Main

Responses to vasoactive influences of the same segment of a regional vasculature vary between species (1). Responses can also vary between segments of different size and location within the same regional vasculature of a single species-for example rabbit (2), cat (3), and sheep (4). Many characteristics have been shown to vary with vascular diameter (5,6). Variation in functional features in some instances can be related to patterns of embryologic development (7), which can provide insight into why transitions occur at a particular anatomical location. Changes can also occur during development and aging (8–10). Parts of some vascular beds exhibit specialized features (11). Although in some instances this variability seems appropriate to function, frequently there is no known rationale for such specialization. The cellular basis of the variation is sometimes understood; for example the sensitivity of a series of rabbit arteries and veins to NE has been related to variation in adrenergic receptor affinity and density (12,13). Recently, branch specific variation in agonist affinity and receptor density and occupancy of α1-adrenoceptors in sheep cerebral arteries has been demonstrated (4). In human blood vessels functional characteristics are altered in many diseases, for example, hypertension (14) (14) and vasospasm (15).

The present report documents in vitro studies of the responsiveness of segments of human cerebral arteries obtained from eight infants aged at birth between 23 wk of gestation and term, to activation of the perivascular innervation and exposure to a number of vasoactive agents acting on the endothelium or smooth muscle. Some of these compounds have physiologic relevance and are of potential pathophysiologic significance. The capacity of cerebral artery segments from the same eight infants to develop spontaneous myogenic tone responses to stretch and to dilate to increases in intralumenal flow has been previously published (16). In general, the vascular responses found in the infant were similar to the adult with the exception that 1) contractions to sympathetic nerve activation and to NE were greater than in adult pial arteries of similar diameter and adult proximal major cerebral arteries. Our findings suggest that this is related to a more dense sympathetic innervation in the infant and the closer proximity of nerves to vascular smooth muscle cells combined with increased responsiveness of the cells to NE; 2) a substantial potentiation of agonist-induced tone by indomethacin; and 3) an indication of the emergence of constrictor responses to PGF2α and 5-HT during the developmental period studied. These observations indicate that some functional features of the human cerebrovascular system change during development.

METHODS

This study was approved by the Ethics Committee of the University of Vermont. After consent from the parents, upon the decease of the infant the cerebral arteries were obtained during autopsy. The cranial contents were removed according to the method of Valdes-Dapena et al. (17). In all cases the removal of the blood vessels was completed within 2 h of death. The leptomeninges at the base of the brain were carefully dissected to expose the major arteries. Portions of the middle cerebral and basilar arteries were removed and placed in oxygenated, cold physiologic saline solution [composition, in mM, NaCl 130, KCl 4.7, KH2PO4 1.18, MgSO4 1.17, NaHCO3 14.9, EDTA 0.026, dextrose 11.0, and 100 µM deferoxamine, 10 U/mL heparin, 50 U/mL penicillin, and 50 µg/mL streptomycin]. They were transferred to the laboratory and prepared for in vitro study. Only segments free of adherent blood were studied. They were either examined on the same day of removal or after storage for 24 h at 4°C in physiologic saline solution.

Isometric myograph technique. Arteries were cut into rings 3-4 mm long and two wires passed through the lumen of each segment which was suspended in a resistance artery myograph (18,19) in Krebs physiologic solution maintained at pH = 7.4 ± 0.1, 37°C, gassed continuously with 95% O2/5% CO2. Each ring was connected to a force transducer (Grass FT03, Quincy, MA), and changes in isometric force were recorded on a strip recorder. Internal diameter was measured using a video camera (Colorado video micrometer). The myograph mounting wires were slowly separated until a just significant change in the force record was observed. Wire separation was taken to be half the unstretched circumference.

Because the segments from different infants varied in their internal diameter and the time required for determination of the optimum length of the smooth muscle cells for functional studies from an active length-tension relationship in each case would compromise the subsequent experimentation, the rings were stretched to an internal diameter determined in prior experiments on a large series (n > 40) of adult human pial arteries (R. Bevan and J. A. Bevan, unpublished data). The rings were then allowed to equilibrate for a further 1 h at this preload optimum before the experimental protocol was initiated. Throughout the experiments bath solution was changed every 20-30 min. Once equilibrium was reached, the effect of NE (3 × 10-6 and 10-5 M) was recorded, or if response was poor to another agonist. At peak tone increase acetylcholine (10-6 to 10-5 M) was added cumulatively. The size of the relaxation seen is an accepted method of assessing endothelial viability. In all protocols drugs were added to the tissue bath and the next step not initiated until the response reached equilibrium. If drugs did not cause a change in the force record, 10-20 min was allowed to lapse before the next step of the protocol was undertaken. At the end of each experiment the arterial segments were maximally contracted with 127 mM KCl-Krebs solution followed by AVP (10 nM). This is referred to as tissue maximum.

Endothelial cells of some segments were removed by gently rubbing the intimal surface with rough plastic tubing. Complete removal or inactivation of endothelium was accepted when the relaxation to ACh (10 µM) was lost.

EFS of intramural nerves. In the infant cerebral artery a pulse duration of 0.3 ms was used because in our experience with vessels from various species this provided optimum selectivity of stimulation of intramural nerves but not smooth muscle cells (20). Voltage parameters for maximum selective activation of perivascular nerve terminals were determined using the "breakthrough method." Pulses were applied as biphasic square waves of 0.3 ms duration at supramaximal voltage through platinum wire electrodes placed on either side of the arterial segment and connected to a Grass stimulator. Thirty minutes after the addition of TTX (3 × 10-7 M), EFS at 8 Hz was applied at 10-min intervals at increasing voltage until a small force response was recorded. This response was assumed to be due to direct stimulation of vascular smooth muscle cells, because TTX selectively blocks neuronal conduction and thus neuronal release of the adrenergic transmitter. This voltage reduced by 2 V was used after TTX washout to assess neuronal function of the segment (see Fig. 1). The neurogenic response commonly measured was that to 8 Hz. Three segments were exposed to phentolamine (10-6 M) 20 min before EFS at 8 Hz was repeated.

Figure 1
figure 1

Infant 7. Responses of segment of middle cerebral artery of infant of 37 wk of gestation to EFS before and in the presence of TTX (3 × 10-7 M), and after wash (W). Note breakthrough at 10 V. Voltage selected for study of responses to EFS is breakthrough minus 2 V.

Full size image

Catecholamine histofluorescence of nerve. To visualize the adrenergic innervation of some vessels, segments adjacent to those used for the functional studies were opened and processed for viewing catecholaminergic perivascular nerves, using the glyoxylic acid method with the addition of pontamine sky blue (0.5% wt/vol) to mask the nonspecific background fluorescence associated with elastin fibers (21). Segments were placed flat on a glass slide with the adventitial surface uppermost and were observed with a Zeiss epifluorescence microscope, with specific excitation band width from an HBO/100 mercury lamp.

Light and electron microscopy. Segments of arteries 1-2 mm in length were immersion fixed in 2.5% glutaraldehyde and 2% paraformaldyhyde in 0.1 mol/L phosphate buffer, pH 7.4, for 4 h at room temperature or overnight at 4°C. After a phosphate buffer rinse, they were postfixed in 2% osmium tetroxide in phosphate buffer for 1 h at room temperature. The segments were then dehydrated in graded alcohols and embedded in Durcupan ACM (Fluka), 1-µm transverse sections stained with toluidine blue were examined by light microscopy. Ultrathin transverse sections were cut on a RMC Ultramicrotome (MT-7) with a diamond knife and stained with uranyl acetate and lead citrate. The sections were viewed in a JEOL 100 CXII electron microscope at 60 kV.

For the quantitation of nerves, the entire tunica adventitia of each artery was examined for the presence of nerve bundles. Low magnification micrographs were taken to measure the perimeter of the adventitiomedial border. Higher magnification micrographs were made of each nerve bundle and were used 1) to measure the distances separating the nerve bundle from the nearest smooth muscle cell; montages were taken for nerve bundles at greater distances, and 2) to determine nerve densities-the number of nerve bundles per unit length of adventitiomedial border. Measurements of all micrographs were made using the Sigma Chemical Co.-Scan measurement system with a Numonics Digitizing Tablet and cursor linked to a computer.

Drugs used. The following drugs were used: acetylcholine hydrochloride, NE bitartrate, 5-HT, indomethacin (Sigma Chemical Co.), ± isoproterenol HCl (Sigma Chemical Co.), DL-propranolol HCl (Sigma Chemical Co.), TTX (Sigma Chemical Co.), AVP (Bachem California). Drugs were dissolved in Krebs solution, prepared freshly every day, and kept on ice.

Data analysis. Contractile responses were expressed as a percentage of the maximum tension produced by 127 mM KCl-Krebs and AVP (1 µM) and relaxation as a percentage of preaddition tone. Concentration response curves based on a Hill relationship were fitted to individual concentration response data by a computer program (least square method).

Statistics. Data are given as mean ± SEM. In all experiments, n refers to the number of patients from whom the blood vessels were obtained. Statistical differences between the means were determined by ANOVA followed by a Scheffe's F test. A probability of 0.05 was accepted as significant for differences between groups.

RESULTS

Observations were made on cerebral artery segments obtained from eight infants aged at birth between 23 wk of gestation and term. In Table 1 the diagnosis of each infant is noted. The majority of segments were studied within 24 h of removal, but a significant number were held at 4°C to the second day if more segments were available than could be examined at 24 h. This is a common practice in our laboratory when studying pial arteries. No distinction is made between results obtained on the first and second day as they were quantitatively and qualitatively similar. The results of studies on the basilar and middle cerebral arteries are pooled.

Table 1 Clinical diagnosis of infants
Full size table

Because of the constraints of experimental resources, although some features could be examined in segments from most subjects, observations on others were made in only a few. In some instances complete agonist dose-response curves were constructed but in others, responses to only one or two doses were recorded. Responses to EFS of intramural nerves and to NE were examined in tissues from most infants as unexpectedly significant results were obtained in segments corresponding to those that poorly responded in the adult (21,22) (Table 2) or because of their potential significance, for example histamine and indomethacin (Table 3). Spontaneous rhythmic activity was commonly seen as a component of the responses to both constrictors and dilators. Patterns, frequency, and size were extremely variable, for example, see Figures 2 and 4. An oscillatory form of rhythmic activity developed in response to a variety of constrictor and dilator agonists and tended to occur at intermediate tone levels.

Table 2 Vasoactivity of human infant cerebral arteries
Full size table
Table 3 Effect of agonists: study on a limited number of infant segments
Full size table
Figure 2
figure 2

Infant 2. The response of middle cerebral artery from infant of 24 wk of gestation to norepinephrine followed by acetylcholine. Note oscillation during the equilibrium phase of the response.

Full size image
Figure 4
figure 4

Infant 5. Histamine (3 × 10-6 M) caused relaxation of contracted middle cerebral artery from infant aged 32 wk of gestation. Tone had been induced by norepinephrine (10-6 M) in the presence of indomethacin (10-5 M).

Full size image

Responses to EFS of perivascular nerves. Using parameters of EFS selective for perivascular nerve activation determined by the breakthrough method, repetitive stimulation was continued until contractile responses reached equilibrium. These varied in size up to 36% of tissue maximum in infant 3, with a mean ± SE (number of segments) of 12.7 ± 3.5% (11) (for example, see Fig. 1). In all infants an EFS-initiated contraction was elicited which was greater than the largest neurogenic contraction observed in a series of 21 human pial arteries in which only nine responded. There was no trend in the size of responses to EFS with maturation in this limited series. In the three instances when it was tested, EFS responses were completely inhibited by phentolamine (10-6 M). In the presence of various levels of tone elicited by PGF2α, 5-HT, or NE, EFS failed to reveal a TTX-sensitive dilation both in the presence and absence of phentolamine. This was tested on an artery segment from each infant.

Agonist responses of all infants. NE. Responses to NE were almost invariably contractile, showing an initial phase followed by an equilibrium response (Figs. 2 and 3). In infant 2, at 24 wk of gestation, NE (10 nM) caused a small smooth muscle depolarization. However, depolarization accompanied by contraction was not seen until 100 nM NE (N. Gokina, personal communication). In a number of instances exposure to higher concentrations than 3 × 10-6 M resulted in dilation. Not infrequently the equilibrium phase of contraction showed rhythmic activity (Fig. 2). The NE ED50 varied from 2 to 16 × 10-7 M (7.6 ± 1.8, 10-7 M) (8) and maximum obtained responses from 28 to 62% of tissue maximum (43.1 ± 5.7%) (8). The maximum contractions to NE in infants numbers 2, 3, 4, and 5 were greater than the largest maximum seen in the adult pial artery series (41%) (22). In several instances exposure to lower NE concentrations resulted in an initial small dilation. Probably this reflects β-adrenoceptor activation in a tissue with a low level of spontaneous tone (16).

Figure 3
figure 3

Infant 2. The response of middle cerebral artery from infant of 24 wk of gestation to norepinephrine (A) before and (B) after endothelium inactivation. Note diminution of dilation to acetylcholine.

Full size image

PGF2α. Single or two-dose sequences of PGF2α up to 3 × 10-6 M were frequently given to increase tone as a prelude to exposure to dilator agents. Constriction occurred in infants 3, and 5-8. Segments from the two youngest subjects did not respond. Responses of the fourth infant were seen only with higher concentrations and were not maintained. The mean maximum response of infants 5-8 was 61.7 ± 21.0% (4) of tissue maximum.

5-HT. 5-HT caused contraction. Concentrations of 10-7 to 3 × 10-7 were used to produce a level of maintained tone to allow testing of dilator agents. Infant number 1 did not respond to concentrations up to 3 × 10-5 M and the response of number 2 to this concentration was 4.6% of tissue maximum. The mean maximum equilibrium response achieved in the older infants at 1-10 × 10-6 M was 64.0 ± 12.9 (4). The ED50 in the older subjects was 7.6 ± 1.48 × 10-7 M (4). The corresponding value in the adult pial artery was 5.3 ± 2.5 × 10-6 M (R. Bevan and J. A. Bevan, unpublished data). In several instances vessels from the younger infants became refractory to further addition of 5-HT added 30 min after washout.

Histamine. The primary effect of histamine on human infant cerebral arteries is dilation; which is similar to that in the adult pial artery (R. Bevan and J. A. Bevan, unpublished data) (Fig. 4). Histamine has been previously reported to reverse spontaneous tone (16). It also reversed tone caused by all agonists used; NE, AVP, PGF2α, and 5-HT. The maximum effective dilator dose was 1-3 × 10-6 M causing 91 ± 5.4% (7) inhibition. In several segments without intrinsic tone, histamine (3 × 10-6 M and higher) caused small transient contractions.

Acetylcholine. In artery segments from all infants acetylcholine (up to 10-5 M) was effective in causing a rapid relaxation of spontaneously occurring tone (16). It also reversed tone due to NE, 5-HT, and PGF2α (87 ± 6.1%) (8). Relaxation of these artery segments was close to maximum in many instances. See, for example, Figures 2 and 3. When tested in two segments, acetylcholine-induced dilation was obtunded by endothelium removal or inactivation by rubbing (Fig. 3). In the middle cerebral artery segment of infant 2 (24 wk of gestation), acetylcholine-induced relaxation at 3 × 10-6 M was associated with smooth muscle membrane hyperpolerization (N. Gokina, personal communication).

Indomethacin. It has previously been reported that indomethacin added to isometrically mounted infant segments caused an often dramatic increased baseline tone in six of seven infants studied 56.8 ± 14% (6,16). In two instances the increase was of the order of 90% of tissue maximum. The contraction was usually maintained until tissue wash. Indomethacin also markedly potentiated responses to NE and UK14308 (α2 agonist) (Fig. 5) decreasing the agonist threshold concentration by several orders of magnitude.

Figure 5
figure 5

Infant 2. Norepinephrine (10-9 to 107 M) (A) and UK14308 (α2-adrenoceptor agonist) (10-10 to 10-9 M) (B) failed to cause an increase in tone in middle cerebral artery segment from infant aged 24 wk of gestation. In the presence of indomethacin (10-5 M) both agonist caused marked dose-dependent increases in force.

Full size image

Agonists studied on a limited number of infant artery segments. In Table 3 is summarized the effects of a variety of agents tested, often over a limited dose range, on segments from a small number of infants. For this reason the data permit only limited conclusions.

Responses were usually observed to drug concentrations of the same order as were effective in adult pial vessels. Angiotensin II and NPY caused contraction. Bradykinin caused relaxation. TTX was without effect on baseline tone.

Catecholamine histofluorescence of nerves. No significance has been attached to the intensity and resolution of the catecholamine histofluorescence in adventitial nerves, as the time from harvesting to application of the technique could not be standardized. However, catecholamine-containing nerves were present in the middle cerebral and basilar arteries from all ages examined (Figs. 6 and 7). At 23-24 wk of gestation, well defined, brightly fluorescent, single fibers with varicosities forming a plexus were present in the middle, anterior cerebral, and basilar arteries and first order branches (Fig. 7). Nerve bundles of different sizes containing fluorescence fibers were also seen in a more superficial plane, particularly in the main trunk of the middle cerebral artery. Insufficient material was available for the basilar artery to comment on the distribution of the nerve bundles.

Figure 6
figure 6

Catecholamine histofluorescence in perivascular nerves. Proximal middle cerebral artery: gestational age (a) 38 wk (b) term. Basilar artery: gestational age (c) 32 wk (d) 35 wk, and (e) 38 wk. (Magnification, × 320.)

Full size image
Figure 7
figure 7

Catecholamine histofluorescence in perivascular nerves at gestational age of 24 wk. (a) Distal middle cerebral artery; (b) branch of basilar artery. (Magnification, × 320.)

Full size image

Light and electron microscopy studies of middle cerebral artery. Segments (1-mm) of proximal and distal cerebral arteries were fixed in vitro and transverse sections (1 µm thick) stained with toluidine blue were examined by light microscopy. At 24 wk of gestation the proximal middle cerebral artery consisted of three well defined layers as in the adult. The component cells in all three layers appeared less elongated than those in the adult, and nuclei were large relative to the cytoplasm and contained prominent nucleoli. The endothelium was closely applied to the internal elastic lamina, which was thick, undulating, and beaded in appearance. Smooth muscle cells in the media were circumferentially arranged in layers separated by extracellular matrix containing collagen and occasional fine elastic fibers. Nerve bundles were present in the outer adventitia. The middle cerebral artery distal to its first major branch differed in that the internal elastic lamina was smoother and much thinner (Fig. 8a). No mitotic figures were seen. By 32 wk the distal middle cerebral artery contained more prominent extracellular matrix between the smooth muscle layers and the media and collagen bundles in the adventitia, both of which were thicker (Fig. 8b). At 37 wk the media smooth muscle cells appeared to have more cytoplasm (Fig. 8c).

Figure 8
figure 8

Transverse sections showing a portion of the wall of human middle cerebral arteries. (a) Distal 24 wk gestation; (b) proximal 32 wk of gestation; (c) distal 37 wk of gestation. E, endothelium; M, media; A, adventitia. (Magnification, × 375.)

Full size image

Electron microscopy. Nerve bundle density or incidence relative to the circumferential length of the outer medial border (number/mm) of infant cerebral arteries was assessed in segments from two patients (Table 4). Neuronal structures were identified in a total length of 2.4 mm of the adventitiomedial junction, with an incidence of approximately 17.5/1000 µm. The mean closest distance was 1.95 µm. Values obtained in an equivalent study of human pial arteries are included for comparison (22).

Table 4 Incidence of nerve bundles and closest nerve muscle separation in adult human pial arteries (PA) and infant middle cerebral arteries (MCA)
Full size table

DISCUSSION

In this study observations on the responses of human infant middle cerebral and basilar arteries to perivascular nerve activation and to a variety of constrictor and dilator drugs are summarized. The infants ages ranged from 23 wk of gestation to term. We have previously published an analysis of the responses to physical stimuli seen in segments from the same infant series-specifically the spontaneous increase in tone probably in response to stretch, and also the effect of lumenal shear stress increase (16).

NE and sympathetic nerve stimulation. Because of the poor responsiveness of isolated adult human pial artery segments to EFS activation of their intramural nerves and to exogenous NE (22), emphasis was given to the study of these features in the infant vessels. The sensitivity of the infant cerebral arteries and adult middle cerebral and pial arteries to NE is not different. The NE ED50 values are 7.6 ± 1.8, 7.9 ± 0.2, and 5.4 ± 2.2 × 10-7 M, respectively. The maximum contractile NE response of the infant vessels relative to tissue maximum, however, was greater. The mean maximum response to NE, 43.1 ± 5.7%, was greater than that of the adult pial (20.8 ± 3.1%) and middle cerebral arteries (20.0 ± 3%).

Catecholamine-containing nerve fibers were seen in the basilar and middle cerebral arteries at all ages studied. Consistent with these observations are the responses to EFS. At 8 Hz the infant cerebral artery response was 12.7 ± 3.5% of tissue maximum and all tissues responded. The corresponding value of the adult pial arteries was 0.96 ± 0.67% obtained from nine responding segments out of 21. Only 20% (7 of 35) of adult proximal middle cerebral artery segments showed that TTX blocked responses to EFS. At 32 Hz these were 6 ± 1.0% of tissue maximum. The proximal segments responded more commonly than did distal segments (21). No evidence of dilator innervation was obtained in this present study nor in the adult autopsy or surgically removed pial arteries (21,22). The density of nerve bundles at the adventitiomedial junction is greater and their distance from smooth muscle cells of the infant middle cerebral artery is less than the adult pial artery. Corresponding values for the adult middle cerebral artery are not available.

There are no published data on agonist-induced vasoreactivity of human premature and newborn infant cerebral arteries, although infant vessels were included in the Edvinsson et al. (23) examination of innervation using histofluorescence. There are relevant studies of premature, newborn, and adult baboon cerebral arteries (8). The NE ED50 for premature newborn vessels was 0.1 of that of the adult. Similar results were obtained using phenylephrine, suggesting that the contractile response was mediated by the α1-adrenoceptor. The maximum tissue response to NE was the same for all groups. The human infant cerebral arteries were contracted by UK14308 (Fig. 6b) considered a selective α2-adrenoceptor agonist. Dilator responses to the β-adrenoceptor agonist isoproterenol of the premature and newborn baboon were greater than that of the adult (8). In the three infants we previously reported (16), isoproterenol was effective in reversing spontaneous tone (maximum 67.3 ± 7.0%). Of five adult pial arteries tested, three were unresponsive to the β-adrenoceptor agonist, and the others responded to maxima of 43 and 100% of preaddition tone (22).

Changes in NE responsiveness during development have been assessed in the rat basilar artery. Using electrophysiologic techniques and iontophoretic application of drugs, Byrne et al. (24) concluded that the majority of the NE response of the basilar artery of the neonatal rat was mediated through excitatory β-adrenoceptors and γ-adrenoceptors. Postnatally the excitatory β-receptors diminish and the γ-receptor population become dominant. A γ-adrenoceptor is a high threshold, low discriminating junctional receptor associated with contraction but resistant to α-adrenoceptor blockade. In the human infants, neurogenic contraction was blocked completely by phentolamine, excluding the involvement of "γ-like" receptors in this tissue. In the human adult proximal middle cerebral artery, concentrations of NE greater than 10-5 M caused a phentolamine-insensitive component of contraction (21), presumably through a γ-like receptor. Such a response was not observed in the infant or in pial arteries. Elliot and Pearce (4) undertook a comprehensive comparison of changes of sensitivity with development in the α-adrenoceptor in the sheep, comparing newborn lambs (5-7 d) and nonpregnant adults (18-24 mo). They observed decreases in sensitivity associated with a decrease in affinity in the fourth but not the second order branch segment of the middle cerebral artery with maturity. It is difficult to detect an overall pattern or trend in all these observations of various species, except that responses of cerebral arteries to sympathetic nerve activity and NE tend to diminish with maturation and that larger vessels are more responsive and have higher maximum NE-induced responses than smaller.

A relevant observation is that basal and agonist but not K+-induced 45Ca2+ influx in the newborn sheep is greater than that in the adult (25). This is consistent with our observed decrease in myogenic activity with maturity and in responsiveness to NE that in the cerebral circulation are both highly dependent on external calcium (26).

We have previously postulated that the poor response of human pial arteries to EFS indicates that sympathetic nerves play little part in the regulation of their tone (22). The small response of the adult proximal middle cerebral artery suggests an influence mostly aggregated at the proximal portions of the large cerebral arteries. If this were sufficient to influence flow, the consequence would extend downstream, probably resulting in constriction due to activation of local shear stress sensitive mechanisms. The findings in this present study strongly suggest a potentially larger role for neuroregulation in the human infant. Despite overwhelming evidence for a strong influence of dilator nerves in many species, this study of artery segments in vitro only adds to the other observations (21,22) that such nerves do not appear to be functional in the human.

Constrictor and dilator agonists. The physiologic relevance of many agonist responses studied on isolated vascular segments is often problematic. In the latter part of gestation segments contracted to PGF2α, 5-HT, angiotensin II, and NPY and dilated to histamine, acetylcholine, and bradykinin in concentrations of the same order that are effective in adult pial arteries. There is insufficient data to make exact comparisons. There is little available data supporting a physiologic role for 5-HT or acetylcholine in the human cerebral vascular system. Nerves containing both substances have been described in this vasculature in other mammals (27–31). There is no evidence that histamine is normally involved in the regulation of cerebrovascular tone; although its pathologic involvement and also that of 5-HT has been proposed (32). PGF2α has yet to be implicated in a normal tone regulating mechanism. NPY is probably a co-transmitter in the sympathetic constrictor innervation of the rabbit middle cerebral artery (33); however, in three infant artery segments, phentolamine completely blocked the contraction to nerve activation. This result also tends to exclude a role for the γ-adrenoceptors in the infant human neurotransmission process.

Indomethacin. Indomethacin (10-5 M) consistently produced a large increase in basal intrinsic tone and augmented agonist tone in the infant, but only occasionally and then to only a smaller extent in the adult pial artery (16,21). In adult pial arteries indomethacin did not unmask a response to EFS when it was previously absent, and neither the size nor the sensitivity of pial artery segments to NE was altered by this treatment (p = 0.34 and 0.26 (8), respectively).

Indomethacin is an inhibitor of cyclo-oxygenase and hence the synthesis of prostaglandins and is very effective in the treatment of patent ductus arteriosus in premature infants. It has been reported to reduce cerebral blood flow by a mean of 24% in premature infants (34). This could be due to an increase in intrinsic tone in the infant cerebral vessels as has been previously described (16). Inhibition of tone by an indomethacin-sensitive system is not limited to myogenic activity as this drug revealed constriction to both NE and UK14308 (α2-adrenoceptor agonist) (Fig. 6, a and b). Prostanoids potentiate the cerebrovascular constriction to oxytocin in piglets (35) and there are many reports of the modification of cerebral blood flow in preterm neonates by indomethacin (36–40).

It is possible that not all agonist responses mature at the same time. Segments from the two youngest infants, aged 23 and 24 wk, were unresponsive to PGF2α and 5-HT, although arteries from the older infants responded. The unresponsiveness is unlikely to be due to damage as the same artery segments responded to other constrictor and dilator agents in a manner similar to older segments.

We have previously argued that the two major mechanisms that underlie the regulation of cerebrovascular tone in the adult pial vessels, intrinsic myogenic pressure, and shear stress-dependent systems are present in the infant, at least at the 23rd wk (16). However, in the infant there is evidence for sympathetic neural regulation related to a significant reactivity to NE possibly a higher nerve density than in the adult. Innervation can influence normal growth and development of blood vessels and vascular size and vessel number (41,42).