Pfliigers Arch (1992) 421:336-342
Joumal of Physiology 9 Springer-Verlag 1992
Regional capillary perfusion in muscles with limited blood supply: effects of torbafylline S. Egginton and O. Hudlicka Department of Physiology, University of Birmingham, Birmingham B15 2TT, UK Received September 27, 1991/Receivedafter revision March 20, 1992/AcceptedMarch 26, 1992 Abstract. Severe limitation of blood supply mainly affects the oxidative regions of skeletal muscles. In mammals, they are located medially and are thus not accessible to direct observation. We therefore investigated capillary perfusion in rat tibialis anterior, which has a predominantly glycolytic cortex and oxidative core, using timed intraarterial injection of the fluorochrome thioflavine S conjugated with serum albumin. Muscles with intact blood supply were compared with those in which the blood supply had been limited for 5 weeks by unilateral ligation of the common iliac artery. The effect of a new xanthine derivative, torbafylline (1% solution, 12.5 rag/ kg, in two daily doses by gavage, 7 days/week), was also studied. The capillary/fibre ratio was estimated for perfused capillaries (those filled with fluorochrome within 7.5 s after injection; Cp) and all capillaries (those subsequently stained for alkaline phosphatase; Ct), from micrographs of cryostat sections. Regional differentiation in relative capillary perfusion was evident in all muscles samples. Cp: Ct was 0.406 -t- 0.086 (mean • 95% CI) in the glycolytic cortex of the contralateral normal muscle, and 0.255 _ 0.071 in the oxidative core. Muscles with limited blood supply had a significantly lower pro, portion of perfused capillaries, 0.119 -t- 0.056 in glycolytic and 0.034 • 0.038 in oxidative regions. Torbafylline treatment nearly doubled perfusion in the glycolytic regions (Cp: Ct = 0.216 • 0.137) and nearly quadrupled it in oxidative (Cp:Ct = 0.121 ___0.151) regions of ischaemic muscles. It also improved perfusion in the contralateral muscles (Cp: Ct = 0.705 + 0.085 in the glycolytic cortex and 0.583 • 0.230 in the oxidative core). Limited blood supply thus affects perfusion of oxidative regions to a greater extent than glycolytic, while torbafylline may reduce the adverse effects. Key words: Fluorochrome - Capillary p e r f u s i o n Muscle ischaemia - Iliac artery ligation - Torbafylline
Correspondence to." S. Egginton
Introduction Severe limitation of blood supply to skeletal muscles, whether produced experimentally by ligation of supplying arteries or occurring as a consequence of arterial obstruction in peripheral vascular diseases, results in altered muscle metabolism and muscle fibre damage. Wholly oxidative muscles, or those regions of mixed muscles predominantly composed of oxidative fibres, are usually most affected. Karpati et al. [19] found a greater number of necrotic fibres in the purely oxidative soleus than in the mixed gastrocnemius muscles of rats 4 days after ligation of the abdominal aorta, and an increased proportion of muscle fibres with disarranged myofibrils has also been described in rat soleus 1 week after ligation of the common iliac artery [2]. Centrally located necrosis was shown in dog gracilis muscle exposed to 3 - 5 h ischaemia and 48 h reperfusion [22]. The size of the vascular bed, estimated by acrylic casts, in rat tibialis anterior (TA) was significantly lower 5 weeks after ligation of the common iliac artery [15], and muscle performance was considerably impaired [16]. Blood flow, measured by radiolabelled microspheres, was almost 50% lower in ischaemic TA [14], but direct observations of microcirculation in the superficial layer of TA, which is composed mainly of glycolytic fibres, revealed only relatively small changes in capillary perfusion [5]. This discrepancy might be due to a differential perfusion ofglycolytic a n d oxidative regions, since changes in microcirculatory parameters (intermittent flow and adhesion of leucocytes to venules) were more extensive in the ischaemic, purelY oxidative soleus than in the glycolytic TA. As oxidative fibres in mixed mammalian striated muscles are usually found in a medial location, however, they are not accessible to direct observation. We have therefore utilised timed perfusion with a fluorochrome, thioflavine S conjugated with bovine serum albumin [11], to determine how severe limitation of blood supplyaffects capillary perfusion in different regions of the TA. We also wished to determine whether torbafylline, a new xanthine derivative [7-ethoxymethyl- l-(5-hydroxy-5-methylhex-
337
yl)-3-methylxanthine] (Hoechst AG, Werk Albert, Wiesbaden [9]), would improve capillary perfusion. Torbafylline does not increase resting blood flow or capillary density in either control or ischaemic muscles [14], but improves performance in muscles with arterial ligation [25], a discrepancy that may be explained by improved capillary perfusion.
>F-
rr
Sprague Dawley rats of 4 0 0 - 6 0 0 g body weight were given torbafylline (12.5 mg kg-1, 1% aqueous solution) by gavage morning and evening, and littermates received a comparable volume of water for 7 days week-1 for 5 weeks. All animals had the right common iliac artery ligated just below the bifurcation from the aorta under halothane (Fluothane, ICI, Macclesfield) anaesthesia and aseptic conditions. Drug treatment started the day after the operation. The animals were anaesthetised with sodium pentobarbitone (Sagatal, RMB, Dagenham), 6 m g (100g body weight) -1 intraperitoneally, which was supplemented as required via a cannulated jugular vein. After intubation of the trachea, the tibialis anterior muscles of both legs were carefully exposed (to permit rapid excision) while the integrity of the circulation was maintained. The ventral abdominal wall was incised from xiphisternum to pubis and the intestines displaced. Exposed visceral organs were covered with gauze soaked in warm saline to minimise evaporative water loss, and a loose ligature was placed around the abdominal aorta just proximal of the aortic bifurcation. The superior mesenteric artery was cannulated for measurement of blood pressure and heart rate via a pressure transducer (Bell and Howell, Wembley, type 44220001), in order to monitor possible cardiovascular effects of the injectate. The carotid artery was cannulated for injection of the fluorochrome. Perfused capillaries were demonstrated as described in detail previously [8, 11] using a saturated thioflavine S (CI 49010; Sigma, Poole) solution in 6% bovine serum albumin and saline. Bolus injections were given via the right carotid artery [20] to maximise the time available for the dye to undergo blood mixing before reaching the muscle, and minimise inhomogeneity of labelling. Previous experiments [11] demonstrated that 7.5 s allowed a good comparison of capillary perfusion when muscle blood flow was increased by various patterns of contractions, and it was therefore decided to study perfusion in resting muscles with limited blood supply under these conditions. Comparison of relative (perfused) capillary density with the interval allowed for dye circulation shows the response of both regions to be similar in the time taken for full perfusion (Fig. 1). Muscles were therefore quickly removed 7.5 s after injection; this reveals the summed pattern of capillaries either actively being perfused when the ligature was closed, or which had been perfused during the previous 3.5 s, given a mean transit time of around 4 s for the dye front, as assessed by intravital microscopy (unpublished observations). The experimental protocol was as follows: 500 IU heparin (Weddel Pharmaceuticals Ltd., London) was given as a bolus via the jugular cannula and the animal was left undisturbed for 1 0 15 min. Thereafter a 1.5- to 2.2-ml bolus of thioflavine S solution was injected in around 2 s into the carotid artery, and flow to both limbs was stopped by ligation of the abdominal aorta at the end of the timed perfusion period. Both TA were then quickly cut free and slices about 3 gm thick coated in OCT mounting medium and immersed in isopentane precooled in liquid nitrogen. The removal and freezing were completed in less than 60 s to minimise extravasation of the dye. Muscle blocks were sectioned and photographed immediately, as localisation of the dye progressively deteriorated with storage, although satisfactory results were obtained from material held in liquid nitrogen overnight when the dye was still primarily in-
0.6
5..A O< O W
Materials and methods
o.8
Z u.I a
0.4
>
b-< rr
0.2
5
10
15
20
PERFUSION INTERVAL (sec)
Fig. 1. Relative capillary density: (capillaries identified by the presence of dye)/(capillaries identified by enzymatic activity), for various perfusion intervals measured from injection of dye to arterial ligation. A similar pattern is seen for both glycolytic and oxidative regions of the tibialis anterior (TA) muscle, filling being approximately proportional to filling time over the range of 7.5--15 s. ..... O, glycolytic TA; 9 0 , oxidative TA
travascular. For analysis, transverse sections 8-- 12 gm thick were cut on a cryostat microtome at - 2 0 ~ Sections were collected onto acetone-cleaned slides at room temperature, the consequent immediate dehydration by sublimation thus preventing additional dye extravasation. Two sets of serial sections, separated by an interval of 600 gm (greater than the maximal length of capillaries in rat TA [3]), were cut from each muscle. In this way two sections with different capillary populations were obtained for analysis of perfusion variability within and among muscles. All sections were cut from the middle third of TA where the muscle is clearly differentiated by colour and histochemical character into a glycolytic fibre cortex and oxidative fibre core [11], permitting comparison of capillary perfusion in two metabolically distinct regions. Sections were protected from light exposure during cutting, and immediately examined under a fluorescence microscope (Zeiss III RS condenser) with 50-W mercury source. The sections were epiilluminated using a 470-nm primary filter, a 510-nm dichroic interference mirror and a 528-nm secondary filter. This caused the thioflavine S to fluorescence at approximately 550 nm, which was observed using a 546- to 580-nm filter combination, when perfused capillaries appeared as blue dots (Fig. 2). Four to six fields in a region of each fibre type were photographed on each section at a magnification of x 125 using 35-mm 1000-ASA black and white film (Kodak 2425). The fields were selected at random by scratching crosses on the underside of the glass slide under both the glycolytic and oxidative regions with a diamond pencil. These fields were then located in transilluminated light by focusing on the lower face of the slide and centering the cross in the field of view. In addition to fluorescence photographs, each field was photographed while transilluminated (bright-field illumination) to record fibre outlines. Following fluorescence microscopy, the same section was stained for alkaline phosphatase using indoxytetrazolium staining [29]; this enzyme is present in capillary endothelium and demonstrates all anatomically present capillaries [1], and staining is unaffected by ligation [2, 14]. The same fields were located and re-photographed using 35-mm 125-ASA film (Kodak Plus-X pan). Each set of matched fields was traced from the negative using a film reader at x 9 magnification. Fibre outlines and the positions of perfused and all capillaries were drawn; single fields contained 6 0 - 1 0 0 glycolytic fibres or 150-200 oxidative fibres. Fibres and capillaries on the edges of each field were samples in an unbiased manner [7], and
338
Fig. 2. Cross-sections of tibialis an-
terior with capillaries stained for alkaline phosphatase depicted as black dots (top). The same cross-section of capillaries perfused with thiofiavine S; capillaries are depicted as blue dots. Crosses in both sections indicate the same muscle fibre oblique sections omitted from analysis. Capillary/muscle-fibre ratios were calculated for perfused and all capillaries. All data are presented as means + 95% confidence intervals; paired or unpaired t-tests and two-way analysis of variance were used for evaluation of statistical significance, where appropriate.
Results
M e a n arterial blood pressure varied little between groups ( 1 2 1 - 1 2 6 m m H g ; 1 6 - 1 7 kPa) but was increased by 2 0 - 50 m m H g (2.7 - 6.7 kPa) for 1 - 2 s during injection of the dye, being thereafter 10 - 20 m m H g (1.3 - 2.7 kPa) lower than initial values. N o difference was seen between experimental groups.
The capillary/fibre ratio for all anatomically present capillaries (Ct) was always higher in the oxidative core than in the glycolytic cortex, and there was no significant difference between Ct in muscles with limited blood supply c o m p a r e d to that in contralateral muscles (Table 1). The capillary/fibre ratio for fluorochromefilled ('perfused') capillaries (Cp) was m u c h lower than Ct, in both the oxidative core and glycolytic cortex in control (contralateral) muscles of water-fed rats. Limited blood supply resulted in a drastic decrease in the proportion of perfused capillaries (Fig. 3), being proportionally worse in the core than in the cortex (Table 1). Torbafylline treatment was without effect on the capillary/fibre ratio for anatomically present capillaries,
339 Table 1. Capillary/fibre ratio for perfused (Cp) and total (Ct) number of capillaries in rat tibialis anterior muscle, 5 weeks after ligation of the right common iliac artery or contralateral controls, with and without concomitant treatment with torbafyllinea Ratio
Water-fed
Torbafylline-fed
Contralateral
Ligated
Contralateral
Ligated
0.069 _+0.085 (6) 1.762 _+0.334 (6) 0.034 + 0.038 (6)
1.071 + 0.614 (6) 1.722 + 0.532 (6) 0.583 _ 0.230 (6)
0.243 _+ 0.323 (8) 1.872 __ 0.164 (8)
Cp:C t
0.481 +_0.141 (10) 1.849 + 0.149 (10) 0.255 _+0.071 (10)
0.121 +__0.151 (8)
Outer glyeolytic cortex Cp Ct Cp:Ct
0.648 _+0.151 (10) 1.575 _+0.147 (10) 0.406 + 0.086 (10)
0.168 + 0.109 (6) 1.374 ___0.262 (6) 0.119 _+0.056 (6)
0.949 _+ 0.190 (6) 1.348 -t- 0.143 (6) 0.705 + 0.085 (6)
0.311 -t- 0.205 (8) 1.432 _+ 0.164 (8) 0.216 _-t-0.137 (8)
Inner oxidative core Cp Ct
a Data given as mean + 95% CI (n sections)
Cp:Ct
60 20
C
A
100
L
LIP (29)
LIP (281
0.8
L/All(28)
0.5
20
0.4
60 ~']
C/P (12)
C/P (12)
0.3
20
0.2 C/AU(12)
60
C/All (12)
0.1 20
B
D
u. 60 20
LIP (37)
LIP (42)
"7
60
L/All (42)
,3o,
20 C/P (25)
Cortex
0.7 0.6
L/All(29)
Core
_i
l dl
L L&T L&'F L Fig. 4. Proportion of perfused capillaries (capillary/fibre ratio for perfused capillaries, Cp, divided by that for total capillaries Ct) in the oxidative core and glycolytic cortex of rat tibialis anterior following ligation (L), and ligation plus torbafylline (L & 7) treatment. Group means were compared by unpaired t-test; asterisk denotes significant difference vs ligated muscle. [] = Contralateral muscle; [] = ischaemic muscle; * = p < 0.05 vs L
C/P (25)
20 C/All (25) 20"
UT7 0.4
0.8
1.2 C/F
1.6
2.0
C/Am(25)
. 2.4
0.4
! 0.8
1.2
1.6
,.. 2.0
2.4
C/F
Fig. 3 A - D . Frequency distribution of capillary/fibre ratio (C/F) for perfused and total capillary supply in rat tibialis anterior. A, B Oxidative core; C, D glycolytic cortex. L, muscle with limited blood supply; C, contralateral control muscle; P, perfused capillaries; All, total anatomical capillary supply. Numbers in parentheses refer to fields analysed. A, C From control, water-fed rats; B, D from experimental, torbafylline-fed rats
but resulted in a dramatic improvement in the proportion of perfused capillaries both in the contralateral and ischaemic muscles (Fig. 4). Cp: Ct increased by around a factor of 2 in both the cortex and core of contralateral muscles. A similar increase was observed in the glycolytic part of ischaemic muscles, and an even more dramatic (3.6-fold) increase in comparison with water fed animals in the oxidative core (Table 1, Fig. 4). Thus, limitation of blood supply resulted in a greater impairment of perfusion in the oxidative than in the glycolytic part of TA, and torbafylline treatment resulted
in an increase in the proportion of perfused capillaries so that the values, although still lower, were no longer significantly different from those of control muscles of water-fed animals (t-test). Interestingly, in all cases there was an increased heterogeneity of perfusion, indicated by a wider distribution of values, as a result of torbafylline treatment (Fig. 3). Analysis of variance, using animal means, showed the effect of either ligation or drug treatment to be highly significant, but that there was no significant interaction, i.e. torbafylline had a similar effect on intact and ischaemic muscle. However, as two experimental groups had few (n = 3) animals, the calculated P values (0.11 and 0.12) are likely to be very conservative. Our aim of obtaining two independent samples of the capillary bed from a muscle, by means of systemic random sections, was validated by showing that the difference between sections for each variable was not significantly different from zero (null hypothesis, one-tail t-test). Hence, subsequent statistical analysis used mean values from 2n independent sections. In this case variance will include both a between-animal component and a non-significant
340 Table 2. Analysis of variance table for two-factor analysis of variance for Cp:Ct"
Region of muscle
Source
df
SS
MS
F
P
Inner oxidative core
Drug (A) Ligation (B) Interaction (AB) Error Drug (A) Ligation (B) Interaction (AB) Error
1 1 1 26 1 1 1 26
0.309 0.836 0.104 0.563 0.282 1.077 0.073 0.366
0.309 0.836 0.104 0.022 0.282 1.077 0.073 0.014
14.297 38.636 4.803
0.0008 0.0001 0.376
20.012 76.439 5.174
0.0001 0.0001 0.0314
Outer glycolyticcortex
" Both ligation and drug effects are highly significant within-animal component, which more realistically described group trends (Fig. 3, Table 2). Discussion
Blood flow in TA and extensor digitorum longus was decreased by around 50% 5 weeks after ligation of the iliac artery [14]. However, there was no change in overall capillary/fibre ratio (using all anatomically present capillaries) either in untreated muscles [16], or in those treated with torbafylline [14]. The present data confirm these previous findings. In addition, they also demonstrate poorer perfusion of capillaries in the oxidative compared to glycolytic regions in control muscles, with a greater impairment of perfusion in the oxidative than in the glycolytic part with ischaemia, and improved perfusion in both regions by torbafylline. Capillary perfusion in ischaemic muscles has so far been studied predominantly after acute limitation of the blood supply. Mikulash and Tyml [24] found a decrease in perfusion pressure and increased heterogeneity of flow immediately after occlusion in frog sartorius, while Tyml and Budreau [28] showed a progressively larger muscleto-muscle heterogeneity of red blood cell velocity with longer (1 s - 30 min) duration of occlusion in the same tissue. In addition, Lindbom and Arfors [23] described a decrease in the number of capillaries with moving red blood cells when perfusion pressure decreased below 50 m m H g (6.7 kPa). Limited blood supply would therefore appear to reduce bulk flow by accentuating the naturally occurring heterogeneity of capillary perfusion, which is due mainly to intermittency of flow within individual capillaries and the proportion of capillaries with continuous or intermittent flow. Previous experiments on chronic limitation of blood supply only addressed the question of capillary perfusion to a limited extent, but of necessity ignored the effects on perfusion of the deeper (medial) oxidative regions. Dawson and Hudlicka [5] studied the microcirculation in predominantly glycolytic and purely oxidative muscles (surface of TA and soleus, respectively) using intravital microscopy. The main differences were in capillary diameters and flow velocity, with calculated flow per capillary being lower in soleus than TA. A parallel situation may be present in the present study, where a lower proportion of perfused capillaries was found in the oxidative region
of TA. Five weeks after ligation of the common iliac artery, capillary diameters were decreased while the flow velocity was unchanged, resulting in a small decrease in TA blood flow [5]. However, this could not explain the 50% decrease in bulk flow measured by microspheres [14]. In soleus, which is more comparable to the oxidative core of TA, ligation resulted in an increased proportion of capillaries with intermittent flow, a significant increase in the time spent stationary by erythrocytes in individual capillaries (greater than 1 min duration), and an increased proportion of white blood cells adhering to venules [5]. These findings may explain the present data on poor perfusion in the oxidative region of TA, which was not improved by extending the period of observation (Fig. 5). Intravital microscopy only allows observation of capillary perfusion in the superficially located capillaries. Since ischaemia affects more the central parts of muscles [22], it was considered important to introduce a method that would enable the study of capillary perfusion in regions not accessible for direct observation [8]. Hargreaves et al. [11] used thioflavine S to study capillary perfusion in both the glycolytic and the oxidative parts of TA. They found that 27% of capillaries were perfused in the core and 39% in the cortex when blood flow was stopped 7.5 s after the bolus injection. These values are very similar to those found in the present experiments for control (contralateral) muscles of water-fed animals. Lower capillary "perfusion" in the oxidative than in the glycolytic regions is in agreement with Renkin's data on timed (7.5 s) India ink perfusion in oxidative and glyeolytic regions of rabbit gastrocnemius and soleus [26]. Prolongation of the perfusion interval to 15 s resulted in more than 80% perfused capillaries in both regions of control muscles [11]. It is unlikely that improved filling would be due to flow intermittency as a result ofvasomotion, since the duration of one cycle of intermittent flow was similar ( 2 5 - 27 s) in capillaries supplying glycolytic (TA cortex) and oxidative (soleus) muscles E3], and was longer than that required for complete filling with fluorochrome [11]. Poorer filling in the oxidative region is more likely to be a result of reduced erythrocyte velocity, and possibly greater capillary tortuosity [3]. Thus, lower capillary perfusion in the oxidative than in the glycolytic region is in agreement with the data of Renkin et al. on timed (7.5 s) India ink perfusion in oxidative and glycolytic regions of rabbit gastrocnemius
341 r LU
E
1.0
CORTEX
CORE
(glycolytic)
(oxidative)
Joumal of Physiology 9 Springer-Verlag 1992
Regional capillary perfusion in muscles with limited blood supply: effects of torbafylline S. Egginton and O. Hudlicka Department of Physiology, University of Birmingham, Birmingham B15 2TT, UK Received September 27, 1991/Receivedafter revision March 20, 1992/AcceptedMarch 26, 1992 Abstract. Severe limitation of blood supply mainly affects the oxidative regions of skeletal muscles. In mammals, they are located medially and are thus not accessible to direct observation. We therefore investigated capillary perfusion in rat tibialis anterior, which has a predominantly glycolytic cortex and oxidative core, using timed intraarterial injection of the fluorochrome thioflavine S conjugated with serum albumin. Muscles with intact blood supply were compared with those in which the blood supply had been limited for 5 weeks by unilateral ligation of the common iliac artery. The effect of a new xanthine derivative, torbafylline (1% solution, 12.5 rag/ kg, in two daily doses by gavage, 7 days/week), was also studied. The capillary/fibre ratio was estimated for perfused capillaries (those filled with fluorochrome within 7.5 s after injection; Cp) and all capillaries (those subsequently stained for alkaline phosphatase; Ct), from micrographs of cryostat sections. Regional differentiation in relative capillary perfusion was evident in all muscles samples. Cp: Ct was 0.406 -t- 0.086 (mean • 95% CI) in the glycolytic cortex of the contralateral normal muscle, and 0.255 _ 0.071 in the oxidative core. Muscles with limited blood supply had a significantly lower pro, portion of perfused capillaries, 0.119 -t- 0.056 in glycolytic and 0.034 • 0.038 in oxidative regions. Torbafylline treatment nearly doubled perfusion in the glycolytic regions (Cp: Ct = 0.216 • 0.137) and nearly quadrupled it in oxidative (Cp:Ct = 0.121 ___0.151) regions of ischaemic muscles. It also improved perfusion in the contralateral muscles (Cp: Ct = 0.705 + 0.085 in the glycolytic cortex and 0.583 • 0.230 in the oxidative core). Limited blood supply thus affects perfusion of oxidative regions to a greater extent than glycolytic, while torbafylline may reduce the adverse effects. Key words: Fluorochrome - Capillary p e r f u s i o n Muscle ischaemia - Iliac artery ligation - Torbafylline
Correspondence to." S. Egginton
Introduction Severe limitation of blood supply to skeletal muscles, whether produced experimentally by ligation of supplying arteries or occurring as a consequence of arterial obstruction in peripheral vascular diseases, results in altered muscle metabolism and muscle fibre damage. Wholly oxidative muscles, or those regions of mixed muscles predominantly composed of oxidative fibres, are usually most affected. Karpati et al. [19] found a greater number of necrotic fibres in the purely oxidative soleus than in the mixed gastrocnemius muscles of rats 4 days after ligation of the abdominal aorta, and an increased proportion of muscle fibres with disarranged myofibrils has also been described in rat soleus 1 week after ligation of the common iliac artery [2]. Centrally located necrosis was shown in dog gracilis muscle exposed to 3 - 5 h ischaemia and 48 h reperfusion [22]. The size of the vascular bed, estimated by acrylic casts, in rat tibialis anterior (TA) was significantly lower 5 weeks after ligation of the common iliac artery [15], and muscle performance was considerably impaired [16]. Blood flow, measured by radiolabelled microspheres, was almost 50% lower in ischaemic TA [14], but direct observations of microcirculation in the superficial layer of TA, which is composed mainly of glycolytic fibres, revealed only relatively small changes in capillary perfusion [5]. This discrepancy might be due to a differential perfusion ofglycolytic a n d oxidative regions, since changes in microcirculatory parameters (intermittent flow and adhesion of leucocytes to venules) were more extensive in the ischaemic, purelY oxidative soleus than in the glycolytic TA. As oxidative fibres in mixed mammalian striated muscles are usually found in a medial location, however, they are not accessible to direct observation. We have therefore utilised timed perfusion with a fluorochrome, thioflavine S conjugated with bovine serum albumin [11], to determine how severe limitation of blood supplyaffects capillary perfusion in different regions of the TA. We also wished to determine whether torbafylline, a new xanthine derivative [7-ethoxymethyl- l-(5-hydroxy-5-methylhex-
337
yl)-3-methylxanthine] (Hoechst AG, Werk Albert, Wiesbaden [9]), would improve capillary perfusion. Torbafylline does not increase resting blood flow or capillary density in either control or ischaemic muscles [14], but improves performance in muscles with arterial ligation [25], a discrepancy that may be explained by improved capillary perfusion.
>F-
rr
Sprague Dawley rats of 4 0 0 - 6 0 0 g body weight were given torbafylline (12.5 mg kg-1, 1% aqueous solution) by gavage morning and evening, and littermates received a comparable volume of water for 7 days week-1 for 5 weeks. All animals had the right common iliac artery ligated just below the bifurcation from the aorta under halothane (Fluothane, ICI, Macclesfield) anaesthesia and aseptic conditions. Drug treatment started the day after the operation. The animals were anaesthetised with sodium pentobarbitone (Sagatal, RMB, Dagenham), 6 m g (100g body weight) -1 intraperitoneally, which was supplemented as required via a cannulated jugular vein. After intubation of the trachea, the tibialis anterior muscles of both legs were carefully exposed (to permit rapid excision) while the integrity of the circulation was maintained. The ventral abdominal wall was incised from xiphisternum to pubis and the intestines displaced. Exposed visceral organs were covered with gauze soaked in warm saline to minimise evaporative water loss, and a loose ligature was placed around the abdominal aorta just proximal of the aortic bifurcation. The superior mesenteric artery was cannulated for measurement of blood pressure and heart rate via a pressure transducer (Bell and Howell, Wembley, type 44220001), in order to monitor possible cardiovascular effects of the injectate. The carotid artery was cannulated for injection of the fluorochrome. Perfused capillaries were demonstrated as described in detail previously [8, 11] using a saturated thioflavine S (CI 49010; Sigma, Poole) solution in 6% bovine serum albumin and saline. Bolus injections were given via the right carotid artery [20] to maximise the time available for the dye to undergo blood mixing before reaching the muscle, and minimise inhomogeneity of labelling. Previous experiments [11] demonstrated that 7.5 s allowed a good comparison of capillary perfusion when muscle blood flow was increased by various patterns of contractions, and it was therefore decided to study perfusion in resting muscles with limited blood supply under these conditions. Comparison of relative (perfused) capillary density with the interval allowed for dye circulation shows the response of both regions to be similar in the time taken for full perfusion (Fig. 1). Muscles were therefore quickly removed 7.5 s after injection; this reveals the summed pattern of capillaries either actively being perfused when the ligature was closed, or which had been perfused during the previous 3.5 s, given a mean transit time of around 4 s for the dye front, as assessed by intravital microscopy (unpublished observations). The experimental protocol was as follows: 500 IU heparin (Weddel Pharmaceuticals Ltd., London) was given as a bolus via the jugular cannula and the animal was left undisturbed for 1 0 15 min. Thereafter a 1.5- to 2.2-ml bolus of thioflavine S solution was injected in around 2 s into the carotid artery, and flow to both limbs was stopped by ligation of the abdominal aorta at the end of the timed perfusion period. Both TA were then quickly cut free and slices about 3 gm thick coated in OCT mounting medium and immersed in isopentane precooled in liquid nitrogen. The removal and freezing were completed in less than 60 s to minimise extravasation of the dye. Muscle blocks were sectioned and photographed immediately, as localisation of the dye progressively deteriorated with storage, although satisfactory results were obtained from material held in liquid nitrogen overnight when the dye was still primarily in-
0.6
5..A O< O W
Materials and methods
o.8
Z u.I a
0.4
>
b-< rr
0.2
5
10
15
20
PERFUSION INTERVAL (sec)
Fig. 1. Relative capillary density: (capillaries identified by the presence of dye)/(capillaries identified by enzymatic activity), for various perfusion intervals measured from injection of dye to arterial ligation. A similar pattern is seen for both glycolytic and oxidative regions of the tibialis anterior (TA) muscle, filling being approximately proportional to filling time over the range of 7.5--15 s. ..... O, glycolytic TA; 9 0 , oxidative TA
travascular. For analysis, transverse sections 8-- 12 gm thick were cut on a cryostat microtome at - 2 0 ~ Sections were collected onto acetone-cleaned slides at room temperature, the consequent immediate dehydration by sublimation thus preventing additional dye extravasation. Two sets of serial sections, separated by an interval of 600 gm (greater than the maximal length of capillaries in rat TA [3]), were cut from each muscle. In this way two sections with different capillary populations were obtained for analysis of perfusion variability within and among muscles. All sections were cut from the middle third of TA where the muscle is clearly differentiated by colour and histochemical character into a glycolytic fibre cortex and oxidative fibre core [11], permitting comparison of capillary perfusion in two metabolically distinct regions. Sections were protected from light exposure during cutting, and immediately examined under a fluorescence microscope (Zeiss III RS condenser) with 50-W mercury source. The sections were epiilluminated using a 470-nm primary filter, a 510-nm dichroic interference mirror and a 528-nm secondary filter. This caused the thioflavine S to fluorescence at approximately 550 nm, which was observed using a 546- to 580-nm filter combination, when perfused capillaries appeared as blue dots (Fig. 2). Four to six fields in a region of each fibre type were photographed on each section at a magnification of x 125 using 35-mm 1000-ASA black and white film (Kodak 2425). The fields were selected at random by scratching crosses on the underside of the glass slide under both the glycolytic and oxidative regions with a diamond pencil. These fields were then located in transilluminated light by focusing on the lower face of the slide and centering the cross in the field of view. In addition to fluorescence photographs, each field was photographed while transilluminated (bright-field illumination) to record fibre outlines. Following fluorescence microscopy, the same section was stained for alkaline phosphatase using indoxytetrazolium staining [29]; this enzyme is present in capillary endothelium and demonstrates all anatomically present capillaries [1], and staining is unaffected by ligation [2, 14]. The same fields were located and re-photographed using 35-mm 125-ASA film (Kodak Plus-X pan). Each set of matched fields was traced from the negative using a film reader at x 9 magnification. Fibre outlines and the positions of perfused and all capillaries were drawn; single fields contained 6 0 - 1 0 0 glycolytic fibres or 150-200 oxidative fibres. Fibres and capillaries on the edges of each field were samples in an unbiased manner [7], and
338
Fig. 2. Cross-sections of tibialis an-
terior with capillaries stained for alkaline phosphatase depicted as black dots (top). The same cross-section of capillaries perfused with thiofiavine S; capillaries are depicted as blue dots. Crosses in both sections indicate the same muscle fibre oblique sections omitted from analysis. Capillary/muscle-fibre ratios were calculated for perfused and all capillaries. All data are presented as means + 95% confidence intervals; paired or unpaired t-tests and two-way analysis of variance were used for evaluation of statistical significance, where appropriate.
Results
M e a n arterial blood pressure varied little between groups ( 1 2 1 - 1 2 6 m m H g ; 1 6 - 1 7 kPa) but was increased by 2 0 - 50 m m H g (2.7 - 6.7 kPa) for 1 - 2 s during injection of the dye, being thereafter 10 - 20 m m H g (1.3 - 2.7 kPa) lower than initial values. N o difference was seen between experimental groups.
The capillary/fibre ratio for all anatomically present capillaries (Ct) was always higher in the oxidative core than in the glycolytic cortex, and there was no significant difference between Ct in muscles with limited blood supply c o m p a r e d to that in contralateral muscles (Table 1). The capillary/fibre ratio for fluorochromefilled ('perfused') capillaries (Cp) was m u c h lower than Ct, in both the oxidative core and glycolytic cortex in control (contralateral) muscles of water-fed rats. Limited blood supply resulted in a drastic decrease in the proportion of perfused capillaries (Fig. 3), being proportionally worse in the core than in the cortex (Table 1). Torbafylline treatment was without effect on the capillary/fibre ratio for anatomically present capillaries,
339 Table 1. Capillary/fibre ratio for perfused (Cp) and total (Ct) number of capillaries in rat tibialis anterior muscle, 5 weeks after ligation of the right common iliac artery or contralateral controls, with and without concomitant treatment with torbafyllinea Ratio
Water-fed
Torbafylline-fed
Contralateral
Ligated
Contralateral
Ligated
0.069 _+0.085 (6) 1.762 _+0.334 (6) 0.034 + 0.038 (6)
1.071 + 0.614 (6) 1.722 + 0.532 (6) 0.583 _ 0.230 (6)
0.243 _+ 0.323 (8) 1.872 __ 0.164 (8)
Cp:C t
0.481 +_0.141 (10) 1.849 + 0.149 (10) 0.255 _+0.071 (10)
0.121 +__0.151 (8)
Outer glyeolytic cortex Cp Ct Cp:Ct
0.648 _+0.151 (10) 1.575 _+0.147 (10) 0.406 + 0.086 (10)
0.168 + 0.109 (6) 1.374 ___0.262 (6) 0.119 _+0.056 (6)
0.949 _+ 0.190 (6) 1.348 -t- 0.143 (6) 0.705 + 0.085 (6)
0.311 -t- 0.205 (8) 1.432 _+ 0.164 (8) 0.216 _-t-0.137 (8)
Inner oxidative core Cp Ct
a Data given as mean + 95% CI (n sections)
Cp:Ct
60 20
C
A
100
L
LIP (29)
LIP (281
0.8
L/All(28)
0.5
20
0.4
60 ~']
C/P (12)
C/P (12)
0.3
20
0.2 C/AU(12)
60
C/All (12)
0.1 20
B
D
u. 60 20
LIP (37)
LIP (42)
"7
60
L/All (42)
,3o,
20 C/P (25)
Cortex
0.7 0.6
L/All(29)
Core
_i
l dl
L L&T L&'F L Fig. 4. Proportion of perfused capillaries (capillary/fibre ratio for perfused capillaries, Cp, divided by that for total capillaries Ct) in the oxidative core and glycolytic cortex of rat tibialis anterior following ligation (L), and ligation plus torbafylline (L & 7) treatment. Group means were compared by unpaired t-test; asterisk denotes significant difference vs ligated muscle. [] = Contralateral muscle; [] = ischaemic muscle; * = p < 0.05 vs L
C/P (25)
20 C/All (25) 20"
UT7 0.4
0.8
1.2 C/F
1.6
2.0
C/Am(25)
. 2.4
0.4
! 0.8
1.2
1.6
,.. 2.0
2.4
C/F
Fig. 3 A - D . Frequency distribution of capillary/fibre ratio (C/F) for perfused and total capillary supply in rat tibialis anterior. A, B Oxidative core; C, D glycolytic cortex. L, muscle with limited blood supply; C, contralateral control muscle; P, perfused capillaries; All, total anatomical capillary supply. Numbers in parentheses refer to fields analysed. A, C From control, water-fed rats; B, D from experimental, torbafylline-fed rats
but resulted in a dramatic improvement in the proportion of perfused capillaries both in the contralateral and ischaemic muscles (Fig. 4). Cp: Ct increased by around a factor of 2 in both the cortex and core of contralateral muscles. A similar increase was observed in the glycolytic part of ischaemic muscles, and an even more dramatic (3.6-fold) increase in comparison with water fed animals in the oxidative core (Table 1, Fig. 4). Thus, limitation of blood supply resulted in a greater impairment of perfusion in the oxidative than in the glycolytic part of TA, and torbafylline treatment resulted
in an increase in the proportion of perfused capillaries so that the values, although still lower, were no longer significantly different from those of control muscles of water-fed animals (t-test). Interestingly, in all cases there was an increased heterogeneity of perfusion, indicated by a wider distribution of values, as a result of torbafylline treatment (Fig. 3). Analysis of variance, using animal means, showed the effect of either ligation or drug treatment to be highly significant, but that there was no significant interaction, i.e. torbafylline had a similar effect on intact and ischaemic muscle. However, as two experimental groups had few (n = 3) animals, the calculated P values (0.11 and 0.12) are likely to be very conservative. Our aim of obtaining two independent samples of the capillary bed from a muscle, by means of systemic random sections, was validated by showing that the difference between sections for each variable was not significantly different from zero (null hypothesis, one-tail t-test). Hence, subsequent statistical analysis used mean values from 2n independent sections. In this case variance will include both a between-animal component and a non-significant
340 Table 2. Analysis of variance table for two-factor analysis of variance for Cp:Ct"
Region of muscle
Source
df
SS
MS
F
P
Inner oxidative core
Drug (A) Ligation (B) Interaction (AB) Error Drug (A) Ligation (B) Interaction (AB) Error
1 1 1 26 1 1 1 26
0.309 0.836 0.104 0.563 0.282 1.077 0.073 0.366
0.309 0.836 0.104 0.022 0.282 1.077 0.073 0.014
14.297 38.636 4.803
0.0008 0.0001 0.376
20.012 76.439 5.174
0.0001 0.0001 0.0314
Outer glycolyticcortex
" Both ligation and drug effects are highly significant within-animal component, which more realistically described group trends (Fig. 3, Table 2). Discussion
Blood flow in TA and extensor digitorum longus was decreased by around 50% 5 weeks after ligation of the iliac artery [14]. However, there was no change in overall capillary/fibre ratio (using all anatomically present capillaries) either in untreated muscles [16], or in those treated with torbafylline [14]. The present data confirm these previous findings. In addition, they also demonstrate poorer perfusion of capillaries in the oxidative compared to glycolytic regions in control muscles, with a greater impairment of perfusion in the oxidative than in the glycolytic part with ischaemia, and improved perfusion in both regions by torbafylline. Capillary perfusion in ischaemic muscles has so far been studied predominantly after acute limitation of the blood supply. Mikulash and Tyml [24] found a decrease in perfusion pressure and increased heterogeneity of flow immediately after occlusion in frog sartorius, while Tyml and Budreau [28] showed a progressively larger muscleto-muscle heterogeneity of red blood cell velocity with longer (1 s - 30 min) duration of occlusion in the same tissue. In addition, Lindbom and Arfors [23] described a decrease in the number of capillaries with moving red blood cells when perfusion pressure decreased below 50 m m H g (6.7 kPa). Limited blood supply would therefore appear to reduce bulk flow by accentuating the naturally occurring heterogeneity of capillary perfusion, which is due mainly to intermittency of flow within individual capillaries and the proportion of capillaries with continuous or intermittent flow. Previous experiments on chronic limitation of blood supply only addressed the question of capillary perfusion to a limited extent, but of necessity ignored the effects on perfusion of the deeper (medial) oxidative regions. Dawson and Hudlicka [5] studied the microcirculation in predominantly glycolytic and purely oxidative muscles (surface of TA and soleus, respectively) using intravital microscopy. The main differences were in capillary diameters and flow velocity, with calculated flow per capillary being lower in soleus than TA. A parallel situation may be present in the present study, where a lower proportion of perfused capillaries was found in the oxidative region
of TA. Five weeks after ligation of the common iliac artery, capillary diameters were decreased while the flow velocity was unchanged, resulting in a small decrease in TA blood flow [5]. However, this could not explain the 50% decrease in bulk flow measured by microspheres [14]. In soleus, which is more comparable to the oxidative core of TA, ligation resulted in an increased proportion of capillaries with intermittent flow, a significant increase in the time spent stationary by erythrocytes in individual capillaries (greater than 1 min duration), and an increased proportion of white blood cells adhering to venules [5]. These findings may explain the present data on poor perfusion in the oxidative region of TA, which was not improved by extending the period of observation (Fig. 5). Intravital microscopy only allows observation of capillary perfusion in the superficially located capillaries. Since ischaemia affects more the central parts of muscles [22], it was considered important to introduce a method that would enable the study of capillary perfusion in regions not accessible for direct observation [8]. Hargreaves et al. [11] used thioflavine S to study capillary perfusion in both the glycolytic and the oxidative parts of TA. They found that 27% of capillaries were perfused in the core and 39% in the cortex when blood flow was stopped 7.5 s after the bolus injection. These values are very similar to those found in the present experiments for control (contralateral) muscles of water-fed animals. Lower capillary "perfusion" in the oxidative than in the glycolytic regions is in agreement with Renkin's data on timed (7.5 s) India ink perfusion in oxidative and glyeolytic regions of rabbit gastrocnemius and soleus [26]. Prolongation of the perfusion interval to 15 s resulted in more than 80% perfused capillaries in both regions of control muscles [11]. It is unlikely that improved filling would be due to flow intermittency as a result ofvasomotion, since the duration of one cycle of intermittent flow was similar ( 2 5 - 27 s) in capillaries supplying glycolytic (TA cortex) and oxidative (soleus) muscles E3], and was longer than that required for complete filling with fluorochrome [11]. Poorer filling in the oxidative region is more likely to be a result of reduced erythrocyte velocity, and possibly greater capillary tortuosity [3]. Thus, lower capillary perfusion in the oxidative than in the glycolytic region is in agreement with the data of Renkin et al. on timed (7.5 s) India ink perfusion in oxidative and glycolytic regions of rabbit gastrocnemius
341 r LU
E
1.0
CORTEX
CORE
(glycolytic)
(oxidative)
