
continuously exposed to circulating pathogen-associated molecular patterns such as
endotoxin lipopolysaccharide (LPS).14, 63 In addition, human pulmonary microvascular endothelial cells and HUVECs respond to TNF-D and LPS similarly in terms of NF-NB
activation and surface ICAM-1 expression,64 suggesting that a DUSP6-dependent
regulatory mechanism might be adopted by two types of endothelia. This hypothesis
was tested by intraperitoneal injection of WT and Dusp6-/-mice with TNF-D (0.1 mg/kg)
or LPS (10 mg/kg); the latter stimulates the expression of ICAM-1 and thus promoting neutrophil infiltration-dependent pulmonary injury through release of TNF-D.65After 24
hours of treatment, lung sections taken from the mice were subjected to staining with
hematoxylin and eosin (H&E) or anti-ICAM-1 antibody. As shown in Fig. 24A, a
significant degree of histologic lung injury, as indicated by notable inflammatory cells
infiltration and inter-alveolar septal thickening, was observed in the WT mice exposed to TNF-D or LPS. Interestingly, these tissue damages were attenuated in TNF-D or
i
42
LPS-treated Dusp6-/- mice (Fig. 24A). Moreover, ICAM-1 staining on the surface of alveolar walls was increased in the WT mice exposed to TNF-D or LPS, whereas such
LPS-treated Dusp6-/- mice (Fig. 24A). Moreover, ICAM-1 staining on the surface of alveolar walls was increased in the WT mice exposed to TNF-D or LPS, whereas such