changes in the blood flow to the digestive organs of

Quarterly Journal of Experimental Physiology (1983) 68, 77-88Printed in Great Britain

CHANGES IN THE BLOOD FLOW TO THE DIGESTIVEORGANS OF SHEEP INDUCED BY FEEDING

R. J. BARNES, R. S. COMLINE AND A. DOBSON*Physiological Laboratory, Downing Street, Cambridge CB2 3EG and *Department of Physiology, New York

State College of Veterinary Medicine, Cornell University, Ithaca, NY 14853, U.S.A.(RECEIVED FOR PUBLICATION 19 MAY 1982)

SUMMARY

The blood flow to the digestive organs of nine sheep was determined by the use of isotopicallylabelled microspheres before, during and at 2 h and 4 h after feeding. Within 3 min of the startof feeding, the blood flow to the salivary glands and to the smooth muscle of the rumen andreticulum increased three-fold. The blood flow to the epithelium of the rumen and reticulumalso increased before any appreciable effect on ruminal fermentation could have occurred. Thisincrease in flow was greater in absolute but smaller in relative terms than that to the muscle.At 2 h after feeding blood flow to the epithelium of the rumen and reticulum was two to fourtimes greater than before food was taken, while the flow to the smooth muscle of these organshad fallen to the level found before feeding. In the more distal parts of the gastrointestinal tract,blood flow changes in response to feeding were less pronounced and, where they occurred atall, consisted of decreases at different times. Thus blood flow to the omasum decreased duringfeeding but recovered thereafter, while the flows to abomasum, duodenum and ileum were notchanged during feeding but were significantly lower at 2 h and 4 h later. In the rest of the smallintestine and in the large intestine there were no significant changes in flow during the periodof observation, nor were there any changes in the blood flow to pancreas or spleen. However,the flow to the omental and mesenteric fat declined abruptly on feeding and reached its minimalvalue 2 h afterwards. These results are in marked contrast to those reported in other speciesin that the subepithelial capillary plexus of the reticulum and rumen was the only regioncontributing to the increased hepatic portal blood flow after feeding.

INTRODUCTION

The blood flow to the rumen is known to be increased by a rise in the luminal concentrationof carbon dioxide or butyrate, which are products of microbial digestion within the organ(Dobson & Phillipson, 1956). It is also known that the flow through a ruminal arteryincreases during eating and for some time thereafter (Sellers, Stevens, Dobson & McLeod,1964). The distribution of blood to the different tissues of the rumen wall has, however,been studied only 18 h after withdrawal of food when the contents of the rumen havereached a steady state between production and absorption. Under these circumstances theblood flow is uniformly distributed to the different compartments of the compound stomachand within each compartment is largely directed to the subepithelial capillary network ratherthan the smooth muscle (Engelhardt & Hales, 1977).

There is, however, no information on the effects of feeding or digestion on the relativeblood flow to the different compartments or its distribution to the different capillarynetworks in each part of the compound stomach. The first objective, therefore, was toascertain by the use of isotopically labelled microspheres whether the blood flow to allregions of the ruminant stomach changed uniformly during or after feeding and how fardown the alimentary canal similar variations in the blood flow could be detected. The secondobjective was to assess the degree to which changes in the perfusion of the cranial divisions

R. J. BARNES, R. S. COMLINE AND A. DOBSON

of the stomach could account for the increase in portal flow some hours after feeding,reported by Bensadoun & Reid (1962).

Preliminary accounts of this work have already been published (Dobson, Barnes &Comline, 1981; Barnes, Comline & Dobson, 1982).

METHODS

Adult Welsh sheep, weighing 27-39 kg were penned individually and fed on 200 g concentrate pellets(B.O.C.M. Silcox 303) and hay ad libitum at the same time each day; water was available at all times.The sheep were trained to consume their ration within two hours by a progressive reduction of thetime for which hay was available. Thereafter the regime was maintained for at least 2 weeks. A weekbefore the experiment the sheep were anaesthetized with halothane and N20 and Teflon catheters(1 mm bore) were introduced through a femoral artery so that their tips lay in the left ventricle andin the caudal aorta immediately above its bifurcation. This procedure had little effect on the dailyfood intake; food was not withheld before the operation.Mock experiments were carried out for several days to accustom each animal to the procedure and

to the experimenter. During each experiment the blood flow was measured by the injection ofmicrospheres into the left ventricle at the following times: 20-30 min before, 3-4 min, at 2 h and at4 h after the start of feeding. A further two observations were made in some sheep, between the sixthand sixteenth hour of the feeding cycle. Mean arterial blood pressure was measured immediately beforeand after each injection using a strain gauge pressure transducer (Endevco Ltd.). After the lastinjection the sheep were anaesthetized with Na pentobarbitone and exsanguinated. The organs andtissues listed in Table 1 were dissected and the placement of the ventricular catheter was checked.Ligatures were placed about the oesophagus 20 cm above the cardia, and on the abdominal rectum;this corresponds to the portal catchment area. The omental and mesenteric fat were removed, thegut contents were expressed and the mucosal surfaces were rinsed with tap water and dried. Thereticulum, the atrium and the dorsal and ventral sacs of the rumen were separated, which left thereticular groove, where it is difficult to strip the epithelium from the muscle. This region was thereforeconsidered along with the terminal oesophagus. In the reticulum and the three divisions of the rumen(the atrium was not divided from the ventral sac in the first animal), the epithelium was dissectedand counted separately from the muscle. Attempts to separate omasal muscle and epithelium wereabandoned as impractical because the muscle penetrates deeply in each leaf of epithelium. The smallintestine was divided into three parts. The first metre was called duodenum, the last metre representedthe caudal ileum and the remainder was jejunum. Except for subcutaneous and perirenal fat, of whichsamples were taken, the whole organ or tissue was counted to eliminate sampling errors. The portalflow was estimated by the addition of the arterial flows of those organs which drain into the hepaticportal vein.Two sheep were anaesthetized with sodium pentobarbitone (60 mg/kg) and a catheter inserted into

the duodenal branch of the hepatic artery; a catheter was also placed in the portal vein through amesenteric tributary. Microspheres were injected into these catheters and the liver, lungs and kidneysthen removed for counting.The method for determining blood flow by the reference organ technique (Makowski, Meschia,

Droegemueller& Battaglia, 1968) has been validated for sheep (Hales, 1973, 1974; Hales & Cliff, 1975).The microspheres injected into the left ventricle (15 ,um diameter; New England Nuclear or 3-M Co.)were labelled with l41Ce, 61Cr, 13Sn, 85Sr, 95Nb or 46Sc. During each injection, blood was withdrawnat a rate of 25 ml/min from the femoral artery catheter. Sufficient spheres were injected to permittheir accurate determination in the reference blood sample. The amounts injected were determinedby difference using a 2 in Nal crystal situated 75 cm from a plastic spiral containing the spheres. Thisprocedure sufficiently reduced the counting efficiency to avoid problems with saturation of the crystal.Tissues and organs were spread over the bottom of weighed parallel-sided polypropylene jars so thateach sample approximated a disc of 3 in diameter but of a thickness proportional to the tissue weightwhich varied from sample to sample. Tissues, blood samples and standards were counted midwaybetween 6 in diameter Nal crystals which were 55 cm apart, in a whole body counter using a 99-channelpulse height analyser.

Calculations were made on an IBM 370/165 computer or a PDP-8E minicomputer. Counts werecombined to give a channel for the peak of each isotope, and corrected for background. Because the

78

PORTAL BLOOD FLOW PATTERNS

Ruminal Ruminal Ruminal

400 Submandibular Parotid Oesophagus Reticulum atrium dorsal sac ventral sac Omasum

and groove 4

200-

100

E 40-

0

0 20-0

10

41

Fig. 1. Patterns of blood flow in organs proximal to the omasal-abomasal junction. Each organ has four pointscorresponding to pre-feeding, feeding, and 2 h and 4 h post-feeding. 0-0, whole organ flow; 0-0, muscleflow; x- x, epithelial flow. Vertical bars +S.E.M.

gamma spectra shift in a complex manner with sample thickness, empirical corrections were made,based on counts of a solution of each isotope diluted with different amounts of water. A quadraticexpression was derived to express the relation between the count in each channel and the sampleweight. This was used to convert the counts of each standard, whose weight was known, to the countsat the weight of a particular sample. This matrix was then inverted and multiplied by the matrix ofcounts in the same channels from the sample to give the activities of each isotope in the sample.

In order to distinguish significant changes of blood flow per 100 g tissue for each region with time,the variance between animals, which commonly was highly significant, was removed followinganalysis of variance. Since the variance appeared proportional to the flow, the logarithms of the datawere analysed. The non-additivity test of Tukey confirmed that analysis of the data following thistransformation was always appropriate, whereas analysis of the untransformed data, especially wherethe range was large, was often inappropriate. In recombining the data to give a clear picture of theportal distribution, the logarithmically transformed data were cumbersome because the means areno longer additive. If, however, the flow to each organ was expressed as a fraction of the total portalflow, the untransformed data could be used to assess the statistical significance of changes. Thisapproach was therefore preferred in the over-all summary of the changes in portal distribution.

RESULTS

No unusual behaviour was apparent during the experimental manipulation. In particularthere was no depression of appetite. Over the five days immediately preceding theobservations each sheep consumed between 240 and 560 + 51 g/d (S.E.M.) of hay in additionto the concentrate pellets. The mean hay intake on the day of the experiment minus itsaverage over the preceding five days for each sheep was 36 + 30 g (S.E.M.). The pellets wereconsumed within the first few minutes of being offered.

79

80 R. J. BARNES, R. S. COMLINE AND A. DOBSON

Table 1. Bloodflow intensity related to time after start of meal

Mean blood flow*Coefficient

Tissue wt. (ml./min 100 g) of varianceTime n (g S.D.) -20 min 4 min 2 h 4 h (%) Pt

Left parotid gland 7 11 2-7 169 365 127 142 25 5 < 0 05Submandibular gland 7 17 3-9 41 206 51 39 31-1 < 0-01Oesophagus+groove 9 29 11 19 33 23 22 8-1 < 0-01Reticulum epithelium 9 33 8-8 103 111 229 175 7 9 < 0-01Reticulum muscle 9 40 7-2 8-9 10-4 5-9 4-2 13 4 < 0-01Atrium epithelium 8 31 9 7 130 171 353 200 9.1 < 0-01Atrium muscle 8 40 18 9-2 13 0 5 6 5 2 16-1 < 0-01Dorsal sac epithelium 9 61 13 74 95 327 222 9-6 < 0-01Dorsal sac muscle 9 97 19 6 3 16 9 5 7 5 2 20 1 < 0 01Ventral sac epithelium 8 63 21 91 130 332 207 11 2 < 0-01Ventral sac muscle 8 93 21 8 2 21 3 6-4 6 5 19.1 < 0-01Omasum 9 62 13 57 35 67 56 11 1 <0-01Abomasum 9 146 26 98 103 67 67 9 1 < 001Duodeunum 9 30 11 93 89 74 77 6 2 < 0-05Jejunum 9 186 31 74 68 70 71 7 6 n.s.Ileum 9 54 29 72 69 48 58 10-1 < 0 05Caecum+colon 9 95 17 59 54 51 49 4 5 n.s.Spiral colon 9 182 61 31 31 32 25 8 1 n.s.Omental fat 8 1090 450 60 3-4 2 0 2 5 14 2 < 0-01Mesenteric fat 8 890 670 6 5 4 8 3 6 3 8 13 6 < 0 05Pancreas 9 38 9 3 381 384 335 343 7 6 n.s.Spleen 9 105 87 152 164 161 125 11 4 n.s.Left liver 9 139 25 4 2 5 2 8 2 8 1 27-7 n.s.Central liver 9 155 15 5.3 5.7 8 8 8 9 25 6 n.s.Right liver 9 128 23 4-4 4-6 8-2 7 9 27-3 n.s.Heart 7 151 33 52 64 66 62 7 8 n.s.Left kidney 9 37 5 2 600 595 646 662 5 3 n.s.Right kidney 9 37 7 0 614 595 643 651 5-6 n.s.Adrenals 9 3 0 7 192 184 217 244 8 3 n.s.Perirenal fat 6 5.9 2-9 3.5 4.4 17 1 n.s.Subcutaneous fat 6 - 5 0 3.3 3 0 2 5 28 0 n.s.

* From logarithmic transformation, i.e., geometric means.t Significance of difference between times.n, number of sheep.

200 Abomasum Duodenum Jejunum Ileum Small Caecum Spiral Largeintestine and colon colon intestine

0

100 -

E

400

20-

Fig. 2. Patterns of blood flow distal to the omasal-abomasal junction. Each organ has four points correspondingto pre-feeding, feeding, and 2 h and 4 h post-feeding. Vertical bars +S.E.M.


At the times that the blood flows were measured the mean blood pressures in six of thesheep were 97*6 mmHg at -20 min, 105 8 mmHg at +3 min, 97 3 mmHg at +2 h and974 + 20 (s.E.M.) mmHg at + 4 h. Thus, in the sheep as in other species, there was a

significant rise in blood pressure (P < 0-001) associated with the act of eating.

Patterns of change of bloodflowThe large body of data generated by simultaneous observation of different regions made

it convenient to use a general statistical criterion to establish whether changes in blood flowwith time were consistent between sheep. A pattern of flow change was considered to existif, after appropriate analysis of variance, there was a significant difference in the flows atthe different sampling times (established by the usual F test). Some synchronizaton ofresponses between sheep is required to achieve significance and, because the time course offeeding responses is bound to vary in different animals, this criterion tends to beconservative. The microsphere method gives flows to the different regions at the sameinstant. Thus the analysis of the differences in flow between two regions, where this promisesto be interesting, provides a sensitive test of difference of both the time course of changesin flow and of the mean flow rate. In considering the patterns of flow per unit weight oftissue, logarithmic transformation of data was necessary to preserve homogeneity ofvariance (see Methods). Testing the differences between two regions after such a transfor-mation of data is equivalent to looking for differences in the ratio of flows in these regions.The statistical tests of the effects of feeding on the blood flow (last column, Table 1)

showed greater significance in those regions ofthe alimentary canal proximal to the pylorus.Distal to this point, or with other organs, the significance of temporal changes became less,not because the variance increased, but because the changes with time were themselvessmaller. Feeding appeared to give rise to five different time-courses of flow, two where theflow increased, and three where it decreased with feeding.

Pattern A characterized by flow stimulation solely during eating, is found in the salivaryglands, the muscle of the rumen and reticulum and the oesophagus+ reticular groove (Fig.1). No differences in this pattern could be detected between the salivary glands althoughthe parotid had a blood flow four times greater per 100 g tissue than the submandibularglands. Again, although the flow through each region of muscle of the rumen and reticulumis at least twenty-fold lower, on a weight basis, than that through the salivary glands, thegeneral pattern of response to feeding is similar. The specific pattern observed in theoesophagus+ groove is similar to that of the parotid gland, but is different from that foundin the muscle of the reticulorumen (P < 0 01).The changes in blood flow in the reticuloruminal muscle in response to feeding appear,

on inspection (Fig. 1), to alter gradually from reticulum to ventral sac. This is confirmedby statistical analysis which demonstrates that, while differences in the pattern of flowcannot be detected between adjacent regions such as the reticulum and atrium, atrium anddorsal sac or dorsal and ventral sac muscle, nevertheless the reticular pattern differs fromthose of the dorsal and ventral sacs (P < 0 01) and the atrial pattern from that of the ventralsac (P < 0O05). In general the mean blood flows to each region of muscle at any one timeare very similar when they are compared on a weight basis, but the blood flow to the ventralsac is 20% higher than that to either the reticulum or the atrium (P < 0 05).

Pattern B shows maximal flow stimulation 2 h after feeding and occurs only in thesubepithelial capillaries of the ruminoreticulum. As with pattern A, there is an enhanced

81


flow during eating but this is much smaller than the subsequent increase noted at 2 h, whichhas decreased at 4 h. While the pattern of flow for each region is statistically unique(P < 0 005 except between the dorsal and ventral sacs where P < 0 05) it is neverthelessthe similarity between responses which is intriguing. At rest and during the feeding andfermentation phases of the feeding cycle the atrial blood flow is 18-28o% higher than thatto the other epithelial regions. The greater flow may be related to the longer papillae whichare found in this region.

Pattern C is the inverse of pattern A, with flow inhibited solely during eating. This isseen only in omasum.

Pattern D shows a decrease in blood flow at 2 and 4 h. This is found in the abomasum,and is shared with the cranial and caudal ends but not the middle of the small intestine(Fig. 2). Between the abomasum, duodenum and ileum no difference in pattern could bediscerned.

Pattern E shows a depression in flow both during and after feeding and is found in theomentum and mesentery which consist mainly of fat. The subcutaneous fat has nosignificant pattern whereas the perirenal fat just fails to have a significant patterp at theP = 0.05 level. The absence of a statistically significant flow pattern may reasonably beattributed to an increase in variability. Differences between the patterns of flow reachsignificance only between the omental fat and the perirenal fat (P = 0005).

Absence ofpattern. There is no detectable flow pattern in jejunum, caecum+ colon, spiralcolon, spleen and pancreas in the portal drainage area, nor in the hepatic, coronary or renalarteries. It is encouraging to discern no pattern in those organs outside the portal regionwhere there is no particular a priori reason to expect a variation with feeding.

Total hepatic bloodflowThe mean portal flow was similar in the pre-feeding, feeding and 4 h samples but was

19O% higher at 2 h post-feeding (Table 2, P < 0 01, t test), although more rigorous statisticalanalysis demonstrated that the pattern of change of portal blood flow with time was notstatistically significant (P = 005). The hepatic arterial flow, measured directly by theretention of spheres within the liver showed no pattern between times in either the wholeorgan or its left, right and central regions (Table 1). Whereas the concentrations of spheresin the left and right segments were similar, in the central region the spheres wereconcentrated 13% more than in the left and 16% more than in the right segments(P < 0 00 1). In the two animals in which microspheres were injected directly into the hepaticarterial and portal vessels, 99.5% of the spheres retained within the liver and lungs werefound in each case in the liver. This established that the spheres were sufficiently large tobe removed quantitatively on a single passage through the liver sinusoids and did not passthrough to be trapped in the lungs.

Sources ofportal bloodflowThe transformed data (Table 1) suitable for establishing patterns of flow change are

inconvenient in developing a composite picture of the sources of portal flow. Fortunately,the flows to each region when expressed as a fraction ofthe total portal flow can be analysedwithout transformation; the mean fractional flows to each region are, therefore, additive.

82


Table 2. Effect offeeding on bloodflow to selected regions

Mean blood flow* (mi./min)

Change in blood flow fromvalues before feeding

Valuebefore Whilefeeding eating At 2h At 4h s.E.M.t Pt

Hepatic portal flow 1143 -23 +218 -11 +63 n.s.Rumen, reticulum, groove 186 +78 +389 + 199 +26 < 0Q005+ oesophagus

Rest of portal inflow (i.e. 938 -91 -193 -201 +40 < 0-01excluding ruminoreticulum)Omental and mesenteric fat 119 -45 -69 -65 +8 <0005Portal, excluding 758 -18 -91 -114 +36 n.s.ruminoreticulum and fatOmasum 35 -14 +4 -2 +3 < 0 005

* Based on geometric means.t S.E.M. calculated from coefficient of variation and mean flow.I Significance of difference between times.

Mesenteric Omental Gut Perirenal Subcutaneous

Fig. 3. Patterns of blood flow in fat. Each organ has four points corresponding to pre-feeding, feeding and 2 hand 4 h post-feeding. Gut fat is the combined mesenteric and omental fat.

Since all the regions contributing to the portal flow are in parallel, fractional flow in anyregion is a measure of relative conductance of the organs draining into the portal vein. Eachmajor fraction of the hepatic portal flow except that from the spleen showed a significantchange with time (Fig. 4). The largest changes were to the ruminoreticulum, whichcontributed 20, 26, 46 and 37%O to the total portal flow at the four sampling times. Thesmall increase in fractional blood flow to the ruminoreticulum during feeding was offsetby decreases in the fractional flow to the small intestine and the sum of the flow to themesentery and omentum. The larger increase in forestomach flow at 2 h was accompaniedby decreased fractions from abomasum, small intestine and large intestine, fat and pancreas.At 4 h, not surprisingly, most of these changes had partially reversed. The small and largeintestine together contributed 43, 41, 31 and 36% to the total portal flow at the four times.The three regions outside the alimentary canal, namely the fat, the pancreas and the spleen

10o

0

0C00

.0E

'E4730:

F 2

83


R & R Om Abo

...................

-pS>oc 0 0

''' ''°oc *:::- ,c 0

H H H H H H~~~

Fig. 4. Fractional distribution of the sources of hepatic portal flow before, during and after feeding. The regionsdifferentiated are the ruminoreticulum, including the distal oesophagus and oesophageal groove (R & R),omasum (Om), abomasum (Abo), small intestine (SI), large intestine (LI), pancreas (Pan), spleen 'Spl) and thecombined mesenteric and omental fat (GuF).

R & R Om Abof............

Si Li Pan Spi GuF

o°] 1Iq.1

II°ILJ1J40!I I1'''. J °°U..1-'-100o .

ia

0 ml/min 100 g liver

I 0 I/min . 40 kg sheep

Fig. 5. Distribution of the sources of hepatic portal flow before, during and after feeding. Regions differentiatedare the ruminoreticulum including the distal oesophagus and oesophageal groove (R & R), omasum (Om),abomasum (Abo), small intesting (SI), large intestine (LI), pancreas (Pan), spleen (Spl) and the combinedmesenteric and omental fat (GuF).

Pre-fee

SI LIFo°o0 0o

Pan Spl

t:,o:aGuF

Fee(

2 1

4 h

± 2 X S.E.M.

0 50 100

Pre-fee(

Feed

2

4 h

200 100

.................... .......... .--,I ......

u (- 0::::::::::::::::::::::::::::::: 0 0 0 0:::::: C) c a

u u0 9

0 0 0 :0 0

0 0 0 00 0.--m

-

........ .......I ..*-I..

84


contributed 37, 33, 23 and 27% to the portal flow at the sampling times. The increases infractional flow in the ruminoreticulum were thus offset by decreases, not only to thegastrointestinal tract, but also to other organs that contribute to portal flow.The fractional portal flow from each organ can be multipled by the total portal flow/100 g

liver to give the absolute contribution to portal flow/100 g liver for each of these organs.Moreover, the same data can be scaled for a 40 kg sheep using the mean liver weight inthese experiments, 12-4 + 1 2 (S.D.) g/kg live body weight (Fig. 5). The values obtained showno statistically significant change with time in the small or large intestine, the pancreas orspleen (Table 1).

Relation to feedingOne sheep was not fed, but was observed at the usual times after it was expecting to be

fed. At 2 h and 4 h the flow to all parts of the portal drainage area tended to be lower thanthe pre-feeding sample, except for the fat, which remained high. The enhanced flows in theforestomach were absent. While the adjacent animals were being fed, this sheep showedincreased blood pressure, and enhanced flow to the salivary glands and muscle andepithelium of the ruminoreticulum. Another sheep which refused to eat its concentratesshowed no rise in blood flow at 2 and 4 h. Two other animals were observed whilst theother sheep were being fed, and although their blood pressures rose, there was no increasein blood flow to the ruminoreticulum indeed flow appeared to be depressed. This type ofsham feeding thus gave erratic responses.Two of the nine sheep were accustomed to feeding at 4 p.m. The results from these two

sheep could not on inspection be distinguished from those of the other seven sheep fed atthe normal time, 9 a.m. This illustrates that the results were related to time of feeding ratherthan time of day: patterns of flow were therefore entrained with feeding.

Isolated observations made at 6, 8 and 16 h after feeding were insufficient in number forsystematic analysis. Within the large variability of individuals they suggested that the trendbetween the 4 h post-feeding and pre-feeding was monotonic. There were no wide swings.The depression of blood flow to the fat associated with feeding had disappeared and theflow had returned to the pre-feeding level by 6 h.

DISCUSSION

The major increases in the blood flow to the organs of the splanchnic area after feedingin the sheep are confined to the rumen and reticulum. Presumably, the relatively slow releaseof food retained within these compartments results in little change in the factors influencingcapillary blood flow in the abomasum and lower intestine. The absence of pattern in thecombined elements of the small intestine, which is largely influenced by the jejunum, andin the large intestine was not unexpected, since feeding effects in digesta flow are knownto have disappeared by the time the food has reached the ileocaecal valve (Bruce, Goodall,Kay, Phillipson & Vowles, 1966). The distribution of the areas of vasodilatation in the sheepis, therefore, quite different from that reported in dogs and primates in which postprandialhyperaemia is most pronounced in the small intestine and pancreas (Vatner, Patrick,Higgins & Franklin, 1974; Gallavan, Chou, Kvietys & Sit, 1980).

Within the ruminoreticulum the relatively greater flow rate through the subepithelialcapillary plexus, in comparison with that of the smooth muscle, dominates the blood flowto these organs and their over-all contribution to portal flow. The difference is so large thateven the greater mass of the muscle fails to compensate. Thus, during eating, when the

85


muscle blood flow doubles, it is the much smaller fractional change in the subepithelial flowwhich contributes over 70% of the total increase in flow from the forestomachs.During the experiments it was essential to reduce the interference with the alimentary

tract to a minimum. While this precluded a definitive analysis of the mechanism of thepostprandial vasodilatation it is still possible to consider a number of mechanisms whichmight be responsible for the observed effects. Thus the pattern of blood flow to the smoothmuscle of the reticulum and rumen during feeding so closely resembled that of the salivaryglands that the effects could be attributed with some confidence to responses mediatedthrough the parasympathetic innervation. It is also tempting to ascribe the immediate effectof feeding on the flow in the subepithelial capillaries to a parasympathetic vasodilatation,for the timing of the response is similar to that in the salivary glands, a network ofcholinergic and adrenergic nerve fibres surrounds the vascular plexus (Message, 1966) and3-4 min is insufficient to make an appreciable difference to the microbial digestion of therumen contents. Nevertheless, an alternative explanation which cannot be excluded is thatthe enhanced movements of the reticulum and the rumen at this time may have stirred theircontents and thereby altered the concentration of carbon dioxide and butyrate in the fluidadjacent to the epithelium, thus indirectly increasing the epithelial blood flow. The results,therefore, provide strong but not unequivocal evidence for a nervous vasodilation at thistime.

Subsequent changes, 2 h and 4 h after feeding, were not associated with enhanced smoothmuscle blood flow and can be attributed to the effects on the subepithelial capillary plexusof an increased absorption of the products of microbial digestion from the rumen,particularly carbon dioxide and butyrate, both of which are effective in stimulating rumenblood flow (Sellers et al. 1964). The intensity of the response varied in different regions ofthe forestomach. At 2 h after eating the blood flow to the epithelium of the ruminoreticulumas a whole increased by 3 1 times; this over-all figure, however, included many differentrates of flow from 2 2 times in the reticulum, 2-7 times in the atrium, 3-6 times in the ventralrumen and 4-4 times in the subepithelial capillaries of the dorsal rumen. The differencesin the responses between the areas can probably be attributed to variations in theirsensitivity to chemical stimuli for Dobson (1979), in anaesthetized sheep, showed a smallerresponse of the blood flow of the reticulum than that of the rumen to graded applicationsof CO2. The subepithelial vascular plexus of the reticulum and rumen is the only site inthe portal area where an increased blood flow could be detected at 2 h, and so must be thesource of the increase in portal blood flow at this time. An increase in portal flow afterfeeding has been established (Bensadoun & Reid, 1962) and the magnitude observed here(19%) is similar to that reported by Webster & White (1973).The effects of the subepithelial vasodilatation in the ruminoreticulum on the portal blood

flow were offset by a reduction in blood flow to other parts of the alimentary canal andthe extremely fatty omentum and mesentery (Table 2). The flow to this fat in the portalarea had fallen by 26% within 3 min of eating and by 58% at 2 h. This fall may relate tometabolic changes, since lipolysis predominates before feeding, whereas lipogenesis takesover when absorption is maximal. Although these changes in blood flow to fatty tissuesare significant, the somewhat larger changes to the rest of the gut are within the variabilityof the observations. Caution is therefore necessary in assuming that a hyperaemia in onepart of the portal supply is compensated by vasoconstriction in another, not the least inlean sheep, where the contribution from fatty tissue would be reduced.Not all the keratinized columnar epithelium of the cranial divisions of the stomach

responded with an increased blood flow to the stimulus of feeding. The exception appeared

86


to be the omasum which has a large surface area covered with this type of epithelium witha morphologically similar subepithelial vascular plexus but in which the total organ flowdecreased by 40% during feeding. The omasum is relatively vascular but as it is small itsabsolute contribution to portal flow is also quite small. Thus, although the omasum is theonly region of the gut in which a significant decrease in flow occurs during feeding, itcontributes only 22% of the combined decrease in flow to all regions other than theruminoreticulum at that time. The remaining 78% is contributed by the summation ofstatistically insignificant changes in other regions of the gastrointestinal tract. It isunfortunate that the epithelium of the omasum cannot be separated quantitatively fromthe muscle. At least there is presumptive evidence that the omasal epithelium behavesdifferently from that in other regions of the stomach, since Engelhardt & Hales (1977) founda similar flow rate in samples of omasal and reticular muscle 18 h after the withdrawal offood, when the contents of the rumen have come into a steady state.The other surprising feature of these results is the low blood flow in the hepatic artery

compared to that of the portal vein. Thus the contribution of the hepatic artery of 2-3%to the total liver blood flow was much smaller than the 21 % reported by Brockman &Bergman (1975), from the clearance ofp-aminohippurate, and similar values obtained fromother species in the literature (for references, see Greenway & Stark, 1971). The possibilityof an experimental artifact which would give low values for the blood flow, namely thatthe relatively small (15 4um) microspheres were not retained in the liver sinusoids, was notsupported by the recovery of a high proportion of the microspheres from the liver afterdirect injection into the hepatic artery. The hepatic arterial blood flow would thereforeappear to vary, at least in the sheep, under different conditions and the relative contributionsof the arterial and portal flow to the total hepatic flow on different diets may give cluesto the understanding of the discrepancy.These experiments extend the results obtained by Engelhardt & Hales (1977) who studied

the capillary blood flow in the forestomachs after food had been withdrawn for about 18 h.It is clear that the flow in the subepithelial plexus is extremely labile after feeding, but theproblems of the mechanisms involved in these responses and the influence of different dietson them, have yet to be resolved.

This work was supported by a grant from the Agriculture Research Council. We wish to thankMr P. Hughes, Miss T. Grimes and Mr K. Richardson for their help during the experiments.

REFERENCES

BARNES, R. J., COMLINE, R. S. & DOBSON, A. (1982). Splanchnic circulatory responses to feeding insheep. Journal of Physiology 327, 53P.

BENSADOUN, A. & REID, J. T. (1962). Estimation of rate of portal blood flow in ruminants: effect offeeding, fasting and anaesthesia. Journal of Dairy Science 45, 540-543.

BROCKMAN, R. P. & BERGMAN, E. N. (1975). Quantitative aspects of insulin secretion and its hepaticand renal removal in sheep. American Journal of Physiology 229, 1338-1343.

BRUCE, J., GOODALL, E. D., KAY, R. N. B., PHILLIPSON, A. T.-& VOWLES, L. E. (1966). The flowof organic and inorganic materials through the alimentary tract of the sheep. Proceedings of theRoyal Society (Series B) 166, 46-62.

DOBSON, A. (1979). The choice of models relating tritiated water absorption to subepithelial bloodflow in the rumen of sheep. Journal of Physiology 297, 111-121.

DOBSON, A., BARNEs, R. J. & COMLINE, R. S. (1981). Changes in the sources of hepatic portal bloodflow with feeding in the sheep. Physiologist 24, 15.

87


DOBSON, A. & PHILLIPSON, A. T. (1956). The influence of the contents of the rumen and of adrenalineupon its blood supply. Journal of Physiology 133, 76P.

ENGELHARDT, W. VON & HALES, J. R. S. (1977). Partition of capillary blood flow in rumen, reticulumand omasum of sheep. American Journal of Physiology 232, E53-56.

GALLAVAN, R. H., CHOU, C. C., KVIETYs, P. R. & SIT, S. P. (1980). Regional blood flow duringdigestion in the conscious dog. American Journal of Physiology 238, H220-225.

GREENWAY, C. V. & STARK, R. D. (1971). Hepatic vascular bed. Physiological Reviews 51, 23-65.HALES, J. R. S. (1973). Radioactive microsphere measurements of cardiac output and regional tissue

blood flow in the sheep. Pflugers Archiv 344, 119-132.HALES, J. R. S. (1974). Radioactive microsphere techniques for studies of the circulation. In Clinicaland Experimental Blood Flow Measurement, ed. HALES, J. R. S. & WHITE, S. W., Clinical andExperimental Pharmacology and Physiology, Supplement 1, pp. 31-46.

HALES, J. R. S. & CLIFF, W. J. (1975). Behaviour of microspheres in the microvascular circulation ofrabbit ear chambers. Proceedings of the Australian Physiological and Pharmacological Society 6,119-120.

MAKOWSKI, E. L., MESCHIA, G., DROEGEMUELLER, W. & BATTAGLIA, F. C. (1968). Measurements ofumbilical arterial blood flow to the sheep placenta and fetus in utero. Circulation Research 23,623-631.

MESSAGE, M. A. (1966). The innervation of the blood supply to the papillae of the rumen of the sheep.Journal of Physiology 187, 17P.

SELLERS, A. F., STEVENS, C. E., DOBSON, A. & MCLEOD, F. D. (1964). Arterial blood flow to theruminant stomach. American Journal of Physiology 207, 371-377.

VATNER, S. F., PATRICK, T. A., HIGGINS, C. B. & FRANKLIN, D. (1974). Regional circulatoryadjustments to eating and digestion in conscious unrestrained primates. Journal of AppliedPhysiology 36, 524-529.

WEBSTER, A. J. F. & WHITE, F. (1973). Portal flow and heat production in the digestive tract of sheep.British Journal of Nutrition 29, 279-292.

88

changes in the blood flow to the digestive organs of

Documents