INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PERIPHERAL VASCULAR DISEASE (PVD) and critical limb ischemia are major health problems in the Western world. Each year, in the US alone, ~100,000 peripheral vascular surgeries are performed, and many patients are not candidates for attempts at limb salvage (8). In these patients, surgical and percutaneous revascularization techniques are often unsuccessful in reversing or limiting the complications of this progressive disease. Furthermore, pharmacological therapy has had relatively little impact on limb-threatening ischemia and PVD.

Recently, several polypeptide growth factors have been shown to activate endothelial cell proliferation and enhance angiogenesis in vivo (11). The angiogenic effects of these factors have led investigators to propose their therapeutic use to improve tissue perfusion (10) in diseases ranging from PVD to myocardial ischemia. As a result, two angiogenic factors, basic fibroblast growth factor (bFGF) (4) and vascular endothelial growth factor (VEGF) (24), have become a major focus of molecular medicine for clinical consideration in therapeutic angiogenesis.

VEGF is unique among angiogenic growth factors in that it has direct endothelial-specific mitogenic effects, presumably due to endothelial cell-specific localization of VEGF receptors (9). VEGF exists in four isoforms, having 121, 165, 189, or 206 amino acids (14, 16), which are produced in cells by alternative splicing. All four isoforms are secreted and thus biologically capable of mediating receptor-specific changes in response to ischemia. The 121-amino acid isoform (VEGF₁₂₁) is the most freely soluble when secreted because it does not contain a heparin-binding domain. The larger isoforms have progressively greater heparin-binding affinity and thus more readily localize in the extracellular matrix (23). Although these differences in heparin affinity among the various VEGF isoforms affect protein localization, the presence of multiple VEGF isoforms in vivo suggests that differential VEGF synthesis, distribution, and bioavailability may affect its potential role in angiogenesis (14).

Expression of VEGF mRNA is increased in ischemic areas of solid tumors (21), heart (5), and retina (17). Little, if any, data exist regarding differential changes among VEGF isoforms in ischemic skeletal muscle. However, oxidative and glycolytic skeletal muscle fibers differ in their basal VEGF expression (2) and capillary density (15). To directly establish what effect ischemia may have on endogenous angiogenic peptide expression in oxidative and glycolytic skeletal muscle, changes in VEGF and bFGF protein levels were assessed in both skeletal muscle fiber types at 1, 5, and 21 days of surgically induced ischemia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal model. All procedures involving animals conformed to the Guidelines for Use of Laboratory Animals published by the United States Department of Health and Human Services and approved by the Duke University Institutional Animal Care and Use Committee. All animals received care in accordance with Guide for the Care and Use of Laboratory Animals, published by the National Research Council.

Hemodynamic assessment of hindlimb ischemia and hypoxia. In a subset (n = 18) of the studied animals, blood flow to the tibialis anterior (TA) and soleus (SOL) muscles was measured with a laser Doppler flowmeter (BPM403; TSI, St. Paul, MN). At 1, 5, and 21 days postischemia, animals were anesthetized as described above, and a longitudinal incision was made along the lateral aspect of the hindlimb to allow proper exposure of the TA and SOL on both hindlimbs. A surface probe was placed on the muscle to be examined and secured with an external fixation device. Six laser Doppler readings were taken along the anterior surface of each TA or SOL to compare flow between similar regions of the ischemic and control muscles. Rectal temperature was monitored throughout this procedure, and core temperatures were maintained between 37 and 38°C by placement of the animals on warming blankets if necessary.

Tissue harvest and protein extraction. At study termination, each animal was euthanized as outlined by Duke University Medical Center Institutional Animal Care Guidelines. The TA and SOL muscles were surgically excised from tendon to tendon. The midportion of the muscle was used for immunohistochemistry to ensure that representative samples were taken from each muscle and adjacent sections were used for protein analysis. For protein extraction, muscle samples were weighed (400-500 mg), pulverized in liquid nitrogen, and homogenized in 3 ml of 10 mM Tris (pH 7.4) and 100 mM NaCl with the use of a Brinkmann polytron (Westbury, NY). The suspension was then centrifuged twice at 8,000 g at 4°C for 15 min. The protein content of the supernatant was determined by Bradford assay (6).

Total VEGF and bFGF detection by ELISA. Total VEGF and bFGF concentrations in muscle supernatant were determined by the Quantikine (R & D Systems, Minneapolis, MN) VEGF or bFGF immunoassay systems, respectively. All techniques and materials used in these experiments were in accordance with the provided protocols.

VEGF₁₆₅ detection. To detect VEGF₁₆₅ (165-amino acid isoform) protein in muscle lysate, 400 µg of total muscle protein lysate (supernatant from tissue harvest) were concentrated on heparin-agarose, subjected to SDS-PAGE, and quantified by immunoblotting with a polyclonal antibody to VEGF₁₆₅ as previously described (2). An immunoreactive band at ~46 kDa was detected by chemluminescence (Amersham, Arlington Heights, IL) and X-ray film (X-AR; Kodak, Rochester, NY). To confirm the specificity of the immunoreactive signal, recombinant human VEGF₁₆₅ (R & D Systems) was preincubated at a concentration of 1 µg/ml with the VEGF antibody. To ensure that the binding of VEGF to heparin-agarose beads was complete and not saturable, 100, 200, and 400 µg of protein lysates from ischemic and control muscle were loaded on heparin agarose beads and quantified by immunoblotting to ensure that corresponding increases in VEGF protein signal were detected. VEGF₁₂₁ does not bind with heparin-agarose (9), so it did not interfere with these studies. Although some molecular interactions with higher-weight isoforms is possible, no change was detected in higher-molecular-weight isoforms in these studies (data not shown). The VEGF₁₆₅ signal was quantified by laser densitometry of the developed film and subsequent NIH Image (1.61) Analysis.

VEGF immunohistochemical staining. Immunohistochemistry was used to visualize VEGF expression in histological sections. The techniques employed were modifications of previously described methods (1). Tissue sections adjacent to those used for protein extraction were thawed in 30% sucrose-PBS, placed in cross section in optimum cutting temperature compound (OCT; Miles Pharmaceuticals, West Haven, CT), and snap frozen in liquid nitrogen. Frozen sections (5 µm) were obtained on microscope slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA). Slides were allowed to reach room temperature, placed in ice-cold acetone for 10 min, and placed in PBS for three 5-min washes. Blocking solution (10% horse serum in PBS) was applied for 1 h at room temperature. A murine anti-human VEGF antibody (clone 26503.11; Sigma Chemical, St. Louis, MO) was diluted (1:250) in 1× PBS and applied to tissue sections for 1 h. Incubation with the primary antibody was followed by sequential incubation with a biotinylated anti-mouse immunoglobulin G (IgG) and ABC reagent, according to the manufacturer's specifications (Vectorstain ABC kit; Vector Laboratories, Burlinghame, CA). Levamisole was added to block endogenous alkaline phosphatase activity, and immune complexes were localized with the use of the chromogenic alkaline phosphatase substrate Vector Red (Vector Laboratories). The sections were counterstained with hematoxylin, dehydrated, and mounted with Permount (Fisher Scientific). In this method, VEGF appears red, and cell nuclei are blue.

Capillary density quantification. To measure capillary density, serial 5-µm frozen tissue sections were stained with nitro blue tetrazolium/5-bromo-4-chloro-3-indoylphosphate-p-toluidine salt (NBT/BCIP; GIBCO, Grand Island, NY), which stains alkaline phosphatase in endothelial cells. The number of capillaries per muscle fiber was evaluated as described by Takeshita et al. (23). Briefly, an independent observer analyzed 20 randomly selected fields, and the number of capillaries and muscle fibers were counted under a ×20 objective to yield the capillary-to-muscle fiber ratio.

Statistical analysis. Results were expressed as means ± SE unless otherwise indicated. Statistical significance within each muscle fiber type was evaluated by use of a paired Student's t-test. Statistical significance between muscle types was calculated by ANOVA. A value of P < 0.05 was interpreted as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Induction of ischemia. Similar degrees of ischemia and hypoxia were present in the TA and SOL muscles after femoral arterial resection. As depicted in Fig. 1, perfusion (solid lines) was decreased to ~70% of the contralateral control in both TA and SOL at all time points in the study. Similarly, surface oxygen tension (Fig. 1, dashed lines) was decreased to <30% of control in both muscles at 1 day (P < 0.05) and to <40% of control at 5 and 21 days (P < 0.05) of ischemia. Overall, from 1 to 21 days of hindlimb ischemia, there was no significant change in perfusion or oximetry in either muscle group.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Perfusion and oximetry of ischemic soleus (SOL) and tibialis anterior (TA) skeletal muscle expressed as percentage of contralateral control muscle. Values are means ± SE perfusion or blood flow (n = 3 at 1, 5, and 21 days) as measured by tissue oximetry at 1, 5, and 21 days of hindlimb ischemia.

Total VEGF expression by ELISA. Expression of VEGF was quantified in ischemic and control muscle samples by ELISA. Total VEGF expression was higher in control SOL than in control TA (see Fig. 3) as previously described (2). To compare VEGF expression after surgically induced ischemia, a ratio of ischemic-to-control VEGF expression was calculated and is depicted as the percentage of control values in Fig. 2. As depicted in the hatched bars in Fig. 2A, VEGF expression by ELISA was significantly increased at 1 and 5 days of ischemia in TA (mean ± SE: 140 ± 20%, n = 7; and 134 ± 55%, n = 7; P < 0.05). At 21 days, total VEGF was also increased over control (169 ± 38%, n = 6; P < 0.06), although this value did not achieve statistical significance.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Mean changes in total combined 121- and 165-amino acid isoforms of vascular endothelial growth factor (VEGF121+165), VEGF165, and basic fibroblast growth factor (bFGF) levels in ischemic SOL and TA expressed as a percentage of contralateral (control) limb values. A: TA skeletal muscle at 1, 5, and 21 days of hindlimb ischemia (1D, 5D, and 21D, respectively). B: SOL skeletal muscle at 1, 5, and 21 days of hindlimb ischemia. * Statistically significant change from control (P < 0.05, Student's t-test). + Significant differences between TA and SOL (P < 0.05, ANOVA).

Expression of the VEGF₁₆₅ isoform. Changes in VEGF₁₆₅ protein expression were quantified by immunoblot (Western) analysis of TA and SOL muscle lysates from the ischemic and control limbs at 1, 5, and 21 days of surgically induced hindlimb ischemia. Figure 3A depicts a typical immunoblot used to quantify VEGF₁₆₅ expression from lysates of ischemic and contralateral control SOL at 1 and 5 days. Figure 3B shows VEGF₁₆₅ expression in the ischemic and control TA muscles of the same animals, demonstrating the difference in both basal and ischemia-induced expression in the two fiber types. In control TA, VEGF was present to a lesser degree than in control SOL, similar to results previously published (2). After 1 day of ischemia, VEGF₁₆₅ was decreased in SOL but was significantly increased in TA, which demonstrates the differential response in the two fiber types to the same degree of ischemia. After 5 days of ischemia, VEGF₁₆₅ expression was increased in TA and SOL; however, the magnitude of the increase was significantly greater in SOL. The ratios of expression in ischemic to control limbs were calculated as described in METHODS and are expressed as percentages of control in Fig. 2. As indicated in Fig. 2A, VEGF₁₆₅ expression (solid bars) was significantly increased at 1 and 5 days in ischemic TA (P < 0.05). The mean ± SE increases in TA were 260 ± 70%, (n = 6), 150 ± 30% (n = 6), and 180 ± 70% (n = 8) of control at 1, 5, and 21 days, respectively. In contrast, at 1 day of ischemia in SOL, VEGF₁₆₅ decreased to 54 ± 8% of control (n = 8) but later rebounded to 250 ± 40% (n = 7) and 155 ± 60% (n = 7) of control at 5 and 21 days, respectively (Fig. 2B, solid bars). Statistically significant differences in the expression of VEGF₁₆₅ were seen between TA (glycolytic muscle) and SOL (oxidative muscle) at 1 and 5 days (Fig. 2B, + symbol). At 1 day, VEGF was decreased to 54 ± 8% in SOL but was increased to 260 ± 70% in TA; however, by 5 days, VEGF₁₆₅ expression was significantly greater in SOL than in TA (250 ± 40 vs. 150 ± 30%).

View larger version (68K): [in this window] [in a new window]   Fig. 3.   Representative immunoblots of 400 µg of SOL and TA muscle lysate from rabbits at 1 and 5 days of hindlimb ischemia. A: VEGF165 expression in control (Con) and ischemic (Isch) SOL lysate (n = 2) at 1 and 5 days of ischemia. Immunoreactive band at ~46 kDa corresponds to VEGF165. At 1 day of ischemia, SOL VEGF levels were decreased (~54%) relative to contralateral control values. By 5 days, VEGF levels were increased (250%) in ischemic SOL. B: representative immunoblots of VEGF expression from control and ischemic TA lysate (n = 2) at 1 and 5 days of ischemia. After 1 day of ischemia, VEGF levels were increased ~twofold, whereas at 5 days, VEGF levels remained elevated to only ~150% of control.

bFGF expression. Small increases in bFGF expression (Fig. 2, open bars) were observed in glycolytic TA muscle in response to ischemia at all times, but no value was statistically significant. Similarly, no statistically significant change in bFGF expression was seen in SOL at 1 and 5 days of ischemia; however, in the 21-day SOL, an increase to 129 ± 29% (P < 0.05) was statistically significant.

Localization of VEGF expression. To determine the site of VEGF expression and whether VEGF was seen within muscle fibers or alternatively at the fiber periphery as we have previously described (2), immunohistochemical detection of VEGF was performed on thin sections. In control SOL (Fig. 4A), VEGF staining was detectable at the fiber periphery and in what appeared to be nonmuscle cells. After 1 day of ischemia, little or no VEGF could be detected in SOL (Fig. 4B). Conversely, in control TA, little or no VEGF was detected (Fig. 4C). Yet, by day 1 of ischemia, VEGF expression was detectable at the periphery of muscle fibers and in interstitial cells (Fig. 4D). In all samples examined by immunohistochemistry, VEGF protein appeared to localize primarily in the extracellular matrix between muscle bundles (Fig. 4, A-D).

View larger version (153K): [in this window] [in a new window]   Fig. 4.   Immunohistochemical anti-VEGF staining of control and ischemic SOL and TA muscle sections and of jugular vein graft. Anti-VEGF staining at ×66 magnification (A-D and F) and ×33 magnification (E) appears as a red precipitate either at periphery of muscle fibers or within the region between fibers. Intensity of staining reflects amount of VEGF detected. A: in control SOL muscle, VEGF expression is detectable between muscle fibers and occasionally at muscle fiber periphery. B: after 1 day of ischemia, VEGF staining is no longer readily detectable in interstitial region surrounding SOL muscle fibers or within fibers themselves. C: in TA control muscle, VEGF is undetectable or present at very low levels. D: after 1 day of ischemia, VEGF staining increases and can be seen in region surrounding muscle fibers or at fiber periphery. E and F: rabbit external jugular vein demonstrates cell-specific VEGF staining and serves as a positive control for VEGF immunohistochemistry, as described in previous reports (see Ref. 2).

Capillary density. In control SOL, capillary density (capillaries/muscle fiber) was ~2.5 times greater (2.9 capillaries/fiber) than in control TA (1.1 capillaries/fiber) (P = 0.05). At 1 and 21 days of ischemia, no significant increase in the capillary-to-fiber ratio was seen in either fiber type (Table 1). Although capillary density grossly appeared to increase at 21 days (data not shown), no change in the number of capillaries per fiber was noted. This is likely due to the observed decrease in fiber diameter, where in ischemic SOL fiber, diameter decreased from 945 to 840 µm by 21 days, and in TA, a decrease from 604 to 561 µm was noted.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Therapeutic angiogenesis is emerging as an enormously powerful treatment strategy for patients with peripheral vascular and ischemic heart disease. Two polypeptides that have been proposed for therapeutic angiogenesis are VEGF and bFGF. An understanding of their endogenous expression under ischemic conditions and elucidating their native effect in potential target tissues are essential to maximize the benefits of administered peptides.

Although various stresses including hypoglycemia and ischemia can increase VEGF mRNA expression in skeletal muscle (21, 22), very few data exist to show that VEGF mRNA expression in any tissue correlates with protein expression. Furthermore, no data exist to address the effect of ischemia on VEGF or bFGF expression in skeletal muscle. The goal of this study was to define the underlying changes in the expression of VEGF and bFGF protein in chronically ischemic skeletal muscle and to determine whether oxidative and glycolytic muscle fiber types respond differently.

The major findings of this study are as follows. First, as indicated by ELISA, both glycolytic and oxidative skeletal muscle fibers increase VEGF expression in response to ischemia; however, the isoforms that contribute to this change differ between the two muscle groups. In glycolytic muscle, VEGF₁₆₅ increases in parallel with the increase in total VEGF expression, whereas, in oxidative muscle, VEGF₁₆₅ expression decreases even while total VEGF expression is increased. Second, total VEGF expression by ELISA is greater in oxidative muscle than glycolytic muscles after chronic ischemia produced by arterial resection. Finally, in contrast to the increases in VEGF expression, little or no change in bFGF expression is seen in the same muscle in response to ischemia.

The ELISA and Western analyses used in this study differ in their ability to detect VEGF subtypes. Because VEGF is present at levels that are too low to be detected by Western blot of muscle lysates (as previously shown, see Ref. 2), a heparin-binding step was used to concentrate VEGF. This step effectively excludes VEGF₁₂₁ and allows assessment of VEGF₁₆₅. Currently, no direct means of quantifying VEGF₁₂₁ exist; however, a combination of VEGF₁₂₁ and VEGF₁₆₅ can be measured by ELISA. In glycolytic TA muscle, total VEGF expression by ELISA was increased to ~150% of control by 1 and 5 days of ischemia and 180% at 21 days. However, the contribution of VEGF₁₆₅ to the observed increases in VEGF expression by ELISA was markedly different at the same time points. At 1 day, VEGF₁₆₅ was increased almost 260% over the control value in the contralateral limb, despite the fact that total VEGF increased to only 140% of control. Taken together, these data suggest that decreases in other VEGF isoforms must have occurred. At 5 and 21 days, the percent increase in VEGF₁₆₅ was similar in magnitude and direction to the increase seen in total VEGF levels (by ELISA), implying that VEGF₁₆₅ is the major VEGF isoform expressed in glycolytic skeletal muscle in response to ischemia at those times.

The changes in VEGF expression in oxidative muscle are strikingly different than in TA. Although ELISA studies demonstrated that total VEGF expression increased at all time points in both muscles, the magnitude of this increase in SOL was significantly greater than in TA at 5 days. In addition, isoform responses differed in the two muscles. For example, at 1 day of ischemia, total VEGF expression was increased in SOL to 140% of control despite a significant decrease in VEGF₁₆₅ expression, whereas, in TA, expression of VEGF₁₆₅ increased along with total VEGF. As VEGF₁₆₅ expression decreased in SOL, other VEGF isoforms appeared to be immediately and significantly increased, presumably to compensate for this observed decrease in VEGF₁₆₅. Together, these findings highlight the importance of analyzing VEGF subtypes within a given tissue in response to a pathological stress.

The observed discrepancies in the direction and/or magnitude of VEGF₁₆₅ and total VEGF protein expression between oxidative and glycolytic skeletal muscle may result from both the intrinsic differences between the muscle types and from functional differences between these fiber types; the different stimuli to these muscles may also contribute to altered gene expression. Previous studies have demonstrated that the levels of VEGF protein per microgram of total cellular protein (2) and overall capillary density are consistently greater in oxidative (type I) than glycolytic (type II) muscles (20). Perhaps under ischemic conditions, molecular changes that reflect these baseline differences in VEGF levels differentially affect fiber-type survival. Data supporting intrinsic fiber-type differences in response to ischemia include those from Regensteiner et al. (19) showing that in the ischemic limbs of patients with PVD, significant type II (glycolytic) muscle fiber atrophy occurs. This observed discrepancy in ischemic type I and type II muscle fiber survival and viability in patients with PVD may, in part, be explained by differences in the expression and concentrations of angiogenic peptides as seen in the present study.

Recently, Charnok-Jones et al. (7) have shown that VEGF mRNA processing may differ among cell and tissue types, so VEGF₁₆₅ may not uniformly be the predominant isoform in all tissues. Yet, studies examining in vivo collateral vessel development have shown similar degrees of neovascularization after exogenous administration of three VEGF isoforms, phVEGF₁₂₁, phVEGF₁₆₅, and phVEGF₁₈₉(189-amino acid isoform), suggesting that the three isoforms are biologically equivalent (24). Those data, when viewed in conjunction with the present study, suggest that future investigations of VEGF isoform-specific regulation and bioactivity within different fiber types are warranted.

Despite the observed changes in the expression of these angiogenic molecules in response to ischemia, no significant change in tissue perfusion, oximetry, or capillary density occurred during the 21 days of this study. This suggests that the observed increases in VEGF expression in response to ischemia are insufficient to produce a significant angiogenic response over 21 days. Very minor, if any, changes in bFGF expression were seen in this model of arterial resection. These findings suggest that VEGF may play a more critical role earlier in the process of new vessel formation in ischemic skeletal muscle or that sustained vascular sprout survival requires increases in expression of both VEGF and bFGF. Perhaps the described hindlimb ischemia model was below a critical threshold for significant bFGF induction (12, 13). Although studies have shown angiogenic synergism between exogenous bFGF and VEGF, significant in vivo coregulation of these two peptides was not obvious under our conditions of surgically induced skeletal muscle ischemia (3).

In conclusion, our results support the hypothesis that angiogenic peptide levels are increased and/or maintained in ischemic skeletal muscle relative to control. Muscle fiber type appears to impact the expression of these molecules, in that VEGF protein expression was greater in oxidative muscle than in glycolytic muscle after ischemia produced by arterial occlusion. Moreover, the contribution of specific VEGF isoforms to the increase in total VEGF expression was significantly different between these two muscles. These different patterns of VEGF expression between muscle fiber types may correlate with the specific physiology and/or function of oxidative and glycolytic skeletal muscle. These data suggest that future approaches in therapeutic angiogenesis may need to consider muscle fiber type and specific VEGF isoforms when attempting to optimize the benefits of these therapies.