492

J. Physiol. (1960), 153, pp. 492-509

With 7 text-ffgures Printed in Great Britain

MAMMARY-GLAND BLOOD FLOW AND OXYGEN, GLUCOSE AND VOLATILE FATTY ACID UPTAKE IN THE CONSCIOUS GOAT

BY J. L. LINZELL From the Agricultural Research Council Institute of Animal Physiology, Babraham, Cambridge

(Received 13 May 1960)

It has been shown that in the isolated perfused mammary gland the blood flow is often lower than that measured through the gland in situ in anaesthetized animals (Linzell, 1950; Hebb & Linzell, 1951). In an "investigation of the secretory activity of the perfused udder of the goat (Cargill-Thompson, Drury, Hardwick, Linzell & Tucker, 1958), it was felt desirable to know the blood flow through the udder of the conscious animal in order to assess the importance of this factor in the reduced level of milk secretion that is characteristic of this preparation. Figures were also needed for 02 consumption and for the uptake of some milk precursors. Apart from one figure quoted by Graham (1937), which was given without the weight of the tissue, the only published figures for udder blood flow in the conscious goat are those of Jung (1932a, b, 1933) who found values of 5-5-7*2 ml./100 g/min and Vladimirova (1955) who gives 4-18-7 for conscious and anaesthetized animals. These values were obtained with stromuhrs in animals held down on their sides and are considerably lower than those obtained earlier in the anaesthetized lactating goat (30-60 ml./ 100 g/min) by Linzell (1954). A number of indirect measurements based on the uptake, from analysis of arterial and venous blood samples, of substances appearing in the milk (reviewed by Folley, 1949) have indicated that the ratio of blood flow to milk yield is about 500: 1 in the cow and goat. This would indicate absolute flows of at least 30 ml./100 g/min. In view of these discrepancies it was decided to make new blood-flow measurements at different stages of the reproductive cycle in goats trained to stand quietly without restraint in the laboratory.
METHODS Two methods, both based on the indicator dilution principle of Stewart, have been found to be applicable to mammary blood-flow measurements in conscious goats subjected to the minimum of restraint (Linzell, 1957). In the first, arterial inflow is measured by injecting Evans Blue dye into the main mammary artery and measuring the concentration-time

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

MAMMARY BLOOD FLOW

493

curve of the dye in the venous blood. In the second, venous outflow is measured by the thermodilution method of Fegler (1954, 1957), in which cold saline is injected into a main vein and the concentration-time curve of the 'cold' is recorded with a thermocouple downstream. The first method is limited in value because it is difficult to make repeated injections into the mammary artery without disturbing the animal or subjecting it to considerable surgical preparation. The venous outflow method is easily repeatable over long periods, but since there are two main veins draining each of the two glands forming the udder, one (extemal pudic) as inaccessible as the artery, total venous outflow cannot be measured easily in all animals. However, in many multiparous goats all the venous blood leaves via the readily accessible caudal superficial epigastric vein in front of the udder, in the standing position, and the majority of measurements quoted in this paper were made in this way. The methods of selecting animals on this basis for long-term experiments are described in the preceding paper (Linzell, 1960). In many of the experiments arterial and mammary venous blood samples were taken immediately after flow measurements for the calculation of mammary uptake of 02, glucose and volatile fatty acids, and for the calculation of blood flow indirectly from the clearance of Ca. General procedures At a preliminary operation under cyclopropane and oxygen anaesthesia a carotid artery and the caudal superficial epigastric vein ('milk vein') on one side were exteriorized as skincovered loops. At the same time the vessels crossing between the two halves of the udder were divided and their regrowth prevented by careful oversewing with the median suspensory ligament or by separating the two halves permanently with simple skin flaps. In early experiments polythene sheets were sewn between the glands. However, in the course of months, the fibrous tissue around the plastic gradually contracted so that the sheet was tightly compressed into a small package and ceased to separate the udder halves entirely. In none of the animals, however, was there regrowth of large anastomotic vessels across the mid line. All measurements were made in goats trained to stand in a stand in the laboratory, where they were held only by a neck yoke which allowed them to move their heads in all directions and to feed. The animals were thoroughly accustomed to the laboratory and the experimenters before the experiment began and walked into the stand quite willingly. They stood quietly (often eating or ruminating) during the measurements and were not held in any way. Their heart rates were not significantly different from when they were standing in the field or animal house. No painful procedures were associated with the laboratory because the vein anid carotid loops were locally anaesthetized with 2 % lignocaine or 2 % procaine in oil before enitering. On a few occasions, when the local anaesthesia was incomplete, puncture of the carotid loop in particular was resented by some animals who reacted by bleating and stamping their feet. Since on all other occasions the animals stood quietly eating whilst the injection, sampling and thermocouple needles were inserted, it can be concluded that the flows were measured in completely undisturbed animals. Some animals did not like being left alone and bleated when the experimenters left the room. The presence of a stranger in addition to the experimenter did not appear to upset them. These precautions were taken because it has been shown in the cow that the disturbance of obtaining arteriovenous samples without local anaesthesia may seriously alter the blood composition, chiefly in respect of haemoglobin and fat (Graham, Kay & McIntosh, 1936; Shaw & Petersen, 1939). The maximum arteriovenous difference for haemoglobin or haematocrit was 3 00 and on average was less than 0-5 % in these experiments. Furthermore, no significant difference in composition was found in two venous samples drawn successively over about 2 min, or between samples taken from mid stream or from the periphery of the vein. As an additional check on the possible effect of the experiments upon the blood flow, the milk yields of the two halves of the udder were recorded separately throughout the entire 32 PHYSIO. CLIII
Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

494

J. L. LINZELL

lactation period, whereas the blood flows were measured on one side only. Blood flow measurements, even when carried out many times a day, had no immediate effect upon the milk yield of the experimental or the control gland. Furthermore, the total yields of the two halves during the lactation differed by no more than was found in unoperated animals from the same herd. Arterial blood flow was measured by the injection of 1 ml. 0 5-1 0 % Evans Blue into the main artery of one gland and recording the concentration-time curve of the dye in successive samples of venous blood drawn from the 'milk' vein. The dye concentration was measured either in the plasma or- whole blood in a spectrophotometer at 620 mp. The method was checked in isolated perfused goats' udders by comparing the estimated flow with the venous outflow measured directly and with the arterial inflow calculated from the calibration of the perfusion pump carried out after the experiment. The accuracy was 104 + 4 % (S.D.). When the glands were perfused in situ, in anaesthetized goats supported in the standing position,

EP
/

~~~~~~~~C.
b

Fig. 1. Method of measuring venous outflow from one mammary gland of the standing goat by the thermo-dilution method, with an exteriorized 'milk' vein (M). (a) Disposition of main vessels; during recording the external pudic (EP) and perineal (P) are clamped manually. (b) Arrangement of injection needle and needle thermocouple. (c) Method of fixing thermocouple to the loop. (d) Tip of injection needle giving three jets.
the accuracy was 103 + 20%. The artery is difficult to puncture without disturbing the animal, and when exteriorized as a skin-covered loop, the tension necessarily involved and the movement of the udder cause the vessel to close down and atrophy. This had no ill effects on the gland because of the many collaterals which quickly enlarged. In one out of three animals the artery was sucoessfully exteriorized by transplanting one half of the udder, with its main vessels and nerves, to the inside of the thigh so that movement was considerably lessened. In other animals repeated intra-arterial injections were made by cannulating the constant side branch of the main artery that supplies the skin at the side of the udder. In spite of repeatedly washing through with heparin, clotting occurred after a few days. Venous blood flow was measured by injecting saline (0X2-0*6 ml.) at room temperature into the 'milk' vein loop and recording the change of temperature downstream with a needle thermocouple (Fig. 1 b). The needle for the injection of saline was blocked at the tip and had

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

MAMMARY BLOOD FLOW

495

three small holes drilled near the end so that a forceful injection mixed the saline as thoroughly as possible with the blood (Fig. 1 d). The thermocouple was made of copper constantan 36 S.W.G. wires and the hot junction was welded to the tip of a hypodermic needle, which was held in the mid stream of the vein 5-9 cm from the injection needle by a movable shank and rubber strap (Fig. 1 c). No solder was used in the construction of the hot or cold junctions of the thermocouple, and the switches, wires and terminals were made of copper. The injection needle, with a Perspex tap of low dead space, and the thermocouple needle and leads were s3 light and attached in such a way that the animal was quite free to move without disturbing itself or the apparatus. The recording galvanometers were of 1 sec period, slightly less than critically damped and had a sensitivity in use of 33-35 mm/' C. They were calibrated every time they were used against N.P.L. thermometers calibrated to 0-1 C. The records were photographed on moving paper and the area under each response to the passage of cooled blood measured with a planimeter accurate to 0-3 %. The blood temperature and that of the injected saline were recorded just before each injection so that the quantity of injected 'cold' could be determined and the flow calculated from this and the concentrationtime curve of 'cold' (Fegler, 1954, 1957). When the method was used to measure the steady flow of water at 36-40 'C through rubber tubing of the same size as the vein, the accuracy was 99 + 5 % (S.D.). In a conscious lactating goat the accuracy was 105 + 11 % as checked by making injections whilst the venous outflow was being collected from a T cannula temporarily inserted into the vein distal to the point of injection. The method was reliable when checked with two needle thermocouples inserted at the same place and when they were a few centimetres apart, but some difficulties were encountered in non-lactating animals, owing to the large size of the superficial epigastric vein. The diameter of the exteriorized vein was 0-8-1-2 cm and during lactation it carried flows of 300-1000 ml./min. The blood temperature was 38.5-39-50 C and did not vary more than 0.10 C in different parts of the stream, even though the flow was laminar as revealed by the injection of radio-opaque fluids. In non-lactating animals the flow was only 50-150 ml./min, but since the vein is below heart level and sometimes carried non-mammary blood its diameter did not decrease. Under these conditions with the lower flows the blood stream did not always completely fill the vein. Radio-opaque fluids sometimes clung to one wall and the flow profile was often erratic. Relatively stagnant pockets were found where the blood temperature was 0.5-3 C below the hottest point, so that the positioning of the thermocouple then became critical. Below 50 ml./min in such large veins the slow transit time allowed heat exchange to occur and the method was unreliable. Validation and comparison of methods. Venous outflow from the 'milk' vein was measured directly in two goats (one lactating and the other dry) via a T cannula inserted in the vein under local anaesthesia. The side arm of the cannula had two parallel vertical tubes one of which was used to measure the venous pressure and the other for collecting the blood. Flow was measured by simultaneously clamping the vein beyond the cannula and opening the collecting tube, making sure that during collection the venous pressure remained at its previous level. In both goats other tests showed that the total venous outflow from one gland was flowing through this vein in the standing position. The recorded flows (45 ml./ 100 g/min for the lactating animal and 28 for the dry one) agree well with the mean values in Table 1. As an additional check the venous outflow (thermodilution) was compared with the arterial inflow (dye dilution) in two lactating goats. In the first, which was selected because total venous outflow was not via the 'milk' vein, the arterial inflow was 31 % greater than the ' milk' vein flow. In the other goat, in which it was believed that the total venous outflow was being measured, the arterial inflow (303 + 37 ml./min) over a period of 2 hr agreed well with the flow measured in the 'milk' vein (329 + 52 ml./min). Gland size. Estimates were made of the weights of the udders by displacement of water after removing the milk as thoroughly as possible with oxytocin, or in the case of the smaller udders of dry animals by taking an impression in calcium alginate (Zelex, Amalgamated

32-2
Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

496

J. L. LINZELL

Dental Co.), making a plaster of Paris cast from the impression and getting the volume from the cast. The specific gravity of udder tissue was found to be 1-032-1-045 for lactating animals and 1-03 and 1-054 in two dry pregnant ones. When checked on seven lactating animals that were killed immediately after the measurements, the maximum error encountered was 12 %, the mean error 3 % and in four cases the estimate was within 1 % of the actual weight. In the case of dry non-pregnant animals three had udder sp.gr. of 1-05, 1F07 and 1416, but in a fourth it was 1*0 because in this, a fat animal, the udder was over 50 % adipose tissue. Although a great amount of fat can be detected by palpation and the animals used in this work were not excessively fat, the estimates of the weight of mammary tissue must be considered less accurate in dry animals. In these the maximum error encountered was 25 % and the mean error 8 %. Analytical methods Blood. This was analysed at once for 02 and CO2 in the Van Slyke manometric apparatus; for haemoglobin as cyanmethaemoglobin in a spectrophotometer at 642 mix; for plasma sugar by Somogyi's (1952) method with Nelson's (1944) precipitant; and for calcium and magnesium, by titration of the oxalate and ammonium phosphate salts with EDTA (Wilson, 1955). Analyses were also sometimes made for volatile fatty acids (VFA) (Annison, 1954), inorganic phosphorus (Fiske & Subbarow, 1925 and later Berenblum & Chain, 1938), and packed cell volume by centrifuging at 15000 g for 15 min. Milk. A portion of the total daily milk yield from the experimental side on the day of flow measurement was analysed at once for lactose (Somogyi's reagent after Grimbleby's (1956) precipitation) and for total solids, ash, fat (Rose-Gottlieb) and calcium from the ash after storage at -20 C.
RESULTS

The rate of blood flow through a single mammary gland was found to vary in fourteen conscious goats from 50 to 1000 ml./min, depending upon the stage of the reproductive cycle. When related to the estimated weight of tissue the range was 15-66 ml./100 g/min (Table 1). The flows during lactation were considerably higher than those previously recorded by Jung (1932 a, b; 1933) and Vladimirova (1955), both of whom held the goat down on its side and inserted stromuhrs into the main external pudic artery to the mammary gland under local anaesthesia (the ethyl chloride used by Jung may not have been fully effective). Although it will become apparent that, considered with average arteriovenous differences of milk precursors, the higher flows recorded in this work are the minimum values that could account for the milk formed, some preliminary experiments were done under the conditions used by Jung and by Vladimirova. It was found that holding a goat on its side lowered the mammary blood flow by at least 50 % in both lactating and dry animals, as compared with the flows measured at the same time on the previous day (Fig. 2). It will be seen that under effective local and epidural anaesthesia the lowered blood flow was related more to the position of the animal than to the surgical procedure. However, under all forms of anaesthesia dissection of the mammary artery caused it to go into spasm. The fall in flow due to the animal being held down seems likely to be due to stimulation of the sympathetic nervous

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

p. 497, Table 1, 8hould read

Lactating
More than B50ml./lOOg/day

A-V duff. Glucose (mg/100 g/min)

Less than 50 ml./100 g/day

1-17 (21) 7+1 0

5-40 (148) 16+ 0-58

Uptake Glucose (mg/lO0g/min)

2*3-19-7 (81) 6-80 37

0*2-3*4 (5)

1F7+0-51

02 uptake: for mg/100 g/min read ml./100 g/min.

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

MAMMARY BLOOD FLOW

497

TABLE 1. Summary of observations made on 13 goats at different stages of the reproductive cycle; range, (no. of observations), mean + S.E. of mean. Each observation was made during one session in the laboratory and is the mean of several blood-flow measurements but only one pair of blood samples Lactati ing Non-lactating

A-V diff.

More than 50 ml./100 g/ day

Less than 50 ml./100 g/

2-4-7-5 (139) 4-88 + 0-085 Glucose 5-53 (148) (mg/ 100 ml.) 19-4+ 0-78 VFA (mM) 0-42-1-65 (20) 0-95+ 0-08 R.Q. 0-85-2-4 (130) 1-24+0-02 Blood flow 32-66 (97) (ml./100 g/min) 47 + 0-92

02 ml./100 ml.

4-6+0-42 1-22 (21) 8-6+ 1-3 0-4-1-3 (5) 0-98 + 0-165 0-96-1-5 (20) 1-24+0-011 20-44 (10) 31-5 + 2-2 0-75-1-86 (6) 1-42 + 0-16

day 2-8-6-6 (13)

Pregnant 1-3-7-2 (38) 3-9+0-18 +9-20 (43)

0-7 + 0-057 0-49-2-1 (31) 1-16+0-06 15-43 (30) 26+ 1-5

Non-pregnant 2-0-5-0 (4) 3-7 +9-5 (4) 1 5+0-71 0-25-0-88 (11) 0-33 (1)
0-78-1-25 (4) 1-01 19-39 (6) 28 + 2-8

Uptake

02

(mg/100 g/min) 9-0 + 0-5 Acetate 0-9-3-7 (12) (mg/100 g/min) 2-3+0-23
Local

1-10-4-47 (76) (mg/100 g/min) 2-34 + 0- 10 Glucose 2-3-26-3 (81)

0-58-2-02 (28) 1-09 + 0-057

0-6-2-2 (5)
1-6+0-31

2-2 + 0-76

0-2-3-6 (5)

+1-8-4-1 (28)
0-85+ 0-25

0-71, 1-27, 1-96 (3) 1-31 0-93, 1-8, + 1-7 (3)

0-34

0-5-1-4 (9)
0-8+0-1

Epidural

General Udder
-

Anaesthetic

I
Vein
-

Vein
-

Operation
Mean blood flow in dry goats

_E
bo

-_ 20

i
0
'a

10

Position

5 4 6 7 Hours Fig. 2. Effect of posture, anaesthesia and operation on mammary blood flow in a non-lactating goat. The flow from one gland was measured directly at the existing venous pressure via a T cannula inserted in the caudal superficial epigastic vein. The animal received chlorpromazine 2 mg/kg 1 hr before the experiment and heparin 1000 i.u./kg at the start. The flow fell markedly only when the animal was held on its side. The minor operation (cannulation of 'milk' vein) in this position was not responsible because a more extensive operation (separation of two halves of udder) in comfortable sitting position had no effect. Notice the slow recovery after the fall in flow. 3
Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

498 J. L. LINZELL system. It is well known that ruminant animals are extremely uncomfortable on their sides, chiefly because they cannot belch so that the rumen quickly becomes distended with the gas constantly produced by fermentation. It has been shown previously that mammary blood vessels are very sensitive to adrenaline and noradrenaline (Hebb & Linzell, 1951) and that the sympathetic adrenergic vasomotor nerves can stop blood flow entirely during stimulation (Linzell, 1950, 1953). In this work it has been found in trained animals standing quietly that intravenous adrenaline 08-1 2,ug/ kg/min reduces mammary blood flow almost to zero. A further finding shown in Fig. 2 was that after being held down, although the animals stood quietly eating as soon as they were released, the blood flow did not recover for 1-1 5 hr. After surgical operations in this position under general anaesthesia (e.g. making a carotid loop), the mammary blood flow remained 40-62 % of normal for as long as 4 hr after recovery from the anaesthetic.

Variations in blood flow Minute-to-minute. The method of venous outflow measurement used integrates the flow over a period of 5-10 sec and has an error in vitro of 5 % (S.D.) of the mean. In vivo this was apparently increased to 10 %, but it is possible that part of this additional fluctuation may be genuine, because the method could only be checked accurately by collecting the venous outflow over a period of 30 sec. In the majority of experiments where flow was measured 2-20 times over periods of 3-60 min the variation was of the order of 10 % (S.D.) In a small number of observations the flow varied 1 % or less or showed larger variations up to a maximum of 20 %. Hour-to hour. No clear evidence of significant variation was obtained. Observations were made in one goat at different hours of the day and night on 6 days between the 23rd and 34th weeks of lactation, when the milk yield was almost stationary. The mean flows for each day were identical but the standard deviation of observations varied from 8 to 38 % of the mean. The highest flows (120 % of the mean) were encountered just before and after the morning and afternoon milkings, at which time the goat was also fed, and the lowest (75 % of the mean) midway between milkings. Therefore the majority of observations reported in this paper were made just after the morning milking (9.30 a.m.) and again about 3.5-4 hr later. However, over this animal's next lactation the morning flow was only great3st on 8 out of 14 days and when the results from all the goats were analysed a-t the end of the experiment no significant difference was found. Day-to-day. The blood flow measured on one day (mean of morning and midday estimates) averaged 99 + 10 % of that measured 2-4 days earlier in the same way.

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

MAMMARY BLOOD FLOW 499 Month-to-month. Since the flows measured twice in one day and on 2 days in one week did not differ significantly, the values were pooled and used as an estimate of the blood flow during that week. Analysis of arterial and mammary venous blood samples obtained at each session gave four estimates of R.Q. and 02 glucose and sometimes volatile fatty acid (VFA) uptake. The average daily milk yield for the week and the empty weight of the gland were also recorded. In five animals this was repeated monthly during the normal lactation (some results had to be discarded because total blood flow was not always measured), a sixth animal was followed for 2 years over 2 pregnancies and 1 lactation, and a number of other estimates were made in the remaining eight goats at all stages of the reproductive cycle. The results are summarized in Table 1. It will be seen, as might be expected, that the blood flow, the arteriovenous differences, the R.Q. and the calculated uptakes of 02, glucose and VFA were all greater in lactating than in dry animals. However, in nonlactating animals there was little significant difference between nonpregnant and pregnant animals, except possibly that the uptake of glucose and the R.Q. were higher in pregnant animals. Observations made during lactation have been divided into those from fully lactating animals (more than 50 ml. milk/100 g tissue/day) and those from animals in declining lactation (less than 50 ml.). This figure was chosen because below this level lactation ceased quickly in most animals. Table 1 shows that the blood flow, 02 consumption and the uptake of glucose and VFA were all lower at this low milk yield. However, to some extent the grouping of the observations in this way is misleading, because it obscures variations occurring at different stages of pregnancy and lactation (Figs. 3 and 4). During lactation in each animal the blood flow, 02, glucose and VFA consumption all varied with the milk yield and a similar dependence was seen between the average daily milk yield and the blood flow and glucose uptake per 100 g tissue in all the goats (Fig. 5).
The ratio of blood flow to milk yield Mammary blood flow has been estimated in the past from the plasma clearance of milk precursors on the Fick principle and the results expressed as the ratio of blood flow to milk yield. Estimates have varied from 200 to 1000:1 but a figure of about 500:1 is generally accepted (see Folley, 1949). In this work it has been possible to calculate the ratio directly from observed blood flows and indirectly from arteriovenous differences. Direct estimates. These show that the ratio of blood flow to milk yield is not fixed but varies inversely with the rate of milk secretion and is thus a measure of efficiency (Fig. 6). In Fig. 6 the curve has been calculated from the regression of Fig. 5a and it will be seen that actual observations

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

500
1.0

J. L. LINZELL

VFA (mM)
0-5-

6I' | mI1ol
v

)j

(mI./100 ml.)
<
~~~~30

02

GlIucose

20 10

(mg/I0

ml1.)

105

Weight of gland (kg)

x _ sXxxz

XxX

0*5

Milk yield (I./day)

0-S

500

400Blood flow
300

(mI/

m)

X200
100

Pregnancy liness

Pregnancy

Months Fig. 3. Variations during the reproductive cycle of one goat (Bessie); right gland only; mean and range of observed data during a week. The fall in milk yield due to a mild illness at the start of lactation is abnormal (see Fig. 4).

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

MAMMARY BLOOD FLOW

501

Acetate uptake (mg/100 g/min)

02 uptake
(ml./100 g/min)
x

Glucose uptake

(mg/100 g/min)

R.Q.

\
Milk yield

(mli/1 00 g/day)
x

Blood flow Milk yield

--

0, x
*x,

Blood flow (ml .1/ 00 g/min)

Pregnancy Illness (a) Months

(b)

Fig. 4(a). Calculated data derived from Fig. 3. The values in early lactation are abnormal owing to mild illness and typical values for two other goats (Joan and Gertrude) are shown in (b). Volatile fatty acid uptake is calculated as acetate.
Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

502 J. L. LINZELL fall reasonably close to this line. At the highest yield recorded in these goats (172 ml./100 g/day) the average ratio of blood flow to milk yield was 460: 1 but at low yield (50 ml./ 100 g/day) it was over 1000: 1 and rapidly approached infinity as milk secretion ceased. This marked decrease in efficiency associated with declining lactation further justifies the separation of these observations in Table 1.
200
-

0
150
x

0 60
0

x~

0V

i~~~~~~O
1lood flo (m./0 g/i)Gucs
x~~~~~~~~~

Cpakm/00gmn

Ix 'C1 ' Blood flow (ml./100 gimin) Glucose uptake (mg/lOog/min) Fig. 5. Variations in blood flow and glucose uptake with the rate of milk secretion. Each point is the mean of 4 estimates in one week; twelve goats. Goat Bessie x, Joan 0, Gertrude A. The lines are calculated from regressions.

It is of interest that the ratio of blood flow to milk yield also varied in the expected way at different stages of lactation in each animal and also changed during a fall in yield owing to illness. The yield of the animal shown in Figs. 3 and 4a fell shortly after parturition (owing to a uterine infection) at a time when the yield normally increases (Fig. 4b). However, the blood flow fell less than the milk yield, the ratio being 880: 1, indicating a decreased efficiency in the secretory ability of the tissue. Indirect estimates. Shaw & Petersen (1940) recommended the use of the arteriovenous difference of Ca, since all milk Ca must come from the plasma, whereas, for example, not all lactose comes from glucose. The average arteriovenous difference of Ca was 0-22 + 0-01 (s.E. of mean) mg/100 ml. for goats giving more than 50 ml. of milk/100 g /day, and 0-15 + 0- 04 for nonlactating animals. These values did not appear to increase with the interval between milkings as Shaw & Petersen (1940) found in the cow, but unfortunately were small in relation to the error of the method (0-28 + 0-01

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

503 MAMMARY BLOOD FLOW mg/100 ml.). Therefore, as might be expected, the ratio of blood flow to milk yield calculated from individual figures varied widely and even the mean of 2-4 observations in a day varied from 30 to 150 % of direct estimates
200 r

150 s
.1

oo Fv-

N
.

.2

50 hx

3000 1500 2000 2500 Blood flow Milk yield Fig. 6. Relation between blood flow/milk yield and level of milk secretion, showing that this is a measure of efficiency. The curve is calculated from the regression line in Fig. 5 and the points are weekly averages (twelve goats).
500

1000

TABLE 2. A quantitative comparison of the major milk constituents with their probable plasma precursors in the goat. Based on the following average figures: milk yield 100 ml./ 100 g/day; blood flow 45 ml./l00 g/min; lymph flow 50 ml./100 g/day; milk composition (%), fat 4-76, lactose 4 7, total N 0 53. The arteriovenous differences for fat and amino N have been taken from Lintzel (1934). The calorific value of milk was calculated by the method of Perrin (1958) Milk output Mammary uptake (g/100 g tissue/day) Substance (g/100 g tissue/day) 0-142 Calcium 0-143 lactose 4-7 Glucose 3-2 4-76 Fat 3-2 Acetate 0-65 total N 0-52 Amino N (see p. 509) 81 Calories (kcal) 3.37.=- 17kcal 02

made at the same time. Nevertheless, the ratio of blood flow to milk yield (645: 1) calculated from the mean figures in Table 2 agrees well with that observed directly (660: 1). Similarly, in declining lactation the calculated ratio from Ca uptake (1700: 1) is of the same order as that observed directly (2000:1).

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

504
E
v

J. L. LINZELL

0
C 0 .

/
4,43
~~~~~~0

301 8~~~~~~
*~~~~~~>

0~~~~~
0L5
1

15

003 1
l

05
5

07

09

0c0.5

0304

0809

Arterial plasma glucose (mg/100 mi.) Arterial plasma VFA (mm) 7. Variations of arteriovenous differences of volatile fatty acids and glucose Fig. with arterial plasma level of goat Bessie during lactationi. The lines are calculated from regresions.
DISCUSSION

The blood flows recorded under these conditions are several times higher than those previously measured directly by Jung (1932a, b, 1933) and Vladimirova (1955). It is believed that the higher values obtained in this work more nearly approach the normal, for the following reasons. The measurements were made in undisturbed animals and it has been shown that the methods used previously reduce the flow. Moreover, the methods used here for long periods did not significantly affect the milk yield of the gland used for measurement as compared with the opposite gland of the same animal, or as compared with other animals of the same herd. When considered with arteriovenous differences measured at the same time (which agree well with previous figures), the calculated uptake of Ca is just enough to account for that appearing in the milk. The Th2 consumption per gram of tissue is similar to that encountered in slices of goat mammary gland by Folley b French (1949) and glucose consumption is similaralo that of slices of guinea-pig mammary gland (VenkataramanJnReithel, 1957). It must be pointed out that, since the mammary blood vessels also hhe figures given here supply the surrounding skin and connective tissue, include the blood flow through redue the f .skin of the udder and the suspensory ligaments and connective tissue capsule. In lactating goats mammary tissue forms 84 + 4 (S.D.)oeof the weight of the udder, so that the recorded blood flow must be mainly that through secretory tissue. However, in non-pregnant dry goats mammary tissue may form only 40-60 % of the total weight of the udder when involution is maximal. The figures for total flow through dry glands therefore should be interpreted with this in mind. In pregnant goats the proportion of mammary tissue increases with the growth of the gland. Previous indirect estimates of blood flow in lactating cows and goats,

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

505 MAMMARY BLOOD FLOW expressed as the number of volumes of blood that flow through the mammary gland to form one volume of milk, more nearly agree with the present values than the older direct estimates already referred to. In this work it has been found that this ratio varies with the amount of milk secreted and thus may be a measure of over-all efficiency. It is well known that the udder decreases in size during lactation. In this herd in nonpregnant animals the glands shrink to 25-40 % of their maximum weight in about 10 months, and this is accompanied by a fall in yield per gram of tissue (Fig. 4). It has been pointed out already that the proportion of mammary tissue is also lower in dry animals. If the skin, teat and connective tissue have a much lower blood flow than the mammary tissue the question arises as to whether an increasing proportion of non-mammary tissue accounts for the apparent fall in efficiency and milk yield per gram of total tissue as the milk yield declines. It seems unlikely that this could cause the whole effect. It has been calculated that the proportion of non-mammary tissue (skin, teat and capsule) of the glands concerned (500-1500 g) might be expected to increase from about 10 to 20 % during the course of the lactation. Figures used in this calculation were an average skin thickness of0 7-1 0 mm and a teat weight of 35-50 g (this changes little during a lactation). The area of skin and capsule was measured from plaster casts of the glands, where available, or calculated from the volume (assuming the gland to be a hemisphere, a cylinder or appropriate geometrical shape). Actual measurements from other goats killed at known stages of lactation were essentially in agreement. The milk yield per gram of total tissue per day decreases from about 1*7 to zero, whilst the proportion of mammary tissue only falls from 90 to 80 % of the total weight. Clearly there is a real decrease in the amount of milk secreted by each gram of tissue as lactation proceeds. Furthermore, even if it is assumed that the blood flow through the nonmammary tissue is zero then there is still a relation between the yield per gram and the blood flow per gram, and of course in this case the ratio of blood flow to milk yield increases even more markedly as the yield decreases. In fact, it may well be that the blood flow through the skin and teat is relatively high. Both structures have a large number of arteriovenous anastomoses (starch grains 22,u pass easily through goats' udders only when the skin and teat are present) and the udder skin temperature in lactating goats is 36-38 C. If this is true then the error due to the inclusion of non-mammary tissue in blood flow measurement is much smaller than is already allowed for. It must be admitted that the changes in mammary gland weight in the goat are partly due to a varying proportion of purely secretory tissue over and above a relatively constant weight of large ducts and cisterns. Although

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

J. L. LINZELL this must partly account for the variations in milk yield and blood flow, the fact that the milk yield can vary greatly whilst mammary weight changes little shows that the rate of secretion by the alveolar tissue can also change (Fig. 3). This occurs at the very end of lactation in all goats. It would seem reasonable to conclude, therefore, that the decline in milk yield that occurs during a lactation is due both to a loss of secretory tissue and to a fall in the rate of secretion per cell. Furthermore, even when allowance has been made for the increasing proportion of non-mammary tissue, there is a parallel decrease in efficiency of milk formation in that less milk is formed from each volume of blood passing through the gland. One may also investigate mammary efficiency by comparing the uptake of milk precursors from the plasma with the amount of the product appearing in the milk in the same time (Table 2). The hazards of equating small arteriovenous differences calculated from samples collected in less than a minute with milk formed during a period of hours must be stressed. This has been already pointed out in the case of Ca, where the arteriovenous difference is very small and, as might be expected, the errors are less in the case of glucose and lactose, where the difference is larger. When calculated from weekly figures of all goats the glucose accounted for as lactoje is 55 + 5 % which agrees well with the value in Table 2 calculated from means. It will be seen that for Ca the efficiency approaches 100 %. However, in the case of daily balances for individual goats in early lactation the Ca uptake was more often insufficient to account for the milk Ca, whereas late in lactation it was more often in excess of that in the milk. This supports the suggestion of Swanson, Monroe, Zilversmit, Visek & Comar (1956) from studies of45Ca excretion in the milk of cows, that the udder has a variable store of calcium. It is less easy to do quantitative balances for the major milk constituents because much modern work with isotopes shows that several plasma precursors help to form each substance (Folley, 1956). Nevertheless, milk proteins and lactose are derived mainly from plasma amino acids and glucose, respectively. In the ruminant acetate is the source of the 10% of short-chain fatty acids (C4-C12) of milk fat and most probably forms acids up to C16 as well (i.e. 50 % of fat). The remainder is largely derived from plasma lipids. The remaining figures in Table 2 show that the mammary gland transfers about 80 % of the total calories and amino N removed from the plasma into the milk. Only 50 % of the glucose can be accounted for as lactose but milk fat could account for 80 % of the neutral fat and acetate taken up. It
506

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

507 MAMMARY BLOOD FLOW would seem likely that some acetate and glucose are oxidized, but it is not possible to decide the proportion of each because the R.Q. of lactating tissue is over 1 and the R.Q. for acetate and glucose are identical. The acetate consumption of non-lactating tissue is similar to that of glucose. This suggests that acetate is being oxidized as in most other ruminant tissues, and sufficient is taken up by the lactating tissue to continue this and transform the remainder into fatty acids. Nevertheless, one of the most striking findings in this work is the very great increase in the uptake of glucose during lactation. Whereas 02 consumption doubles, and acetate consumption trebles, glucose uptake goes up nine times. This is more than sufficient to account for all the lactose and the entire 02 consumption. Gaps in our knowledge of glucose metabolism in ruminants have been pointed out by Lindsay (1959). The estimated daily turnover of glucose in a sheep is only 100 g and yet a lactating ewe gives about 80 g of lactose daily in the milk. 200 g of lactose daily would not be unusual for a lactating goat and the present figures indicate that at maximum efficiency at least 270 g of glucose would be taken up by the udder to produce this. These figures emphasize the very great demand for glucose that the onset of lactation suddenly throws on the organism. Although the arterial plasma glucose was slightly higher in lactating (60 mg %) than in dry animals (53 mg %) the figures were not statistically significant. However, as Graham, Jones & Kay (1936) found in the cow, the arteriovenous difference of glucose increased with the arterial level (Fig. 7). Similarly the level of arterial plasma VFA was higher in lactating animals (1.3 mM) than in dry ones (0.89 mM) and, as McClymont (1951) found in the cow, the arteriovenous difference was also related to this level. These figures emphasize the avid uptake of glucose and acetate by the mammary glands. It is not surprising that the present figures show that the metabolic rate of the lactating mammary gland is similar to that of the brain and approaches that of the liver. This is to be expected in an organ that is actively synthesizing. The fact that the blood flow and the uptake of milk precursors do not fluctuate widely during the 24 hr supports the recent view that ordinarily milk secretion is a steady continuous process (see Linzell, 1959). Even in dry animals the total udder blood flow is equivalent to that through skin during maximal vasodilatation, suggesting that mammary tissue is always metabolically relatively active.
SUMMARY

1. Mammary-gland blood flow, weight, milk yield and arteriovenous differences of 02, Ca, glucose and volatile fatty acids have been measured at all stages of the reproductive cycle in goats trained to stand in the laboratory.

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

J. L. LINZELL 508 2. The mean blood flow was 28 ml./100 g/min in dry goats and 45 ml./ 100 g/min in lactating animals giving 100 ml. milk/l00 g/day. Holding goats on their sides, as was done by previous workers who recorded much lower figures, significantly reduced the mammary blood flow. 3. Blood flow variation from minute to minute was 10 % of the mean (S.D.), and was not significantly greater from hour to hour or from day to day. During lactation the blood flow varied with the milk yield. 4. The mean arteriovenous differences and instantaneous R.Q. agree well with those of previous workers. The mean ratio of blood flow to milk yield (650:1) calculated from the Ca difference and the milk Ca agrees with earlier estimates and with directly observed values, but was found to increase with a fall in milk yield and is thus a measure of efficiency. 5. During lactation 02 consumption doubles, the uptake of volatile fatty acids trebles but glucose uptake increases nine times. 6. On the average 100 % of the Ca, and 80 % of the N and calories removed from the blood appear in the milk. Similarly 80 % of the neutral fat and volatile fatty acids taken up are accounted for in milk fat but only 50 % of the glucose appears as lactose. 7. The importance of glucose in the metabolism of lactating ruminants is stressed.
I am grateful to Mr I. R. Fleet for efficient and enthusiastic technical assistance, to Mr G. Bull for help with blood and milk analyses and to Dr D. C. Hardwick for many helpful

discussions.
REFERENCES

E. M. (1958). The blood flow and milk yield of goats' udders perfused with artificial media. J. Physiol. 143, 74-75P. FEGLER, G. (1954). Measurement of cardiac output in anaesthetized animals by a thermodilution method. Quart. J. exp. Physiol. 39, 153-164. FEGLER, G. (1957). The reliability of the thermodilution method for the determination of the cardiac output and the blood flow in central veins. Quart. J. exp. Physiol. 42, 254-266. Fisxs, C. H. & SuBBAROW, Y. (1925). The colorimetric determination of phosphorus. J. biol. Chem. 66, 375-400. FOLLEY, S. J. (1949). Biochemical aspects of mammary gland function. Biol. Rev. 24. 316-354. FOLLEY, S. J. (1956). The Physiology and Biochemistry of Lactation. London and Edinburgh: Oliver and Boyd. FOLLEY, S. J. & FRENCH, T. H. (1949). The intermediary metabolism of the mammary gland. I. Respiration of lactating mammary gland slices in presence of carbohydrates. Biochem. J. 45, 117-125. GRAHAM, W. R. (1937). The utilization of lactic acid by the lactating mammary gland. J. biol. Chem. 122, 1-9. GRAHAm, W. R., JoNEs, T. S. G. & KAY, H. D. (1936). Precursors in cows' blood of milk fat and other milk constituents. Proc. Roy. Soc. B, 120, 330-346.

ANNIsoN, E. F. (1954). Studies on the volatile fatty acids of sheep blood with special reference to formic acid. Biochem. J. 58, 670-680. BERENBLUM, I. & CHAI, E. (1938). An improved method for the colorimetric determination of phosphate. Biochem. J. 32, 295-298. CARGILL-THomPsoN, H. E. C., DRURY, A. N., HARDWICK, D. C., LINZELIL, J. L. & TucKEcJR,

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010

MAMMARY BLOOD FLOW

509

GRAHAM, W. R., KAY, H. D. & MCINTOSH, R. A. (1936). A convenient method for obtaining bovine arterial blood. Proc. Roy. Soc. B, 120, 319-329. GRIMBLEBY, F. H. (1956). The determination of lactose in milk. J. Dairy Res. 23, 229-237. HEBB, C. 0. & LINZELL, J. L. (1951). Some conditions affecting the blood flow through the perfused mammary gland, with special reference to the action of adrenaline. Quart. J. exp. Physiol. 36, 159-175. JUNG, L. (1932a). Importance de la circulation a travers la mamelle au repos ou en lactation chez la chevre. C.R. Soc. Biol., Paris, 109, 737-739. JUNG, L. (1932b). Particularit6s de la circulation mammaire chez la chevre en lactation. C.R. Soc. Biol., Paris, 109, 1052-1053. JUNG, L. (1933). Intensit6 de la circulation mammaire. Lait, 13, 307-315. LINDSAY, D. B. (1959). The significance of carbohydrate in ruminant metabolism. Vet. Rev. Annot. 5, 103-128. LINTZEL, W. (1934). Le chimisme de la formation du lait. Lait, 14, 1125-1130. LINZELL, J. L. (1950). Vasomotor nerve fibres to the mammary glands of the cat and dog. Quart. J. exp. Physiol. 35, 295-319. LINZELL, J. L. (1953). Internal calorimetry in the measurement of blood flow with heated thermocouples. J. Physiol. 121, 390-402. LINZELL, J. L. (1954). Some observations on the use of the perfused lactating mammary gland. Rev. canad. Biol. 13, 291-298. LINZELL, J. L. (1957). The measurement of udder blood flow in the conscious goat. J. Physiol. 137, 75-76P. LINZELL, J. L. (1959). Physiology of the mammary glands. Physiol. Rev. 39, 534-576. LINZELL, J. L. (1960). Valvular incompetence in the venous drainage of the udder. J. Physiol. 153, 481-491. MCCLYMONT, G. L. (1951). Volatile fatty acid metabolism of ruminants with particular reference to the lactating bovine mammary gland and the composition of milk fat. Aust. J. agric. Res. 2, 158-180. NELSON, N. (1944). A photometric adaptation of the Somogyi method for the determination of glucose. J. biol. Chem. 153, 375-380. PERRIN, D. R. (1958). The calorific value of milk of different species. J. Dairy Res. 25, 215-220. SHAW, J. C. & PETERSEN, W. E. (1939). Blood volume changes in the mammary gland. Proc. Soc. exp. Biol., N.Y., 42, 520-524. SHAW, J. C. & PETERSEN, W. E. (1940). The fat metabolism of the mammary gland. J. Dairy Sci. 23, 1045-1056. SOMOGYI, M. (1952). Notes on sugar determination. J. biol. Chem. 195, 19-23. SWANSON, E. W., MONROE, R. A., ZILVERSMIT, D. B., VISEK, W. J. & COMAR, C. L. (1956). A study of the variations in secretion of Ca45 by the mammary glands of dairy cows. J. Dairy Sci. 39, 1594-1608. VENKATARAMAN, R. & REITHEL, F. J. (1957). Studies on lactose synthesis in mammary gland slices. Arch. Biochem. Biophys. 70, 205-209. VLADIMIROVA, A. D. (1955). Reflex regulation of blood supply to the udder. J. gen. Biol., Moscow, 16, 141-155. WILSON, A. A. (1955). The determination of magnesium and calcium in serum or plasma with ethylene diamine tetra-acetic acid (EDTA). J. comp. Path. 65, 285-290.
Note added in proof There is some doubt whether Lintzel (1934. Z. Zucht. B. 29, 219-242) reported total amino-acid N. If he did not, then the mammary uptake may be greater than that indicated, and the over-all N-efficiency less.

33

PHYSIo. CLII

Downloaded from J Physiol (jp.physoc.org) by guest on December 29, 2010