INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE PROCESS OF ANGIOGENESIS can be divided into two categories: capillary growth, referring to the sprouting of capillaries from preexisting vessels (6), and blood vessel remodeling, referring to the enlargement of new and existing vessels, primarily arterioles. Recent studies have provided information on the molecular events of angiogenesis, and, under the appropriate conditions, capillaries can be induced to grow in culture from endothelial cells (6). In contrast, the mechanisms involved in the growth and remodeling of arterioles and larger vessels are generally unknown. In addition, the relationship, if any, between capillary development and arteriolar growth has not been elucidated.

Angiogenesis does not normally occur in the adult heart. However, vascular synthesis and remodeling does occur in response to ischemia and pressure-overload hypertrophy (28). In addition, many investigations have shown that chronic endurance training induces coronary vascular adaptations, including increased coronary blood flow (CBF) reserves and increased coronary transport reserves (CTR; see Refs. 14, 15). The physiological basis for these changes includes structural adaptations in the size and number of the blood vessels, as well as alterations in systemic, neurohumoral, and local vascular control (13). The effects of endurance training on neovascularization in the heart are complex. In young animals, training generally increases capillary density, whereas, in adult animals, training appears to have little effect on capillary density but does increase the density and diameter of arterioles (see Ref. 8 for review). However, most studies have examined the exercised subjects at the end point of the training. Thus information on the progressive nature of the changes induced by exercise was not obtained.

In this report, we have used a well-documented model of chronic exercise in pigs (1, 17, 24, 32, 34) to quantify capillary and arteriolar growth after 1, 3, 8, and 16 wk of endurance treadmill training. Pigs have many physiological characteristics in common with humans, including similarities in the coronary vascular anatomy and the development of coronary collateral vessels in response to ischemia (31) and in O₂ consumption (O₂) and blood flow distribution during exercise (7). Angiogenesis was measured by assessing sprouting and vessel diameter and by quantifying endothelial and smooth muscle cell DNA labeling. Our results show that the appearance of new capillaries was most significant within the first 3 wk of exercise, culminating in an increase in coronary capillary density at 3 wk. After 3 wk, capillary density returned to control levels while capillary diameter increased. Similarly, the density of the smallest class of arterioles (10-19.9 µm) also increased at 3 wk and diminished thereafter. This occurred concomitantly with an increase in the density of arterioles in the 20- to 30-µm-diameter class. At 16 wk, the diameter of the largest arterioles was also increased. These results indicate that exercise-induced neovascularization occurred in a stepwise manner. In addition, our calculations indicate that the increase in size and number of arterioles was a more important contributor to increased blood flow than was angiogenesis. Finally, these observations suggest that capillaries develop into small arterioles; thus capillary angiogenesis potentially contributes to increased blood flow by providing a source of new arterioles.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animal models. Studies were carried out in 38 Yucatan male minipigs that had an initial mean body weight (BWt) of 34 ± 1 kg. The animals were 3-6 mo of age during the study. The animals were tested to determine whether they would submit to a maximal stress test. Four pigs were determined not to be suitable for further testing because of large variations in maximal O₂ consumption (O_{2 max}) or because they refused to reach maximum heart rate on a stress test. Four other animals were eliminated because of surgical complications at the termination of training. The 30 remaining animals were divided at random into five groups of six animals each. One group was randomly selected as controls. All animals were familiarized with treadmill running for 1 wk before the beginning of training. This preliminary exercise was limited to 5-10 min. The four experimental groups were treated as follows: group 1 (1-W), animals exercised for 1 wk; group 2 (3-W), animals exercised for 3 wk; group 3 (8-W), animals exercised for 8 wk; and group 4 (16-W), animals exercised for 16 wk. In the control group, final experiments were conducted on two animals at the 3-W, 8-W, and 16-W times.

Exercise training and stress test. All animals underwent four stress tests, two at the beginning of the program and two at the end, and the values of O_{2 max} measured during these stress tests were averaged for each pair. These stress tests were designed to elicit a maximal effort and have been previously described by this laboratory (32). At 1 wk, the exercise training protocol consisted of running the pigs on a treadmill for 30 min at 70-80% of maximal heart rate. Heart rates were monitored during training as described below. Each week thereafter, the exercise effort was increased by 5 min until 8 wk. From 8 until 16 wk, the pigs maintained the same work effort of 70 min/day, 5 days/wk.

O_{2 max}. Training adaptations were determined by measurements of O_{2 max} during treadmill exercise. A progressive ramp protocol was used to exhaust the pigs in 15-20 min (32). O₂ was determined from the Fick equation (O₂ = cardiac output × arteriovenous O₂ differences). Cardiac output was measured by thermodilution with the use of a 7-Fr Swan-Ganz thermodilution catheter that was passed via the external jugular into the coronary artery. The catheter was interfaced with a model COM-IRS cardiac output computer (American Edwards Laboratory). Bolus injections of ice-cold saline were repeated until duplicate measurements of cardiac output varied by <10%. Simultaneous blood samples were obtained in duplicate from the pulmonary artery via the Swan-Ganz catheter and via a Silastic catheter positioned in the carotid artery. O₂ content of the blood samples was determined by using an IL-282 CO-oximeter. Heart rates were determined from a bipolar surface electrocardiogram (ECG). The skin was shaved and cleaned with alcohol, and fluid-column electrodes were then placed on the sternum, on the back at the same level as the sternum electrode, and adjacent to the posterior spine. The ECGs were recorded at rest and at the end of each minute of exercise with a Hewlett-Packard 1511B single-channel ECG recorder.

Surgical preparation and measurements of capillary diffusion capacity and CBF. One day after the last stress test, the animals underwent a procedure to measure CBF capacity and capillary diffusion capacity by using the procedures previously described (14, 15). These protocols are identical to those used by Laughlin and colleagues (14). The pigs were anesthetized with ketamine (30 mg/kg im) and pentobarbital sodium (10 mg/kg iv), heparinized (10 mg/kg iv), and ventilated with a positive-pressure respirator by using room air (or supplemented with 95% 0-5% CO₂ when necessary to maintain arterial PO₂). A midsternal incision was made, and the heart was suspended in a pericardial cradle. The coronary sinus was then cannulated with a Silastic catheter passed into the sinus via the right atrial appendage. The anterior descending branch of the left coronary artery (LAD) was cannulated at the aorta; the LAD was perfused by a roller pump (Cole-Palmer) with a Servodyne mechanism that used blood from the pig's femoral artery. The pressure was controlled by the servomechanism of the pump at ~105 mmHg. The single-injection indicator-diffusion method was used as described previously (14) to measure capillary diffusion. An injection of blood containing a reference tracer (¹²⁵I-serum albumin) and a diffusible solute (⁵¹Cr-EDTA) was made on the arterial side of the coronary bed, and continuous sequential venous samples were obtained from the coronary sinus with the aid of a Cole-Palmer roller pump withdrawing blood at a rate of 30 ml/min. The venous samples were taken continuously at a rate of 1 sample/s. Venous blood samples (0.2 ml) were placed in counting vials, and the radioactivity was determined in a model 5912A gamma counter (Packard). Extraction was then calculated for each sample as (C_r  C_d)/C_r, where C_r and C_d are the normalized venous concentrations of the reference (C_r) and diffusible (C_d) tracers, respectively. The extraction value (E) that was chosen to represent the capillary bed at the time of measurement was the maximal value seen at the peak or 2-4 s before the peak of the concentration-time curve as described in detail previously (15). The capillary diffusion coefficient or capillary permeability surface and product (PS) was calculated from the E and plasma flow values (F) as PS = F ln(1  E). E was also used to calculate the capillary clearance (C) as C = E × F. Capillary CTR is then defined as the difference between the measured PS under baseline conditions and the maximal PS (measured during maximal adenosine vasodilatation under constant-pressure conditions).

Morphometric image analysis. We measured coronary arterioles, both external and internal diameters, and identified vessels directly from microscope slides by using an automated image-analysis system. Our system used a Hitachi model KP 140 high-resolution closed-circuit television black-and-white video camera attached to a model 100 Image-Analysis System (Analytical Imaging Concepts, Irvine, CA). This system is interfaced with an AT/386 computer for data storage, calculations, and analysis by using LOTUS 1-2-3 (Lotus Development, Cambridge, MA). This imaging system consists of an image-capture and -display board with 512 × 512 × 8-bit resolution and 256 gray levels.

Capillary and arteriolar numerical densities. Capillary densities (number of capillaries/mm² myocytes) were obtained from photographs of 1-µm-thick sections stained with toluidine blue. Sections used for analysis of capillary density were cross sections from the endomyocardial and epimyocardial regions. From each animal, five tissue blocks were selected and two photographs per block were used, so that each animal was analyzed on the basis of 10 photographs. All photographs were taken at ×160 magnification with the use of a Zeiss Ultrashot microscope with a 35-mm camera. Negatives were enlarged full frame to a magnification of ×1,200. Capillary identification was facilitated by perfusion-fixation of capillaries in an open position. Criteria for capillary identification were as follows: vessel diameter was not to exceed 10 µm, vessels presented no outer layers, and vessel lumen presented no more than one endothelial nucleus. Myocyte area was determined from each photograph by using the point-counting technique to determine the volume density of myocytes. This value was then multiplied by the actual area per photograph. Arteriolar densities (number of arterioles/mm² of myocytes) were obtained in a similar fashion as described above for capillary densities. Photographs of 1-µm-thick sections stained with toluidine blue were obtained from cross sections of the endomyocardial, midmyocardial, and epimyocardial regions of the left ventricle. Ten blocks were selected from each tissue region, and one photograph was taken from each section. Thus the arteriolar density for each animal was based on 30 photographs. With a Zeiss Ultrashot microscope, each section was examined, and the number of arterioles per section was counted. At a magnification of ×16, photographs of the entire section were taken, and the resulting 35-mm negatives were printed full frame to a final magnification of ×120. Vessels were identified as arterioles if the vessel had one or two layers of smooth muscle. The arterioles were divided into four groups (in µm): 20-30, 31-40, 41-70, and 71-120. Vessels 10-19.9 µm in diameter were considered to be arterioles and venules. An estimate of the ratio of venules to arterioles was made in two animals by simultaneously injecting colored silicone into arteries and veins. Myocyte area was calculated as described above.

DNA synthesis with tritium-labeled thymidine. The DNA in the nuclei of the dividing cells was labeled with [³H]thymidine. To accomplish this, we injected an ear vein with 0.3 mCi/kg [³H]thymidine (3.7 Ci/mmol; New England Nuclear) at 12 and 21 h before euthanasia. Label was injected at two separate times to label the maximum number of dividing cells, because not all of the dividing cells would synthesize DNA at the same time. Sedentary control animals were labeled similarly. Cross sections 3 µm thick were cut and autoradiographed by coating the slides in Ilford emulsion (diluted 3:2 with distilled water) and exposing them for 60 days in a light-tight box at 20°C. After they were developed in Dektol, the slides were lightly stained with hematoxylin and eosin, and then coverslips were applied. Silver grains superimposed on the visible nuclei were counted on labeled mitotic cells. Visualization of the grains was enhanced by the use of ultraviolet reflected light. The labeling index data were accrued while the slides were viewed at ×400 and ×1,000, using both transmitted light and epiillumination. Nuclei with four or more grains were considered labeled. Background grain counts were <1 grain per nuclear area. A percentage labeling index was calculated as follows: (number of endothelial or smooth muscle cells labeled/total number of endothelial or smooth muscle cells examined) × 100. Cells were defined as endothelial or smooth muscle cells by the criteria of their position within the vessel wall and their nuclear shapes. The total numbers of endothelial cells and smooth muscle cells observed to calculate this index were ~3,000 capillary endothelial cells and 1,000 smooth muscle cells for each animal.

Statistical analysis. Data are expressed as means ± SE. Statistical analysis was carried out by using either the Student's mean t-test for intergroup comparisons, with the Bonferroni correction for multiple comparisons, or repeated-measures analysis of variance for intragroup comparisons by using Rao's r to test statistical comparison. Values were considered significantly different at P < 0.05. Calculations were carried out by using a STATS+ statistical package (Statsoft, Tulsa, OK).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Coronary vascular adaptations during progressive exercise. Several parameters, including cardiac hypertrophy, blood flow, capillary diffusion capacity, and total O_{2 max}, were measured to document the efficacy of the training program. The effects of exercise on cardiac hypertrophy are described in Table 1. On the basis of using the standard ratio of left ventricular weight/body weight (LV/BWt in g/kg), hypertrophy was present at 8 and 16 wk of exercise (16.5 and 24%, respectively). However, the BWt of the exercised animals did not increase at the same rate as did the BWt of the sedentary control animals. If the BWt of the control animals is used to estimate hypertrophy, then a significant increase in LV/BWt ratio in controls occurred at 16 wk but not at 8 wk (data not shown). The effects of exercise on total O₂, maximal CBF, and CTR are shown in Fig. 1. O_{2 max} progressively increased through 8 wk and then leveled off, presumably due to the plateauing of the exercise effort at 8 wk. Maximal CBF during adenosine infusion increased, beginning at 3 wk, and continued increasing throughout the training duration. Capillary transport increased between 1 and 8 wk of training and remained elevated at 16 wk.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Graph showing effects of exercise training on maximal coronary blood flow (CBF), capillary transport reserve (CTR), and maximal whole body O2 consumption (O2 max). cont, Control; 1W, 3W, 8W, and 16W: groups after 1, 3, 8, and 16 wk of exercise training, respectively. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. * P < 0.05, compared with control.

Exercise induces capillary angiogenesis. Several indexes were examined to measure capillary angiogenesis. These included sprouting, proliferation of endothelial cells in existing capillaries, and capillary density. It is clear from the results described below that exercise rapidly induced a transient upregulation of capillary development.

View larger version (148K): [in this window] [in a new window]   Fig. 2.   Photomicrographs of typical high-power fields. A: venule (V) with 2 capillary sprouts (S); C, capillaries. Arrows, DNA labeling of endothelial cells. B: labeled V and C. Bar represents 5 µm.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Effects of exercise training on sprouting of new capillaries (Sprout #) and labeling of sprouts (Sprout %) during exercise training. C, control. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. Sprout % is percentage of sprouts that showed DNA labeling. Sprout # is no. of sprouts in 50 high-power fields. * P < 0.05. compared with control.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effects of exercise training on DNA labeling in capillaries and in venules (VEN) and arterioles (ART; 10- to 19-µm diameter). Amount of labeling in control animals (C) was so small that value is not visible as a bar. First visible bar on left for both labeled capillaries as %total capillaries (CAP) and venules and arterioles (10- to 19-µm in diameter) labeled as %total venules and arterioles (VEN/ART) is for 1-wk animals. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. * P < 0.05 compared with control.

View larger version (44K): [in this window] [in a new window]   Fig. 5.   No. of capillaries/mm2 of myocyte and capillary diameter. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. * P < 0.05 compared with control.

Exercise induced the growth of arterioles. Vessels larger than capillaries were divided into five groups having diameters of 10-19.9, 20-30, 31-40, 41-70, and 71-120 µm. The latter four groups were easily identified as arterioles. The 10- to 19.9-µm group of vessels is composed of both arterioles and venules, which are not easily distinguished in this system. To estimate the fraction of each, a coronary artery and a vein were simultaneously injected with different silicone dyes, and then the colored vessels were counted. In two control animals, venules constituted 80% of the dyed vessels, and arterioles constituted the remaining 20%. The ratio was not determined for the exercised hearts in this study. However, in other studies, the ratio of venules to arterioles was maintained even in the presence of extensive hypertrophy (9). Therefore, in this study, it seems likely that the proportion of arterioles in the 10- to 19.9-µm size class in the exercised hearts is 20% of the total vessel count. Thus, for the sedentary control animal, we estimate that, in this size group, arterioles numbered 73.2/mm². Although studies in other species (19, 20, 23) have reported far fewer vessels in this size range, the value calculated above is consistent with detailed morphometric studies in the pig heart (11, 12). These studies have shown that only ~10-15% of capillaries (i.e., 300 capillaries) branch off the smallest arterioles and that there are approximately five capillaries connected to each small vessel. Thus structural data predicts ~60 arterioles in the 10- to 19.9-µm range. We know of no evidence that terminal arterioles in the pig connect to ~10-30 capillaries, as would be predicted by the lower number of arterioles which is often reported for this size class of vessels in other species. For the trained animals, we estimate that, in the 1-, 3-, 8-, and 16-wk groups, there were, respectively, 79, 96.8, 88, and 80 arterioles/mm². Thus the pattern of growth in the smallest class of arterioles was similar to that observed for the capillaries; namely, there was a transient increase in vessel density at 3 wk, followed by a decrease to near control levels.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Graph showing DNA labeling in nuclei of smooth muscle cells and endothelial cells in arterioles. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. SMC, smooth muscle cells (in arterioles >20 µm); ART ENDO, endothelial cells in arterioles. * P < 0.05 compared with control.

View larger version (47K): [in this window] [in a new window]   Fig. 7.   Graph showing arteriole cross-sectional areas (CSA; µm2/mm2) for each size group at each time point. Arterioles were divided into 4 size groups: 20-30, 31-40, 41-70, and 71-120 µm. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. * P < 0.05 compared with control.

View larger version (35K): [in this window] [in a new window]   Fig. 8.   Density for each group of arterioles: 20-30, 31-40, 41-70, and 71-120 µm/mm2. No. of arterioles/mm2 was only increased in the 20- to 30-µm group. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. * P < 0.05 compared with control.

View larger version (43K): [in this window] [in a new window]   Fig. 9.   Arteriole diameters. Arterioles were divided into four groups: 20-30, 31-40, 41-70, and 71-120 µm diameter. SE bars are included, but, in some instances, either bar or SE bar may be too small to see. * P < 0.05 compared with control.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Exercise and angiogenesis. In previous studies, exercise-induced moderate hypertrophy has been shown to be associated with a proportional increase in capillary number (27, 32). Similarly, in this study, angiogenesis was manifest in the maintenance of capillary density at control levels in the heart at 16 wk of training. Increased capillary density has generally not been associated with exercise in large species and adult animals, although it has been shown to occur with exercise in young rodents. However these measurements were usually made at the end of a long course of training. In contrast, an increase in capillary density was detected in this study after only 3 wk of training. The increase in capillary density compared with that in control animals occurred only at this early time and would have been missed if only the final 16-wk time point had been considered. Consistent with the increase in capillary density, an increase in sprouting at the early time points was observed. Early work by Clark and Clark (5) has previously shown that sprouting was necessary to achieve new capillaries.

CTR and maximal CBF. CTR can be viewed as a measurement of the ability of the vascular bed to transport nutrients to and metabolites away from cardiac myocytes, and this process has been shown to be augmented by exercise. CTR is composed of two features: the transport of blood to the capillaries and the capillary-mediated exchange between blood and tissue (13). Several possible mechanisms have been suggested to account for the training-induced increase in CTR, including altered capillary permeability, redistribution of blood flow and greater anatomic area (13). The results in this study clearly show that training resulted in anatomic changes in the capillary bed. The exercise protocol and CTR measurements used were modeled after Laughlin et al. (14) and Laughlin and Tomanek (15). These investigators reported a 51% increase in CTR after 12 wk of training with the use of the pig model (14). Similarly, we observed a maximal increase in CTR of 59% at 16 wk, which was preceded by incremental increases at the earlier times. Our observations suggest that structural adaptations contribute to the rising CTR. In the early phase of training, the surface area of the capillary bed was enhanced by both increased sprouting and increased capillary density while at later times the surface area of the bed was enlarged by an increase in the average capillary diameter.

Model for the progressive growth of small vessels into larger vessels. This study is unique in that we have investigated the effects of exercise on angiogenesis at several points in a time course. It is clear that the vascular changes induced by exercise are not uniform over the course of training used in this study. Rather, it appears that the adaptations occurred in a stepwise manner, such that an increase in capillary and small arteriolar density preceded both an increase in the number of the next size class of arterioles and an increase in the diameter of the larger vessels.