Matrix metalloproteinase activity is required for activity ...

4 downloads 208014 Views 473KB Size Report
(MMPs) constitute the major class of proteases responsible ..... tic Instruments) and Adobe Photoshop software. Compila- ..... Exercise training-induced coronary ...
Am J Physiol Heart Circ Physiol 279: H1540–H1547, 2000.

Matrix metalloproteinase activity is required for activityinduced angiogenesis in rat skeletal muscle T. L. HAAS,1 M. MILKIEWICZ,2 S. J. DAVIS,1 A. L. ZHOU,2 S. EGGINTON,2 M. D. BROWN,2 J. A. MADRI,1 AND O. HUDLICKA2 1 Department of Pathology, Yale University, New Haven, Connecticut 06515; and 2Department of Physiology, University of Birmingham, Birmingham B15 2TT, United Kingdom Received 22 September 1999; accepted in final form 17 March 2000

Haas, T. L., M. Milkiewicz, S. J. Davis, A. L. Zhou, S. Egginton, M. D. Brown, J. A. Madri, and O. Hudlicka. Matrix metalloproteinase activity is required for activityinduced angiogenesis in rat skeletal muscle. Am J Physiol Heart Circ Physiol 279: H1540–H1547, 2000.—Proteolysis of the capillary basement membrane is a hallmark of inflammation-mediated angiogenesis, but it is undetermined whether proteolysis plays a critical role in the process of activity-induced angiogenesis. Matrix metalloproteinases (MMPs) constitute the major class of proteases responsible for degradation of basement membrane proteins. We observed significant elevations of mRNA and protein levels of both MMP-2 and membrane type 1 (MT1)-MMP (2.9 ⫾ 0.7and 1.5 ⫾ 0.1-fold above control, respectively) after 3 days of chronic electrical stimulation of rat skeletal muscle. Inhibition of MMP activity via the inhibitor GM-6001 prevented the growth of new capillaries as assessed by the capillary-tofiber ratio (1.34 ⫾ 0.08 in GM-6001-treated muscles compared with 1.69 ⫾ 0.03 in control 7-day-stimulated muscles). This inhibition correlated with a significant reduction in the number of capillaries with observable breaks in the basement membrane, as assessed by electron microscopy (0.27 ⫾ 0.27% in GM-6001-treated muscles compared with 3.72 ⫾ 0.65% in control stimulated muscles). Proliferation of capillary-associated cells was significantly elevated by 2 days and remained elevated throughout 14 days of stimulation. Capillary-associated cell proliferation during muscle stimulation was not affected by MMP inhibition (80.3 ⫾ 9.3 nuclei in control and 63.5 ⫾ 8.5 nuclei in GM-6001-treated animals). We conclude that MMP proteolysis of capillary basement membrane proteins is a critical component of physiological angiogenesis, and we postulate that capillary-associated proliferation precedes and occurs independently of endothelial cell sprout formation. proteolysis; endothelium; extensor digitorum longus; chronic electrical stimulation; extracellular matrix

has been accomplished largely through studying the growth of new capillaries during wound healing, cytokine or growth factor-initiated responses, or tumor growth. Inflammation and recruitment of immune cells play a major role in each of these processes through a release of endothelial cell-activating factors, such as cytokines ELUCIDATION OF THE STAGES OF ANGIOGENESIS

Address for reprint requests and present address of T. Haas: York Univ., Dept. of Kinesiology and Health Sciences, 4700 Keele St., Toronto, ON Canada M3J 1P3 (E-mail: [email protected]). H1540

and proteases. In the current models of inflammatorymediated angiogenesis, it is thought that the capillary endothelial cells are activated first, leading to the production of proteases that degrade basement membrane proteins surrounding the capillary. Proteolysis allows endothelial cell sprouts to migrate into the interstitial matrix, but sprout elongation also requires proliferation of the endothelial cells at the base of the sprout. A lumen is formed, and the new capillary matures by recruitment of pericytes and reconstruction of a complete basement membrane (1, 8, 30). Activity-induced angiogenesis occurs in adult skeletal and cardiac muscles, and this response is essential for maintaining adequate tissue function in the face of chronic increases in metabolic demand (18, 31). The predominant involvement of hemodynamic and mechanical stimuli, and the lack of evidence for a significant inflammatory response (18, 23), distinguish activity-induced angiogenesis from angiogenesis occuring during tumor growth or wound healing. Chronic indirect electrical stimulation of the muscle provides a good model of activity-induced angiogenesis because hemodynamic and mechanical stimuli are thought to provide the major angiogenic stimuli, whereas an inflammatory component is minimal (10, 27). Evidence to date suggests that not only the initial stimuli involved but also the course of events that occurs in activityinduced angiogenesis may differ significantly from the course of inflammatory-mediated angiogenesis. First, in skeletal muscles of which activity is increased by chronic electrical stimulation, endothelial cell proliferation occurs at a very early stage (17) that precedes detectable sprout formation (16) rather than being associated with sprout elongation. Second, proteolysis of the basement membrane, detected by transmission electron microscopy as breaks in the basal lamina, has been observed very rarely (16), whereas, during tumor growth and wound healing, venules and capillaries often are completely devoid of electron-dense basal lamina (9, 15, 30, 34). These observations raise key questions as to whether proteolysis is essential in the process of physiological angiogenesis and whether, in The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

0363-6135/00 $5.00 Copyright © 2000 the American Physiological Society

http://www.ajpheart.org

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS

this case, endothelial cell proliferation precedes or follows proteolysis and/or sprout formation. Proteolysis of basement membrane proteins is most frequently performed by matrix metalloproteinases (MMPs), a diverse family of proteases each with a specificity for multiple matrix proteins (3). MMP activity is known to be involved in tumor and inflammationmediated angiogenesis (11, 31), but it is not known whether MMP production and/or activity is modulated during activity-induced angiogenesis. Two particular MMPs, MMP-2 (gelatinase A) and membrane type 1 (MT1)-MMP, were shown to play an essential role in an endothelial cell culture model of extracellular matrixinduced angiogenesis (13) and have been implicated in tumor angiogenesis (20, 33, 35). Coordinated expression of the two MMPs is frequently observed, and MT1-MMP is known to activate MMP-2 (28, 33). On the basis of these findings, we used the chronic stimulation model to address the questions as to whether MMP-2 and MT1-MMP are upregulated during activity-induced angiogenesis and whether MMP activity is a requisite for this type of adaptive angiogenesis. We found significantly increased production of both MMPs as a result of chronic muscle stimulation and that chronic inhibition of MMP activity prevented the growth of new capillaries. This study defines a critical role for MMP activity in the activity-induced formation of new capillaries and suggests that endothelial cell proliferation occurs independently of proteolysis and sprout formation. METHODS

Animal model. Rat hindlimb skeletal muscles [ankle flexors: the tibialis anterior (TA); extensor digitorum longus (EDL); and extensor hallucis proprius (EHP) muscles] were subjected to increased activity induced by chronic electrical stimulation as previously described (27). Briefly, animals were implanted unilaterally with electrodes in the vicinity of the common peroneal nerve, providing stimulation of the TA, EDL, and EHP muscles via the peroneal nerve. The implanted electrodes never came in direct contact with the muscle groups studied and thus did not cause trauma to the analyzed muscles. The implantation was performed under fluothane anesthesia in accordance with the United Kingdom Animals (Scientific Procedures) Act of 1986. Animals received analgesic (buprenorphine, 2.5 mg in 0.1 ml) twice daily for the first 2 days after the operation. Electrical stimulation, 8 h/day for up to 14 days, commenced ⬃24 h after the operation using a frequency of 10 Hz with 0.3-ms pulse width and voltage sufficient to elicit strong contractions (up to 6 V) without causing discomfort. One group (n ⫽ 4) of 7-day-stimulated animals were given the MMP inhibitor N-[2(R)-2-(hydroxamido carbonylmethyl)4-methylypentanoyl]-L-tryptophane methylamide, referred to as GM-6001 (AMS Scientific). This particular MMP inhibitor was selected based on the commercial availability and the numerous studies reporting effective use of GM-6001 in a range of animal models (11, 12, 36). Each animal received daily intraperitoneal injections of 30 mg of the drug (dissolved in 150 ␮l of DMSO) during the 7 days of stimulation, with the last injection received 24 h before euthanization. This dose corresponds to the previously reported doses that resulted in the inhibition of MMP activity without reported

H1541

side effects (2, 36). A second group (n ⫽ 4) of 7-day-stimulated animals received daily intraperitoneal injections of 150 ␮l DMSO. Control unstimulated rats received daily doses of DMSO (n ⫽ 4) or GM-6001 (n ⫽ 4) for 7 days at the same doses used for stimulated rats. Neither treatment regimen exerted any apparently adverse effects on gross body weight, wet muscle mass, or muscle ultrastructure. Tissue preparation. Muscles were excised ⬃16 h after termination of stimulation under pentobarbital sodium anesthesia (60 mg/kg). The middle portion of each EDL muscle was mounted in optimum cutting tissue medium, immediately frozen in isopenthane cooled in liquid nitrogen, and used for immunohistology. The remainder of the muscle, used for RNA and protein extraction, was weighed and snap frozen in liquid nitrogen. In a second set of experiments, strips from the middle portion of the EDL muscle or the whole EHP muscle were fixed by superfusion with 2.5% glutaraldehyde in 0.05 M phosphate buffer (pH 7.2, ⬃350 mosM) for electron microscopy. Electron microscopy. Muscle strips were sliced into four blocks, each 1–2 mm3, and placed in fresh fixative for 24 h at 4°C. Blocks were then treated as described previously (40). Sections were viewed using a Jeol 100 CXII electron microscope and photographed onto 35-mm film (magnifications ⫻1,700 to ⫻2,500) to score the frequency of vessels with disturbed basement membranes, defined as any region of the normally contiguous capillary basement membrane that appeared to be broken or missing. Blocks from each animal were analyzed (control unstimulated rats, n ⫽ 4 blocks; 7-day-stimulated rats, n ⫽ 4 blocks; DMSO-treated, 7-daystimulated rats, n ⫽ 4 blocks; and GM-6001-treated, 7-daystimulated rats, n ⫽ 4 blocks). Approximately 40 capillaries per section were selected using systematic random sampling, where a random point on each section was chosen, and micrographs of the 10 closest capillaries were taken. Ten additional capillaries were photographed at each corner of a predefined lattice. Two or three sections from each block were analyzed in this manner. Representative capillaries also were photographed (magnifications ⫻4,800 to ⫻29,000) using quarter-plate film. Histology and immunohistochemistry. Serial cryostat sections (8 ␮m) were stained for alkaline phosphatase using an indoxyltetrazolium method to depict all anatomically present capillaries (41) and for proliferating cell nuclear antigen (PCNA; clone pc10, DAKO) using standard immunohistochemical techniques. The counting of capillaries and PCNApositive nuclei was performed using an Olympus microscope equipped with a drawing arm. Two to four fields (0.20 mm2 each at ⫻200 magnification) were used for counting capillaries in each muscle, and four fields (0.05 mm2 at ⫻400) were used for counting PCNA-positive nuclei in each muscle. Three to four muscles were analyzed for each intervention. The same fields identified in the alkaline phosphatasestained serial sections were used to assess whether the position of PCNA-positive nuclei colocalized with the alkaline phosphatase staining and, hence, the site of a capillary. This enabled calculation of the density of proliferating nuclei linked with capillaries (PCNAcap) or other interstitially located nuclei not associated with capillaries (PCNAint) (measured in PCNAcap/mm2 or PCNAint/mm2, respectively). The number of muscle fibers also was counted, and capillary supply was expressed as the capillary-to-fiber ratio (C/F). MMP analysis. Frozen EDL muscles were powdered and then homogenized in TRIzol (GIBCO BRL), and RNA was isolated. Total cellular RNA (20 ␮g) from each sample was size fractionated and Northern blotted as previously described using cDNA probes specific for rat MMP-2 (from D.

H1542

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS

Table 1. Proliferation precedes increase in capillary supply during chronic stimulation Days of Stimulation

C/F ratio PCNAcap PCNAint

Control

2

3

7

8

14

1.29 ⫾ 0.04 5.1 ⫾ 3.4 10.3 ⫾ 1.3

1.43 ⫾ 0.09 107.8 ⫾ 4.4* 46.2 ⫾ 2.2*

1.43 ⫾ 0.06 65.5 ⫾ 20* 46.6 ⫾ 19*

1.67 ⫾ 0.05* 61.6 ⫾ 7.7* 60.4 ⫾ 5.1*

1.86 ⫾ 0.11* 38.5 ⫾ 3.8* 19.5 ⫾ 10

2.07 ⫾ 0.12* 44.0 ⫾ 4.2* 14.0 ⫾ 3.4

Values are means ⫾ SE and are expressed as number of positive cells per millimeter squared; n ⫽ 3 rats. C/F ratio, capillary-to-fiber ratio; PCNAcap and PCNAint, proliferating cell nuclear antigen (PCNA) associated with nuclei linked to capillaries and interstially located nuclei not linked to capillaries, respectively. *P ⬍ 0.05 compared with control.

Lovett, Univ. of CA San Francisco) and mouse MT1-MMP (from S. Apte, Washington Univ., St. Louis, MO) (13). Proteins were isolated by homogenization of the powdered EDL muscle in Triton X-100-containing lysis buffer. The protein concentration of the homogenate supernates was quantitated by bicinchoninic acid solution (Pierce). Western blotting using anti-MMP-2 and anti-MT1-MMP (Chemicon) antibodies was performed on 20 ␮g of each sample, using methods previously described (13). Gelatin zymography analysis was performed on 15 ␮g of each sample as previously described (13). Films and zymographic gels were scanned, and band densitometry performed using Biomax software. Northern blots were normalized to 28S rRNA, and Western blots were normalized to vimentin to control for loading. Immunofluorescence staining was performed on 10-␮m-thick frozen sec-

tions of EDL and TA muscles using polyclonal antibodies against MMP-2 (Chemicon) and MT1-MMP (Chemicon; provided by S. Weiss, Univ. of MI, Ann Arbor, MI), detected using Texas red secondary antibodies, and viewed with a Zeiss microscope using a ⫻40 objective. Images were captured using a SPOT charge-coupled device camera (Diagnostic Instruments) and Adobe Photoshop software. Compilations were made using Freehand software. Statistical analysis. Data are presented as means ⫾ SE, with samples from a minimum of three rats included in each data point. Capillary density and proliferating cell data were analyzed by factorial ANOVA, with significance ascribed as P ⬍ 0.05. MMP mRNA and protein levels were analyzed using Student’s unpaired t-test, assuming equal variance among samples, with significance also ascribed as P ⬍ 0.05.

Fig. 1. Basement membrane disturbances in chronically stimulated muscle. A: electron micrograph of a capillary from a control unstimulated extensor hallucis proprius (EHP) muscle showing thin endothelial cell layer with smooth luminal and abluminal surfaces surrounded by an intact basement membrane (open arrows) that appears as a thin electron-dense band encircling the capillary. The cellular process of a pericyte (*) is apparent on the inside of the capillary basement membrane (bar ⫽ 0.5 ␮m). B: capillary from 3-day-stimulated EHP muscle showing abluminal cytoplasmic projection into the interstitium (sprout). The basement membrane is intact around the capillary (open arrow) but absent at the tip of the sprout (filled arrows). The adjacent cell (*) likely is a fibroblast (bar ⫽ 0.2 ␮m).

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS RESULTS

Chronic electrical stimulation of EDL muscle induced a significant 30% increase in the C/F ratio after 7 days of stimulation, and this was increased further at the 14-day-stimulation time point (Table 1), which was similar to previous findings (27). The increased C/F ratio strongly indicates a growth of new vessels because the methodology used identifies all anatomically present capillaries. The number of proliferating capillary-associated nuclei was increased 20-fold in muscles that had been stimulated for only 2 days, and, although this number gradually decreased, it remained significantly elevated at 14 days of stimulation (Table 1). In contrast, the number of proliferating interstitially located nuclei (those not associated with capillaries, most likely fibroblasts) was increased four- and sixfold at 2 and 7 days of stimulation, respectively, but returned to control levels by 14 days of stimulation. Basement membrane disruptions were detected by electron microscopy as breaks in the normally contiguous basement membrane surrounding the capillary endothelium. In control unstimulated muscles, there were no capillaries with a sprouting endothelium or a loss of basement membrane (Fig. 1A). In 3- and 7-daystimulated muscles, sprouting endothelial cells were apparent (Fig. 1B). Basement membrane disruption did not occur around the parent capillaries except in locations showing signs of budding (irregularity of the endothelial cell abluminal surface and microfilament reorganization). After 3 days of stimulation, 1.2 ⫾ 0.7% of all capillaries examined exhibited budding sprouts with a disrupted basement membrane. After 7 days of

H1543

stimulation, 2.5 ⫾ 1.0% (P ⬍ 0.05 compared with control) of all capillaries had budding sprouts with a detectable loss of basement membrane investiture. Examination of muscle sections by both light and electron microscopy showed no evidence for accumulation of inflammatory cells surrounding the capillaries or in the interstitial spaces of stimulated muscles, nor was there morphological evidence of muscle damage, a finding consistent with previous observations using the same experimental system (5, 16, 27). MMP-2 mRNA was detectable in control and contralateral unstimulated EDL muscles but was increased significantly by 3 days of stimulation. Levels peaked at 3 days of stimulation, remained elevated at 8 days of stimulation, and were not significantly higher than control levels by 14 days of stimulation (Fig. 2A and Table 2). The MMP-2 protein level, as assessed by Western blotting, increased slightly over the time course of stimulation, and a greater proportion of MMP-2 was detectable in the 62-kDa (active) form (Fig. 2B). MT1-MMP mRNA, also detectable in control and unstimulated muscles, was increased significantly after 4 days of stimulation, with levels continuing to increase until 7 days of stimulation, and remaining elevated after 14 days of stimulation (Fig. 2A and Table 2). The MT1-MMP protein levels increased over the time course of stimulation and also remained elevated at 14 days of stimulation (Fig. 2B). Immunofluorescent staining localized the MMP-2 and MT1-MMP proteins to capillaries as well as to larger arterioles/ venules. Although detectable in control unstimulated muscles (Fig. 2, C and E), higher intensities of both

Fig. 2. Matrix metalloproteinase (MMP)-2 and membrane type 1 (MT1)-MMP mRNA and protein increase with chronic stimulation. A: 2 representative Northern blots show a time courserelated increase for MMP-2 and MT1-MMP mRNA in stimulated extensor digitorum longus (EDL) muscle, with minimal levels detectable in unstimulated control (day 0) or contralateral muscles. Detection of the ribosomal 28S RNA [by radioactive probe (left) and by ethidium bromide staining (right)] was used to normalize band intensities according to loading. B: Western blotting detects increases in MMP-2 and MT1-MMP protein over a time course of EDL stimulation. A longer exposure of the MMP-2 blot shows that the amount of active MMP-2 (62 kDa) also increases over time of stimulation. Vimentin staining was used to demonstrate equivalent protein loading in each lane. C-F: immunofluorescence of MMP-2 (C and D) and MT1-MMP (E and F) shows much higher intensities of staining in 7-day-stimulated EDL (D and F) than in corresponding unstimulated contralateral muscles (C and E). Arrows, staining that appears to colocalize with capillaries (scale bars ⫽ 5 ␮m).

H1544

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS

Table 2. Increased MMP-2 and MT1-MMP during chronic stimulation Days of Stimulation Control

2

3

MMP-2 MT1-MMP

1.1 ⫾ 0.1 1.0 ⫾ 0.1

1.4 ⫾ 0.2 1.3 ⫾ 0.2

2.9 ⫾ 0.7* 1.5 ⫾ 0.1*

Total MMP-2 % Active MMP-2 100-kDa MMP

0.9 ⫾ 0.1 29 ⫾ 3 0.9 ⫾ 0.1

NT NT NT

NT NT NT

4

7

8

14

2.1 ⫾ 0.5 1.8 ⫾ 0.4*

2.3 ⫾ 0.1 2.4 ⫾ 0.1*

1.7 ⫾ 0.3* 1.6 ⫾ 0.3*

1.4 ⫾ 0.2 2.4 ⫾ 0.8*

1.2 ⫾ 0.1 47 ⫾ 2* 2.5 ⫾ 0.7*

1.5 ⫾ 0.1 63 ⫾ 5* 3.1 ⫾ 0.5*

NT NT NT

1.3 ⫾ 0.1* 52 ⫾ 1* 3.7 ⫾ 0.4

mPNA

Protein

Values are means ⫾ SE. mRNA and protein are expressed as relative densitometry compared with control, except for % active matrix metalloproteinase (MMP-2), which is expressed as a ratio of 62 kDa to 72 kDa for each time point. MT1-MMP, membrane type1-MMP. NT, not tested. For control unstimulated, 2-day-, and 3-day-stimulated rats, n ⫽ 5 rats; for 4-day-, 7-day-, and 8-day-stimulated rats, n ⫽ 4 rats; and for 14-day-stimulated rats, n ⫽ 3 rats. * P ⬍ 0.05 compared with control.

MMP-2 and MT1-MMP staining were observed in 7-day-stimulated muscles (Fig. 2, D and F). The pattern of staining observed is consistent with production of MMP-2 and MT1-MMP by vascular and/or perivascular cells. Gelatin zymography was used to quantitate the levels of pro- and active MMP-2 protein. The total MMP-2 protein levels increased 1.5-fold over the duration of stimulation, and, as was observed in Western blots, the percentage of active MMP-2 also increased twofold (Fig. 3 and Table 2). Stimulation also significantly increased the production of a 100-kDa MMP (Fig. 3 and Table 2). Interestingly, this high molecular weight MMP does not appear to correspond to rodent MMP-9 (gelatinase B), on the basis of the comparison of the molecular weight of this protein with purified murine MMP-9 using gelatin zymography (data not shown). The broad-specificity MMP inhibitor GM-6001 inhibited new capillary formation. In GM-6001-treated 7-day-stimulated animals, the C/F ratio in EDL muscle was not increased significantly above the C/F ratio for vehicle-treated unstimulated muscles (1.34 ⫾ 0.08 and 1.24 ⫾ 0.06, respectively) (Fig. 4). In contrast, EDL muscle from 7-day-stimulated rats treated with vehicle showed the expected significant increase in the C/F ratio (1.69 ⫾ 0.03). By electron microscopy, extending capillary sprouts were not detectable in EHP muscles of GM-6001-treated rats. Furthermore, the capillaries

Fig. 3. Increased production of multiple MMPs is observable by gelatin zymography. The time course of stimulation results in increases in MMP-2 protein, both latent (72 kDa) and active (62 kDa) forms, detectable as bands of clearing within the stained gel. A highmolecular-weight (MW) MMP-2 (⬃100 kDa) also increases during EDL stimulation. MW standards (Std) are indicated on the left side of the gel.

in GM-6001-treated muscles had significantly fewer focal regions (budding sprouts) with a disturbed basement membrane compared with capillaries in DMSOtreated muscles (0.27 ⫾ 0.27 vs. 3.72 ⫾ 0.65%, respectively; P ⬍ 0.05, n ⫽ 3 rats). Capillaries in GM-6001treated muscles also appeared to be invested with a thicker than normal basement membrane (Fig. 5). Endothelial cell activation, marked by cell thickening and the increased appearance of synthetic organelles, is thought to occur early in the angiogenic response. There was no observed difference between DMSO- and GM-6001-treated animals in the number of visibly activated endothelial cells, implying that the responsiveness of the endothelium to the angiogenic stimulus was not altered by inhibition of MMP activity. Similarly, despite the lack of new sprout formation associated with GM-6001 treatment, chronic stimulation induced proliferation of both capillary-associated and interstitial cells, with no significant differences between GM-6001- and vehicle-treated rats (Table 3). Thus MMP inhibition blocked stimulation-induced increases in the C/F ratio but did not exert an inhibitory effect on cellular activation and proliferation.

Fig. 4. Mean capillary-to-fiber (C/F) ratio does not increase in EDL muscle treated with MMP inhibitor GM-6001. Comparisons were made of C/F ratio in EDL from stimulated and nonstimulated rats treated with DMSO (vehicle) or GM-6001. C/F ratio is expressed as means ⫾ SE (n ⫽ 4 rats). *Only C/F ratio significantly different (P ⬍ 0.05) from nonstimulated, DMSO-treated animals.

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS

H1545

Fig. 5. Abluminal sprouting does not occur in GM-6001-treated muscle. A: capillary from DMSO-treated, 7-daystimulated EHP muscle has activated endothelium that is thick, has an irregular luminal surface, and increased cytoplasmic organelle content. Proliferation is indicated by the increased no. of cells/circumference (4 instead of an average of 2 in normal capillaries). The basement membrane (appearing as a thin electron-dense band) is intact around most of the capillary (open arrows), but there is a focal area of amorphous basement membrane (closed arrows) at what appears to be a budding endothelial sprout (bar ⫽ 0.5 ␮m). B: capillary from GM-6001-treated, 7-day-stimulated EHP muscle has activated endothelium similar to A with thick cells and increased number of cytoplasmic organelles. However, the basement membrane is completely intact (filled arrows) with no indications of endothelial cell sprouting. There is an abundance of interstitial collagen fibrils (*) surrounding the capillary (L, lumen; bar ⫽ 0.5 ␮m). DISCUSSION

Chronic electrical stimulation of hindlimb muscles induces a robust angiogenic response detectable as an increase in the C/F ratio. We present the novel findings that the process of activity-stimulated angiogenesis in skeletal muscle involved an increased production of MMP-2 and MT1-MMP and, furthermore, that MMP activity was a requisite for new capillary formation. Notably, capillary-associated cell proliferation preceded and occurred independently of MMP activity and new capillary growth. Although production of proteolytic enzymes is well known to be associated with inflammatory responses, the involvement and the requirement for proteolysis of basement membrane components during activity-induced angiogenesis previously had not been tested. We first demonstrated that indications of the regional absence of the basement membrane, possibly due to proteolysis, are detectable at sites of budding endothelial cells but not observed around the parent capillaries (Fig. 1). A similar observation was reported, but not quantified, in an earlier study (16). Notably, we found that inhibition of MMP activity through use of the broad-spectrum MMP inhibitor GM-6001 prevented Table 3. Proliferation is not affected by inhibition of MMP activity Unstimulated

PCNAcap PCNAint

7-Day Stimulated

DMSO

GM-6001

DMSO

GM-6001

16.4 ⫾ 1.85 7.7 ⫾ 0

11.6 ⫾ 2.7* 6.7 ⫾ 2.4*

80.3 ⫾ 9.3 44.9 ⫾ 5.0

63.5 ⫾ 8.5* 47.2 ⫾ 5.1*

All values expressed as means ⫾ SE; n ⫽ 4 rats. * Values not significantly different from DMSO-treated animals.

the growth of capillaries, as assessed by the C/F ratio. The failure of capillary growth in these muscles may be attributable to the lack of visible sprout formation (Fig. 5), implying that proteolysis of the basement membrane by MMPs is a critical element in sprout formation. The significant increases in MMP-2 mRNA, MT1MMP mRNA, and protein detectable after 4 days of stimulation are consistent with these particular proteases playing a major role in the process of angiogenesis. The involvement of MMP-2 and MT1-MMP has been demonstrated previously (13, 20) using cell culture and tumor growth models but has not been shown for activity-induced angiogenesis. It is generally considered that production of MT1-MMP is required for activation of MMP-2 (28), and, in our study, we observed a correlation between increased levels of MT1MMP and increased processing of MMP-2 to the active form (Fig. 2 and Table 2). Although it is clear from our data that MMP-2 and MT1-MMP are produced, we do not yet know whether the distribution of these MMPs correlates with the electron microscope-visualized pattern of basement membrane disruption. However, proteolysis and the localization of the membrane-bound MT1-MMP and MMP-2 (28) to the extending tips of migrating cells have been noted previously (26, 29), indicating that proteolysis can be finely localized to subregions of the cellular membrane. Further studies may provide evidence as to whether the tight control of proteolytic enzymes is required for the appropriate organization of capillary networks. The MMP-2 and MT1-MMP proteins were detected in capillaries as well as in larger blood vessels using immunofluorescence, but the cell type(s) producing these proteases remains to be determined. Both pro-

H1546

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS

teases are produced not only by endothelial cells but also by vascular smooth muscle cells, fibroblasts, and (perhaps) pericytes, and it is quite possible that during chronic stimulation more than one of these cell types participates in MMP production. It also is possible that activity of MMPs other than MMP-2 and MT1-MMP, such as the 100-kDa MMP protein detected by zymography, may be required for the process of sprout formation. Similarly, it has been reported (6) that endothelial cell-stimulating angiogenesis factor (ESAF), which can activate prostromelysin-1 and procollagenase (25), is elevated during chronic electrical stimulation and thus may play a role in the initiation of proteolysis. Because there currently are no inhibitors specific for particular MMPs, it is not yet possible to test the roles of individual MMPs using this model. Early increases in proliferation were seen both for capillary-associated and interstitial nuclei, as has been noted previously using 5-bromo-2⬘-deoxyuridine (BrdU) labeling (17). Interestingly, PCNAcap staining remained elevated during the 14-day stimulation period (suggesting early as well as continuing phases of proliferation), whereas PCNAint increased rapidly but declined to baseline by 14 days of stimulation (suggesting only an early proliferative response in the interstitial cells). PCNAcap may represent not only endothelial cell proliferation but may also include those fibroblasts or pericytes in close association with the capillaries; a previous study (39), utilizing in vivo BrdU labeling coupled with electron micrographic analysis, reported equal numbers of BrdU-positive endothelial cells and capillary-associated fibroblasts after 2 days of stimulation but almost entirely endothelial cell labeling at day 7. The significant finding of the present study is that the MMP inhibitor GM-6001 did not inhibit cell proliferation (either in capillary-associated or interstitial cells), arguing that MMP activity is not required for, but rather occurs downstream of, cellular activation and proliferation processes. This finding is consistent with observations made by a previous study (2) that reported GM-6001 successfully inhibited the migration of vascular smooth muscle cells in vivo without affecting the proliferation rate. Thus, on the basis of the early increases in capillary-associated cell proliferation (Table 1) and because GM-6001 successfully inhibited sprout formation without inhibiting cell proliferation, we postulate that endothelial (and possibly fibroblast or pericyte) cell activation and proliferation temporally precede sprout formation in this model of activityinduced angiogenesis. Increased blood flow to the contracting muscle increases shear stress and transmural wall pressure, and the contracting myocytes themselves cause stretch and deformation of the interstitial proteins and the neighboring capillaries. These stimuli are thought to play major roles in the initiation of angiogenesis during chronic muscle activity. The intracellular signaling events coupling these stimuli with endothelial cell activation in vivo have not been defined, although several mediators are likely to be involved. The production of

vascular endothelial growth factor, stimulated either by hypoxia or by mechanical stretch (4, 24, 32), has been correlated with the initiation of angiogenesis in coronary and skeletal muscles (37, 38) and is a strong candidate for signaling the initiation of endothelial cell proliferation. Hemodynamic and mechanical forces also are known to stimulate the production and release of nitric oxide (7), and the production of nitric oxide has been shown to play an important role in activity-induced angiogenesis (19), indicating a potential involvement of this signaling molecule as well. Our report documents a time course of early increases in proliferation with a temporally delayed increase in production of MMP-2 and MT1-MMP. This suggests that, although proliferation likely is mediated by immediate-early growth response genes (21), enhanced production of MMPs may lie downstream of the initial signaling events. Endothelial cell transcriptional upregulation of MT1-MMP in response to changes in the extracellular matrix environment is mediated by the transcription factor egr-1 (14). Egr-1 is known to be upregulated in endothelial cells as a result of both fluid shear stress and mechanical stretch (22), thus implicating this transcription factor as a likely candidate for inducing the production of MT1-MMP during activityinduced angiogenesis in vivo. This study points to several potential differences between the processes of activity-induced angiogenesis and inflammation-mediated angiogenesis. Electron microscopic observations indicate that proteolysis during adaptive angiogenesis appears localized to the sprouting tips of endothelial cells. In contrast, proteolysis is reported to occur around tumor capillaries and the sprouts of tumor capillaries, both of which frequently appear entirely devoid of the electron-dense basement membrane (30). Proliferation has been noted to occur only in sprouts during inflammation-mediated angiogenesis, whereas we have seen clear evidence of capillary-associated proliferation before the formation of sprouts. Indeed, this proliferation is maintained in the absence of sprout formation. On the basis of these differences, it is likely that the intracellular signaling mechanisms driving angiogenesis also differ, perhaps as a result of disparate initiating stimuli. Further analysis of the signaling pathways responsible for initiating the responses of proliferation and proteolysis during activity-induced angiogenesis will be useful in understanding the physiological maintenance of the capillary number. Identification of the signaling pathways used for the hemodynamic or mechanical stimulation of angiogenesis may have a therapeutic use in some pathologies, when ischemia alone fails to elicit compensatory capillary growth. This work was supported by National Heart, Lung, and Blood Institute Grant F32-HL-09983 and by a Research Career Enhancement Award from the American Physiological Society (to T. L. Haas); National Heart, Lung, and Blood Institute Grant R01-HL-51018 (to J. A. Madri); and a donation from Dr. J. Barclay (to O. Hudlicka).

MMPS AND ACTIVITY-INDUCED ANGIOGENESIS REFERENCES 1. Ausprunk D and Folkman J. Migration and proliferation of endothelial cells in preformed and newly formed blood vessels during tumor angiogenesis. Microvasc Res 14: 53–65, 1977. 2. Bendeck MP, Irvin C, and Reidy MA. Inhibition of matrix metalloproteinase activity inhibits smooth muscle cell migration but not neointimal thickening after arterial injury. Circ Res 78: 38–43, 1996. 3. Birkedal-Hansen H. Proteolytic remodeling of extracellular matrix. Curr Opin Cell Biol 7: 728–735, 1995. 4. Breen EC, Johnson EC, Wagner H, Tseng HM, Sung LA, and Wagner PD. Angiogenic growth factor mRNA responses in muscle to a single bout of exercise. J Appl Physiol 81: 355–361, 1996. 5. Brown MD, Cotter M, Hudlicka O, and Vrbova G. The effect of different patterns of muscle activity on capillary density, mechanical properties and structure of slow and fast rabbit muscles. Pflu¨gers Arch 361: 241–250, 1976. 6. Brown MD, Hudlicka O, Makki RF, and Weiss JB. Low molecular mass endothelial cell stimulating angiogenic factor in relation to capillary growth induced in rat skeletal muscle by low-frequency electrical stimulation. Int J Microcirc Clin Exp 15: 111–116, 1995. 7. Busse R and Fleming I. Pulsatile stretch and shear stress: physical stimuli determining the production of endotheliumderived relaxing factors. J Vasc Res 35: 73–84, 1998. 8. Chiquet M, Matthisson M, Koch M, Tannheimer M, and Chiquet-Ehrimann R. Regulation of extracellular matrix synthesis by mechanical stress. Biochem Cell Biol 74: 737–744, 1996. 9. Cliff WJ. Observations in healing tissue: a combined light and electron microscopic investigation. Trans R Soc Lond Ser B246: 305–325, 1963. 10. Dawson JM and Hudlicka O. Can changes in microcirculation explain capillary growth in skeletal muscle? Int J Exp Pathol 74: 65–71, 1993. 11. Galardy RE, Grobelny D, Foellmer HG, and Fernandez LA. Inhibition of angiogenesis by the matrix metalloproteinase inhibitor N-[2(R)-2-(hydroxyamidocarbonylmethyl)-4-methylpentanoyl]-L-tryptophan methylamide. Cancer Res 54: 4715–4718, 1994. 12. Graesser D, Mahooti S, Haas T, Davis S, Clark RB, and Madri JA. The interrelationship of ␣4-integrin and matrix metalloproteinase-2 in the pathogenesis of experimental autoimmune encephalomyelitis. Lab Invest 78: 1445–1458, 1998. 13. Haas TL, Davis SJ, and Madri JA. Three-dimensional type I collagen lattices induce coordinate expression of matrix metalloproteinases MT1-MMP and MMP-2 in microvascular endothelial cells. J Biol Chem 273: 3604–3610, 1998. 14. Haas TL, Stitelman D, Davis SJ, Apte SS, and Madri JA. Egr-1 mediates extracellular matrix-driven transcription of membrane type 1 matrix metalloproteinase in endothelium. J Biol Chem 274: 22679–22685, 1999. 15. Hammersen F, Endrich B, and Messmer K. The fine structure of tumor blood vessels. I. Participation of nonendothelial cells in tumor angiogenesis. Int J Microcirc Clin Exp 4: 31–43, 1985. 16. Hansen-Smith FM, Hudlicka O, and Egginton S. In vivo angiogenesis in adult rat skeletal muscle: early changes in capillary network architecture and ultrastructure. Cell Tissue Res 286: 123–136, 1996. 17. Hudlicka O. Is physiological angiogenesis in skeletal muscle regulated by changes in microcirculation? Microcirculation 5: 7–23, 1998. 18. Hudlicka O, Brown MD, and Egginton S. Angiogenesis in skeletal and cardiac muscle. Physiol Rev 72: 369–417, 1992. 19. Hudlicka O, Brown MD, and Silgram H. Inhibition of capillary growth in chronically stimulated rat muscles by NG-nitroL-arginine, nitric oxide synthase inhibitor. Microvasc Res 59: 45–51, 2000. 20. Itoh T, Tanioka M, Yoshida H, Yoshioka T, Nishimoto H, and Itohara S. Reduced angiogenesis and tumor progression in gelatinase A-deficient mice. Cancer Res 58: 1048–1051, 1998.

H1547

21. Kerr LD, Inoue JI, and Verma IM. Signal transduction: the nuclear target. Curr Opin Cell Biol 4: 496–501, 1992. 22. Khachigian LM and Collins T. Inducible expression of egr-1dependent genes. A paradigm of transcriptional activation in vascular endothelium. Circ Res 81: 457–461, 1997. 23. Laughlin MH and McAllister RM. Exercise training-induced coronary vascular adaptation. J Appl Physiol 73: 2209–2225, 1992. 24. Li J, Hampton T, Morgan JP, and Simons M. Stretchinduced VEGF expression in the heart. J Clin Invest 100: 18–24, 1997. 25. McLaughlin B and Weiss JB. Endothelial cell stimulating angiogenesis factor (ESAF) activates progelatinase A (72-kDa type IV collagenase), prostromelysin 1 and procollagenase and reactivates their complexes with tissue inhibitors of metalloproteinases: a role for ESAF in noninflammatory angiogenesis. Biochem J 317: 739–745, 1996. 26. Monsky WL, Kelly T, Lin CY, Yeh Y, Stetler-Stevenson WG, Mueller SC, and Chen WT. Binding and localization of Mr 72,000 matrix metalloproteinase at cell surface invadopodia. Cancer Res 53: 3159–3164, 1993. 27. Myrhage R and Hudlicka O. Capillary growth in chronically stimulated adult skeletal muscle as studied by intravital microscopy and histological methods in rabbits and rats. Microvasc Res 16: 73–90, 1978. 28. Nagase H. Cell surface activation of progelatinase A (proMMP-2) and cell migration. Cell Res 8: 179–186, 1998. 29. Nakahara H, Howard L, Thompson EW, Sato H, Seiki M, Yeh Y, and Chen WT. Transmembrane/cytoplasmic domainmediated membrane type 1-matrix metalloproteinase docking to invadopodia is required for cell invasion. Proc Natl Acad Sci USA 94: 7959–7964, 1997. 30. Paku S and Paweletz N. First steps of tumor-related angiogenesis. Lab Invest 65: 334–346, 1991. 31. Pepper MS. Manipulating angiogenesis. From basic science to the bedside. Arterioscler Thromb Vasc Biol 17: 605–619, 1997. 32. Richardson RS, Wagner H, Mudaliar SR, Henry R, Noyszewski EA, and Wagner PD. Human VEGF gene expression in skeletal muscle: effect of acute normoxic and hypoxic exercise. Am J Physiol Heart Circ Physiol 277: H2247–H2252, 1999. 33. Sato H, Okada Y, and Seiki M. Membrane-type matrix metalloproteinases (MT-MMPs) in cell invasion. Thromb Haemost 78: 497–500, 1997. 34. Schoefl GI. Studies on inflammation. III. Growing capillaries: their structure and permeability. Virchows Arch 337: 97–141, 1963. 35. Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J Clin Invest 103: 1237–1241, 1999. 36. Strauss BH, Robinson R, Batchelor WB, Chisholm RJ, Ravi G, Natarjan MK, Logan RA, Mehta BR, Levy DE, Ezrin AM, and Keeley FW. In vivo collagen turnover following experimental balloon angioplasty injury and the role of matrix metalloproteinases. Circ Res 79: 541–550, 1996. 37. Tsurumi Y, Kearney M, Chen D, Silver M, Takeshita S, Yang J, Symes JF, and Isner JM. Treatment of acute limb ischemia by intramuscular injection of vascular endothelial growth factor gene. Circulation 96: 382–388, 1997. 38. Zheng W, Brown MD, Brock TA, Bjercke RJ, and Tomanek RJ. Bradycardia-induced coronary angiogenesis is dependent on vascular endothelial growth factor. Circ Res 85: 192–198, 1999. 39. Zhou AL and Egginton S. Immunogold labeling of proliferating cells during skeletal muscle angiogenesis. J Vasc Res 15: 386, 1997. 40. Zhou A, Egginton S, Hudlicka O, and Brown MD. Internal division of capillaries in rat skeletal muscle in response to chronic vasodilator treatment with ␣1-antagonist prazosin. Cell Tissue Res 293: 293–303, 1998. 41. Ziada AM, Hudlicka O, Tyler KR, and Wright AJ. The effect of long-term vasodilatation on capillary growth and performance in rabbit heart and skeletal muscle. Cardiovasc Res 18: 724–732, 1984.