R-Ras is a global regulator of vascular regeneration thatsuppresses intimal hyperplasia and tumor angiogenesisMasanobu Komatsu1,2 & Erkki Ruoslahti1

R-Ras is a small GTPase of the Ras family that regulates cell

survival and integrin activity. Despite a number of in vitro

studies, the in vivo function of R-Ras remains unclear. Here,

we used R-Ras–null mice to explore the in vivo function of this

small GTPase. Our results show a role for R-Ras as a

regulator of vascular differentiation that primarily affects the

remodeling of blood vessels. We show that R-Ras–null mice,

although otherwise phenotypically normal, mount excessive

vascular responses. We found that in vivo R-Ras expression

is largely confined to fully differentiated smooth muscle cells,

including those of blood vessels, and to endothelial cells.

Challenging the R-Ras–null mice with arterial injury or tumor

implantation showed exaggerated neointimal thickening in

response to the injury and increased angiogenesis in the

tumors. In wild-type mice, R-Ras expression was greatly

reduced in hyperplastic neointimal smooth muscle cells and

in angiogenic endothelial cells. Forced expression of activated

R-Ras suppressed mitogenic and invasive activities of growth

factor–stimulated vascular cells. These results establish an

unexpected role for R-Ras in blood vessel homeostasis and

suggest that R-Ras signaling may offer a target for therapeutic

intervention in vascular diseases.

The Ras family of small GTPases includes intracellular signalingmolecules that function as binary switches regulated by GTP. Theoncogenic Ras proteins, H-Ras and K-Ras in particular, are centralplayers in cellular signaling networks; they are activated by mostgrowth factors as well as by integrins. R-Ras is a member of the Rasprotein family that antagonizes H-Ras signaling1,2. Because of itssequence homology with the prototypic Ras proteins, the geneencoding R-Ras has been often described as a transforming oncogene3.The transforming activity of R-Ras is, however, quite low comparedwith that of H-Ras and K-Ras3, and there is no evidence thatactivating mutations would occur in spontaneous malignancies.Moreover, R-Ras differs from the other members of the Ras familyin that it contains a proline-rich SH3 domain binding site. It can alsobe phosphorylated by Eph receptors and Src; both SH3 domainbinding and phosphorylation regulate R-Ras activity4–6.

R-Ras and H-Ras exert opposite effects on cell–extracellular matrixadhesion; R-Ras enhances integrin-mediated cell adhesion by elevating

the affinity and avidity of integrins7, whereas H-Ras inhibits integrinactivities8. There is also a notable contrast in R-Ras and H-Rasactivities in cell differentiation. R-Ras promotes the differentiationof myoblasts and the fusion of these cells to myotubes, a process thatrequires cell-cycle arrest and entry into the G0 state9, whereas H-Rasinhibits these processes10. These observations suggest that the balancebetween R-Ras and H-Ras signaling may govern the decision betweengrowth and differentiation. Despite these in vitro studies, however, thein vivo function of R-Ras remains unclear.

Here, we wished to explore the in vivo function of R-Ras usingR-Ras–null mice. Heterozygous intercrosses produced viable offspringwith a mendelian distribution of the genotypes. The R-Ras–null micewere fertile and showed no obvious abnormalities, and their tissuesappeared normal upon histological examination. We confirmed thedisruption of Rras mRNA (which encodes R-Ras) and the absence ofprotein product in null mice by RT-PCR, real-time RT-PCR andR-Ras–specific immunoblotting (Supplementary Fig. 1 online) as wellas by immunohistostaining assays. The viability of the null mice showsthat R-Ras is not essential for mouse embryonic and postnataldevelopment, suggesting the existence of alternative pathways tocompensate for the loss of R-Ras.

The lack of any obvious phenotypic abnormalities in the R-Ras–nullmice prompted us to examine their responses to stress. To determinewhich tissues to focus on, we examined R-Ras expression in mousetissues. Early mRNA studies indicated that Rras is widely expressedthroughout various tissues and organs11. But antibody stainingshowed a notably restricted tissue distribution for R-Ras, which wasprimarily confined to smooth muscle in various tissues and organs(Supplementary Fig. 2 online). Vascular smooth muscle cells (VSMC)in small arterioles and in major arteries showed intense R-Ras staining,whereas veins were not as strongly positive (Supplementary Fig. 2online). R-Ras was distributed along the plasma membrane in thesmooth muscle cells (SMC; Supplementary Fig. 2 online). Tissuesfrom R-Ras–null mice were negative for R-Ras (Supplementary Fig. 2online). Endothelial cells of lung capillaries also expressed R-Ras at ahigh level (Supplementary Fig. 2 online). We detected lower levels ofexpression in renal glomeruli (data not shown) and the venousendothelium of the spleen (Supplementary Fig. 2 online). All othertypes of cells, including hepatocytes, neuronal, epithelial, hematopoi-etic, cardiac muscle and skeletal muscle cells did not or only weakly

Received 13 June; accepted 6 October; published online 13 November 2005; doi:10.1038/nm1324

1Burnham Institute for Medical Research, Cancer Research Center, 10901 North Torrey Pines Road, La Jolla, California 92037, USA. 2Present address: Division ofMolecular and Cellular Pathology, Department of Pathology, School of Medicine, University of Alabama at Birmingham, 1670 University Boulevard, VH660,Birmingham, Alabama 35294-0019, USA. Correspondence should be addressed to E.R. (ruoslahti@burnham.org).

1346 VOLUME 11 [ NUMBER 12 [ DECEMBER 2005 NATURE MEDICINE

L E T T ERS©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

expressed R-Ras (Supplementary Fig. 2 online). Immunoblotting oftissue extracts with R-Ras–specific antibodies (data not shown) gaveresults consistent with the immunohistology studies; we found thehighest R-Ras protein content in extracts of the aorta, lungs andintestines. The ubiquitous R-Ras expression detected in previousnorthern blot and PCR analyses11 and in our immunoblottingexperiments is probably the result of the presence of SMCand endothelial cells in every tissue. Potential roles of R-Rashave been examined in vitro in neuronal cells12 and hematopoieticcells13; however, our observation indicates that physiological functionof this small GTPase would be best characterized in SMC andendothelial cells.

To obtain further clues on R-Ras functionin vivo, we determined the developmentalpattern of R-Ras expression. All embryonictissues were negative for R-Ras from embryo-nic day (E)8 (the earliest time tested) to E16(Supplementary Fig. 3 online). The gastro-intestinal smooth muscle of E17–18 embryoswas the first tissue in which R-Ras becamedetectable (data not shown). At birth, theSMC of most organs and tissues expressedlow levels of R-Ras, and these levels increasedover the first few days (Supplementary Fig. 3online). R-Ras was also present in neonatalcardiac and skeletal muscle, particularly atmyotendinous junctions, whereas adult mus-cle was negative for R-Ras. R-Ras expressionin VSMC and lung capillary endothelial cellswas low throughout late embryonic life, withonset of higher expression at birth. Conver-sion of developing VSMC into fully differen-tiated cells is not complete until early inpostnatal life14, and the temporal pattern ofR-Ras expression we observed was similar tothat of VSMC differentiation markers, such assmooth muscle myosin heavy chain andsmoothelin14 (Supplementary Fig. 3 online).These results show that abundant vascular

expression of R-Ras is restricted in the adult to differentiated SMCand to endothelial cells of mature vessels.

The observed spatiotemporal pattern of R-Ras expression raised thepossibility that R-Ras might have a regulatory role in the growth and/or homeostasis of the adult vasculature. We studied the expression ofR-Ras in a model that is commonly used to mimic the restenoticlesions that can develop after angioplasty15. We found that neointimalthickening was greatly increased in R-Ras–null mice 5–6 weeks afterinjury (P o 1 � 10–8; Fig. 1a). There was a sustained proliferativeresponse to the arterial injury in the R-Ras–null mice: significantnumbers of neointimal cells continued to proliferate in the null mice3–4 weeks after the injury, whereas in agreement with earlier

Wt

Wt Null140120

Brd

-U+ c

ells

/un

it ne

oint

ima

area

100806040200

nn

m m

NullUninjured

3 weeks 6 weeks R-Ras–nullat 6 weeks

9 d

n

n

n

n

n

m

m

m

m

m

50 µm

30 µm

20 µm

50 µm 50 µm

30 µm

2 weeks

6 *

*

5

4

Neo

intim

a/m

edia

3

2

Wild

-type

R-Ras

–null

Wild

-type

R-Ras

–null

1

0

a c

b

Figure 1 Enhanced experimental restenosis and sustained neointimal proliferation in R-Ras–null mice.

(a) Comparison between wild-type (left) and R-Ras–null (center) restenosis at 5–6 weeks after injury

is presented as neointima/media area ratio (right). *P o 1 � 10–8; n ¼ 14 mice per group. Scalebar, 50 mm. (b) Brd-U staining of 3-week-old lesions and quantification of Brd-U–positive cells

in neointima. Data are presented as a cumulative result of 3- and 4-week-old lesions. Unit

neointima area, 0.05 mm2. Arrowheads, internal elastic lamina. *P o 0.02. Scale bar, 20 mm.

(c) Spatiotemporal pattern of R-Ras expression after arterial injury was determined by immunostaining

cross-sections of the injured femoral artery (wild-type) at different time points after injury (9 d to

6 weeks). R-Ras–null lesion provided a control for staining specificity. Open arrowheads, internal

elastic lamina; filled arrowheads, external elastic lamina; m, tunica media; n, neointima.

WtWt

Null Null

125 3.53.0

2.52.01.51.00.50.0

Lung metastasisTumor-free

area

Mic

rove

ssel

s/m

m2

Mic

rove

ssel

s/ao

rtic

rin

g

Rel

ativ

e nu

mbe

rof

ves

sels

100

* *

75

50

50

40

30

20

10

00 2 4

*

Days

ME

CA

32R

-Ras

6 8

25

0

Wild

-type

R-Ras

–null

Wild

-type

R-Ras

–null

a d eb c

Figure 2 Angiogenesis is enhanced in R-Ras–null mice. (a) Fluorescence microscopy of CD31 (red) and DAPI (blue) staining of B16F10 mouse melanoma

implant at 10 d after implantation. (b) Microvessel density in the tumor implant was quantified. Tumor angiogenesis was significantly enhanced in R-Ras–

null mice compared with wild-type host mice. *P ¼ 0.001. (c) Microvessel infiltration into Matrigel plug was significantly enhanced in R-Ras–null mice in

response to vascular endothelial growth factor (VEGF). *P o 0.001. (d) R-Ras–null endothelial cells showed an elevated angiogenic response to VEGF in

ex vivo aortic ring assays. Representative micrographs of sprouting endothelial cells from wild-type (Wt) and R-Ras–null (Null) aortic rings at day 6 and

quantification as number of sprouting microvessels/ring. Solid line, R-Ras–null; dashed line, wild-type control. Error bars indicate s.e.m. *P o 0.02.

(e) R-Ras expression was not detectable in the tumor neovasculature. The expression of R-Ras in angiogenic vessels was determined by immunostaining4T1 mouse mammary tumor lung metastases (upper left). An immediately adjacent section was stained for the endothelial marker, MECA32, to identify

intratumor microvessels (lower left). The dashed lined areas in the left panels are shown at a higher magnification (center). Endothelium of the

tumor-free area of the same lung is shown (right). Scale bar, 50 mm for left panels; 20 mm for center and right panels.

L E T T ERS

NATURE MEDICINE VOLUME 11 [ NUMBER 12 [ DECEMBER 2005 1347

©20

05 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

results16,17, the proliferation had essentially ceased in the wild-typemice (P o 0.02; Fig. 1b). Notably, R-Ras expression was lost in thewild-type lesions during the hyperplastic phase, and the subsequentmitotic quiescence temporally correlated with resumed R-Ras expres-sion in neointima and media (Fig. 1c). These results suggest thatR-Ras signaling in the regenerating arterial wall negatively regulatessmooth muscle hyperplasia. The R-Ras effect on vascular injury wasstrong relative to other negative regulators of vascular proliferation. Ina similar arterial injury model, inactivationof the gene encoding heme oxygenase-1increased neointimal thickening twofoldand Brd-U incorporation sixfold18. The cor-responding changes in the R-Ras–null micewere greater than threefold and tenfold,respectively, showing that R-Ras has a pro-found effect on how vessels respond toinjury. Recently, a potential role for R-Rasto enhance inflammatory responses to ather-osclerotic and/or restenotic lesions wasreported2. F4/80 staining of the arteriallesions, however, showed little macrophageassociation with neointima in both wild-typeand R-Ras–null mice (Supplementary Fig. 4online), indicating that the involvement ofinflammatory cells in the intimal expansionprocess is insignificant. Thus, our resultssuggest that effects of R-Ras deficiencyon inflammatory responses, if any, areoverridden by the enhanced neointimalproliferation and VSMC invasion in theR-Ras–deficient lesions.

We next tested the effect of absence ofR-Ras on angiogenesis. Neovascularizationof mouse melanoma B16F10 tumor implantswas greatly enhanced in R-Ras–null host micecompared to littermate controls (P ¼ 0.001;Fig. 2a,b). Tumor size was not affected in theR-Ras–null mice (data not shown). An in vivoMatrigel plug angiogenesis assay also showedhigher density of microvessels in the plugs ofR-Ras–null mice (P o 0.001; Fig. 2c).

Increased endothelial sprouting from ex vivo aortic ring cultures furtherestablished the enhanced angiogenic response of the R-Ras–null mice(P o 0.02; Fig. 2d). In wild-type mice, neointimal VSMC andangiogenic endothelial cells in tumor vasculature lost their R-Rasexpression, as shown for lung metastases of 4T1 mouse mammarytumor (Fig. 2e). Similar results were obtained from the 4T1 tumorgrown in the spleen (data not shown). As the availability of a bloodsupply is a limiting factor in tumor growth19, we would have expected

R-Ras

45 1.25

1.00

0.75

0.50

0.25

0.00

40353025201510

50

GAPDH

Per

cent

cyc

ling

cells

Rel

ativ

e m

igra

tion

CASMC

Aorta

tissu

e

Contro

l

Control

Pha

seF

-act

in

Contro

l

R-Ras

38V

R-Ras

38V

R-Ras38V

R-Ras

87L

R-Ras

87L

R-Ras

43N

R-Ras

43N

Contro

l

R-Ras

38V

R-Ras

87L

R-Ras

43N

a c db

Figure 3 R-Ras signaling inhibits proliferation and migration and induces extensive morphological differentiation of human coronary artery smooth muscle

cells. (a) Activated (R-Ras38V and R-Ras87L) and inactive (R-Ras43N) forms of R-Ras were transduced into cultured CASMC, and their expression levels

were compared with endogenous levels in aorta tissue and in mock-transduced cells (control). (b–d) Activated forms of R-Ras inhibited cell-cycle progression

(b) and induced morphological differentiation (phase-contrast images and phalloidin staining of filamentous actin (F-actin)) (c) of CASMC cultured inmitogen-rich growth media smGM-2. R-Ras signaling also resulted in the inhibition of cell motility (d). Fraction of cycling cells was determined by Ki-67

staining. PDGF-BB–induced (10 ng/ml) motility was determined using Transwell culture insert (Costar) coated with 10 mg/ml collagen I. Data are presented

relative to mock-transduction control. Error bars indicate s.e.m.

R-RasGAPDH

HUVEC

Control

R-Ras38V

70

60

50

40

30

20

10

0

1.1 1404.0

3.53.0

2.52.0

1.51.0

0.5

0.00 1 2

Days3

Control

4 5 6

120

100

80

60

40

20

0

1.00.90.80.70.60.50.40.30.20.10.0

0 0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

123456789

Per

cent

apo

ptot

ic c

ells

Rel

ativ

e tu

be fo

rmat

ion

Per

cent

rel

ativ

eN

umbe

r of

ves

sels

Rel

ativ

e nu

mbe

r of

ves

sels

Fol

d in

crea

se

Rel

ativ

e in

vasi

on

Lung

tissu

e

Contro

l

Contro

l

Contro

l

Contro

l

Contro

l

Contro

l

R-Ras

38V

R-Ras

38V

R-Ras

38V

R-Ras

38V

R-Ras

38V

R-Ras

38V

R-Ras

87L

R-Ras

87L

R-Ras

87L

R-Ras

43N

R-Ras

43N

R-Ras

43N

R-Ras

43N

R-Ras

43N

R-Ras38VR-Ras87L

a b c d

gfe

Figure 4 R-Ras signaling inhibits angiogenic activities. (a) Activated and inactive forms of R-Ras were

transduced into cultured HUVEC, and expression levels were compared with endogenous levels in lung

tissue and in mock-transduced cells (control). (b–d) Activated forms of R-Ras inhibited apoptosis,

proliferation and invasion of HUVEC. (b) Apoptosis induced by serum starvation, (c) fold increasesin 3-d culture, (d) HUVEC invasion through Matrigel toward 10ng/ml bFGF. (e) R-Ras signaling

resulted in the inhibition of tube formation of HUVEC plated on Matrigel. Left panels, representative

micrographs. Right panel, the length of the tubes was measured and expressed relative to mock-

transduced control. (f) Lentivirus-mediated in vivo delivery of activated Rras suppressed VEGF-induced

angiogenesis in R-Ras–null mice. Microvessel infiltration into Matrigel plugs was determined by

MECA32 staining and quantified (six mice per group). Control, mock infection with empty vector.

(g) VEGF-induced microvessel sprouting from R-Ras–null aortic rings was suppressed ex vivo by

infection of the lentivirus carrying activated Rras. Control, mock infection control. Data are presented

as microvessel count relative to the control at day 4.

L E T TERS

1348 VOLUME 11 [ NUMBER 12 [ DECEMBER 2005 NATURE MEDICINE

©20

05 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

larger tumors in the R-Ras–null group; however, this was not the case.A possible explanation is that the new vessels in the R-Ras–null micewere immature to an extent that made them functionally defective20.Thus, enhanced proliferative endothelial cell responses also correlatedwith a lack of R-Ras expression.

We analyzed the effects of R-Ras expression in primary cultures ofhuman coronary artery smooth muscle cells (CASMC). These cellsshow dedifferentiated features in the presence of growth factorsupplements (basic fibroblast growth factor, endothelial growth factorand insulin-like growth factor), whereas their growth and migratorybehaviors mimic those of VSMC in an atherosclerotic environment21.R-Ras expression in cultured VSMC and human umbilical veinendothelial cells (HUVEC) was 2–5% of the levels in intact tissues(Supplementary Fig. 5 online). The expression increased B3-fold inCASMC when cultured in differentiation-promoting conditions (Sup-plementary Fig. 5 online). We used retrovirus-mediated transductionto express activated R-Ras in proliferating CASMC at a level compar-able to that in quiescent cells in tissues (Fig. 3a and SupplementaryMethods online). Two forms of activated R-Ras (R-Ras38V and R-Ras87L) both inhibited cell-cycle progression and promoted entry toG0, even in mitogen-rich growth media (Fig. 3b). For reasons that arenot clear, R-Ras87L was consistently more active in this assay than R-Ras38V, but not in other assays. R-Ras also caused increased cellspreading and altered actin organization, both features associated withmorphological differentiation of VSMC (Fig. 3c). The effect wassimilar to but even stronger than that of heparin, a well-knownVSMC-differentiating agent22. Migration induced by plateletderivedgrowth factor was also substantially inhibited by R-Ras (Fig. 3d).These effects were dependent on R-Ras signaling and not dependenton R-Ras protein expression per se, as we observed no effects when aninactive R-Ras variant (R-Ras43N) was expressed. In contrast to R-Ras, H-Ras has been shown to enhance both the proliferation ofVSMC and neointimal thickening23, suggesting that R-Ras and H-Rasexert opposing effects on VSMC.

Transduction of activated R-Ras also induced changes in HUVEC(Fig. 4a); the cells spread extensively, becoming extremely flat (datanot shown) and their survival in low-serum cultures was enhanced(Fig. 4b). As with VSMC, R-Ras inhibited the proliferation andinvasion of HUVEC (Fig. 4c,d). Furthermore, tube formation(an in vitro approximation of angiogenesis) by HUVEC placed onMatrigel was greatly diminished by the expression of activatedR-Ras (Fig. 4e).

To determine whether restored R-Ras signaling inhibits angiogen-esis, we incorporated a lentiviral vector carrying an activated Rras intoMatrigel plugs. The R-Ras–expressing plugs showed greatly reducedmicrovessel infiltration (Fig. 4f), indicating that exogenous R-Ras canreverse the R-Ras–null vascular phenotype in vivo. Supporting thisconclusion, microvessel outgrowth from the ex vivo culture of R-Ras–null aortic ring was suppressed by infection with the Rras-carryinglentivirus (Fig. 4g).

The enhanced neointimal growth and tumor angiogenesis causedby the loss of R-Ras expression may be related to the ability of R-Rasto promote cellular quiescence and differentiation. Freed of therestraint of R-Ras, the VSMC and endothelial cells of the R-Ras–nullmice responded more strongly to stimuli. Similarly, the transientlyreduced R-Ras expression we observed in growing and regeneratingvessels may permit proliferation and migration of VSMC andendothelial cells in such vessels. Nascent vessels mature through aprocess that involves inhibition of endothelial cell proliferation andmigration, production of basement membrane and formation of asheath of pericytes or VSMC20. This maturation process may be

inefficient in the absence of R-Ras. Conversely, genetic or pharmaco-logical enhancement of R-Ras activity might improve the functionalityof immature vessels. Our finding on R-Ras function in vascularregeneration and remodeling shows a notable contrast to the role ofprototypic Ras proteins in these processes: it is well established thatH-Ras enhances angiogenesis and intimal hyperplasia23,24. It may bethat the balance of R-Ras and H-Ras acts as a switch that controlsproliferation and invasion versus quiescence decisions in vascular cells.It is, therefore, a possibility that a crosstalk mechanism may existbetween R-Ras and H-Ras–Raf downstream pathways, as suggestedpreviously1. Other R-Ras–regulated pathways may also have a role.Such potential downstream pathways include integrin-mediated sig-naling. In this regard, it is notable that both affinity for and expressionof b1 integrins are downregulated in endothelial cells and VSMC ofhyperplastic arterial lesions25 and that this downregulation temporallycoincides with the loss of R-Ras expression observed here. Furtherstudies on R-Ras signaling may lead to new treatments for proliferativediseases of the vasculature and for cancer.

METHODSMice. R-Ras-null mice OST24882 were generated for us by Lexicon Genetics as

previously described26. The inactivation of Rras in these mice is caused by an

insertion of the gene-trap vector VICTR20 (ref. 26) between exons 4 and 5 of

Rras on chromosome 7. The insertion disrupted Rras gene expression as

determined by RT-PCR, real-time RT-PCR and R-Ras–specific immunoblotting

of tissue extracts (Supplementary Fig. 1 online), as well as immunohisto-

chemistry of tissue sections (Supplementary Fig. 2 online). We analyzed Rras

mRNA by RT-PCR of lung and kidney extracts using a primer set that flanks

the knockout vector insertion site between exons 4 and 5 (5¢-AGGCAGAGTTT

CAATGAGGTGGGCAAGCTC-3¢ (forward), 5¢-CTCATCGACATTCAGACG

CAGTTTG-3¢ (reverse)). Real-time RT-PCR was performed at the Burnham

Institute’s Gene Analysis Facility, using a different primer set that is specific for

exons 3 and 4 (5¢-ACAGGCAGAGTTTCAATGAG-3¢ (forward), 5¢-GTTC

TCCAGATCTGCCTTG-3¢ (reverse)). We amplified Ppia mRNA using specific

primers (5¢-CACCGTGTTCTTCGACATC-3¢ and 5¢-ATTCTGTGAAAGGAG

GAACC-3¢) to normalize real-time RT-PCR results, and we determined the

relative expression of Rras mRNA. We backcrossed heterozygous founder mice

with the original mixed genetic background (C57BL6/129Sv) three to five times

to C57BL6 (Harlan), and intercrossed the offspring to produce homozygous

and wild-type littermates. All animal experiments were approved by the

Burnham Institute’s Institutional Animal Research Committee.

Immunological detection. We determined R-Ras expression by standard

immunohistochemistry, immunofluorescence and immunoblotting methods.

We raised rabbit polyclonal antibodies against full-length human R-Ras protein

prepared as a GST-fusion protein in bacteria6. We affinity-purified antibodies

by positive and negative selection on R-Ras– and H-Ras–affinity columns,

respectively. Rabbit antiserum against amino acids 11–31 of mouse R-Ras27 was

a gift from J.C. Reed (Burnham Institute for Medical Research). We fixed mouse

tissues with Bouin solution (Sigma), embedded them in paraffin and sectioned

them for histological analyses. We obtained sections of 8–16 d mouse embryos

from Novagen. We produced lung metastases by injection of 1 � 105 mouse

mammary tumor 4T1 cells into the tail vein of syngeneic (BALB/c) mice. We

collected the metastasis-bearing lungs 12 d later and fixed them in Bouin

solution. We used rat monoclonal antibody against the pan-endothelial cell

antigen MECA32 (Pharmingen) to visualize microvessels in tumor. To examine

R-Ras expression in cultured cells, we cultured cells under different conditions

and determined the expression by immunoblot analysis and densitometry. We

used pan-actin–specific or GAPDH-specific (Chemicon) blot to normalize total

protein loading. We used heparin (Sigma) to promote differentiation of VSMC

at 200 mg/ml in media supplemented with 5% FCS for 3 d. Preparation of

histological samples was performed at the Institutional Histological Facility.

Experimental restenosis model. We produced neointimal hyperplasia in

mouse femoral artery as previously described17. After the induction of the

L E T T ERS

NATURE MEDICINE VOLUME 11 [ NUMBER 12 [ DECEMBER 2005 1349

©20

05 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

vascular injury, we collected the injured vessels at the times indicated. We

calculated areas of the media and neointima in the cross-sections of restenotic

arteries by image analysis using Spot System (Universal Imaging Corporation).

We presented levels of intimal hyperplasia as neointima/media area ratio. We

performed statistical analysis using Student t-test (n ¼ 14). Proliferation of

neointimal cells was assessed by Ki-67–specific or Brd-U staining after the mice

received 1 mg Brd-U/d for 2 d through intraperitoneal injection before

collecting the arterial lesion.

Angiogenesis assays. We performed tumor angiogenesis assay, Matrigel plug

assay and ex vivo microvessel outgrowth assay as described previously28,29.

Briefly, for subcutaneous tumor implantation, we implanted 1 � 106 B16F10

mouse melanoma cells into the flank of the mice and excised the tumors at 10 d

after implantation for analyses. We counted CD31-positive infiltrating micro-

vessels in five different fields per section from six mice per group. For Matrigel

plug assay, we implanted growth factor–reduced Matrigel (Becton Dickinson)

containing vascular endothelial growth factor and heparin into the flank of the

mice. We harvested the Matrigel plugs 7 d later and quantified the microvessel

infiltration . We carried out an assay for endothelial tube formation on Matrigel

using HUVEC as described previously28,29. We quantified the efficiency of tube

formation by measuring the total length of the tubes in triplicate wells.

Preparation of lentivirus for in vivo gene delivery. We produced the lentiviral

vector carrying activated Rras (R-Ras38V and R-Ras87L) and the virus with an

empty vector (control) according to the manufacturer’s manual (Invitrogen).

We prepared concentrated virus stocks by ultracentrifugation of conditioned

medium from transfected packaging cells at 50,000g for 2 h and resuspension of

the pellets in PBS using 0.5% of the starting volume. To improve the infection

and expression of the transgene, we incubated half of the virus stock with

dNTPs for in vitro reverse transcription as described previously30. We com-

bined the reverse-transcribed and original virus stocks and we concentrated

with a second ultracentrifugation (50,000g for 90 min). We resuspended the

final pellet in 0.1% of the starting volume of PBS containing 0.5% BSA and

4 mg/ml of polybrene, titrated the titer of the virus stock and used 5 � 106

transduction units for in vivo experiments.

Note: Supplementary information is available on the Nature Medicine website.

ACKNOWLEDGMENTSWe thank E. Engvall, D. Hanahan and Y. Yamaguchi for discussions andcomments on the manuscript, S. Krajewski for help with immunohistology,M. Sata for providing a tutorial video on arterial wire injury, and R. Varghesefor editing. This work was supported by US National Cancer Institute grantsPO1 CA82713, RO1 CA79984 and RO1 CA098162 (to E.R.); T32 CA09579(to M.K.); and P30 CA 30199 (Cancer Center Support Grant).

COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.

Published online at http://www.nature.com/naturemedicine/

Reprints and permissions information is available online at http://npg.nature.com/

reprintsandpermissions/

1. Sethi, T., Ginsberg, M.H., Downward, J. & Hughes, P.E. The small GTP-binding proteinR-Ras can influence integrin activation by antagonizing a Ras/Raf-initiated integrinsuppression pathway. Mol. Biol. Cell 10, 1799–1809 (1999).

2. Cole, A.L., Subbanagounder, G., Mukhopadhyay, S., Berliner, J.A. & Vora, D.K.Oxidized phospholipid-induced endothelial cell/monocyte interaction is mediated bya cAMP-dependent R-Ras/PI3-kinase pathway. Arterioscler. Thromb. Vasc. Biol. 23,1384–1390 (2003).

3. Ehrhardt, A., Ehrhardt, G.R., Guo, X. & Schrader, J.W. Ras and relatives–jobsharing and networking keep an old family together. Exp. Hematol. 30, 1089–1106(2002).

4. Wang, B., Zou, J.X., Ek-Rylander, B. & Ruoslahti, E. R-Ras contains a proline-rich sitethat binds to SH3 domains and is required for integrin activation by R-Ras. J. Biol.Chem. 275, 5222–5227 (2000).

5. Zou, J.X. et al. An Eph receptor regulates integrin activity through R-Ras. Proc. NatlAcad. Sci. USA 96, 13813–13818 (1999).

6. Zou, J.X., Liu, Y., Pasquale, E.B. & Ruoslahti, E. Activated SRC oncogene phosphor-ylates R-ras and suppresses integrin activity. J. Biol. Chem. 277, 1824–1827(2002).

7. Zhang, Z., Vuori, K., Wang, H., Reed, J.C. & Ruoslahti, E. Integrin activation by R-ras.Cell 85, 61–69 (1996).

8. Hughes, P.E. et al. Suppression of integrin activation: a novel function of a Ras/Raf-initiated MAP kinase pathway. Cell 88, 521–530 (1997).

9. Suzuki, J., Kaziro, Y. & Koide, H. Positive regulation of skeletal myogenesis by R-Ras.Oncogene 19, 1138–1146 (2000).

10. Ramocki, M.B. et al. Signaling through mitogen-activated protein kinase and Rac/Rhodoes not duplicate the effects of activated Ras on skeletal myogenesis. Mol. Cell. Biol.17, 3547–3555 (1997).

11. Saez, R., Chan, A.M., Miki, T. & Aaronson, S.A. Oncogenic activation of human R-rasby point mutations analogous to those of prototype H-ras oncogenes. Oncogene 9,2977–2982 (1994).

12. Oinuma, I., Ishikawa, Y., Katoh, H. & Negishi, M. The Semaphorin 4D receptor Plexin-B1 is a GTPase activating protein for R-Ras. Science 305, 862–865 (2004).

13. Arai, A., Nosaka, Y., Kohsaka, H., Miyasaka, N. & Miura, O. CrkL activates integrin-mediated hematopoietic cell adhesion through the guanine nucleotide exchange factorC3G. Blood 93, 3713–3722 (1999).

14. Ratajska, A., Zarska, M., Quensel, C. & Kramer, J. Differentiation of the smooth musclecell phenotypes during embryonic development of coronary vessels in the rat.Histochem. Cell Biol. 116, 79–87 (2001).

15. Carmeliet, P., Moons, L. & Collen, D. Mouse models of angiogenesis, arterial stenosis,atherosclerosis and hemostasis. Cardiovasc. Res. 39, 8–33 (1998).

16. Clowes, A.W., Reidy, M.A. & Clowes, M.M. Kinetics of cellular proliferation after arterialinjury. I. Smooth muscle growth in the absence of endothelium. Lab. Invest. 49,327–333 (1983).

17. Sata, M. et al. A mouse model of vascular injury that induces rapid onset of medial cellapoptosis followed by reproducible neointimal hyperplasia. J. Mol. Cell. Cardiol. 32,2097–2104 (2000).

18. Duckers, H.J. et al. Heme oxygenase-1 protects against vascular constriction andproliferation. Nat. Med. 7, 693–698 (2001).

19. Hanahan, D. & Folkman, J. Patterns and emerging mechanisms of the angiogenicswitch during tumorigenesis. Cell 86, 353–364 (1996).

20. Jain, R.K. Normalization of tumor vasculature: an emerging concept in antiangiogenictherapy. Science 307, 58–62 (2005).

21. Moiseeva, E.P. Adhesion receptors of vascular smooth muscle cells and their functions.Cardiovasc. Res. 52, 372–386 (2001).

22. Weissberg, P.L., Grainger, D.J., Shanahan, C.M. & Metcalfe, J.C. Approaches to thedevelopment of selective inhibitors of vascular smooth muscle cell proliferation.Cardiovasc. Res. 27, 1191–1198 (1993).

23. Jin, G. et al. Effects of active and negative mutants of Ras on rat arterial neointimaformation. J. Surg. Res. 94, 124–132 (2000).

24. Meadows, K.N., Bryant, P., Vincent, P.A. & Pumiglia, K.M. Activated Ras induces aproangiogenic phenotype in primary endothelial cells. Oncogene 23, 192–200 (2004).

25. Koyama, N. et al. Regulation and function of an activation-dependent epitope of thebeta 1 integrins in vascular cells after balloon injury in baboon arteries and in vitro.Am. J. Pathol. 148, 749–761 (1996).

26. Zambrowicz, B.P. et al. Disruption and sequence identification of 2,000 genes inmouse embryonic stem cells. Nature 392, 608–611 (1998).

27. Wang, H.G. et al. R-Ras promotes apoptosis caused by growth factor deprivation via aBcl-2 suppressible mechanism. J. Cell Biol. 129, 1103–1114 (1995).

28. Ishida, T. et al. Targeted disruption of endothelial cell-selective adhesion moleculeinhibits angiogenic processes in vitro and in vivo. J. Biol. Chem. 278, 34598–35604(2003).

29. Reynolds, L.E. et al. Enhanced pathological angiogenesis in mice lacking beta3integrin or beta3 and beta5 integrins. Nat. Med. 8, 27–34 (2002).

30. Naldini, L., Blomer, U., Gage, F.H., Trono, D. & Verma, I.M. Efficient transfer,integration, and sustained long-term expression of the transgene in adult ratbrains injected with a lentiviral vector. Proc. Natl Acad. Sci. USA 93, 11382–11388 (1996).

L E T TERS

1350 VOLUME 11 [ NUMBER 12 [ DECEMBER 2005 NATURE MEDICINE

©20

05 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine