iology 30 (2014) 1444e1451

Canadian Journal of Card

Basic Research

Perlecan Heparan Sulfate Proteoglycan Is a CriticalDeterminant of Angiogenesis in Response to Mouse

Hind-Limb IschemiaBeiping Qiang, MD, PhD,a,c,* Sang Yup Lim, MD, PhD,a,f,* Michael Lekas, MD,a,c

Michael A. Kuliszewski, BSc(Hon),b,c Rafael Wolff, MD,a,c Azriel B. Osherov, MD,a,c

Dmitriy Rudenko, BSc(Hon),b,c Howard Leong-Poi, MD,b,c Hossein Noyan, PhD,c,d

Mansoor Husain, MD,c Kiet Tran, MD, PhD,e Karl Tryggvason, MD, PhD,e

Ulf Hedin, MD, PhD,e Karin Tran-Lundmark, MD, PhD,e and Bradley H. Strauss, MD, PhDa,c

a Schulich Heart Center, Sunnybrook Health Sciences Center, Toronto, Ontario, CanadabKeenan Research Centre in The Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Ontario, Canada

cUniversity of Toronto, Toronto, Ontario, CanadadToronto General Research Institute, Toronto, Ontario, Canada

eKarolinska Institute, Stockholm, SwedenfKorea University Ansan Hospital, Ansan, Korea

ABSTRACTBackground: Perlecan is a heparan sulfate proteoglycan (HSPG) con-stituent of the extracellular matrix with roles in cell growth, differen-tiation, and angiogenesis. The role of the HS side chains in regulatingin vivo angiogenesis after hind-limb ischemia is unknown.Methods: Heparan sulfate (HS)-deficient perlecan (Hspg2D3/D3) mice(n ¼ 35), containing normal perlecan core protein but deficient in HSside chains, and wild-type (n ¼ 33) littermates underwent surgicalinduction of hind-limb ischemia. Laser Doppler perfusion imaging(LDPI) and contrast-enhanced ultrasonography (CEU) provided serialassessment of hind-limb perfusion. Harvested muscles underwentimmunostaining for endothelial cell density (CD31), real-time reversetranscription polymerase chain reaction RT-PCR for vascular endothe-lial growth factor (VEGF) mRNA expression and western blot analysis

Received for publication October 31, 2013. Accepted June 4, 2014.

Corresponding author: Dr Bradley H. Strauss, Schulich Heart Center,Sunnybrook Health Sciences Center, 2075 Bayview Avenue, Toronto,Ontario, Canada. Tel.: þ1-416-480-6066; fax: þ1-416-480-6174.

E-mail: bradley.strauss@sunnybrook.ca*These authors contributed equally to this work.See page 1451 for disclosure information.

http://dx.doi.org/10.1016/j.cjca.2014.06.0030828-282X/� 2014 Canadian Cardiovascular Society. Published by Elsevier Inc. A

R�ESUM�EIntroduction : Le perl�ecane est un prot�eoglycane de type h�eparane-sulfate (HSPG : heparan sulfate proteoglycan), constituant de lamatrice extracellulaire, qui joue un rôle dans la croissance cellulaire, ladiff�erenciation et l’angiogenèse. On ignore le rôle des chaînes lat�eralesd’h�eparane-sulfate (HS) dans la r�egulation de l’angiogenèse in vivoaprès l’isch�emie des membres post�erieurs.M�ethodes : Les souris (n ¼ 35) d�eficientes en h�eparane-sulfate(Hspg2 D3/D3), contenant la prot�eine centrale du perl�ecane normal,mais d�eficient dans les chaînes lat�erales d’HS, et la port�ee de sou-riceaux de type sauvage (n¼ 33) ont subi une induction chirurgicale del’isch�emie des membres post�erieurs. L’imagerie de perfusion au laserDoppler (IPLD) et l’�echographie rehauss�ee par un agent de contraste(ERAC) ont fourni une �evaluation s�erielle de la perfusion des membres

Perlecan is a large heparan sulfate proteoglycan (HSPG) that isabundant in the extracellular matrix, with preferred localizationto vascular basement membranes.1 Perlecan has been shownto play an important role in cell growth, differentiation,

development, and inflammation.2 Perlecan is composed of 5distinct protein domains (I-V) with up to 4 heparan sulfate(HS) glycosaminoglycans (GAGs) bound to domains I and V.3

Perlecan appears to be involved in vasculogenesis or denovo blood vessel formation during embryogenesis, in addi-tion to cardiac development, because the perlecan-null(Hspg2�/�) mouse exhibits embryonic lethality (embryonicdays 10-12) characterized by vascular modeling defects.4

Beyond embryonic development, perlecan appears to playan important role in angiogenesis, the process of neovesselformation from pre-existing blood vessels, as demonstrated byin vitro studies including tumour angiogenesis.5 Perlecan core

ll rights reserved.

for VEGF and fibroblast growth factor (FGF)2 protein expression at days2 and 28.Results: Serial LDPI showed significantly greater perfusion recovery inischemic limbs of wild-type compared with Hspg2D3/D3 mice. CEUshowed that normalized microvascular perfusion was increased inwild-type compared with Hspg2D3/D3 mice at day 28 (0.67 � 0.12 vs0.26 � 0.08; P ¼ 0.001). CD31-positive cell counts were significantlyhigher in wild-type compared with Hspg2D3/D3 mice on day 28 (122 �30 cells vs 84 � 34 cells per high-power field [HPF]; P < 0.05).Endogenous VEGF mRNA expression (P < 0.05) and VEGF proteinexpression (P < 0.002) were significantly decreased in the ischemiclimbs of Hspg2D3/D3 mice compared with wild-type mice at day 2 andday 28, respectively. FGF2 protein expression showed no significantdifferences.Conclusions: These results suggest that the HS side chains in perlecanare important mediators of the angiogenic response to ischemiathrough a mechanism that involves upregulation of VEGF expression.

post�erieurs. Les muscles pr�elev�es ont subi l’immunocoloration de ladensit�e cellulaire endoth�eliale (CD31), la transcription inverse de lar�eaction en chaîne par polym�erase en temps r�eel (TI-RCP) de l’ex-pression de l’ARNm du facteur de croissance endoth�elial vasculaire(FCEV) et l’analyse du buvardage de western pour l’expression desprot�eines du FCEV et du facteur de croissance fibroblastique (FCF) detype 2 au 2e jour et au 28e jour.R�esultats : L’IPLD s�erielle a montr�e un r�etablissement de la perfusiondes membres isch�emiques des souris de type sauvage significative-ment plus grand que les souris Hspg2D3/D3. L’ERAC a montr�e que laperfusion microvasculaire normalis�ee augmentait chez les souris detype sauvage par rapport aux souris Hspg2D3/D3 au 28e jour (0,67 �0,12 vs 0,26 � 0,08; P ¼ 0,001). Les d�enombrements cellulairespositifs CD31 ont �et�e significativement plus �elev�e chez les souris detype sauvage que chez les souris Hspg2D3/D3 au 28e jour (122 � 30cellules vs 84 � 34 cellules par champ à fort grossissement [HPF :high-power field]; P < 0,05). L’expression endogène de l’ARNm duFCEV (P < 0,05) et l’expression de la prot�eine du FCEV (P < 0,002) ontsignificativement diminu�e dans les membres isch�emiques des sourisHspg2D3/D3 par rapport aux souris de type sauvage au 2e jour et au28e jour, respectivement. L’expression des prot�eines du FCF2 n’amontr�e aucune diff�erence significative.Conclusions : Ces r�esultats suggèrent que les chaînes lat�erales d’HSdu perl�ecane sont d’importants m�ediateurs de la r�eponse angiogène àl’isch�emie par un m�ecanisme qui implique la r�egulation à la hausse del’expression du FCEV.

Qiang et al. 1445Angiogenesis in Hspg2 D3/D3 Perlecan Mice

protein abundance has been demonstrated in many primaryhuman tumours, including prostate, breast, colon, and ma-lignant melanoma, in which it correlates with enhancedmetastatic potential.6,7 Perlecan knockdown in metastaticprostate cells reduces responses to fibroblast growth factor(FGF) and vascular endothelial growth factor (VEGF) andresults in decreased in vivo tumour growth.8 In addition to astructural role in the basement membrane, perlecan maypromote angiogenesis through the HS chains binding growthfactors, including FGF2 and VEGF, and promote receptoractivation resulting in enhanced angiogenesis.9-11 A mutantmouse with targeted deletion of 3 possible perlecan HSattachment sites in domain I (Hspg2D3/D3) resulted in areduced amount of perlecan-bound HS chains and decreasedmatrix binding of FGF2.12 The HS-deficient animals wereviable but displayed increased narrowing of injured carotidarteries, aberrant wound healing, and impairment ofangiogenesis.13

Although much information regarding the role of perlecanin angiogenesis has been obtained largely from in vitro studies,the importance of perlecan-HSPG on in vivo angiogenesis inthe postnatal response to tissue ischemia is unknown.Accordingly, we investigated whether HS side chain deficiencyof perlecan resulted in impaired recovery and reduced angio-genesis in response to hind-limb ischemia and, if so, whetherthis had any effect on growth factors such as VEGF or FGF2.

Methods

Animals

Transgenic mice. The animal care protocol was approved bythe St. Michael Hospital Institutional Animal Care and Use

Committee. Hspg2D3/D3 mice were generated by deletion ofexon 3 of the mouse perlecan gene (Hspg2), thus lacking theability to encode the sites for attachment of the 3GAG chains indomain I.14 Hspg2D3/D3 mice were back-bred 9 generations totheC57BL/6 background.C57BL/6mice were purchased fromScanbur-BK(Sollentuna, Sweden). Wild-type and homozygousmutant offspring fromN9/F2 (second-generation homozygousbreeding between littermates) animals were studied. Genotypewas determined by RT-PCR analysis of ear notch samples.

In vivo hind-limb ischemia angiogenesis model. A total of68 mice (wild-type, n ¼ 35; Hspg2D3/D3, n ¼ 33) underwentsuccessful hind-limb ischemia surgery. Animals were anaes-thetized with intraperitoneal ketamine/xylazine 200 mg/kgand 10 mg/kg, respectively, for the surgical procedure as wellas for laser Doppler measurements of limb perfusion. Hind-limb ischemia was created using a previously publishedmethod.15 Briefly, after skin incision, the entire femoral arterywas dissected and all branches were exposed. Ischemia wasinduced in the left hind limb by ligation of the distal externaliliac artery, all branches of the femoral artery, and the prox-imal popliteal and saphenous arteries. After ligation, the entirefemoral arterial segment was excised.

Assessment of hind-limb blood flow

Laser Doppler perfusion imaging. The measurement of theischemic-normal hind-limb blood perfusion ratio was doneusing a laser Doppler perfusion image (LDPI) analyzer (MoorInstruments, Wilmington, DE), which is a well-known sen-sitive method to assess differences in hind-limb blood perfu-sion, as described previously.16 In this method, a color-coded

1446 Canadian Journal of CardiologyVolume 30 2014

image representing blood perfusion is displayed: low bloodperfusion is displayed as dark blue, whereas the highestperfusion level is shown as red. LDPI was performed in 51mice (wild-type, n ¼ 23; Hspg2D3/D3, n ¼ 28) and was usedto record perfusion of both right and left limbs preoperativelyand at predetermined time points postoperatively at days 0, 7,14, and 28 (and in several mice at day 2 as well). Consecutivescans were performed on legs and feet of anaesthetized animalsplaced on a heating pad kept at 37

�C. To minimize variations

caused by ambient light, calculated blood perfusion wasexpressed as the ischemic-normal limb blood perfusion ratio.

Contrast-enhanced ultrasonographic imaging. Contrast-enhanced ultrasonography (CEU) of the distal hind-limbmuscles (gastrocnemius/soleus) was performed 7 days afterthe initial surgery with gated pulse inversion imaging(HDI5000, Philips Healthcare, Andover, MA) at a mechan-ical index of 1.0 and a transmission frequency of 3.3 MHz.Background images were acquired at baseline for subtractionof tissue signal. During continuous intravenous infusion oflipid microbubbles (Definity, Lantheus Medical Imaging,Billerica, MA), intermittent imaging was performed by pro-gressive prolongation of the pulsing interval (PI) from 0.2 to20 seconds, using the internal timer. Averaged backgroundframes were digitally subtracted from averaged contrast-enhanced frames at each PI. PI vs signal intensity were fit tothe function, y ¼ A(1 � e�bt),17 where y is acoustic intensityat pulsing interval t, A is plateau signal intensity representingmaximum blood volume, and b is the rate constant reflectingtime to plateau signal intensity, representing blood flow velocity.Blood flow was calculated by the product of A and b. CEU wasperformed in 7 mice (3 wild-type and 4 Hspg2D3/D3) at day 2and in 7 mice at day 28 (4 wild-type and 3 Hspg2D3/D3). Themice were killed after the CEU study.

Histologic and immunohistochemical evaluation. Histo-logic and immunohistochemical analysis was performed inmice killed at day 28 after completing LDPI. The ischemicand contralateral nonischemic gastrocnemius muscles were

Figure 1. Ischemic ulcer and loss of digit at day 14 (arrow, upper panel) asulfateedeficient perlecan (Hspg2D3/D3) mutant mice. These changes were

harvested immediately after death. Tissue samples were fixedin 10% formalin and processed for histologic examination. Toidentify endothelial cells, CD31 immunostaining was per-formed on skeletal muscles from both the ischemic and thenonischemic limbs, which were excised and snap frozen inliquid nitrogen. Frozen tissue samples were fixed in acetone at1:500 dilution for 1 hour and stained with antibodies directedagainst CD31 (BD Pharmingen, San Diego, CA, catalogueNo. 550274) at 1:500 dilution for 1 hour. A secondaryantirabbit antibody was then used. Each muscle specimen wascut into 4-mm sections. CD31-stained cells were countedunder high-power field (HPF) (�400 magnification) at 20randomly selected sites.

VEGF mRNA expression. Tissues from hind-limb adductorand gastrocnemius muscles in both ischemic and nonischemiclimbs were excised and snap frozen in liquid nitrogen. Sampleswere removed from 24 mice at day 2 (11 wild-type and 13Hspg2D3/D3) and 18 mice (11 wild-type and 7 Hspg2D3/D3) atday 28. Samples were subsequently sonicated in TRIzol 1 mL(Life Technologies, Grand Island, NY) using an ultrasonichomogenizer. RNA was extracted with an Aurum Total RNAmini kit (Bio-Rad Laboratories, Hercules, CA) and 1 mg wasreverse transcribed. Real-time PCR for endogenous VEGFexpression was performed on isolated cDNA with mouseHPRT used as the housekeeping gene. Data was normalized tothe contralateral nonischemic hind-limb muscle.

The mouse primers used were hypoxanthine-guanine phos-phoribosyltransferase (HPRT) Fwd: 50-GCTGGTGAAAAGGACCTCT-30; HPRT Rev: 50-CACAGGACTAGAACACCTGC-30; VEGF Fwd: 50TGTACCTCCACCATGCCAAGT-30; and VEGF Rev: 50TGGAAGATGTCCACCAGGGT-30

Western blot analysis for VEGF and FGF2 proteins.Snap-frozen muscle samples were homogenized in a wholeprotein extraction buffer containing protease inhibitors, andthe extracted proteins (20 mg) were run on a 12% Tris-Glycine gel (Novex, Life Technologies). Proteins were thentransferred to PVDF membranes (PerkinElmer, Waltham,

nd loss of nails and digits at day 21 (arrow, lower panel) in heparannot observed in wild-types (right side, upper and lower panels).

Figure 3. Microvascular blood flow ratio (ischemic-nonischemic) ofcontrast-enhanced ultrasonography (CEU) perfusion images in wild-typeand mutant mice at day 2 and day 28 after femoral artery ligation.Normalized microvascular blood flow by CEU in the ischemic leg wassignificantly higher in the wild-type mice at day 28 compared with theheparan sulfateedeficient perlecan (Hspg2D3/D3) mice. *P ¼ 0.001.

Figure 2. (A) Hind-limb perfusion monitored serially in vivo by laserDoppler perfusion imaging (LDPI). Immediately after ligation (day 0)and at day 2, perfusion of the ischemic hind limb was markedlydecreased (blue) in both the wild-type and the heparan sulfateedeficient perlecan (Hspg2D3/D3) mice compared with the nonischemichind limb (red). Perfusion improved (depicted) in the ischemic hindlimb in the wild-type mouse beginning at day 7. The Hspg2D3/D3

mouse continued to show reduction in the ischemic hind-limb perfu-sion at days 7, 14, and 28 compared with the wild-type mouse. (B)Serial changes of LDPI ratio (ischemic/nonischemic) in the wild-typemice group (grey line; n ¼ 23) and Hspg2D3/D3 mice group (blackline; n ¼ 28) after hind-limb ischemia surgery. Significant differencesin perfusion between groups were present at day 7, day 14, and day28. Isch, ischemic hind limb; Nonisch, nonischemic hand limb.*P <

0.002 vs the Hspg2D3/D3 mice group.

Qiang et al. 1447Angiogenesis in Hspg2 D3/D3 Perlecan Mice

MA) and blocked with 5% nonfat milk for 1 hour at roomtemperature, followed by incubation overnight at 4

�C with

either rabbit polyclonal anti-VEGF antibody (ab46154;1:1000; Abcam, Cambridge, MA) or rabbit polyclonal anti-FGF2 antibody (ab8880; 1:4000Abcam) in 5% milk and0.05% Tris Buffered Saline and Tween 20. Horseradishperoxidaseeconjugated antirabbit IgG secondary antibody(1:10000; BioRad Laboratories) and an enhanced chem-iluminescence detection method (Amersham; GE HealthcareLife Sciences, Pittsburgh, PA) were used to visualize theantigen-antibody complexes. Glyceraldehyde-3-phosphatedehydrogenase (GAPDH) served as loading control. Densi-tometric quantification of fold changes in GAPDH-normalized expression of VEGF and FGF2 proteins was

performed and data were shown as mean � standard error ofthe mean with n ¼ 4 per group.

Statistical analysis. Differences between the experimentalgroups analyzed in vitro and in vivo were tested by the Studentt test and c2 test, respectively. Data were expressed as mean �standard deviation as indicated, except for western blot anal-ysis, which used mean � standard error of the mean. A Pvalue < 0.05 was considered significant.

Results

Gross pathologic examination

Ischemic ulcers and loss of digits were observed in severalof the Hspg2D3/D3 mutant mice at day 14 and at day 28, butnot in the wild-type mice (Fig. 1).

Ischemic hind-limb perfusion

Laser Doppler perfusion imaging. Serial blood flow mea-surements with LDPI showed significantly higher blood flowin ischemic limbs in wild-type compared with Hspg2D3/D3

mice at day 7 (34.6% � 17.4 % vs 17.9% � 5.4%;P < 0.001), day 14 (56.2% � 28.6% vs 32.3% � 13.8%;P ¼ 0.02), and day 28 (61.7% � 16.3% vs 41.1% � 11.7 %;P ¼ 0.002) (Fig. 2, A and B).

CEU imaging. Normalized microvascular blood perfusionby CEU in the ischemic leg was significantly increased inwild-type mice at day 28 compared with Hspg2D3/D3 mice(P ¼ 0.001) (Fig. 3).

Endothelial cell immunostaining

At day 28, CD31-positive cells were significantly higher ingastrocnemius muscle samples from ischemic limbs in wild-

Figure 4. (A) Histogram showing the numbers of CD31-positive cells in high-power fields. At day 28, CD31-positive cells were significantly higher inwild-type compared with heparan sulfateedeficient perlecan (Hspg2D3/D3) mice. *P < 0.05 vs Hspg2D3/D3 mice group. (B) Immunohistochemicalstaining for CD31 in ischemic hind-limb muscle in wild-type mice and Hspg2D3/D3 mice. Brown staining represents CD31-positive cells (�400magnification).

1448 Canadian Journal of CardiologyVolume 30 2014

type compared with Hspg2D3/D3 mice (122 � 30 vs 84 � 34cells per HPF; P < 0.05) (Fig. 4, A and B).

VEGF expression

mRNA levels. Hspg2D3/D3 mice showed a significant decreasein VEGF mRNA levels in ischemic hind-limb adductormuscles at day 2 and day 28 after ligation and in ischemicgastrocnemius muscles at day 2 after ligation compared withwild-type mice (Fig. 5).

Protein levels. At day 2, there were no differences in VEGFprotein expression in the ischemic hind-limb muscles be-tween wild-type and Hspg2D3/D3 mice. There was a signifi-cant decrease in VEGF protein expression in the ischemichind-limb muscles in Hspg2D3/D3 mice compared withwild-type mice at day 28 after ligation (P ¼ 0.002)(Fig. 6A).

FGF2 protein

There were no differences in FGF2 protein expression in theischemic limbs betweenHspg2D3/D3 mice and wild-type mice ateither day 2 or day 28 after hind-limb surgery (Fig. 6B).

DiscussionThe present study is the first demonstration that perlecan-

HSPG plays an important in vivo role in the process of adultangiogenesis in response to ischemia. HS side chainedeficientperlecan results in an impairment of angiogenesis in themouse ischemic hind-limb model. Furthermore, we reveal thenovel finding that HS deficiency is associated with decreasedVEGF mRNA and protein expression in the ischemic limb,implicating VEGF as a possible mediator of proangiogeniceffects of the perlecan proteoglycan.

Compared with wild-type littermates possessing perlecanwith an intact complement of HS domains, mice with HSside chainedeficient perlecan (Hspg2D3/D3) displayed poor

Figure 5. Endogenous vascular endothelial growth factor (VEGF) mRNA levels measured by real-time reverse transcriptase polymerase chainreaction. Heparan sulfateedeficient perlecan (Hspg2D3/D3) mice showed significantly decreased VEGF mRNA levels in the gastrocnemius muscles(day 2) and in the hind-limb adductor muscles (day 2 and day 28) compared with wild-type mice. *P< 0.05 vs wild-type. At day 2, n¼ 5/group and atday 28, n ¼ 6/group.

Qiang et al. 1449Angiogenesis in Hspg2 D3/D3 Perlecan Mice

spontaneous recovery of hind-limb perfusion as assessed by 2independent techniques, LDPI and CEU. Both techniquesshow that perfusion restoration was markedly impaired intransgenic mice expressing HS-deficient perlecan. In addition,HS deficiency was associated with stigmata of critical limbischemia, including decreased limb use and frank digit/footnecrosis. The perfusion data were consistent with the CD31immunostaining for endothelial cells, a morphologic measureof capillary formation, which revealed that wild-type miceformed a greater number of capillaries in ischemic musclecompared with Hspg2D3/D3 mice. Of note, the Hspg2D3/D3

mice also showed differences in VEGF expression. There wasa marked and significant decrease in VEGF mRNA levels inboth gastrocnemius and adductor muscles in the ischemichind limb at day 2 after ligation, which was persistentlydecreased at day 28 in the hind-limb adductor muscles. VEGFprotein levels were significantly reduced at day 28 in theischemic hind-limb muscles of Hspg2D3/D3 mice comparedwith wild-type mice. Of note, we could not detect differencesin protein expression of a second heparin-binding growthfactor, ie, FGF2, between wild-type and Hspg2D3/D3 mice ateither of the 2 time points.

Previous in vitro studies revealed that perlecan maymodulate angiogenesis through interactions of the proteincore or the HS side chains with several growth factors,including VEGF and members of the FGF family, which arepowerful promoters of growth and angiogenesis found in areasof active tissue remodeling and tumour growth.18,19 Perlecancan also bind other proangiogenic factors, such as platelet-

derived growth factor-B, transforming growth factor-b, b1and b3 integrins, and fibronectin,19 supporting a crucial roleof perlecan in growth factor regulation. The perlecan HSPGcan modulate angiogenesis by binding, concentrating,sequestering and protecting the heparan binding growth fac-tors from proteolysis or presenting ligand to its cognatereceptor in a biologically active form, or both.20 Although therole of the HS side chains of perlecan has been widelyinvestigated regarding regulation of FGF2 activity, few studieshave reported on possible perlecan-VEGF interactions. Thisstudy extends the perlecan HSPG-growth factor interactionsto include in vivo evidence of impaired angiogenesis alongwith decreased mRNA and protein expression of VEGF inresponse to ischemia in the absence of the HS side chains.

Several recent in vivo and in vitro studies have indicated arelationship between intact perlecan, VEGF, and VEGFreceptors. A study in zebrafish showed that perlecan regulatesdevelopmental angiogenesis by modulating the VEGF-VEGFreceptor (VEGFR2) axis.21 Perlecan knockdown inhibitedangiogenic blood vessel development throughout the trunk andtail with findings of angiogenic sprouts that were nonfunc-tional. In addition, there was improper localization andderanged VEGF-A distribution in the developing perlecan-deficient embryo. Using an overlay assay, perlecan was shownto bind VEGF-A through the HS side chains, in addition toFGF2 binding. The binding between perlecan and angiogenicproteins seems to be mediated through the HS side chains. Ithas been reported that the HS chain of perlecan enhances theheparin-mediated phosphorylation of VEGFR2, the functional

Figure 6. (A) At day 2 (left panel), there were no significant differences in vascular endothelial growth factor (VEGF) protein expression in theischemic hind-limb muscles between wild-type and heparan sulfateedeficient perlecan (Hspg2D3/D3) mice. At day 28 (right panel), there was asignificant decrease in VEGF protein expression in the ischemic hind-limb muscles in Hspg2D3/D3 mice compared with wild-type mice. *P ¼ 0.002compared with wild-type. (B) At day 2 (left panel) and day 28 (right panel) after hind-limb ischemia surgery, there were no significant differences infibroblast growth factor 2 (FGF2) protein expression in the ischemic limbs between Hspg2D3/D3 mice and wild-type mice. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

1450 Canadian Journal of CardiologyVolume 30 2014

consequence of the perlecan-VEGF complex formation.22

Perlecan domain I (PD-I), the major site of HS-GAG attach-ment, has been shown to increase VEGF-R2 tyrosine phos-phorylation and Akt activation.23 PD-I, either alone or incombination with VEGF-165, has been shown in vitro toincrease endothelial cell tube-like structures, and this effect ofPD-I occurs in a HS-chainedependent manner. Also, HIP/RPL29, an HS-binding protein with heparanase activity, ant-agonizes VEGF- and FGF2-stimulated angiogenesis by inter-fering with HS-dependent responses.24 In total, these studiessuggest that perlecan binding sequesters VEGF in a tissue-dependent manner and thereby regulates VEGF distribution,availability, and functional activity through VEGF-R2 duringangiogenesis. These perlecan/VEGF-A interactions enhanceVEGF-R2 activity, thereby promoting endothelial cell migra-tion and proliferation during angiogenesis.

An additional important finding in our study is that incontrast to VEGF, a second heparan-binding growth factor,FGF2, did not show any differences in Hspg2D3/D3 mice andwild-type mice at 2 time points after hind-limb ischemia

surgery. Both FGF2 and VEGF have been implicated asimportant growth factors for angiogenesis.9-11,24 However, anangiogenic role of FGF2 was not evident is several experimentalstudies.25-28 Endogenous FGF2 was not required for angio-genesis after experimental ischemic retinopathy27 and ocularchoroidal injury.28 In a mouse hind-limb ischemic modelsimilar to our study, Sullivan et al.29 demonstrated thatendogenous FGF2 was not required for an angiogenic vascularresponse. These results are consistent with the findings of ourcurrent study in which FGF2 did not seem to be an importantfactor in perlecan-mediated angiogenesis after hind-limbischemia. Future work on perlecan and interactions withgrowth factors such as VEGF or FGF2 merits furtherinvestigation.

There were some limitations in our study. We used thefemoral artery ligation model to induce hind-limb ischemia asa stimulus to angiogenesis. In this particular animal model,the ischemia is induced by a surgical intervention, which maydiffer from a more chronic form of ischemia and injury thatoccurs over time with peripheral vascular disease in patients.

Qiang et al. 1451Angiogenesis in Hspg2 D3/D3 Perlecan Mice

Our mouse model in which mice are fed a normal diet alsolacks the metabolic changes characteristic of patients withperipheral artery disease and atherosclerotic risk factors such asa high-fat diet, smoking, and diabetes mellitus.

ConclusionsMice expressing HS-deficient perlecan (Hspg2D3/D3) dis-

played marked impairment in ischemic limb perfusion resto-ration and endothelial cell density compared with wild-typemice in a hind-limb ischemia model. Furthermore, HS-deficient perlecan was associated with decreased VEGFmRNA and protein expression. These results suggest that theHS side chain of perlecan is an important mediator of theangiogenic response to ischemia through a mechanism thatinvolves upregulation of VEGF expression.

Funding SourcesThis work was supported by the Canadian Institutes of

Health Research.

DisclosuresThe authors have no conflicts of interest to disclose.

References

1. Iozzo RV, Zoeller JJ, Nystrom A. Basement membrane proteoglycans:modulators par excellence of cancer growth and angiogenesis. Mol Cells2009;27:503-13.

2. Iozzo RV. Matrix proteoglycans: from molecular design to cellularfunction. Annu Rev Biochem 1998;67:609-52.

3. Jiang X, Couchman JR. Perlecan and tumor angiogenesis. J HistochemCytochem 2003;51:1393-410.

4. Costell M, Carmona R, Gustafsson E, et al. Hyperplastic conotruncalendocardial cushions and transposition of great arteries in perlecan-nullmice. Circ Res 2002;91:158-64.

5. Cohen IR, Murdoch AD, Naso MF, et al. Abnormal expression of per-lecan proteoglycan in metastatic melanomas. Cancer Res 1994;54:5771-4.

6. Sharma B, Handler M, Eichstetter I, et al. Antisense targeting of perlecanblocks tumor growth and angiogenesis in vivo. J Clin Invest 1998;102:1599-608.

7. Aviezer D, Hecht D, Safran M, et al. Perlecan, basal lamina proteoglycan,promotes basic fibroblast growth factor-receptor binding, mitogenesis,and angiogenesis. Cell 1994;79:1005-13.

8. Savore C, Zhang C, Muir C, et al. Perlecan knockdown in metastaticprostate cancer cells reduces heparin-binding growth factor responsesin vitro and tumor growth in vivo. Clin Exp Metastasis 2005;22:377-90.

9. Nugent MA, Iozzo RV. Fibroblast growth factor-2. Int J Biochem CellBiol 2000;32:115-20.

10. Iozzo RV, Murdoch AD. Proteoglycans of the extracellular environment:clues from the gene and protein side offer novel perspectives in moleculardiversity and function. FASEB J 1996;10:598-614.

11. Kadenhe-Chiweshe A, Papa J, McCrudden KW, et al. Sustained VEGFblockade results in microenvironmental sequestration of VEGF bytumors and persistent VEGF receptor-2 activation. Mol Cancer Res2008;6:1-9.

12. Zhou Z, Wang J, Cao R, et al. Impaired angiogenesis, delayed woundhealing and retarded tumor growth in perlecan heparan sulfate-deficientmice. Cancer Res 2004;64:4699-702.

13. Tran PK, Tran-Lundmark K, Soininen R, et al. Increased intimal hyp-erplasia and smooth muscle cell proliferation in transgenic mice withheparan sulfate-deficient perlecan. Circ Res 2004;94:550-8.

14. Rossi M, Morita H, Sormunen R, et al. Heparan sulfate chains of per-lecan are indispensable in the lens capsule but not in the kidney. EMBOJ 2003;22:236-45.

15. Couffinhal T, Silver M, Zheng LP, et al. Mouse model of angiogenesis.Am J Pathol 1998;152:1667-79.

16. Sarkar K, Fox-Talbot K, Steenbergen C, Bosch-Marce M, Semenza GL.Adenoviral transfer of HIF-1alpha enhances vascular responses to criticallimb ischemia in diabetic mice. Proc Natl Acad Sci U S A 2009;106:18769-74.

17. Pascotto M, Leong-Poi H, Kaufmann B, et al. Assessment of ischemia-induced microvascular remodeling using contrast-enhanced ultrasoundvascular anatomic mapping. J Am Soc Echocardiogr 2007;20:1100-8.

18. Knox SM, Whitelock JM. Perlecan: how does one molecule do so manythings? Cell Mol Life Sci 2006;63:2435-45.

19. Gustafsson E, Almonte-Becerril M, Bloch W, Costell M. Perlecanmaintains microvessel integrity in vivo and modulates their formationin vitro. PLoS One 2013;8:e53715.

20. Theocharis AD, Skandalis SS, Tzanakakis GN, Karamanos NK. Pro-teoglycans in health and disease: novel roles for proteoglycans in malig-nancy and their pharmacological targeting. FEBS J 2010;277:3904-23.

21. Muthusamy A, Cooper CR, Gomes RR Jr. Soluble perlecan domain Ienhances vascular endothelial growth factor-165 activity and receptorphosphorylation in human bone marrow endothelial cells. BMC Bio-chem 2010;11:43.

22. Zoeller JJ, Whitelock JM, Iozzo RV. Perlecan regulates developmentalangiogenesis by modulating the VEGF-VEGFR2 axis. Matrix Biol2009;28:284-91.

23. Ashikari-Handa S, Habuchi H, Kariya Y, Kimata K. Heparin regulatesvascular endothelial growth factor 165-dependent mitogenic activity,tube formation, and its receptor phosphorylation of human endothelialcells. J Biol Chem 2005;280:31508-15.

24. D’Souza S, Yang W, Marchetti D, Muir C, Farach-Carson MC,Carson DD. HIP/RPL29 antagonizes VEGF and FGF2 stimulatedangiogenesis by interfering with HS-dependent responses. J Cell Biochem2008;105:1183-93.

25. Hanneken A, deJuan E, Lutty GA, et al. Altered distribution of basicfibroblast growth factor in diabetic retinopathy. Arch Ophthalmol1991;109:1005-11.

26. Ohira A, de Juan E, Tano Y, Wilson CA. Immunohistochemical distri-bution of basic fibroblast growth factor in experimental retinal ischemiaand reperfusion in the rat. Histochem J 1996;28:607-11.

27. Ozaki H, Okamoto N, Ortega S, et al. Basic fibroblast growth factor isneither necessary nor sufficient for the development of retinal neo-vascularization. Am J Pathol 1998;153:757-65.

28. Tobe T, Ortega S, Luna JD, et al. Targeted disruption of the FGF2 genedoes not prevent choroidal neovascularization in a murine model. Am JPathol 1998;153:1641-6.

29. Sullivan CJ, Doetschman T, Hoying JB. Targeted disruption of the Fgf2gene does not affect vascular growth in the mouse ischemic hindlimb.J Appl Physiol 2002;93:2009-17.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具