Ying Wang,1 Amy Pei-Ling Chiu,1 Katharina Neumaier,1 Fulong Wang,1 Dahai Zhang,1 Bahira Hussein,1

Nathaniel Lal,1 Andrea Wan,1 George Liu,2 Israel Vlodavsky,3 and Brian Rodrigues1

Endothelial Cell HeparanaseTaken Up by CardiomyocytesRegulates Lipoprotein LipaseTransfer to the Coronary LumenAfter DiabetesDiabetes 2014;63:2643–2655 | DOI: 10.2337/db13-1842

After diabetes, the heart has a singular reliance on fattyacid (FA) for energy production, which is achieved byincreased coronary lipoprotein lipase (LPL) that breaksdown circulating triglycerides. Coronary LPL originatesfrom cardiomyocytes, and to translocate to the vascularlumen, the enzyme requires liberation from myocytesurfaceheparan sulfate proteoglycans (HSPGs), an activitythat needs to be sustained after chronic hyperglycemia.We investigated themechanism by which endothelial cells(EC) and cardiomyocytes operate together to enablecontinuous translocation of LPL after diabetes. EC werecocultured with myocytes, exposed to high glucose, anduptake of endothelial heparanase into myocytes wasdetermined. Upon uptake, the effect of nuclear entry ofheparanasewas also investigated. A streptozotocinmodelof diabeteswasused toexpandour in vitro observations. Inhigh glucose, EC-derived latent heparanase was taken upby cardiomyocytes by a caveolae-dependent pathway using HSPGs. This latent heparanase was converted into anactive form inmyocyte lysosomes, entered the nucleus, andupregulatedgeneexpressionofmatrixmetalloproteinase-9.The net effect was increased shedding of HSPGs from themyocyte surface, releasing LPL for its onwards trans-location to the coronary lumen. EC-derived heparanaseregulates the ability of the cardiomyocyte to send LPLto the coronary lumen. This adaptation, although acutelybeneficial, could be catastrophic chronically because

excess FA causes lipotoxicity. Inhibiting heparanasefunction could offer a new strategy for managing car-diomyopathy observed after diabetes.

In diabetes, because glucose uptake and oxidation areimpaired, the heart is compelled to use fatty acid (FA)exclusively for ATP generation (1). Multiple adaptivemechanisms, either whole-body or intrinsic to the heart,operate to make this achievable, with hydrolysis oftriglyceride-rich lipoproteins being the major source ofFA to the diabetic heart (2). This critical reaction is cata-lyzed by the vascular content of lipoprotein lipase (LPL),and we were the first to report significantly higher coro-nary LPL activity after diabetes (3). In the heart, LPL issynthesized by cardiomyocytes, transported to heparansulfate (HS) proteoglycan (HSPG) binding sites on themyocyte surface, and from this temporary reservoir, theenzyme is transferred across the interstitial space to reachendothelial cells (EC) (4,5). Before this transfer, liberationof HSPG-sequestered LPL is a prerequisite and is facili-tated by heparanase, an EC endoglycosidase that cancleave HS side chains on HSPGs in the extracellular ma-trix and on the cell surface to release bound proteins (6).

Heparanase is synthesized as a latent 65-kDa precursor.After its secretion and reuptake (7), heparanase enters

1Faculty of Pharmaceutical Sciences, The University of British Columbia, Vancouver,BC, Canada2Institute of Cardiovascular Sciences and Key Laboratory of Molecular Cardiovas-cular Sciences, Peking University, Beijing, China3Cancer and Vascular Biology Research Center, Rappaport Faculty of Medicine,Technion, Haifa, Israel

Corresponding author: Brian Rodrigues, rodrigue@mail.ubc.ca.

Received 5 December 2013 and accepted 3 March 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1842/-/DC1.

© 2014 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

See accompanying article, p. 2600.

Diabetes Volume 63, August 2014 2643

METABOLISM

the lysosome and is cleaved into a 50-kDa active form bycathepsin L (8,9). Active heparanase is stored in this organ-elle until secreted. In addition to an extracellular function,by entry into the nucleus to regulate histone acetylation/methylation, heparanase is also capable of modulating genetranscription (10,11). Thus, tumor cells, by expressing higherlevels of heparanase, are more invasive because secretedheparanase can break down extracellular matrix in addi-tion to promoting gene expression related to an invasivephenotype (12). For example, matrix metalloproteinase-9(MMP-9) gene expression induced by nuclear heparanasecan further degrade extracellularmatrix and basementmem-branes (13). In addition, MMP-9 also causes acceleratedshedding of HSPGs, such as syndecan-1 (14,15), therebyliberating HSPG-bound ligands, including angiogenic andgrowth factors. At present, whether heparanase can in-fluence gene transcription in the heart is unclear.

In the heart, heparanase is synthesized predominantlyin EC, and after diabetes, we reported a high-glucose–inducedsecretion of active heparanase (16). That high glucose alsostimulated secretion of latent heparanase, whose function inthe heart is unknown, should be noted. Interestingly, humanfibroblasts that do not express heparanase can take up exog-enous latent heparanase and convert it into active heparanase(17). Thus, an immediate response to hyperglycemia is an ECsecretion of active heparanase to facilitate translocation ofLPL from the myocyte surface to the coronary lumen (18).Given that the active heparanase pool in EC is limited andmay not meet the demand for a sustained LPL transfer afterdiabetes, it is possible that latent heparanase could be takenup and converted to active heparanase in cardiomyocytes.Such an effect could turn on MMP-9 expression, therebystimulating shedding of HSPGs that bind LPL and thus main-taining vascular LPL increase chronically. Results from thepresent studies suggest that high-glucose–induced secretionof EC heparanase can regulate gene expression of MMP-9 incardiomyocytes, an effect that may sustain transfer of LPL tothe coronary lumen after chronic diabetes.

RESEARCH DESIGN AND METHODS

This investigation conformed to the Guide for the Care andUse of Laboratory Animals published by National Institutesof Health and the University of British Columbia.

Experimental AnimalsMale Wistar rats (250–320 g) were injected intravenouslywith 55 mg/kg streptozotocin (STZ). These animals (D55)were kept for 4 days before isolation of cardiomyocytes. Toinduce acute hyperglycemia, diazoxide (100 mg/kg, intra-peritoneally) was injected in male Wistar rats (250–320 g),and they were kept for 4 h.

MaterialsRat aortic EC (RAOEC) were obtained from Cell Applica-tions. STZ, filipin, genistein, and chloroquine were obtainedfrom Sigma-Aldrich. Heparin (HEPALEAN; 1,000 units/mL)was from Organon Canada, Ltd. Heparinase III (IBEXTechnologies) was purified from recombinant Flavobacterium

heparinum. Purified latent heparanase was prepared as de-scribed (19). Small interfering (si)RNA for rat MMP-9 wasobtained from Qiagen (SI02004247). Control siRNA con-sisted of a scrambled sequence purchased from SantaCruz Biotechnology. Lipofectamine RNAiMAX was fromInvitrogen. Purified MMP-9 and p-aminophenylmercuric ac-etate (AMPA) were from RayBiotech (Norcross, GA). Anti-LPL 5D2 antibody was a gift from Dr. J. Brunzell, Universityof Washington, Seattle. Anti-heparanase antibody mAb 130,which recognizes the active (50 kDa) and latent (65 kDa)forms of heparanase, was from InSight (Rehovot, Israel). Anantibody that recognizes the cleaved form of HSPGs waspurchased from Seikagaku (Tokyo, Japan). Rat MMP-9 an-tibody was from Millipore. All other antibodies wereobtained from Santa Cruz Biotechnology. A rat syndecan-1ELISA kit was obtained from MyBioSource (San Diego, CA).A histone acetyltransferase (HAT) activity kit was obtainedfrom BioVision (Milpitas, CA).

Isolation of CardiomyocytesVentricular calcium-tolerant myocytes were prepared bya previously described procedure (20). Cardiomyocyteswere isolated from adult male Wistar rats (250–320 g),were plated on laminin-coated culture dishes, and allowedto settle for 3 h. Before treatment, unattached cells werewashed away using Medium 199.

Coronary LPL ActivityTo measure coronary LPL, hearts were perfused retro-gradely with heparin (5 units/mL) (3). Coronary effluentswere collected (for 10 s/fraction) at different time pointsover 5 min, and LPL activity in each fraction was deter-mined by measuring the in vitro hydrolysis of a sonicated[3H]triolein substrate emulsion (3). Coronary LPL activitywas calculated as the area under curve of LPL activity ineach fraction over time.

EC CultureRAOEC were cultured at 37°C in a 5% CO2 humidifiedincubator alone or cocultured with adult rat cardio-myocytes. RAOEC from the fifth to the eighth passagewere used.

TreatmentsTo test whether exogenous heparanase can bind and betaken up by cardiomyocytes, isolated myocytes weretreated with 500 ng/mL recombinant latent heparanasefor different intervals at 37°C. This experiment was alsoconducted at 15°C to inhibit heparanase uptake. In thepresence of 10 IU/mL heparin to compete for HSPGsbinding sites, or 10 IU/L heparinase III to digest HSside chains (2 h), the contribution of myocyte surfaceHSPGs in heparanase uptake was determined. For study-ing the mechanism of internalization, 350 mmol/L su-crose or 1 mg/mL filipin was applied to myocytes for15 min before latent heparanase to block clathrin-coatedpits and caveolae-dependent internalization, respectively(21,22). Dynasore (0–25 mmol/L) or genistein (0–100mmol/L) for 1 h were also used to inhibit dynamin and

2644 EC-Cardiomyocyte Cross Talk in Metabolism Diabetes Volume 63, August 2014

tyrosine kinase activity, respectively, which could be in-volved in internalization of exogenous heparanase (23,24).Chloroquine (200 mmol/L) was used to inhibit lysosomalenzymes, and myocytes were incubated for 2 h. Anacardicacid (10 mmol/L) was used to inhibit HAT activity. PurifiedMMP-9 proenzyme was activated by 10 mmol/L AMPA at37°C for 1 h (25) and added into the myocyte culturemedium at a final concentration of 1, 5, and 10 ng/mLfor 30 min to validate the effect of MMP-9 on syndecan-1shedding and LPL release from the cardiomyocyte surface.To knockdown MMP-9 expression, cardiomyocytes were cul-tured in a six-well plate and transfected with 100 pmolsiRNA using lipofectamine.

Uptake of Exogenous HeparanaseMyocytes were incubated with 500 ng/mL purified latentheparanase and washed three times with cold PBS. Todetermine binding and uptake of heparanase by myocytes,total cell lysates were collected to detect heparanase byWestern blot. For measuring internalization of hepa-ranase, plasma membrane was removed by a proceduredescribed previously (24). Briefly, myocytes were lysed in50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 0.2 mol/Lsucrose, 2 mmol/L EDTA, 2 mmol/L EGTA, and proteaseinhibitor cocktail (Roche) and centrifuged at 10,000g at 4°Cfor 10 min. The supernatant was further centrifuged at100,000g at 4°C for 1 h to spin down the plasma membrane.The resulting supernatant (devoid of plasma membrane)containing the cytosolic fraction was used to monitor hep-aranase internalization (24).

RT-PCR

Total RNA was extracted from isolated myocytes orRAOEC using Trizol reagent (Invitrogen), and 1 mg totalRNA was used for RT-PCR. PCR primers were as follows(26,27):

Heparanase (forward, 59-CAAGAACAGCACCTACTCACGAAGC-39; reverse, 59-CCACATAAAGCCAGCTGCAAAGG-39;616 bp product);

MMP-9 (forward, 59-CCCCACTTACTTTGGAAACGC-39;reverse, 59-ACCCACGACGATACAGATGCTG-39; 686 bpproduct);

18S rRNA (forward, 59-CGGCTACCACATCCAAGGAA-39;reverse, 59-GCTGGAATTACCGCGGCT-39; 187 bp product).

Isolation of Lysosomes and Nuclear FractionsLysosome-enriched fractions were isolated using a kitfrom Sigma-Aldrich. Nuclear and cytosolic fractions wereseparated using the nuclear/cytosol fractionation kit fromBioVision. To validate the purity of proteins, we usedcytosolic (GAPDH) and nuclear (histone H3) protein markersto detect their presence in cytosolic and nuclear fractions.

Western BlotWestern blotting was done as described previously (28).In some experiments, cell culture media was concentratedby trichloroacetic acid precipitation or Amicon centrifugefilter (Millipore) before detection of heparanase or LPL.

Statistical AnalysisValues are means 6 SE. Wherever appropriate, one-wayANOVA, followed by the Bonferroni test, was used todetermine differences between group mean values. Thelevel of statistical significance was set at P , 0.05.

RESULTS

Heparanase Present in Cardiomyocytes OriginatesPredominantly From an Exogenous SourceIn the heart, heparanase is expressed predominantly byEC. Undeniably, our RT-PCR results indicated thatheparanase mRNA was abundant in RAOEC when totalRNA from EC was used as a template. However, when thesame amount of total RNA was used, heparanase mRNAwas not measurable in freshly isolated cardiomyocytes,implying low transcription activity (Fig. 1A, right panel ).Regardless of this indeterminable gene expression, the65-kDa latent and the 50-kDa active forms of heparanaseprotein were both detected in isolated myocytes (Fig. 1A,left panel). We confirmed that the presence of heparanasein cardiomyocytes was not due to EC contamination be-cause we failed to detect the EC marker CD31 in this prep-aration (Fig. 1A, left panel). Interestingly, unlike RAOEC,the dominant heparanase present in cardiomyocytes wasthe active form (Fig. 1A, left panel). Culturing cardiomyo-cytes for 36 h reduced the content of both heparanaseforms, suggesting that the protein cannot be efficientlysynthesized de novo when being turned over (Fig. 1B). In-troducing latent heparanase into the culture medium rap-idly increased the level of latent heparanase in the total celllysate of myocytes (within 5 min), an effect that continuedover time. This increase was inhibited when the tempera-ture was lowered to 15°C. Unlike latent heparanase, theincrease of active heparanase was gradual and only signif-icant after 4 h (Fig. 1C), suggesting an intracellular conver-sion of latent to active heparanase in cardiomyocytes. Ourdata imply that the heparanase content in cardiomyocytesoriginates largely from uptake of exogenous protein.

Internalization of Latent Heparanase byCardiomyocytes Is Through a Caveolae-DependentPathway That Requires HSPGs, Dynamin, and TyrosineKinase ActivationHeparanase in the total cell lysates could consist of twoparts: heparanase bound to the myocyte surface or thatwhich had been internalized. Samples devoid of plasmamembrane were prepared by ultracentrifugation andvalidated by the absence of membrane protein Na+-K+

ATPase (Fig. 2A, right panel ). After incubation of myo-cytes with latent heparanase, a robust increase of thisexogenous protein was observed in the plasma mem-brane free fraction, indicating that it had been internal-ized (Fig. 2A). Using heparin to competitively inhibit thebinding of latent heparanase to HSPGs, we were able toreduce the amount of heparanase that was internalized(Fig. 2A). Because similar results were seen with heparinaseIII, which digests the HS chains of HSPGs (Fig. 2B), ourdata imply that exogenous heparanase may be taken up by

diabetes.diabetesjournals.org Wang and Associates 2645

myocytes by binding to the HS on the cell surface. Down-stream, HSPG-dependent internalization can occur throughcaveolae or clathrin-coated pits. Because the internaliza-tion of latent heparanase was blocked by filipin, but in-sensitive to sucrose (Fig. 3A), HSPGs-mediated endocytosisof heparanase is likely a caveolae-dependent rather thana clathrin-mediated event. It should be noted that at theconcentration of sucrose used, the endocytosis of epi-dermal growth factor receptor (EGFR), which is typicallyinternalized through formation of clathrin-coated pits(29), was blocked (Fig. 3A, right panel ). The involve-ment of dynamin and tyrosine kinase in this endocytoticprocess was apparent because dynasore (Fig. 3B) and

genistein (Fig. 3C) reduced the amount of heparanaseinternalized.Internalized Latent Heparanase Is Activated inLysosomes and Enters the NucleusOn incubation of cardiomyocytes with latent heparanase, wewere able to detect this enzyme in lysosomal fractions within30 min. With time, latent heparanase in this lysosomalfraction progressively increased. Unlike latent heparanase,lysosomal active heparanase only increased at a later time(after 3 h; Fig. 4A). Because this increase in lysosomal activeheparanase was prevented by chloroquine, which inhibitslysosomal proteases, our data suggest lysosomal conversionof latent to active heparanase in cardiomyocytes (Fig. 4B).

Figure 1—Heparanase present in cardiomyocytes relies on exogenous uptake. A: Total cell lysates of RAOEC (EC) and cardiomyocytes(Myo) were extracted to Western blot for latent heparanase (L-HEPA) and active heparanase (A-HEPA). To validate the purity of cardio-myocytes, equal amounts of protein from Myo and RAOEC were used to determine the presence of CD31. RT-PCR used 1 mg total RNAfrom Myo and RAOEC (left panel). Amplified DNA products from PCR were run on an agarose gel in line with a DNA ladder (M), anda negative (N) control (using water to replace PCR product) to detect heparanase gene expression (right panel). B: Myo from control ratswere isolated, plated, and allowed to equilibrate. Total cell lysates were then collected at 0 or 36 h, and probed for L-HEPA and A-HEPAusing Western blot. Purified L-HEPA (500 ng/mL) was added to isolated Myo, and heparanase content in these cells determined at theindicated times by Western blot. C: To inhibit internalization, Myo were also incubated with L-HEPA at 15°C for 15 min. Results are themean 6 SE of three repeated experiments using different animals. *Significantly different from 0 time point, P < 0.05. **Significantlydifferent from 0 time point, P < 0.01 . #Significantly different from results at 15 min at 37°C for L-HEPA, P < 0.05.

2646 EC-Cardiomyocyte Cross Talk in Metabolism Diabetes Volume 63, August 2014

Nuclear Entry of Heparanase Is Accompanied byIncreased MMP-9 ExpressionAssuming that conversion of latent to active heparanasein the lysosomes is to fulfill a biological function, wedetermined the nuclear content of heparanase andobserved the presence of active heparanase in thisorganelle even in the absence of exogenous heparanaseaddition. However, after the addition of latent hep-aranase into the culture medium, there was a dramatic

increase of latent and active heparanase in the nucleusafter 4 h (Fig. 5A). In agreement with previous studies(13), addition of active heparanase to nuclear fractions ofcardiomyocytes increased HAT activity (Fig. 5B). Importantly,nuclear entry of active heparanase was associated with anincrease in MMP-9 expression 18 h after heparanasetreatment (Fig. 5C). This increase in MMP-9 could beinhibited using the HAT activity inhibitor anacardicacid (Fig. 5D). Using purified MMP-9, this sheddase

Figure 2—Internalization of heparanase by cardiomyocytes is mediated by HSPGs. A: Isolated cardiomyocytes were treated with 500 ng/mLpurified latent heparanase (L-HEPA) in the absence or presence of 10 IU/mL heparin for 15 min. Cell lysates devoid of plasma membrane (2PM)were used to monitor internalization of L-HEPA. To determine the purity of these cell lysates, 10 mg protein from2PM or total cell lysates (Total)was used to Western blot for the PMmarker Na+-K+-ATPase (right panel ). B: Isolated myocytes were untreated (UT) or pretreated with 10 IU/Lheparinase III (III) for 2 h and subsequently incubated with 500 ng/mL purified L-HEPA for 15 min. Internalization of L-HEPA was determinedusing the –PM fraction. Results are the mean 6 SE of three repeated experiments using different animals. **Significantly different from 0 timepoint, P < 0.01. ##Significantly different from 15 min without heparin, P < 0.01. *Significantly different from untreated myocytes, P < 0.05.

diabetes.diabetesjournals.org Wang and Associates 2647

was capable of detaching syndecan-1 from the myocytesurface (Fig. 6A, upper right panel ), together with LPL(Fig. 6A, lower right panel). In line with its augmentedexpression, MMP-9 secretion into the medium also in-creased after latent heparanase incubation (Fig. 6A), whichwas accompanied by an accelerated shedding of syndecan-1(Fig. 6B) and release of HSPG-bound LPL (Fig. 6C). Becauseknockdown of MMP-9 effectively damped these effects(Fig. 6B and C), our data suggest that the heparanase/

MMP-9 axis may play an important role in facilitatingLPL translocation from cardiomyocytes to EC.

Increased Coronary LPL Activity After Diabetes IsRelated to Endothelial Heparanase-Induced MMP-9Expression in CardiomyocytesIn the heart, heparanase is predominantly expressed inEC. As we previously reported (18), glucose stimulatedsecretion of the latent and active forms of heparanasefrom EC (Fig. 7A). In addition, coculturing of cardiomyocytes

Figure 3—Internalization of latent heparanase (L-HEPA) is caveolae-dependent and requires dynamin and tyrosine kinase activity.A: Isolated cardiomyocytes were incubated with 350 mmol/L sucrose (Suc; to inhibit clathrin-coated pit formation), or 1 mg/mL filipin(Fili; to inhibit caveolae-dependent internalization) for 15 min before 500 ng/mL purified L-HEPA was added to the medium. Internalizationof L-HEPA was measured after 15 min using the fractions devoid of plasma membrane. The inhibitory effect of sucrose on clathrin-coatedpit-dependent internalization was validated by testing the internalization of epidermal growth factor (EGF) receptor (EGFR) in myocytestreated with EGF in the presence or absence of Suc (right panel ). In a different experiment, isolated myocytes were pretreated withdynasore (to inhibit dynamin activity) (B) or genistein (to inhibit tyrosine kinase activity) (C ) for 1 h, and internalization of L-HEPA after15 min was determined using Western blot. Results are the mean 6 SE of three repeated experiments using different animals. Significantlydifferent from untreated myocytes, *P < 0.05 and **P < 0.01.

2648 EC-Cardiomyocyte Cross Talk in Metabolism Diabetes Volume 63, August 2014

with EC in the presence of high glucose increased theamount of latent heparanase in cardiomyocytes. Becausethis effect was absent when EC were removed from thetop chamber (Fig. 7B), our data suggest that cardiomyo-cytes can take up endothelial-derived heparanase, a pro-cess accelerated by high glucose. We reasoned that afterprolonged hyperglycemia, such an uptake would result inaugmented conversion of latent to active heparanase, to-gether with its nuclear entry. Indeed, cardiomyocytes

from D55 animals demonstrated a 1.5-fold increase inthe nuclear content of active heparanase (Fig. 7C) thatwas associated with cleavage of nuclear HSPGs (Supple-mentary Fig. 1A). Notably, similar to our in vitro obser-vations, MMP-9 expression (Fig. 7D and SupplementaryFig. 1B) and shedding of syndecan-1 (Fig. 7E) in D55myocytes was significantly higher compared with the con-trol. Because increased coronary LPL activity was ob-served in D55 hearts (Fig. 7F), our data imply that ECpromote sustained translocation of LPL from cardiomyo-cytes to the coronary lumen by a heparanase/MMP-9 axis.

DISCUSSION

To maintain its energy supply after impaired glucoseutilization during diabetes, the heart has a higher demandfor FA, and much of it is provided by accelerated LPL-mediated hydrolysis of plasma triglyceride at the coronarylumen. Although functional at the apical side of EC, heartLPL is synthesized in cardiomyocytes, resides at a tempo-rary reservoir on myocyte surface HSPGs, and is fast-tracked to the vascular lumen by high-glucose–inducedrelease of endothelial heparanase, which cleaves theseHSPGs (30). Results from the current study suggest thatin addition to this rapid adaptation, cardiomyocytes cantake up endothelial heparanase, which activates gene ex-pression to maintain LPL trafficking during chronic hy-perglycemia (Fig. 8).

Of the multiple cell types in the heart, the EC is likelythe major source of heparanase. Surprisingly, in thepresence of undetectable heparanase gene expression incardiomyocytes, we detected protein in these cells thatwas predominantly active heparanase. Given that thehalf-life of active heparanase is ;30 h (7) and that invitro culturing of cardiomyocytes for this time decreasedits heparanase content, our data suggest that the presenceof heparanase in cardiomyocytes is not reliant on de novosynthesis but likely due to extracellular uptake. We con-firmed this process using exogenous latent heparanaseand observed that the uptake was rapid and likelyHSPG-dependent. In CHO cells, latent heparanase triggersclustering of both syndecan-1 and -4, followed by rapid in-ternalization of the heparanase-HSPG complex (31). Becausewe have previously reported clustering of syndecan-4 whencardiomyocytes are exposed to latent heparanase (18), it ispossible that such a heparanase-HSPG complex is also re-sponsible for endocytosis of this enzyme in cardiomyocytes.Nevertheless, we cannot rule out the contributive effects ofother receptors, such as LDL receptor–related proteins andmannose 6-phosphate receptors, in this heparanase-uptakeprocess (32). Of the two novel endocytotic pathways,HSPGs are involved more with a caveolae-dependentrather than clathrin-mediated endocytosis (33,34). Forexample, syndecan-1 has been shown to mediate apolipo-protein E-VLDL uptake in the human fibroblast cell lineGM00701, a process that was clathrin-independent, butinhibited by nystatin, an inhibitor of the lipid raft–caveolaeendocytosis pathway (35). Our data demonstrate that

Figure 4—After uptake, exogenous latent heparanase (L-HEPA) isactivated in lysosomes. A: Isolated myocytes were incubated with500 ng/mL purified L-HEPA for the indicated times. Fractions rich inlysosomes were isolated to detect L-HEPA and active heparanase(A-HEPA) by Western blot. Lysosomal-associated membrane pro-tein 1 (Lamp1) was used as a marker to indicate equal loading oflysosomal protein. B: Myocytes were preincubated with 200 mmol/Lchloroquine before L-HEPA was added to the medium. A-HEPA andL-HEPA in the lysosomal fraction were measured after 3 h. Resultsare the mean 6 SE of three repeated experiments using differentanimals. *Significantly different from 0 time point or untreated myo-cytes, P< 0.05. **Significantly different from 0 time point, P < 0.01.

diabetes.diabetesjournals.org Wang and Associates 2649

internalization of latent heparanase by cardiomyocytesrelies on caveolae to form endocytotic vesicles and dyna-min for fission of these vesicles from the plasma mem-brane. The participation of tyrosine kinase in this processhas also been reported in the syndecan family–mediated

endocytosis (36,37). Clustering is believed to lead to re-distribution of syndecans to lipid rafts, where their cyto-plasmic domain can be phosphorylated by tyrosine kinase(34,38). Recruitment of cortactin to phosphorylated syn-decans promotes actin polymerization at the actin cortex,

Figure 5—Nuclear entry of heparanase increases MMP-9 expression. A: Isolated myocytes were incubated with 500 ng/mL purified latentheparanase (L-HEPA) for the indicated times. Nuclear (Nu) fraction was separated from cytosolic (Cyto) protein to detect heparanase byWestern blot. To validate the purity of the Nu fraction, equal amount of protein from both Cyto and Nu fractions were Western blotted forGAPDH and histone H3 (bottom right panel ). B: Nu fraction from control myocyte was incubated with active heparanase (A-HEPA) orwithout (CON) 200 ng/mL A-HEPA for 8 h, and HAT activity was measured. C: Total RNA was extracted from cardiomyocytes treatedwith 500 ng/mL L-HEPA to measure gene expression of MMP-9 using RT-PCR. D: In another experiment, cardiomyocytes were treated500 ng/mL L-HEPA in the presence or absence of 10 mmol/L anacardic acid for 24 h, and MMP-9 gene expression was determined.Cardiomyocytes without any treatment were used as the control. Significantly different from 0 time point or control, *P < 0.05,**P < 0.01. #Significantly different from L-HEPA in the absence of anacardic acid, P < 0.05.

2650 EC-Cardiomyocyte Cross Talk in Metabolism Diabetes Volume 63, August 2014

thus bringing HSPGs-mediated endocytotic vesicles intothe cells (39). Collectively, our data imply that exoge-nous latent heparanase enters cardiomyocytes through theHSPG-dependent caveolae pathway.

Comparable to cardiomyocytes, cells such as fibro-blasts, which do not express heparanase, can also take uplatent heparanase (17). Interestingly, in fibroblasts, thislatent heparanase can be converted into active hepa-ranase, likely through lysosomal processing (8). Our resultsalso indicate a lysosome enzyme–dependent conversion of

latent to active heparanase in cardiomyocytes. In EC, ac-tive heparanase stored in lysosomes can be secreted inresponse to high glucose or enter the nucleus in the pres-ence of FA (40). Data from the current study demonstratecardiomyocyte nuclear entry of both latent and activeheparanase. Two potential nuclear localization signals(residues 271–277; PRRKTAK and residues 427–430;KRRK) are present in the human heparanase sequence,which could directly mediate its nuclear entry (41). Addi-tionally, Hsp90 has been proposed as a chaperone that

Figure 6—MMP-9 releases LPL from myocyte surface by shedding HSPGs. A: Purified MMP-9 proenzyme was activated by incubationwith 10 mmol/L AMPA at 37°C for 1 h and added to myocyte culture medium at the indicated concentrations. Syndecan-1 and LPL releasedinto the medium were both determined after 30 min (upper and lower right panels, respectively). B: MMP-9 secretion into the medium wasdetermined 24 h after cardiomyocytes were incubated with latent heparanase (L-HEPA). siRNA (100 pmol) for MMP-9 (siRNA) was trans-fected into cardiomyocytes in a six-well plate, and the efficacy of transfection was validated 24 h after transfection by Western blot (rightpanel ). Transfection of a scrambled sequence was used as a negative control (SCR). These transfected cardiomyocytes were incubated inthe presence or absence (CON) of 500 ng/mL L-HEPA for 24 h, and shedding of syndecan-1, together with release of LPL into the medium,were determined (B and C, respectively). Results are the mean 6 SE of three repeated experiments using different animals. Significantlydifferent from control, *P < 0.05, **P < 0.01. #Significantly different from SCR, P < 0.05.

diabetes.diabetesjournals.org Wang and Associates 2651

mediates nuclear translocation of heparanase (42). How-ever, the exact mode of heparanase translocation is cur-rently unknown but of crucial importance given thatnuclear protein transport, especially proteins that serve

as transcriptional regulators, have been implicated in thepathogenesis of certain diseases (43). In HL-60 cells (42)and bovine coronary artery EC (40), active heparanase inthe nucleus, by regulating gene expression, influences cell

Figure 7—After diabetes, EC influence coronary LPL activity through a heparanase/MMP-9 axis. A: RAOEC were incubated with 5.5 mmol/Lnormal glucose (NG), or 25 mmol/L high glucose (HG) Dulbecco’s modified Eagle’s medium (DMEM) for 30 min. The amount of latent (L-HEPA)and active heparanase (A-HEPA) released into the medium was determined by Western blot. Results are the mean 6 SE of three repeatedexperiments. *Significantly different from NG-treated RAOEC, P < 0.05. B: A coculture was done using cardiomyocytes in the well (bottompanel ) with (+EC) RAOEC on the insert (top panel ), and the coculture was exposed to NG or HG for 2 h. The amount of L-HEPA incardiomyocytes was measured by Western blot. Cardiomyocytes were also treated with NG or HG DMEMwithout RAOEC (2EC) as a control.Results are the mean6 SE of three repeated experiments. *Significantly different from NG-treated coculture system, P < 0.05. Animals weremade diabetic by injecting 55 mg/kg STZ and kept for 4 days (D55). C: The amount of A-HEPA in the nuclear fractions of myocytes isolatedfrom control and D55 heart was measured by Western blot. D: MMP-9 gene expression was also determined in these myocytes. E: Sheddingof syndecan-1 into the culture medium over 24 h was detected by ELISA. LPL activity at the coronary lumen of control (CON) and D55 heartswas measured by perfusing the hearts with 5 units/mL heparin. F: Coronary effluents were collected (for 10 s) at different time points over 5 min,and LPL activity in each fraction was determined. The results are presented as area under the curve for heparin-released LPL activity over 5 min.Results are the mean 6 SE of three repeated experiments using different animals. Significantly different from control, *P < 0.05, **P < 0.01.

2652 EC-Cardiomyocyte Cross Talk in Metabolism Diabetes Volume 63, August 2014

differentiation and glucose metabolism, respectively. Like-wise, nuclear entry of active heparanase in myocytes wasaccompanied by an increase in MMP-9 gene expression. Itis possible that by cleaving nuclear HSPGs, active hepa-ranase relieves the suppression that HSPGs have on HATactivity, leading to acetylation of histone proteins re-quired to promote gene transcription of MMP-9 (11,13).This effect has been observed in several cancer cell linesupon transfection of heparanase (13,44). Because MMP-9is capable of shedding syndecan-1 and -4 to generate sol-uble ectodomains carrying HSPGs-bound ligands (45),upregulating its expression could conceivably releaseLPL from the myocyte surface for translocation to thecoronary lumen. Our studies illustrate a novel mechanismby which latent heparanase, by modulating myocyteMMP-9 expression, can have a long-term effect on LPLtrafficking (Fig. 8).

Although multiple sources could explain the presenceof heparanase in cardiomyocytes, we assumed that it issecondary to uptake of latent heparanase after itssecretion from EC. Indeed, high-glucose–induced secre-tion of latent heparanase from EC was efficiently takenup by cardiomyocytes in our coculture system. To extendthese observations to a model of diabetes, we used STZanimals that were hyperglycemic for 72 h (D55) and

observed accumulation of active heparanase in the myo-cyte nucleus, together with a robust increase in MMP-9expression and an accelerated shedding of HSPGs. Theseanimals also exhibited increased coronary LPL activitywith heparin perfusion. Our results suggest that in addi-tion to the high-glucose–induced secretion of active hep-aranase to initiate an immediate transfer of LPL from themyocyte surface to the vascular lumen, activation ofMMP-9 expression by latent heparanase taken up intothe myocytes is pivotal for sustaining trafficking of LPLduring chronic hyperglycemia (Fig. 8). Interestingly, wehave previously shown that when compared with D55,animals with acute hyperglycemia (28.2 6 2.3 mmol/Lcompared with 5.9 6 1.0 mmol/L in controls) inducedby diazoxide (4 h) are unable to maintain their highLPL activity at the vascular lumen after enzyme releaseby heparin (46). Because cardiomyocytes from diazoxidehearts do not increase MMP-9 expression (SupplementaryFig. 2), the sustained increase of coronary LPL seen withD55 can likely be attributed to upregulation of MMP-9 bylatent heparanase.

Overall, results from the present study suggest that inresponse to hyperglycemia, EC-derived heparanase is akey mediator that controls the ability of the cardiomyo-cyte to send LPL to the coronary lumen. Although thisadaptation might be beneficial in the short-term to com-pensate for energy deficit after diabetes, it is potentiallycatastrophic over a protracted duration because superflu-ous FA causes lipotoxicity in the heart, ultimately leadingto diabetic cardiomyopathy (Fig. 8). Inhibiting heparanasesecretion or function could offer a new strategy for man-aging the cardiomyopathy after diabetes.

One limitation of this study is the lack of mousemodels supporting the role of MMP-9 in regulatingcoronary LPL. MMP-9 knockout mice are available andcould be potentially used in this study. However, coronaryLPL activity remains unchanged in both type 1 and type 2diabetic mouse hearts. This could be a consequence ofgenetic adaptation or the excessive heart rate in controlanimals (;600 bpm), which, unlike in humans (parasym-pathetic tone), is under sympathetic control. This permitsprior translocation of LPL from the cardiomyocyte to thecoronary lumen to saturate all of the LPL binding sites.Hence, we have largely depended on rats when studyingcardiac metabolism and LPL.

Funding. The current study is supported by an operating grant from theCanadian Diabetes Association to B.R., and in part by a grant to I.V. from theIsrael Ministry of Health and a grant to G.L. from the National Natural ScienceFoundation. Y.W. and D.Z. are the recipients of Doctoral Student ResearchAwards from the Canadian Diabetes Association.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. Y.W. conceived the idea, generated most ofthe data, and wrote the manuscript. A.P.-L.C., K.N., F.W., D.Z., B.H., N.L.,and A.W. helped with obtaining some of the data. B.R. helped with writingthe manuscript. G.L. provided valuable suggestions and technical support

Figure 8—EC affect cardiomyocyte metabolism through heparanase-induced release of LPL. In the presence of hyperglycemia, there isrelease of both active heparanase (A-hepa) and latent heparanase(L-hepa) from EC. A-hepa can cause an immediate release of myo-cyte HSPG-bound LPL. However, inactive L-hepa can be taken upby cardiomyocytes by HSPGs and converted into A-hepa in thelysosomes. Upon entry into the nucleus, A-hepa, by modulatingHAT activity, promotes gene expression of MMP-9, which cleavesthe ectodomain of myocyte surface HSPGs such as syndecan-1.By this mechanism, HSPGs-bound LPL is liberated from myocytesurface to translocate to the coronary lumen. The ultimate effectwould be increased triglyceride (TG) hydrolysis, over-abundant de-livery of FA to the cardiomyocytes, and eventually lipotoxicity.

diabetes.diabetesjournals.org Wang and Associates 2653

on this work. I.V. assisted with valuable suggestions and the preparation ofthe highly purified latent and active heparanase. B.R. is the guarantor ofthis work and, as such, had full access to all the data in the study and takesresponsibility for the integrity of the data and the accuracy of the dataanalysis.

References1. Pulinilkunnil T, Rodrigues B. Cardiac lipoprotein lipase: metabolic basis fordiabetic heart disease. Cardiovasc Res 2006;69:329–3402. Voshol PJ, Rensen PC, van Dijk KW, Romijn JA, Havekes LM. Effect ofplasma triglyceride metabolism on lipid storage in adipose tissue: studies usinggenetically engineered mouse models. Biochim Biophys Acta 2009;1791:479–4853. Sambandam N, Abrahani MA, Craig S, Al-Atar O, Jeon E, Rodrigues B.Metabolism of VLDL is increased in streptozotocin-induced diabetic rat hearts.Am J Physiol Heart Circ Physiol 2000;278:H1874–H18824. Camps L, Reina M, Llobera M, Vilaró S, Olivecrona T. Lipoprotein lipase:cellular origin and functional distribution. Am J Physiol 1990;258:C673–C6815. O’Brien KD, Ferguson M, Gordon D, Deeb SS, Chait A. Lipoprotein lipase isproduced by cardiac myocytes rather than interstitial cells in human myocardium.Arterioscler Thromb 1994;14:1445–14516. Pillarisetti S, Paka L, Sasaki A, et al. Endothelial cell heparanase modulationof lipoprotein lipase activity. Evidence that heparan sulfate oligosaccharide is anextracellular chaperone. J Biol Chem 1997;272:15753–157597. Gingis-Velitski S, Zetser A, Kaplan V, et al. Heparanase uptake is mediatedby cell membrane heparan sulfate proteoglycans. J Biol Chem 2004;279:44084–440928. Zetser A, Levy-Adam F, Kaplan V, et al. Processing and activation of latentheparanase occurs in lysosomes. J Cell Sci 2004;117:2249–22589. Abboud-Jarrous G, Atzmon R, Peretz T, et al. Cathepsin L is responsible forprocessing and activation of proheparanase through multiple cleavages ofa linker segment. J Biol Chem 2008;283:18167–1817610. He YQ, Sutcliffe EL, Bunting KL, et al. The endoglycosidase heparanaseenters the nucleus of T lymphocytes and modulates H3 methylation at activelytranscribed genes via the interplay with key chromatin modifying enzymes.Transcription 2012;3:130–14511. Chen L, Sanderson RD. Heparanase regulates levels of syndecan-1 in thenucleus. PLoS ONE 2009;4:e494712. Chen J, Zheng D, Shen J, et al. Heparanase is involved in the pro-liferation and invasion of nasopharyngeal carcinoma cells. Oncol Rep 2013;29:1888–189413. Purushothaman A, Hurst DR, Pisano C, Mizumoto S, Sugahara K, SandersonRD. Heparanase-mediated loss of nuclear syndecan-1 enhances histone ace-tyltransferase (HAT) activity to promote expression of genes that drive an ag-gressive tumor phenotype. J Biol Chem 2011;286:30377–3038314. Himelstein BP, Canete-Soler R, Bernhard EJ, Dilks DW, Muschel RJ.Metalloproteinases in tumor progression: the contribution of MMP-9. InvasionMetastasis 1994–1995;14:246–25815. Brule S, Charnaux N, Sutton A, et al. The shedding of syndecan-4 andsyndecan-1 from HeLa cells and human primary macrophages is accelerated bySDF-1/CXCL12 and mediated by the matrix metalloproteinase-9. Glycobiology2006;16:488–50116. Wang F, Kim MS, Puthanveetil P, et al. Endothelial heparanase secretionafter acute hypoinsulinemia is regulated by glucose and fatty acid. Am J PhysiolHeart Circ Physiol 2009;296:H1108–H111617. Nadav L, Eldor A, Yacoby-Zeevi O, et al. Activation, processing and traf-ficking of extracellular heparanase by primary human fibroblasts. J Cell Sci 2002;115:2179–218718. Wang Y, Zhang D, Chiu AP, et al. Endothelial heparanase regulates heartmetabolism by stimulating lipoprotein lipase secretion from cardiomyocytes.Arterioscler Thromb Vasc Biol 2013;33:894–902

19. Blich M, Golan A, Arvatz G, et al. Macrophage activation by heparanase ismediated by TLR-2 and TLR-4 and associates with plaque progression. Arte-rioscler Thromb Vasc Biol 2013;33:e56–e6520. Sambandam N, Chen X, Cam MC, Rodrigues B. Cardiac lipoprotein lipase inthe spontaneously hypertensive rat. Cardiovasc Res 1997;33:460–46821. Payne CK, Jones SA, Chen C, Zhuang X. Internalization and trafficking ofcell surface proteoglycans and proteoglycan-binding ligands. Traffic 2007;8:389–40122. Reilly JF, Mizukoshi E, Maher PA. Ligand dependent and independent in-ternalization and nuclear translocation of fibroblast growth factor (FGF) receptor1. DNA Cell Biol 2004;23:538–54823. Bass MD, Williamson RC, Nunan RD, et al. A syndecan-4 hair trigger ini-tiates wound healing through caveolin- and RhoG-regulated integrin endocytosis.Dev Cell 2011;21:681–69324. Li X, Herz J, Monard D. Activation of ERK signaling upon alternative proteasenexin-1 internalization mediated by syndecan-1. J Cell Biochem 2006;99:936–95125. Marcy AI, Eiberger LL, Harrison R, et al. Human fibroblast stromelysincatalytic domain: expression, purification, and characterization of a C-terminallytruncated form. Biochemistry 1991;30:6476–648326. Mendes O, Kim HT, Stoica G. Expression of MMP2, MMP9 and MMP3 inbreast cancer brain metastasis in a rat model. Clin Exp Metastasis 2005;22:237–24627. Zhang Y, Yeung MN, Liu J, Chau CH, Chan YS, Shum DK. Mapping hep-aranase expression in the spinal cord of adult rats. J Comp Neurol 2006;494:345–35728. Pulinilkunnil T, An D, Ghosh S, et al. Lysophosphatidic acid-mediatedaugmentation of cardiomyocyte lipoprotein lipase involves actin cytoskeletonreorganization. Am J Physiol Heart Circ Physiol 2005;288:H2802–H281029. Vieira AV, Lamaze C, Schmid SL. Control of EGF receptor signaling byclathrin-mediated endocytosis. Science 1996;274:2086–208930. Kim MS, Wang Y, Rodrigues B. Lipoprotein lipase mediated fatty acid de-livery and its impact in diabetic cardiomyopathy. Biochim Biophys Acta 2012;1821:800–80831. Levy-Adam F, Feld S, Suss-Toby E, Vlodavsky I, Ilan N. Heparanase facil-itates cell adhesion and spreading by clustering of cell surface heparan sulfateproteoglycans. PLoS ONE 2008;3:e231932. Vreys V, Delande N, Zhang Z, et al. Cellular uptake of mammalian hep-aranase precursor involves low density lipoprotein receptor-related proteins,mannose 6-phosphate receptors, and heparan sulfate proteoglycans. J Biol Chem2005;280:33141–3314833. Gleizes PE, Noaillac-Depeyre J, Amalric F, Gas N. Basic fibroblast growthfactor (FGF-2) internalization through the heparan sulfate proteoglycans-mediatedpathway: an ultrastructural approach. Eur J Cell Biol 1995;66:47–5934. Williams KJ, Fuki IV. Cell-surface heparan sulfate proteoglycans: dynamicmolecules mediating ligand catabolism. Curr Opin Lipidol 1997;8:253–26235. Wilsie LC, Gonzales AM, Orlando RA. Syndecan-1 mediates in-ternalization of apoE-VLDL through a low density lipoprotein receptor-relatedprotein (LRP)-independent, non-clathrin-mediated pathway. Lipids Health Dis2006;5:2336. Fuki IV, Kuhn KM, Lomazov IR, et al. The syndecan family of proteoglycans.Novel receptors mediating internalization of atherogenic lipoproteins in vitro.J Clin Invest 1997;100:1611–162237. Fuki IV, Meyer ME, Williams KJ. Transmembrane and cytoplasmic domainsof syndecan mediate a multi-step endocytic pathway involving detergent-in-soluble membrane rafts. Biochem J 2000;351:607–61238. Tkachenko E, Simons M. Clustering induces redistribution of syndecan-4core protein into raft membrane domains. J Biol Chem 2002;277:19946–1995139. Chen K, Williams KJ. Molecular mediators for raft-dependent endocytosis ofsyndecan-1, a highly conserved, multifunctional receptor. J Biol Chem 2013;288:13988–13999

2654 EC-Cardiomyocyte Cross Talk in Metabolism Diabetes Volume 63, August 2014

40. Wang F, Wang Y, Zhang D, Puthanveetil P, Johnson JD, Rodrigues B. Fattyacid-induced nuclear translocation of heparanase uncouples glucose metabolismin endothelial cells. Arterioscler Thromb Vasc Biol 2012;32:406–41441. Schubert SY, Ilan N, Shushy M, Ben-Izhak O, Vlodavsky I, Goldshmidt O. Humanheparanase nuclear localization and enzymatic activity. Lab Invest 2004;84:535–54442. Nobuhisa T, Naomoto Y, Okawa T, et al. Translocation of heparanase intonucleus results in cell differentiation. Cancer Sci 2007;98:535–54043. Chahine MN, Pierce GN. Therapeutic targeting of nuclear protein import inpathological cell conditions. Pharmacol Rev 2009;61:358–372

44. Purushothaman A, Chen L, Yang Y, Sanderson RD. Heparanase stimulationof protease expression implicates it as a master regulator of the aggressive tumorphenotype in myeloma. J Biol Chem 2008;283:32628–3263645. Manon-Jensen T, Multhaupt HA, Couchman JR. Mapping of matrix metal-loproteinase cleavage sites on syndecan-1 and syndecan-4 ectodomains. FEBS J2013;280:2320–233146. Pulinilkunnil T, Abrahani A, Varghese J, et al. Evidence for rapid “metabolicswitching” through lipoprotein lipase occupation of endothelial-binding sites.J Mol Cell Cardiol 2003;35:1093–1103

diabetes.diabetesjournals.org Wang and Associates 2655