OPEN
RIPK1/RIPK3 p omo es ascula pe meabili y o allow
umo cell ex a asa ion independen o i s nec op o ic
unc ion
Kay Hänggi
1
, Laza os Vasilikos
1
, Aida F ei e Valls
2,3
, Rosa io Ye bes
2
, Janin Knop
1
, Lisanne M Spilgies
1
, K is y Rieck
1
, T isha Mis a
1
,
John Be in
4
, Pe e J Gough
4
, Thomas Schmid
3
, Ca men Ruiz de Almodò a
2
and W Wei-Lynn Wong*
,1
Nec op osis is an in lamma o y o m o p og ammed cell dea h equi ing ecep o -in e ac ing p o ein kinase 1, 3 (RIPK1, RIPK3)
and mixed lineage kinase domain-like p o ein (MLKL). The kinase o RIPK3 phospho yla es MLKL causing MLKL o o m a po e-like
s uc u e, allowing in acellula con en s o elease and cell dea h o occu . Al e na i ely, RIPK1 and RIPK3 ha e been shown o
egula e cy okine p oduc ion di ec ly in luencing in lamma o y immune in il a es. Recen da a sugges ha nec op osis may
con ibu e o he malignan ans o ma ion o umo cells in i o and we asked whe he nec op osis may ha e a ole in he umo
mic oen i onmen al e ing he abili y o he umo o g ow o me as asize. To de e mine i nec op osis in he umo
mic oen i onmen could p omo e in lamma ion alone o by ini ia ing nec op osis and he eby in luencing g ow h o me as asis o
umo s, we u ilized a syngeneic umo model o me as asis. Loss o RIPK3 in he umo mic oen i onmen educed he numbe o
umo nodules in he lung by 46%. Loss o he kinase ac i i y in RIPK1, a membe o he nec osome also educed umo nodules in
he lung by 38%. Howe e , he loss o kinase ac i i y in RIPK3 o he loss o MLKL only ma ginally al e ed he abili y o umo cells
o o m in he lung. Using bone ma ow chime as, he dec ease in umo nodules in he Ripk3
−/−
appea ed o be due o he s omal
compa men a he han he hema opoie ic compa men . T ansmig a ion assays showed dec eased abili y o umo cells o
ansmig a e h ough he ascula endo helial laye , which co ela ed wi h dec eased pe meabili y in he Ripk3
−/−
mice a e umo
injec ion. In esponse o pe meabili y ac o s, such as ascula endo helial g ow h ac o , RIPK3 null endo helial cells showed
dec eased p38/HSP27 ac i a ion. Taken oge he , ou esul s sugges an al e na i e unc ion o RIPK1/RIPK3 in ascula
pe meabili y leading o dec eased numbe o me as asis.
Cell Dea h and Disease (2017) 8, e2588; doi:10.1038/cddis.2017.20; published online 2 Feb ua y 2017
Cell dea h is conside ed o be one o he hallma ks o
malignancy –ei he by up egula ion o an i-apop o ic genes o
down egula ion o silencing o p o-apop o ic genes.
1
Nec op-
osis is a p og ammed cell dea h pa hway igge ed by umo
nec osis ac o (TNF), Fas, Toll-like ecep o ligands and ype I
in e e on upon loss o inhibi ion o caspase-8.
2–6
Unde -
s anding whe he nec op osis is pa o he cell dea h hallma k
o malignancy has become an a ea o in ense esea ch.
Recep o -in e ac ing p o ein kinase 1 (RIPK1), a mul i-
unc ional p o ein ha con ains an N- e minal Se /Th kinase
domain, is known o be a c i ical egula o a he decision poin
o cy okine induced NF-kB ac i a ion o su i al o cell dea h
by ei he apop osis o nec op osis.
7
Nec op osis is igge ed
by RIPK1 in e ac ion wi h RIPK3, RIPK3 dime iza ion,
au ophospho yla ion o RIPK3 and subsequen phospho yla-
ion o mixed lineage kinase domain-like p o ein (MLKL).
8–10
Upon phospho yla ion, MLKL is belie ed o o m a po e
complex ha can comp omise cell memb ane in eg i y
11,12
esul ing in he elease in acellula con en s. The esul is an
in lamma o y esponse o he elease o dange -associa ed
molecula pa e ns and/o he di ec egula ion o cy okines by
RIPK1/RIPK3 dime iza ion.
13
Recen e idence sugges s RIPK3 is a umo supp esso .
RIPK3 has been ound o be epigene ically silenced in b eas
and panc ea ic cance issue
14,15
and in melanoma cell line.
16
In umo models, loss o RIPK3 aided a TAK1-induced
in lamma ion model o hepa oca cinogenesis,
17
whe eas loss
o RIPK3 in combina ion wi h in e nal andem duplica ion
mu a ions o FMS-like y osine kinase-3 led o an inc ease in
leukemia in i o, which was because o he absence o
in lammasome ac i a ion.
18
Exp ession o RIPK3 in pa ien
samples also has been associa ed wi h disease ou come. In
ce ical cance s, low RIPK3 p o ein exp ession in pa ien
biopsies co ela ed o a educed esponse o PolyIC-based
adju an immuno he apeu ic app oaches.
19
In pa ien s su e -
ing om in es inal bowel disease and p esen ing wi h colo -
ec al cance , loss o RIPK3 exp ession was de ec ed in
neoplas ic issue compa ed wi h non-neoplas ic and he loss
o exp ession was co ela i e o a poo p ognosis.
20
These
indings sugges a ole o RIPK3 and po en ially nec op osis
in umo igenesis, as well as p ognosis.
By con as , he dele ion o RIPK1, RIPK3 o MLKL in
b eas cance cell lines educed he umo igenic po en ial o
he cell o o m colonies and subcu aneous umo s in
1
Ins i u e o Expe imen al Immunology, Uni e si y o Zu ich, Zu ich, Swi ze land;
2
Biochemis y Cen e , Heidelbe g Uni e si y, Heidelbe g, Ge many;
3
Depa men o
Gene al, Visce al and T ansplan a ion Su ge y, Heidelbe g Uni e si y, Heidelbe g, Ge many and
4
Pa e n Recogni ion Recep o Disco e y Pe o mance Uni , Immuno-
In lamma ion The apeu ic A ea, GlaxoSmi hKline, College ille, PA, USA
*Co esponding au ho : WW-L Wong, Ins i u e o Expe imen al Immunology, Uni e si y o Zu ich, Win e hu e s asse 190 Y44 J55, Zu ich 8057, Swi ze land. Tel: +41 44
635 3720; Fax: +41 44 635 6883; E-mail: [email protected]
Recei ed 12.10.16; e ised 22.12.16; accep ed 03.1.17; Edi ed by J Zhang
Ci a ion: Cell Dea h and Disease (2017) 8, e2588; doi:10.1038/cddis.2017.20
O ficial jou nal o he Cell Dea h Di e en ia ion Associa ion
www.na u e.com/cddis
immunocomp omised mice.
21
To add o he complexi y,
inc eased exp ession in RIPK3, RIPK1 and MLKL was ound
in pa ien panc ea ic duc al adenoca cinoma (PDA) compa ed
wi h no mal issue and he loss o RIPK3 in a V12K-Ras-
induced PDA model inc eased su i al by educing he
exp ession o CXCL-1/Mincle pa hway and by al e ing he
p esence o umo -supp essi e immune cell.
22
In addi ion, ou
wo k and o he s show ha RIPK1/RIPK3 d i e cy okine
p oduc ion ei he in conjunc ion wi h ini ia ing nec op osis o
comple ely independen ly.
23,24
This makes i di icul o
in e p e in lamma ion d i en in i o models o umo igenesis
in RIPK3-de icien mice.
We sough o di e en ia e he ole o RIPK3 o nec op osis in
umo igenesis e sus he umo mic oen i onmen by u ilizing
syngeneic mouse umo models. These models can be used
o simula e me as asis, wha 90% o cance pa ien s die om
a he han he p ima y umo . Simila o ecen ly published
esul s, we ound he ole o RIPK3 in he umo mic oen i on-
men was no because o he al e ed immune esponse bu he
abili y o umo cells o ex a asa e in o he lung.
25
Howe e ,
we ound he kinase ac i i y o RIPK1 bu no RIPK3 kinase
ac i i y was impo an in he ex a asa ion s ep o me as asis.
Loss o MLKL showed a educ ion bu no signi ican numbe
o umo nodules o med in he lung compa ed wi h Ripk3
−/−
.
Ou da a shows ha in addi ion o he possibili y o ac i a ing
nec op osis h ough a dea h ecep o 6-media ed manne as
p e iously ound, RIPK3 has a c i ical ole as signaling pla o m
downs eam o s imuli p omo ing pe meabili y h ough p38/
HSP27. These da a suppo he ole o RIPK1/RIPK3 as a
signaling pla o m in p omo ing ascula pe meabili y equi ed
o umo cell ex a asa ion.
Resul s
Loss o RIPK3 esul s in dec eased umo nodules in he
lung. To de e mine whe he umo cells we e able o o m a
solid umo , Ripk3
−/−
mice we e injec ed subcu aneously wi h
B16-F10 cells. In ag eemen wi h ecen ly published da a,
22
solid umo s o med a he same a e and olume in bo h wild-
ype and Ripk3
−/−
mice (Figu e 1a). This sugges ed he loss
o RIPK3 in he umo mic oen i onmen did no a ec
es ablished umo cells om g owing. To de e mine i RIPK3
in he umo mic oen i onmen could ha e a ole in a umo
me as asis model, B16-F10 umo cells we e injec ed ail ein
in o wild- ype mice. Lung homogena es we e made 2, 6 and
12 h a e ail ein injec ion and RIPK3 le els we e assayed
0
50
100
150
200
250
Numbe o umo nodules
**
Ripk3-/-
W
0
10
20
30
Numbe o umo nodules
40
Ripk3-/-
W
W
Ripk3-/-
50
100
Pe cen age
0
75
25
Ripk3-/-
W
La ge
Medium
Small
Tumo olume (mm3)
200
300
400
500
100
0
02 141210864
days pos injec ion o B16-F10
Numbe o umo nodules
Ripk3
0
50
10
150
200
+/+ +/- -/-
*
RIPK3
β-Ac in
53
RIPK1
68
41
123456789
PBS 6h 12h
B16-F10
Mouse
MLKL
53
B16-F10 (h)
Change o RIPK3
p o ein exp ession
012
1.0
2.0
3.0
0
*
**
6
kDa
W
Ripk3-/-
*
Figu e 1 Loss o RIPK3 in he umo mic oen i onmen educes umo nodules in he lung in a umo model o ex a asa ion. (a)Wild- ype (W ) and Ripk3
−/−
mice we e
injec ed subcu aneously wi h 1 × 10
5
B16-F10 umo cells and umo size was measu ed o e ime (pooled da a o n=2 expe imen s; 5–7 mice used pe g oup in each
expe imen ). (b) RIPK3 bu no RIPK1 o MLKL p o ein le els we e ele a ed in o al lung lysa es o wild- ype mice 6 and 12 h a e injec ion o B16-F10 cells shown by
ep esen a i e immunoblo and co esponding quan i ica ion o pixel densi y om h ee mice pe imepoin . Fold change ela i e o un ea ed W con ol is shown om alues
no malized o ac in (s a is ical analysis by one-way ANOVA). (c) Tumo nodules we e coun ed 14 days pos ail ein injec ion (dpi) o B16-F10 cells (2 × 10
5
) in o Ripk3 li e ma e
con ol mice (le panel, n=1; 3–4 mice used pe g oup) and umo nodules we e coun ed 14 dpi o B16-F10 cells (2 × 10
5
) in o W and Ripk3
−/−
mice ( igh panel, pooled da a
shown o n=3 expe imen s; 3–5 mice used pe g oup in each expe imen ). (d) Co esponding size is shown o umo nodules classi ied mac oscopically in o la ge (41 mm),
medium (0.5 mm–1 mm) and small (o0.5 mm) shown (pooled, n=3), and co esponding ep esen a i e mac oscopic pic u es o he lungs we e shown. (e) Tumo nodules we e
coun ed 20 dpi o MC-38 cells (3 × 10
5
) in o W and Ripk3
−/−
mice (n=1; 6 mice used pe g oup; P=0.14). Each do ep esen s a mouse excep s a ed o he wise; s a is ical
analysis by - es excep s a ed o he wise. *Po0.05, **Po0.01
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
2
Cell Dea h and Disease
by immunoblo ing. A e 12 h, he le el o RIPK3 p o ein was
app oxima ely wo old highe (Figu e 1b), sugges ing a
po en ial ole o RIPK3 in umo o ma ion in he lung.
In e es ingly, no changes in RIPK1 o MLKL le els we e seen
(Figu e 1b). Using li e ma es, mice we e injec ed ail ein
wi h B16-F10 cells and 14 days pos injec ion (dpi) lungs
we e coun ed o umo nodules. The numbe o umo
nodules we e educed signi ican ly by app oxima ely 50%
(Figu es 1c; Po0.05). In e es ingly, Ripk3
+/−
mice showed
an in e media e pheno ype be ween wild- ype and Ripk3
−/−
.
Addi ional wild- ype and Ripk3
−/−
mice we e hen injec ed ail
ein wi h B16-F10 cells and 14 dpi, a e blinding, he lungs
we e coun ed o umo nodules. The e was a 46% (±7.6%)
dec ease in umo nodules in he RIPK3 null lungs compa ed
wi h wild- ype (Figu es 1d; Po0.01). The size o he umo
nodules was ca ego ized in o la ge, medium o small based
on wid h o he nodule om pic u es o he lungs and al hough
no signi ican , he e was a endency owa d smalle nodules
in Ripk3
−/−
mice (Figu e 1e). To ensu e he pheno ype was
no es ic ed o B16-F10 umo model, MC-38 colon
ca cinoma cells we e injec ed in a enously and umo
nodules in he lung we e coun ed 20 dpi. Consis en wi h
he B16-F10 umo model, a dec eased numbe o nodules in
he Ripk3
−/−
mice we e ound compa ed wi h wild- ype mice
(Figu es 1d; P=0.14) sugges ing he dec ease in umo
nodules in he Ripk3
−/−
was no dependen on he umo
cell ype.
Loss o RIPK1 kinase ac i i y bu no RIPK3 kinase
ac i i y educe umo nodule load in he lung. The
nec osome is composed o RIPK1 and RIPK3 and he kinase
ac i i y o RIPK1 and RIPK3 is belie ed o be impo an and/o
essen ial o ac i a ing nec op osis by phospho yla ing MLKL.
Cu en ly, MLKL is conside ed o be he essen ial and he only
W Ripk3K51A/K51A
200
100
0
150
50
Numbe o umo nodules
W
K45A/K45A
W /K51A
Ripk3
Numbe o umo nodules
600
200
0
400
K51A/K51A
0h
W
Ripk3-/-
Lung: DAPI/TUNEL
6h 12h
Posi i e C l.
B16-F10 (hou s)
W
Numbe o umo nodules
W Ripk3
-/-
Mlkl
-/-
Mlkl-/-
Ripk3-/-
0
100
200
300
400
500
**
*
**
300
100
0
200
Numbe o umo nodules
Ripk1K45A/K45A
W
Ripk1
Figu e 2 RIPK1 kinase ac i i y bu no RIPK3 kinase ac i i y o MLKL in he umo mic oen i onmen is needed o umo nodule o ma ion in he lung. Tumo nodules we e
coun ed 14 dpi o B16-F10 cells in o (a)W and Ripk1
K45A/K45A
mice (pooled da a o n=2 expe imen s; 4–7 mice we e used pe g oup in each expe imen ) and ep esen a i e
pic u es o lungs a e shown, (b)W and Ripk3
K51A/K51A
mice (pooled da a o n=2 expe imen s; 5–8 mice we e used pe g oup in each expe imen ) and (c) in o Ripk3
W /K51A
and
Ripk3
K51A/K51A
mice (pooled da a o n=4 expe imen s; 3–4 mice we e used pe g oup in each expe imen ). (d) Tumo nodules we e coun ed 14 dpi o B16-F10 cells in o W ,
Mlkl
−/−
and Ripk3
−/−
mice and ep esen a i e pic u es a e shown (da a om o n=2 expe imen s; 2–6 mice we e used pe g oup in each expe imen ). (e) Pa a in-embedded
lung sec ions om mice injec ed ail ein wi h B16-F10 a indica ed imepoin s we e s ained o TUNEL (FITC) and DAPI. In es ine o ciap1
−/−
mice injec ed wi h TNF was used as
a posi i e con ol and ed a ows indica e TUNEL-posi i e cells. Each do ep esen s a mouse; s a is ical analysis by - es ; all scale ba s, 50 μm. **Po0.01
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
3
Cell Dea h and Disease
known e ec o p o ein o nec op osis.
10,11
To de e mine i
p o eins composing he nec osome we e impo an in he
educ ion o umo nodules in he lung, RIPK1 kinase dead
(K45A) mice we e injec ed wi h B16-F10 cells ( ail i. .) and
a e 14 dpi, he lungs we e ha es ed, blinded and umo
nodules we e coun ed. Tumo nodules in he Ripk1
K45A/K45A
mice we e dec eased by 38.3% (±8.5%) compa ed o wild-
ype mice (Figu es 2a; Po0.05). Nex , RIPK3 kinase dead
(K51A) mice we e injec ed ail ein wi h B16-F10 cells and
again, a e 14 dpi, lungs we e blinded and nodules we e
coun ed in he lung. No di e ence in he numbe o umo
nodules we e ound (Figu e 2b). As p o ein le els o he kinase
dead RIPK3 K51A mu an ha e been shown o be dec eased
because o p o ein ins abili y, we assessed he le els o p o ein
p esen in lung homogena es and ound he exp ession o he
RIPK3 mu an o be signi ican ly dec eased (Supplemen a y
Figu e 1a).
26
To con ol o he dec eased change in
exp ession, we compa ed he Ripk3
w /K51A
wi h he
Ripk3
K51A/K51A
mice. No di e ence in he numbe o umo
nodules we e o med in he lung (Figu e 2c). To e i y ha loss
o RIPK3 kinase ac i i y was able o block nec op osis, isola ed
endo helial cells (ECs) om he lungs o Ripk3
K51A/K51A
mice
we e es ed wi h ypical ac i a o s o nec op osis, TNF, zVAD
and/o Smac mime ics (Supplemen a y Figu e 1b). As
expec ed, compa ed wi h wild- ype he RIPK3-de icien and
Ripk3
K51A/K51A
-exp essing cells we e esis an o nec op osis
caused by he combina ion o TNF, zVAD and Smac mime ic
ea men . In e es ingly, al hough he kinase ac i i y o RIPK1
was equi ed, i appea ed he loss o kinase ac i i y om
RIPK3 was no equi ed in umo o ma ion in he lung. Finally,
o assess i nec op osis was indeed he eason o he
dec ease in umo nodules in he lung, we injec ed B16-F10
cells ail ein in o wild- ype, Mlkl
−/−
and Ripk3
−/−
mice. Only a
sligh dec ease in he numbe o umo nodules was seen in
Mlkl
−/−
mice, whe eas he educ ion o umo nodules was
seen again in Ripk3
−/−
mice (Figu e 2d). The combined esul s
om he kinase dead RIPK3 and MLKL-de icien mice
sugges ed an al e na i e ole o RIPK3 in ol ing he kinase
ac i i y o RIPK1.
As he loss o RIPK3 in Vucu e al.
17
led o an inc ease in
caspase-8-dependen apop osis in a TAK1 in lamma o y-
d i en hepa oca cinoma, we assessed o he p esence o
cell dea h in he lung a e B16-F10 cells we e injec ed in he
mice. TUNEL s aining was used o de e mine cell dea h, as
his echnique will pick up bo h apop osis o nec osis/
nec op osis.
27
Fo his, lung sec ions o mice injec ed wi h
B16-F10 a di e en imepoin s (0, 6 and 12 h) we e analyzed
o TUNEL-posi i e cells. The posi i e con ol in es inal
150
100
50
0
CD45.1 BM
ecipien
Numbe o umo
nodules
To al lux
[p/s] (x104)
Time a e
B16-F10 luc (h) 61224
0
10
15 W
Ripk3-/-
W
Ripk3-/-
B16-F10 (h)
06240624
0
0.6
0.4
0.2
0.8
1
Neu ophils
To al cell coun
pe g lung (x107)
To al cell coun
pe g lung (x107)
B16-F10 (h)
0.25
0
0.50
0.75
1
06240624
In lamma o y
Monocy es
[pg/ml]
0
200
400
600
800
CCL-2/MCP-1
B16-F10 (h)
06240624
[pg/ml]
0
100
200
300
400
CXCL-1/G o-α
B16-F10 (h)
06240624
W
Ripk3-/-
CD45.2
BM ecipien
Ripk3
-/-
Ripk3-/- BM dono
CD45.1 BM dono
Numbe o umo
nodules
0
100
200
300
5
0
15
20
10
06240
6
24
B16-F10 (h)
[pg/ml]
IL-1β
***
***
*** ***
*** ***
Figu e 3 Loss o RIPK3 in he mic oen i onmen does no al e ea ly immune and cy okine esponse upon B16-F10 injec ion and bone ma ow econs i u ion shows he
s omal compa men o be esponsible. (a) B16-F10 luci e ase cells home o he lung in bo h W and Ripk3
−/−
mice measu ed by IVIS luminescen imaging ( ep esen a i e
shown, 4–5 mice pe g oup, n=2). (b) Cy okine le els in o al lung lysa es we e measu ed by mul iplex bead assay and le els o CCL-2, CXCL-1, IL-1βand TWEAK a e shown
(n=2). (c) Co esponding immune cell in il a ion analysis by low cy ome y shows in lamma o y monocy es (SiglecF
−
Ly6G
−
Ly6C
hi
CCR2
+
) and neu ophil g anulocy e
(SiglecF
−
CD11b
+
Ly6G
+
) popula ions. Popula ions we e p e-ga ed on single s, li e cells, Te 119
−
and CD45
+
cells. (d) Tumo nodules we e coun ed 14 dpi o B16-F10 in o
CD45.1 ecipien mice econs i u ed wi h CD45.2 W and CD45.2 Ripk3
−/−
bone ma ow cells (n=3) and CD45.2 o Ripk3
−/−
ecipien mice econs i u ed wi h CD45.1 o
Ripk3
−/−
bone ma ow cells (n=2). Each do ep esen s a mouse; s a is ical analysis done by one-way ANOVA and Bon e oni pos - es . ***Po0.001
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
4
Cell Dea h and Disease
epi helial cells om wild- ype mice injec ed wi h TNF we e
s ained posi i e o TUNEL, whe eas lungs om wild- ype o
Ripk3
−/−
mice did no s ain TUNEL posi i e (Figu e 2e). In
addi ion, li le o no caspase-3 ac i i y was measu ed in lung
homogena es o mice injec ed wi h B16-F10 a di e en
imepoin s (Supplemen a y Figu e 1c). This sugges s he
kinase ac i i y o RIPK3 and he e o e he abili y o
nec op osis o occu in he umo mic oen i onmen did no
al e umo nodule o ma ion.
Tumo cells home o he lung in bo h wild- ype and
Ripk3
−/−
mice wi h simila immune cell in il a es. To
de e mine he addi ional ole RIPK3 may ha e in me as asis,
we nex assessed whe he he B16-F10 cells we e able o
home o he lung in he Ripk3
−/−
mice. B16-F10 luci e ase
cells we e injec ed and moni o ed a e injec ion o mo e-
men o he lung o e ime. Equi alen luminescence was
measu ed be ween wild- ype and Ripk3
−/−
mice wi hin he
i s 24 h (Figu e 3a). This sugges ed he B16-F10 cells we e
su i ing and homing o he lung he same in he Ripk3
−/−
mice compa ed wi h wild- ype.
Chemokines such as C-C mo i chemokine ligand 2 (CCL-2)
ha e been shown o induce signaling o ec ui in lamma o y
monocy es, as well as induce endo helial ex a asa ion.
28,29
To de e mine i RIPK3 was c i ical in p oducing p o-
in lamma o y media o s o ec ui myeloid in il a es pos
injec ion o B16-F10 cells, cy okines/chemokines we e ana-
lyzed in lungs o wild- ype and Ripk3
−/−
mice. We examined
whe he chemokines/cy okines we e al e ed in Ripk3
−/−
mice
compa ed wi h wild- ype mice a e B16-F10 injec ion using
lung homogena es. An inc ease in CCL-2, CXCL-1 and IL-1β
was seen in bo h wild- ype and Ripk3
−/−
mice (Figu e 3b). TNF
was below he limi o de ec ion. This sugges s he abili y o
p oduce cy okines o a ac monocy es was unc ional in he
lung o Ripk3
−/−
mice. We hen assessed he immune
in il a es in he lung by low cy ome y. A e 6 h pos B16-
F10 injec ion, a signi ican inc ease in he numbe o
in lamma o y monocy es (Ly6C
hi
CCR2
+
) we e seen in
Ripk3
−/−
mice compa ed wi h wild- ype (Figu e 3c). Simila ly,
CD11b
+
dend i ic cells inc eased, whe eas he CD103
+
dend i ic cell popula ion dec eased in wild- ype and Ripk3
−/−
lungs a e B16-F10 injec ion (Supplemen a y Figu e 2).
Na u al kille T cells also esponded in a simila manne in
bo h wild- ype and Ripk3
−/−
lungs a e B16-F10 injec ion
(Supplemen a y Figu e 2).
Loss o RIPK3 in he s oma a ec s umo nodule
o ma ion in he lung. To de e mine he po en ial o he
hema opoie ic o s omal compa men o ha e a ole in he
educ ion in umo nodules ound in Ripk3
−/−
lungs, we
pe o med bone ma ow chime ic expe imen s. CD45.1
ecipien mice we e i adia ed and econs i u ed wi h ei he
wild- ype o RIPK3-de icien bone ma ow. Only mice wi h
econs i u ion e iciency 495% we e used o B16-F10
injec ion (da a no shown). A e 14 dpi, umo nodules in
he mice wi h RIPK3-de icien hema opoie ic compa men
Pe cen age o
EC popula ion
Cell Dea h assay:
Co-cul u e B16-F10/Endo helial cells (8h)
Cell Dea h
o EC
Cell Dea h
o B16-F10
0
25
50
75
100 Annexin V+
Li e-Dead+/Annexin V-
Li e-Dead+/Annexin V+
W Ripk3
-/- Mlkl-/-
nd.
-
B16-F10:
TNF/SM:
zVAD:
-
+++ ++ ++
-- --- -
++
-
--- --- ---
++++++
+++
--- --- ------
To al B16-F10 cells
ansmig a ed pe inse
0
50
100
150
200
+W +Ripk3
-/-
Monocy es: -
EC
W EC
Ripk3K51A/K51A
Ripk3-/- EC
Figu e 4 RIPK3 p omo es umo cell ansendo helial mig a ion independen o i s abili y o block nec op osis. (a) Boyden chambe ansendo helial mig a ion assay using
B16-F10 cells and p ima y CD31
+
ECs isola ed om lungs o W ,Ripk3
−/−
o Ripk3
K51A/K51A
mice. P ima y monocy es we e isola ed on he day o seeding B16-F10 cells and
20 h la e , answells we e ixed, s ained and imaged. Image analysis was blinded and umo cells we e coun ed pe ield o iew. The da a a e pooled om n=2 expe imen s; 1–3
answell inse s we e assayed pe g oup in each expe imen and 10–17 ields pe iew we e analyzed pe answell inse . To al numbe o B16-F10 ansmig a ed cells pe
answell inse is shown. Each do ep esen s an inse . S a is ical analysis by one-way ANOVA and Bon e oni pos - es . (b) EC monolaye s om W ,Ripk3
−/−
and Mlkl
−/−
mice
we e co-incuba ed wi h PKH26
+
B16-F10 (+ =3×10
4
;++=9×10
4
) cells and assayed o cell dea h 8 h la e by low cy ome y (n=3; nd =no de e mined). Cells we e s ained
o annexin Vand PI (li e/dead) o iden i y ea ly apop o ic cells (Li e/Dead
−
annexin V
+
) and dead cells (annexin V
+
Li e/Dead
+
annexin V
+
o Li e/Dead
+
annexin V
−
). S a is ical
analysis done by one-way ANOVA and Bon e oni pos - es
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
5
Cell Dea h and Disease

we e simila in numbe compa ed wi h mice wi h wild- ype
hema opoie ic compa men (Figu e 3d, le panel). This
sugges s ha he hema opoie ic compa men does no
in luence he o al numbe o umo nodules in he lung. We
hen i adia ed Ripk3
−/−
and CD45.2 mice and econs i u ed
he bone ma ow wi h wild- ype o RIPK3-de icien bone
ma ow. A e 14 dpi, umo nodules we e coun ed. In
compa ison wi h wild- ype mice econs i u ed wi h Ripk3
−/−
bone ma ow, he e was a dec easing amoun o umo
nodules in he Ripk3
−/−
mice econs i u ed wi h wild- ype
(17.6% ±23) o Ripk3
−/−
bone ma ow (38.6% ±22.6),
espec i ely (Figu e 3d, igh panel). Taken oge he , he da a
sugges ed ha he educ ion in umo nodules in he Ripk3
−/−
mice was because o he s omal compa men and no
because o he hema opoie ic compa men .
RIPK3 p omo es umo cell ex a asa ion independen o
i s kinase ac i i y. To o m umo s in he lung, he umo
cells mus ex a asa e pas he ascula endo helial ba ie .
To de e mine i he e was a umo cell ansmig a ion de ec ,
B16-F10
E ans blue pe
lung [ug/ml]
-+-+
0
5
10
15
20
W
Ripk3-/-
FGF-basic
0 6 24 0 6 24
B16-F10 (h)
0.5
0
1.0
2.0
1.5
[ng/ml]
20
0
60
80
40
[pg/ml]
B16-F10 (h)
TWEAK
**
***
**
06240
6
24 0 6 24 0
6
24
B16-F10 (h)
2
0
6
4
[ng/ml]
MMP-9
***
** **
W
Ripk3-/-
0.5
0
1.5
2.5
2.0
1.0
[ng/ml]
06240
VEGF-A
6
24
B16-F10 (h)
VEGF-A
164
To al B16-F10 cells
ansmig a ed pe inse
100
200
300
-+-+
VEGF-A164
++
Fold change (RFU) o
dex an-FITC
0
0.50
1.00
1.50
W
Ripk3-/-
--
**
0.25
1.25
0.75
W
Ripk3-/-
siC l
siRipk1
VEGF-A
164
-+-+
Sp ou leng h
(Fold change)
0
1.0
2.0
1.5
2.5
0.5
siC l
siRipk3
**
ns
**
*
+VEGF-A mock
siC l siRipk3
+VEGF-A mock
siC l siRipk1
0
0.5
1.5
1.0
2.5
2.0
0
0.5
1.5
1.0
2.0
Sp ou leng h
(Fold change)
DMSO
Nec-1
VEGF-A
164
-+-+
***
-+-+
VEGF-A
164
Figu e 5 RIPK1 kinase ac i i y and RIPK3 p omo es VEGF-A-dependen ascula pe meabili y and sp ou ing. (a) Le els o angiogenic- and emodeling ac o s in o al lung
lysa es we e measu ed by mul iplex bead assay and le els o VEGF-A, FGF-basic, TWEAK and MMP-9 a e shown (da a pooled o n=2 expe imen ; S.E.M.; 4–6 mice we e used
pe g oup; each do ep esen s a mouse). (b) E ans blue blood essel pe meabili y assay o W and Ripk3
−/−
lungs 20 h a e B16-F10 umo cell injec ion i. . (pooled da a o
n=2 expe imen s, 4–5 mice we e used pe g oup in each expe imen ; each do ep esen s a mouse). (c) P ima y lung endo helial monolaye on answell inse s we e ea ed
wi h VEGF-A
164
(100 ng/ml) o 4 h and dex an-FITC pe meabili y assay was pe o med. Da a show ela i e luo escence uni (RFU) o dex an-FITC ha passed EC ba ie
no malized o un ea ed wild- ype (pooled da a o n=5 expe imen s; S.E.M.; 2–3 answell inse s we e used pe g oup in each expe imen ; each do ep esen s a answell
inse ). (d) Subsequen B16-F10 ansendo helial mig a ion assay was pe o med a e dex an-FITC pe meabili y assay and 20 h a e B16-F10 (PKH26
+
) cells we e seeded in o
inse s, answells we e ixed, s ained and imaged. Image analysis was blinded and umo cells we e coun ed pe ield o iew. The da a a e pooled om n=2 expe imen s; 2–3
answell inse s we e assayed pe g oup in each expe imen and 10–17 ields pe iew we e analyzed pe answell inse . S.E.M. and o al numbe o B16-F10 ansmig a ed
cells pe answell inse is shown. Each do ep esen s an inse . S a is ical analysis by one-way ANOVA and Bon e oni pos - es . (e) HUVECs we e ans ec ed wi h siRNAs
a ge ing RIPK3 o ( ) RIPK1 (le panel) o ea men wi h nec os a in-1 (Nec-1, igh panel) and sp ou ing assay was pe o med. Rep esen a i e luo escnece mic oscopy images
a e shown. Da a om n=3 expe imen s; S.E.M.; s a is ical analysis by - es ; all scale ba s, 100 μm. *Po0.05, **Po0.01, ***Po0.001
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
6
Cell Dea h and Disease
we isola ed CD31
+
cells om he lungs o wild- ype, Ripk3
−/−
o Ripk3
K51A/K51A
mice and pe o med ansmig a ion assays.
A dec eased numbe o B16-F10 cells ansmig a ed h ough
Ripk3
−/−
ECs compa ed wi h wild- ype o Ripk3
K51A/K51A
ECs
(Figu e 4a). We also de e mined i wild- ype monocy es could
acili a e inc eased umo cell ansmig a ion. Two old mo e
B16-F10 cells ansmig a ed h ough wild- ype endo helial
monolaye in he p esence o monocy es (Figu e 4a).
S ikingly, nei he he addi ion o wild- ype no Ripk3
−/−
monocy es inc eased he abili y o B16-F10 cells o ansmi-
g a e h ough a Ripk3
−/−
endo helial monolaye . By con as ,
he addi ion o Ripk3
−/−
monocy es did enhance B16-F10
ansmig a ion h ough wild- ype endo helial monolaye .
Mo eo e , he Ripk3
K51A/K51A
ECs allowed B16-F10 cells o
ansmig a e h ough simila o wild- ype wi h ei he wild- ype
o Ripk3
−/−
monocy es. These esul s suppo he idea ha
he hema opoie ic compa men does no ha e a ole in he
loss o umo o ma ion in he Ripk3
−/−
mice simila o indings
in Figu e 3d and u he suppo s obse a ion ha he kinase
ac i i y o RIPK3 in he ECs is no equi ed o he abili y o
umo cells ansmig a e and umo nodules o o m in he
lung (Figu es 2b and c).
Cell dea h in p ima y ECs upon co-incuba ion o umo
cells is independen o RIPK3 o MLKL loss. To de e mine
i nec op osis occu ed o allow o ansmig a ion, we
isola ed CD31
+
cells om he lungs o wild- ype, Ripk3
−/−
and Mlkl
−/−
mice. P ima y lung EC monolaye s we e
incuba ed wi h ei he 3 × 10
4
o 9 × 10
4
B16-F10 cells. Cell
dea h was assessed 8 h la e by low cy ome y as ou in i o
imaging sugges ed he umo cells homed o he lung wi hin
6 h (Figu e 4b). B16-F10 cells we e s ained o iden i ica ion
wi h he memb ane dye PKH26 and we e emo ed om he
popula ion o cell dea h assessmen o ECs. Cells we e hen
s ained using Annexin V FITC and p opidium iodide (PI) o
disc imina e li e, ea ly apop o ic, o la e apop o ic o nec o ic
cells. The o e all pe cen age o dead wild- ype ECs (ei he
ea ly apop o ic o la e apop o ic/nec o ic cells) inc eased by
1–3% wi h he inc easing amoun o umo cells, whe eas
Ripk3
−/−
o Mlkl
−/−
ECs only showed an inc ease in 1–2%
dead cells upon 9 × 10
4
umo cell addi ion. Howe e , he
pe cen age o la e apop o ic o nec o ic ECs inc eased upon
umo cell addi ion was independen o he geno ype o he
ECs. This sugges ed RIPK3 may ha e an addi ional ole o
ha o he p oposed umo cell induced nec op osis as
p oposed.
25
Loss o RIPK1 kinase ac i i y o comple e loss o RIPK3
al e s pe meabili y and essel sp ou ing in esponse o
VEGF-A. Vascula pe meabili y has been p oposed as a
model o umo me as asis whe e ascula endo helial
g ow h ac o (VEGF), ib oblas g ow h ac o (FGF) o
CCL-2 may in luence pe meabili y.
30–32
Recen da a show
ha ac i a ion o VEGF leads o c-S c ac i a ion and change
in VE-cadhe in dis ibu ion allowing o inc eased ascula
pe meabili y and subsequen umo me as asis.
33
To de e -
mine i ac o s in ol ed in ascula pe meabili y was a ec ed,
we assayed o VEGF-A, FGF-b o TWEAK in lung homo-
gena es o wild- ype and Ripk3
−/−
mice injec ed wi h B16-F10
cells. In e es ingly, he loss o RIPK3 showed no inc ease in
VEGF-A a 6 h a e B16-F10 cell injec ion compa ed wi h
wild- ype (Figu e 5a). A dec ease in me allop o einase 9
(MMP-9) was de ec ed a 24 h a e B16-F10 cell injec ion in
he lungs o Ripk3
−/−
compa ed wi h wild- ype (Figu e 5a).
VEGF-A and FGF-b was also de ec ed in supe na an o
ansmig a ion assays o ei he wild- ype o Ripk3
−/−
ECs wi h
monocy es and umo cells a e 20 h. MMP-9 was dec eased
in he supe na an o ansmig a ion assays o Ripk3
−/−
ECs
compa ed wi h wild- ype (Supplemen a y Figu e 3a). These
da a sugges ed ha ac o s in ol ed in ascula pe meabili y
we e p esen wi h he excep ion o VEGF, which was educed
a 6 h in he lung o Ripk3
−/−
a e B16-F10 cell injec ion. We
he e o e assessed i pulmona y ascula pe meabili y a e
umo injec ion was a ec ed in he Ripk3
−/−
mice using he
E ans blue pe meabili y assay. E ans blue dye binds o
albumin and up ake o he dye in o gans is a measu e o
issue pe meabili y. E ans blue dye was injec ed in o wild-
ype and Ripk3
−/−
mice 20 h pos injec ion o B16-F10 cells.
The amoun o E ans blue dye in wild- ype lung homogena e
inc eased sligh ly when B16-F10 cells we e injec ed, bu no
inc ease in E ans blue dye was obse ed in Ripk3
−/−
mice
(Figu e 5b).
To u he assess he issue o ascula pe meabili y, we
ea ed p ima y ECs o 4 h wi h VEGF-A
164
( e e ed he ea e
as VEGF-A) and assayed o endo helial leakiness using
dex an-FITC. Al hough a signi ican inc ease in app oxi-
ma ely 15% o dex an-FITC de ec ed a e ea ing wild- ype
ECs wi h VEGF-A, he e was no inc ease in pe meabili y seen
in ea ed Ripk3
−/−
ECs (Figu e 5c). We assessed whe he he
loss o RIPK1 kinase ac i i y al e ed endo helial ba ie
pe meabili y in esponse VEGF-A and ound he loss o RIPK1
kinase ac i i y led o no inc ease in endo helial pe meabili y in
esponse o VEGF-A (Supplemen a y Figu e 3b). We hen
asked i p e-s imula ion wi h VEGF-A could inc ease he
numbe o B16-F10 cells mig a ing h ough an EC ba ie .
Al hough he numbe o B16-F10 cells inc eased by
app oxima ely 25% o a wild- ype endo helial ba ie , no
inc ease in B16-F10 umo cells was seen o a Ripk3
−/−
endo helial ba ie (Figu e 5d). To de e mine i o he unc ional
oles o VEGF-A we e also comp omised by he lack o RIPK1
o RIPK3, we assessed essel sp ou ing. In e es ingly, when
RIPK3 was down egula ed in human umbilical ein ECs
(HUVECs), hese cells showed a wo old inc ease in basal
essel sp ou ing. Impo an ly, he addi ion o VEGF-A induced
he ou g ow h o essel sp ou s in wild- ype HUVECs
(siRNA con ol) bu no in HUVECs whe e RIPK3 was silenced
by siRNA (Figu e 5e). Consis en wi h p e ious epo s,
HUVECs did no exp ess RIPK3 p o ein a high le els and
e en a sligh educ ion in RIPK3 le els led o a di e ence in
VEGF-A esponse in essel sp ou ing (Supplemen a y
Figu e 3c). We hen assessed he ole o RIPK1 in essel
sp ou ing and simila ly ound ha he loss o RIPK1 by siRNA
o loss o kinase ac i i y o RIPK1 by use o Nec-1 led o a
dec ease in essel sp ou ing in esponse o VEGF-A
compa ed wi h wild- ype HUVECs (Figu e 5 ). Taken oge he ,
hese esul s show ha RIPK1 o RIPK3-de icien ECs
ail o espond o umo -induced ascula pe meabili y o
VEGF-A-induced pe meabili y and angiogenesis, esul ing in
a de ec in B16-F10 ansendo helial mig a ion.
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
7
Cell Dea h and Disease
Loss o RIPK1 kinase ac i i y and comple e loss o RIPK3
in ECs dampens p38/HSP27 signaling in esponse o
pe meabili y ac o s. To de e mine he po en ial mechanism
o how RIPK3 may a ec ascula pe meabili y, we examined
downs eam signaling in esponse o VEGF-A, VEGF-B o
FGF-b. In esponse o VEGF-A, VEGF-B o FGF-b, wild- ype
ECs showed an inc ease in phospho-p38 ollowed by he
downs eam phospho yla ion o HSP27 p o ein, whe eas
phospho-ERK1/2 dec eased in signal (Figu e 6a and
Supplemen a y Figu e 4a). The in e se s a us o phospho-
p38 and phospho-ERK1/2 in esponse o VEGF-A has been
p e iously shown in HUVECs whe e ac i a ion o p38 leads o
dec eased le els o ERK1/2 ac i a ion.
34
By con as , RIPK3
null ECs s imula ed wi h ei he VEGF-A, VEGF-B o FGF-b
esul ed in educed phospho-p38 ac i a ion compa ed wi h
wild- ype. Mo e s ikingly, he downs eam phospho yla ed
p o ein, HSP27, was u he educed (Figu e 6a and
Supplemen a y Figu e 4a). Con e sely, phospho yla ed
ERK1/2 in he Ripk3
−/−
ECs was inc eased compa ed wi h
wild- ype e lec ing he dampened p38 ac i a ion in he RIPK3
knockou s. In e es ingly, Ripk3
−/−
ECs wi hou s imula ion,
had an inc eased basal le el o phospho-p38 p esen
compa ed wi h wild- ype cells. No majo di e ences in he
ac i a ion o p38/HSP27 be ween wild- ype and Ripk3
−/−
ECs
was seen when ea ed wi h TNF (Supplemen a y Figu e 4a
and 4b). The inhibi ion o RIPK1 ac i i y by nec os a in-1
educed HSP27 phospho yla ion simila ly o he loss o
RIPK3 when ECs we e ea ed wi h VEGF-A (Figu e 6b). To
de e mine i he signaling di e ences occu ed in i o a e
umo cell injec ion, we injec ed mice wi h umo cells and
a e 2, 6 and 12 h, he lungs we e homogenized o p o ein
lysa es. Consis en wi h ou immo alized ECs, basal le els o
phospho-p38 occu ed in he Ripk3
−/−
lungs compa ed wi h
wild- ype. A 2 h, phospho-p38 was up egula ed in wild- ype
mice bu no in Ripk3
−/−
mice (Figu e 6c). Taken oge he , ou
da a show ha RIPK3 posi i ely egula es he ac i a ion o
p38 and HSP27 upon pe meabili y ac o ea men (VEGF-A,
VEGF-B, FGF-b) and he e o e p omo ing pe meabili y in he
umo cells o me as asize (Figu e 6d).
Discussion
The cu en dogma is ha RIPK3 causes in lamma ion
because o he induc ion o nec op osis. Howe e , ou
Time (min)
41
41
41
41
DMSO Nec-1
23
23
phospho-p38 (S)
p38
phospho-HSP27 (L)
0515
30
60
120
VEGF-A
0515
30
60
120
phospho-p38 (L)
β-Ac in
phospho-HSP27 (S)
kDa
kDa
VEGF-B
phospho-p38
p38
phospho-ERK1/2
ERK1/2
W Ripk3
-/-
phospho-HSP27
PARP
41
125
41
41
41
32
93
41
53
RIPK3
β-Ac in
41
53
41
W Ripk3
-/-
0261202612B16-F10
phospho-p38
RIPK3
Time (hou s)
β-Ac in
kDa
0515
30
60
120
0515
30
60
120
Time (min)
Figu e 6 RIPK1 kinase ac i i y and RIPK3 p omo es VEGF-dependen ac i a ion o p38/HSP27 MAP kinase signaling axis. (a) SV40 la ge T immo alized ECs isola ed om
W o Ripk3
−/−
mice we e ea ed wi h VEGF-B (10 ng/ml) and assayed o signaling by immunoblo analysis as indica ed. Rep esen a i e immunoblo is shown o n=3
expe imen s. (b) SV40 la ge T immo alized ECs isola ed om W mice we e ea ed wi h DMSO o nec os a in-1 (Nec-1) 1 h be o e VEGF-A (10 ng/ml) ea men and assayed
o signaling by immunoblo analysis as indica ed. (c) Tumo cells we e injec ed in o mice a indica ed imepoin s and lungs we e lysed and immunoblo ed as indica ed. (S =sho
exposu e, L =long exposu e). (d) P oposed model: RIPK1 kinase ac i i y and RIPK3 p omo e phospho yla ion o HSP27 in ECs upon pe meabili y ac o s imula ion
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
8
Cell Dea h and Disease
p e ious esul s sugges ha RIPK3 may al e cy okine
elease, pa icula ly TNF in esponse o loss o IAPs.
24
We
he e o e se ou o de e mine i RIPK3 may ha e a ole in he
umo mic oen i onmen by p omo ing cy okine p oduc ion o
causing nec op osis leading o umo p og ession. Ou esul s
show RIPK3 did no al e umo g ow h o es ablished umo s.
In addi ion, changes in cy okine p oduc ion in Ripk3
−/−
mice
compa ed wi h wild- ype mice a e umo injec ion we e
minimal. The loss o kinase ac i i y o RIPK1 and he comple e
loss o RIPK3 we e ound o be impo an in he abili y o he
umo cells o o m lung nodules. Su p isingly, we did no ind
Ripk3
K51A/K51A
o Mlkl
−/−
mice exhibi ed a educ ion o umo
nodules in he lung simila o Ripk3
−/−
mice. Bone ma ow
chime as showed he in ol emen o he s omal bu no he
hema opoie ic compa men was c i ical o he loss o umo
nodules in he lung in Ripk3
−/−
mice. We hen ound he abili y
o umo cells o ansmig a e h ough Ripk3
−/−
endo helial
ba ie o be de icien and ha cell dea h in he EC laye was
independen o he geno ype when umo cells we e co-
incuba ed. We unexpec edly ound he loss o RIPK3 o he
kinase ac i i y o RIPK1 in he endo helial compa men we e
de ec i e in p38/HSP27 signaling in esponse o pe meabili y
ac o s such as VEGF-A. This lack o esponse esul ed in
dec eased pe meabili y and essel sp ou ing in esponse o
VEGF-A. These esul s sugges he ole o he nec osome is
minimal in he ex a asa ion p ocess o umo cells and
p esen s a no el ole o RIPK3 as a signaling pla o m
downs eam o angiogenic ac o s such as VEGF and FGF.
The cu en di icul y in assessing he ole o nec op osis in
ela ed pa hologies o human in lamma o y disease is he
absence o iden i ying ma ke s. The use o RIPK3 gene ic
dele ion, pha macological inhibi ion o RIPK1 and/o co ela-
i e e idence o RIPK3 p o ein o RNA up egula ion has been
used o associa e he disease s a e wi h nec op osis.
35,36
Howe e , as RIPK1 has been shown o ha e al e na i e
unc ion in egula ing apop osis and nec op osis
37–39
and also
di ec ly in luencing in lamma ion,
24,40,41
u he downs eam
e ec o s o nec op osis such as he use o Mlkl
−/−
mice a e
equi ed o p o e he ole o nec op osis.
42
In ou expe imen s,
he numbe o pulmona y umo nodules in he Mlkl
−/−
mice
we e sligh ly educed compa ed wi h wild- ype mice bu no o
he same ex en as in he Ripk3
−/−
mice. I is unclea wha may
be he eason o he di e ence in he umo nodule numbe s in
he MLKL knock-ou mice we ob ained compa ed wi h S ilic
e al.
25
, excep o he sou ce o he MLKL null mice. Ou da a
also show ha he kinase ac i i y o RIPK1 has a ole in he
abili y o umo nodules o o m in he lung. Ac i a ion o
VEGFR2 by VEGF-A o VEGF-B causes phospho yla ion o
c-S c and has been linked o inc eased pe meabili y by
al e ing adhe ens junc ion o ma ion, VE-cadhe in.
33,43
c-S c
has been shown o inhibi Fas-induced caspase-8 ac i a ion
by phospho yla ion a Ty 380.
44
Fu he s udies a e equi ed o
de e mine i RIPK1 and RIPK3 al e umo cell ex a asa ion
as a complex wi h o wi hou caspase-8 in ol emen o i hese
p o eins egula e di e en pa hways o esul in ascula
pe meabili y.
Vascula pe meabili y equi es issue emodeling simila o
wound healing. Impai ed wound healing was ound in Ripk3
−/−
mice compa ed wi h wild- ype mice, cha ac e ized by a
dec eased MMP-9 p o ein exp ession and delayed CD31
+
s aining and VEGF p oduc ion.
45
O he cy okines and g ow h
ac o s we assayed, only MMP-9 was educed a 24 h in he
Ripk3
−/−
mice compa ed wi h wild- ype upon umo cell
injec ion. The dec ease in MMP-9 p o ein le els co espond
o an inc ease in neu ophil in il a ion o he lung a 24 h in he
Ripk3
−/−
, sugges ing an a emp o he immune sys em o
compensa e o he lack o esponse in he endo helial
compa men . The abili y o Ripk3
−/−
monocy es o enhance
ansmig a ion o umo cells h ough a wild- ype endo helial
laye u he sugges s a de ec in he Ripk3
−/−
ECs o espond
o ex acellula cues. Indeed, he absence o RIPK3 educed
signaling in esponse o VEGF-A, VEGF-B and FGF-b bu no
TNF in he p38/HSP27 pa hway, a key pa hway in pe me-
abili y. Ou esul s show ha he educed umo nodule
o ma ion in RIPK3-de icien mice canno be a ibu ed o he
ailu e o espond o one s imuli alone. In e es ingly, in Sei e
e al.
22
, he loss o Mincle combined wi h he V12K-Ras PDA
did no p o ec , as well as he loss o RIPK3, sugges ing
addi ional ac o s may ha e a ole in he abili y o RIPK3 o
o m PDA in a V12K-Ras model. Al e na i ely, he ole o
RIPK3 may be mo e p e alen in issues ha e-model
equen ly o issues ha mus unc ion as a ba ie o
in ec ion. Fu he in es iga ion o o he pe meabili y ac o s
in i o and whe he hese ligands can o m complexes o
RIPK1/RIPK3 will be o in e es o de e mine i hese
in acellula p o eins can be a ge ed o block se e al pa h-
ways in ol ed in umo me as asis.
Ripk3
−/−
mice ha e no been epo ed o ha e any ascula
de ec s du ing no mal de elopmen . In e es ingly, we epea -
edly de ec inc eased basal phospho-p38 le els in un ea ed
immo alized ECs and lung homogena es sugges ing he loss
o RIPK3 is compensa ed du ing de elopmen and egula
homeos asis. I appea s only du ing s ess such as wound
healing o umo challenge, does RIPK3 become equi ed o
he esponse o he addi ional s imuli. Changes a he basal
le el in s uc u e o he essels o lung essel s uc u e
ha e no been assessed and may p o ide insigh o why he
Ripk3
−/−
mice de elop wi h li le o no o e pheno ype.
In summa y, ou s udy shows he kinase ac i i y o RIPK1
and RIPK3 has a physiological ole in he umo mic oen i on-
men , in pa icula umo cell ex a asa ion and emodeling by
al e ing he downs eam signaling pa hways o pe meabili y
ac o s. This is a no el ole o he kinase ac i i y o RIPK1/
RIPK3 in addi ion o i s ole in egula ing nec op osis. In he
con ex o in lamma ion and cell dea h, his suppo s p e ious
epo s sugges ing he loss o kinase ac i i y o RIPK1 o he
comple e loss o RIPK3 canno be used as an indica o o
nec op osis and he po en ial o RIPK1/RIPK3 o be a
signaling pla o m will need o be assessed in each disease
model indi idually.
Ma e ials and Me hods
Animal wo k. Animals we e main ained unde op imized hygiene condi ions
(OHB), and expe imen s we e app o ed by Zü ich Can onal Ve e ina y Commi ee in
acco dance o he guidelines o he Swiss Animal P o ec ion Law (License 119/2012
and 186/2015). C57BL/6 mice we e pu chased om Jan ie Labs (F ance),
Ripk3
−/−
mice we e a kind gi om V Dixi ,
46
SNP analyzed a N=7, Ripk1
K45A/K45A
,
Ripk3
K51A/K51A
mice we e a kind gi om John Be in and Pe e J Gough om
GlaxoSmi hKline (PA, USA),
26,40
Mlkl
−/−
mice we e ob ained om J
Mu phy and W Alexande
10
(WEHI, Melbou ne, Aus alia). Mice we e injec ed wi h
B16-F10 (2 × 10
5
cells) o MC-38 (3 × 10
5
cells) in a enously in 200 μl PBS o
RIPK1/RIPK3 p omo es ascula pe meabili y
K Hänggi e al
9
Cell Dea h and Disease