blood and milk prolactin and the rate of milk synthesis in women

Experimental Physiology (1996), 81, 1007-1020Printed in Great Britain

BLOOD AND MILK PROLACTIN AND THE RATEOF MILK SYNTHESIS IN WOMEN

DAVID B. COX, ROBYN A. OWENS* AND PETER E. HARTMANN*

Departments ofBiochemistry and * Computer Science, The University of Western Australia, Nedlands,WA 6907, Australia

(MANUSCRIPT RECEIVED 18 APRIL 1996, ACCEPTED 16 JULY 1996)

SUMMARY

In women, the concentration of prolactin in the plasma increases in response to nipple stimulation.This response has led to the assumption that prolactin influences the rate of milk synthesis. Toinvestigate this hypothesis we have measured 24 h milk production, the short-term (betweenbreastfeeds) rates of milk synthesis and the concentration of prolactin in the blood and breastmilk,from 1 to 6 months of lactation in eleven women. Over the long term, the 24 h milk productionremained constant (means + S.E.M.): 708 + 54-7 g/24 h (n = 11) and 742 + 79-4 g/24 h (n = 9) at1 and 6 months, respectively. The average short-term rate of milk synthesis (calculated from theincrease in breast volume between breastfeeds; means + S.E.M.) did not change: 23 + 3 5 ml/h(n = 23) and 23 + 3 4 ml/h (n = 21) at 1 and 6 months, respectively. However, significantvariation in the short-term rate of milk synthesis (from <5-8 to 90 ml/h) was found both betweenbreasts, measured concurrently (coefficient of variation, c.v. = 72 %), and within the same breast,measured over consecutive breastfeeds (c.v. = 85 %). The basal and suckling-stimulatedconcentrations of prolactin in the plasma (means + S.E.M.) declined from 1 to 6 months (basal,from 119 + 93 to 59 + 29 ,ug/l; peak, from 286 + 109 to 91 + 44 ug/1). In contrast, theconcentration of prolactin in milk was much lower than in plasma, and decreased only slightlyfrom 1 to 6 months of lactation (fore-milk, from 26.4 + 10 to 23 3 + 988ug/l; hind-milk, from18-9 + 5.1 to 13 2 + 6-3 jtg/l). The concentration of prolactin in the milk was related to the degreeof fullness of the breast, such that the concentration was highest when the breast was full. Wefound no relationship between the concentration of prolactin in the plasma and the rate of milksynthesis in either the short or long term. However, the relationship between the concentration ofprolactin in milk and the degree of fullness of the breast suggests that the internalization ofprolactin, after binding to its receptor, may be restricted when the alveolus is distended with milk.

INTRODUCTION

There is strong evidence for the involvement of the endocrine system, and prolactin inparticular, in milk synthesis in women and other mammals (Cowie, Forsyth & Hart, 1980). Inwomen, suppression of prolactin secretion with bromocriptine immediately after the deliveryof the infant inhibited lactogenesis II (the onset of copious milk secretion) (Kulski, Hartmann,Martin & Smith, 1978), whereas augmentation of the concentration of prolactin in the plasmawith sulpiride on day 2 postpartum increased total milk production (suckling plus additionalexpressed milk) over the subsequent 3 days of lactation (Aono, Shioji, Aki, Hirota, Nomura &Kurachi, 1979).

Prolactin is secreted into the blood in response to the suckling stimulus (Noel, Suh & Frantz,1974; Tyson, Khojandi, Huth & Andreassen, 1975) and increased suckling frequency duringgalactopoiesis (established lactation) results in an elevation of the basal concentration ofprolactin in the plasma (Delvoye, Demaegd, Delogne-Desnoeck & Robyn, 1977; Gross &Eastman, 1983). In addition, increased suckling frequency is also associated with increasedmilk production (Hennart, Delogne-Desnoeck, Vis & Robyn, 1981; Rattigan, Ghisalberti &

) by guest on August 22, 2009ep.physoc.orgDownloaded from Exp Physiol (

http://ep.physoc.org/

D. B. COX, R. A. OWENS AND P. E. HARTMANN

Hartmann, 1981; De Carvalho, Robertson, Friedman & Klaus, 1983) and is recommended towomen as a practical means of increasing an inadequate supply of milk (Phillips, 1991). Inthis connection, drugs which increase prolactin secretion also have been recommended forwomen with infants who are failing to thrive (Lawrence, 1989). Clinically, it has beenassumed from this circumstantial evidence that the suckling-evoked release of prolactinprovides the stimulus to the breast to make more milk (Renfrew, Fisher & Arms, 1990).

Studies in Western countries have shown that the magnitude of the suckling-evoked increasein the concentration of prolactin in the plasma declines over the first 6 months of lactation(Noel et al. 1974; Tyson et al. 1975). However, this has not been related to concomitantmeasurements of milk production. In selected Chinese mothers with 'adequate' milkproduction at 40 and 60 days of lactation, Huang, Zheng & Qian (1987) reported nosignificant difference in milk production, although there was a significant decrease in theprolactin response to suckling from 40 to 60 days.

Prolactin also has been measured in milk, and while Healy, Rattigan, Hartmann, Herington& Burger (1980) found significant relationships between the concentration of prolactin in themilk and the concentrations of lactose, a-lactalbumin and total protein in milk, there are noreports on either the relationship between the concentration of prolactin in the milk and milkproduction, or the prolactin response to suckling and the short-term rates of milk synthesis inwomen.

In order to determine the importance of prolactin in the control of milk synthesis, we haveinvestigated the relationships between the concentrations of prolactin in the blood or milk andmilk synthesis, in women who were breastfeeding on demand at 1, 2, 4 and 6 months oflactation.

METHODS

SubjectsMothers (n= 11) of healthy infants breastfed on demand (Table 1) provided informed consent to

participate in the study, which was approved by The University of Western Australia, Committee forHuman Rights. The mothers were recruited through either the Nursing Mothers' Association of Australia,or private health care centres. The mothers were familiarized with the study and the operation of theequipment before any measurements were made. Measurements of the short-term rates of milk synthesisand the collection of blood began between 09.00 and 10.00 h, and continued for two to three breastfeeds,depending on the frequency of breastfeeding and the mother's time constraints. The measurements wereall performed within 1 week of 1, 2, 4 and 6 months postpartum, except for mother H, who was studiedat 3 months instead of 4 months. In addition, measurements were not made on mothers D and I at6 months, since their infants were completely weaned by this time. Blood samples were collected, and theshort-term rates of milk synthesis were measured, at the Department of Biochemistry, while the 24 hmilk production was measured in the volunteers' homes. The volunteers maintained their normal patternof breastfeeding at all times.

Plasma prolactinSoft TouchTm Lancets (Boehringer Mannheim Australia, Castle Hill, NSW, Australia) were used to

collect finger prick blood samples into heparinized haematocrit tubes (Chase Instruments Corp., GlensFalls, NY, USA) before, and 45 min after (to give an approximation of the maximum concentration ofprolactin in the plasma following nursing; Noel et al. 1974), the start of a breastfeed. The tubes werecentrifuged to separate the plasma, which was collected and frozen at -20 °C until analysis. The prolactinconcentration was determined using the RIABEAD II kit (Abbott Australasia, Diagnostics Division,North Ryde, NSW, Australia). Recovery of a known amount of prolactin added to plasma samples was1024 + 058 % (mean + S.E.M.; n = 10).

1008



PROLACTIN AND MILK SYNTHESIS IN WOMEN

Table 1. Details of the subjects and patterns of infant milk intake

Time after Fully or partlySubject Parity birth breastfeeding

(months)

A 1 1246

B 1 1246

C 3 1246

D 1 124

E 3 1246

F 2 1246

G 2 1246

H 1 1236

I 2 124

J 2 1246

K 1 1246

FullyFullyFullyPartly

FullyFullyFullyFully


FullyFullyPartly





FullyFullyPartly


FullyFullyPartlyPartly

Feedingfrequency

(breasts/24

7698

13161313

1091214

11812

78

77

888

7

12111110

910710

7

68

11119

11

11999

h)

24 h individual breastmilk production

Left breast Right breast(g/24 h) (g/24 h)

330 603335 320366 587487 363

400 527498 558357 569346 558

206 470191 263331 549462 697

332 539357 395219 516

329 517321 400382 446551 485

439 430492 398417 310551 627

452 602418 532325 530255 335

425 508344 653483 546545 450

211 190312 708289 328

224 380248 218329 298237 307

591 372373 582258 469136 265

1009




Milk productionThe 24 h milk production was measured by maternal test weighing (Arthur, Hartmann & Smith, 1987).

Due to the difficulty in defining a breastfeed for demand-fed infants, we have expressed the frequency ofbreastfeeding in terms of the number of times the infant suckled on a breast that resulted in milk removal(breasts/24 h). Mothers usually fed their infants from one breast during a breastfeeding episode (over80 %); however, if the baby had been fed from both breasts during a breastfeeding episode, we classifiedthis as two breasts, for the calculation of feeding frequency.

Short-term rates of milk synthesisThe Computerized Breast Measurement (CBM) system (Daly, Kent, Huynh, Owens, Alexander, Ng &

Hartmann, 1992; Daly, Owens & Hartmann, 1993 b) was used to determine changes in breast volume,allowing calculation of the short-term rates of milk synthesis. In addition to the apparatus described byDaly et al. (1992), a JVC-1200 Colour Special Effects Generator (Television Communications, Malaga,WA, Australia) was introduced to assist the mothers to resume the same position for each measurement,by allowing a stored and a live video image, each at half-density, to be superimposed. At the beginning ofeach session, an image of the breast was captured and stored. This image was then retrieved to a monitorwith live video input. The mothers then moved their bodies until the painted lines encircling the breast inthe live image were exactly superimposed on the same lines in the stored image.The breast volume of mothers A-G was determined before and after each breastfeed, and, when time

permitted, during the interval between breastfeeds. Short-term rates of milk synthesis were calculatedfrom the changes in breast volume over time (Daly et al. 1992, 1993 b). The accuracy of thesemeasurements was determined by comparing the changes in breast volume over a breastfeed with thechanges in maternal weight over the same breastfeed, corrected for evaporative water loss (Arthur et al.1987). The regression coefficients ranged from 0 83 to 0 96 for individual mothers with the pooledregression (0.92) being similar to that described previously (Daly et al. 1993 b). Short-term rates of milksynthesis were not calculated for mothers H-K because the volume of their breasts could not be measuredreproducibly by the CBM system, due to large occlusion zones (Daly et al. 1992).

Milk prolactinMothers A-G expressed milk samples from each breast, before (fore-milk) and after (hind-milk)

feeding their infants, both during the measurements of the short-term rates of milk synthesis and duringthe 24 h period over which milk production was measured. These samples were stored at -20 °C untilanalysis. Milk samples were defatted by centrifugation at 10000g for 10 min. Milk prolactin wasmeasured in the defatted milk samples, using the RIABEAD II kit, by directly substituting defatted milkfor plasma in the company-supplied protocol. Recovery of a known amount of prolactin added to milksamples was 102 7 + 1.6% (mean + S.E.M.; n = 7).

Milk fat contentSince 98 % of milk fat is present as triacylglycerols (Jensen, Bitman, Carlson, Couch, Hamosh &

Newburg, 1995), we determined the fat content of breast milk as the esterified fatty acid content (Stern &Shapiro, 1953; Atwood & Hartmann, 1992; Daly, Di Rosso, Owens & Hartmann, 1993a). Samples ofbreast milk were thawed, warmed to 37 °C and mixed thoroughly. Aliquots (2-5 ,ul) were taken from eachsample, placed into 600 ,ul of redistilled ethanol and mixed vigorously. This mixture and standards (2.5 ,ultriolein (Sigma T-7140) in 600#,1 of redistilled ethanol) were assayed in duplicate. The interassaycoefficient of variation was 7-26 % (n = 20) and the detection limit was 3.26 g/l.

Degree offullness of the breastFat content was determined in fore- and hind-milk collected at each breastfeed over a 24 h period. The

relationship between the milk fat content and the degree of emptying of the breast (Daly et al. 1993a)was used to calculate the degree of fullness of the breast. Milk samples containing the maximum andminimum milk fat concentration were designated as having degrees of emptying one and zero,respectively (see Daly et al. 1993 a). By fitting the curve described by Daly et al. (1993 a), the fat contentof a milk sample could be used to determine the degree of emptying of the breast at the time the samplewas collected. This was converted to the degree of fullness of the breast by subtracting the degree ofemptying from one.

1010




Statistical analysisMixed model regression analysis was performed using SAS release 6.07 (SAS Institute Inc., Cary, NC,

USA, 1992). Student's paired t test and linear regression analysis were performed using the AbacusConcepts, StatView SE+GraphicsTM software package (Abacus Concepts, Inc., Berkeley, CA, USA,1987). Coefficient of variation (c.v.) is the standard deviation expressed as a percentage of the mean. Allvalues are expressed as means + S.E.M. unless otherwise stated.

RESULTS

SubjectsThe eleven women recruited for this study all fully breastfed their infants to 2 months of age.

By 6 months only four of the mothers were fully breastfeeding their infants (Table 1). A long-term medical condition may have influenced the feeding behaviour of infant I throughlactation, while a routine immunization made infant A unsettled at 2 months and may haveinfluenced the measurement of milk production.

Plasma prolactin concentrationAfter statistical analysis to control for individual variation, there was a significant decline in

the basal concentration of prolactin in the plasma (prolactin concentration at least 90 min aftera breastfeed) from 119 + 19-1 jug/l (n = 24) and 121 + 18 3 jug/l (n = 32) at I and 2 months,respectively, to 71 + 81 ,ug/l (n = 31) at 4 months and 59 + 5.7 jug/l (n = 26) at 6 months(mixed model analysis; P = 0.0001; Fig. 1).At 1 month postpartum the mean concentration of prolactin in the plasma 45 min after the

commencement of feeding (suckling-stimulated concentration) was 286 + 22 6 jtg/l (n = 23)and by 2, 4 and 6 months, this had decreased significantly, to 218 + 24.1 jug/l (n = 31),139 + 16 4 jug/l (n = 30) and 91 + 8-7 jug/l (n = 26), respectively, after statistical analysis tocontrol for individual variation (mixed model analysis; P = 0.0001; Fig. 1).

Similarly, following statistical analysis to control for individual variation, the magnitude ofthe increase in the suckling-stimulated concentration of prolactin in the plasma declinedsignificantly, from 172 + 29-0 j/tg/l (n = 24) at 1 month to 98 + 21-5 jug/l (n = 32),70 + 131 ,ug/l (n = 29) and 31 + 5.0 jtg/l (n = 25) at 2, 4 and 6 months, respectively (mixedmodel analysis; P = 0-000 1; Fig. 1).

Milk productionThe frequency of breastfeeding did not change, being 9 6 + 0-65 breasts/24 h (n = 11) at

1 month, 9 3 + 0-84 breasts/24 h (n = 11) at 2 months, 9 6 + 0 64 breasts/24 h (n = 11) at4 months and 9 9 + 0-82 breasts/24 h (n = 9) at 6 months of lactation. The mean duration ofthe breastfeeds at 1 month was not significantly different to the mean duration of thebreastfeeds at 6 months: 30 3 + 1 6 min (n = 113) and 25-0 + 4 1 min (n = 99), respectively.The mean milk production at 1, 2, 4 and 6 months was 708 + 54-7 g/24 h (n = 11),732 + 53.2 g/24 h (n = 11), 735 + 34.7 g/24 h (n = 11) and 742 + 79.4 g/24 h (n = 9),respectively. Controlling for variation due to individual mothers, no significant difference wasfound in milk production between different stages of lactation (mixed model analysis;P = 0 77; Fig. 1). A positive relationship was found between total milk production and feedingfrequency (linear regression analysis; correlation coefficient, r2 = 0.40, P = 0 008, n = 42).

1011



D. B. COX. R. A. OWENS AND P. E. HARTMANN

400 800

(22) I300 T600£_ ~~~~~~~~(31)

E200 4000 Q~~~~~~~~~~~(0(24) (32) (30)

E 0

(26) -100 (2)200 2

0 LJ01 2 4 6

Time postpartum (months)

Fig. 1. Immunoreactive prolactin determined in plasma samples collected from the eleven mothers (at 1, 2 and 4 months),and from nine mothers (at 6 months), immediately before suckling (U) and 45 min after the commencement of suckling(LO). Number of observations is shown in parentheses. Twenty-four hour milk production (ml/24 h) of the same mothersdetermined by test weighing (e). Results are mean values + S.E.M.

Short-term rates of milk synthesisThe rate of milk synthesis in individual breasts ranged from a minimum detectable rate of

5 8 ml/h to 90 ml/h. At 1, 2, 4 and 6 months postpartum, the mean rate of milk synthesis was23 + 3.5 mI/h (n = 23), 20 + 3.0 ml/h (n = 26), 20 + 3.8 ml/h (n = 24) and 23 + 3 4 ml/h(n = 21), respectively. Controlling for variation due to the individual breast and the mother,there was no significant change in the average rate of milk synthesis over the duration of thestudy (mixed model analysis; P = 0.95). The rates of milk synthesis within a breast could begrouped into two categories: first, those where the rates of milk synthesis remained the samefollowing consecutive breastfeeds, and second, those where the rates of milk synthesischanged following consecutive feeds. Out of the thirty-three pairs of consecutivemeasurements of the rates of milk synthesis within the same breast (unilateral), eleven pairsexhibited significant changes over time (95% confidence intervals for the rates of milksynthesis did not overlap). Similarly, out of twenty-four pairs of rates of milk synthesisdetermined simultaneously for both breasts (bilateral), eleven exhibited significant differencesbetween breasts (see examples in Fig. 2). The c.v. for the absolute differences between theunilateral rates of milk synthesis was 85 %, and between bilateral rates of milk synthesis c.v.was 72 %.

Milk prolactin concentrationThe concentration of prolactin in milk was measured in seven mothers on samples collected

in conjunction with the measurements of the short-term rates of milk synthesis. Theconcentration of prolactin in fore-milk was 26 4 + 2 29 jug/l (n = 19) at 1 month, 26 4 +3.43 ,ug/l (n = 22) at 2 months, 18 7 + 1 38 ,tg/l (n = 24) at 4 months and 23-3 + 2-25 ,ug/l

1012




30 - Mo

201nv1- 10E:1o .

)ther B,T~~ID30 -

10[.

10.00 12.00 14.00Time of day (h)

30

20

10

10.

50

40 -n30

20

10-

0.10.00 12.00 14.00 16.00

Time of day (h)16.00

30 Mother D

_ 20 -

10

-z.w

ar--4

- -l.00 12.00 14.00 16.00

Time ofday (h)

Lili 1I30~

20 4 a10I

10.00 12.00 14.00Time of day (h)

16.00

Fig. 2. Examples of' the range of within-day variation in the rate of milk synthesis for the left (C1) and right (a) breastmeasured at 4 months of lactation. The height of the histogram bars represents the average rate of milk synthesisbetween measurements of breast volume (width of histogram bars). The bars with arrows indicate the start, duration andfinish of' breastfeeding episodes within a breast. In mothers A (right breast), B (right breast) and D (left breast), breastvolume also was measured in the interval between breastfeeds, and the rate of milk synthesis was calculated for eachinterval. No measurement was made for mother C on the left breast between 10.30 and 14.00 h, and mother D's infantrequired breastfeeding before the initial breast volume could be determined for the left breast.

(n = 19) at 6 months (Fig. 3). The concentration of prolactin in hind-milk was 18 9 + 1.23 jug/l(n = 17) at 1 month, 147 + 198,ug/l (n = 20) at 2 months, 12.9 + 071 ug/l (n = 19) at4 months and 13 2 + 1 45 lug/l (n = 19) at 6 months (Fig. 3). The concentration of prolactinin fore-milk was significantly higher than that in hind-milk (Student's paired t test;P = 0 0001), and when the variation due to the mothers was controlled for, the concentrationof prolactin in the milk declined from 1 to 6 months of lactation (mixed model analysis; fore-milk, P = 0 049; hind-milk, P = 0.012).

Comparative analysisNo significant relationship was found between the volume of milk removed during a feed

and the rate of milk synthesis after that feed (linear regression analysis; r2 = 0 042, P = 0.75).No relationship was found between the basal or the suckling-stimulated concentration of

40 -Mother A

30-

20-

4

1013

.4

v:

*-5




30 (20) (23)

~~~~ i ~~~~~~~~(20)

i (18) (27)120 2

(21)(22) ~(21)

I-5

010

2 ~~~46Time postpartum (months)

Fig. 3. Immunoreactive prolactin determined in milk saimples collected from seven mothers (at 1, 2 and 4 months; mothersA-G), and six mothers (at 6 months; mothers A-C and E-G). immediately before (U) and immediately after suckling(O) on the same day as the plasma samples depicted in Fig. 1. Number of observations in parentheses. Results are meanvalues + S.E.M.

prolactin in plasma and the short-term rates of milk synthesis, in either breast, immediatelyafter that breastfeed (mixed model analysis; basal, P = 0-55; suckling-stimulated, P = 0-93) orafter the next breastfeed (mixed model analysis; basal, P = 0-23; suckling-stimulated,P = 0.48).No significant relationships were found between the concentration of prolactin in the milk

and the short-term rate of milk synthesis before or after a breastfeed. In addition, nosignificant relationship was found between the concentration of prolactin in plasma and theconcentration of prolactin in milk, controlled for variation due to the individual breast, themother and the stage of lactation (mixed model analysis; fore-milk vs. basal concentration ofprolactin in the plasma, P = 0-45; hind-milk vs. basal concentration of prolactin in the plasma,P = 0.65).The concentration of prolactin in milk samples collected simultaneously with the

measurement of milk production was negatively related to the fat content and positively relatedto the degree of fullness of the breast, when both were included in the same statistical model,which controlled for variation due to the individual breast and the stage of lactation (mixedmodel analysis; fat content, P = 0.0001; degree of fullness, P = 0.0001). Although the degreeof fullness was calculated from the fat content of the sample and the maximum and minimumfat content over the 24 h period, using the relationship described by Daly et al. (1993 a), it isimportant to note that the statistical model demonstrated a relationship between the degree offullness of the breast and the concentration of prolactin in the milk, independent of the milkfat content. We have analysed mothers and individual breasts separately by linear regressionanalysis (Table 2). For six of the mothers, the degree of fullness of the right breast explained37-67 % of the variation in the concentration of prolactin in milk (Table 2), while for the leftbreasts of five of the mothers, 21-56 % of the variation in the concentration of prolactin inmilk was explained. Of the seven women, only mother D consistently showed no relationship

1014



PROLACTIN AND MILK SYNTHESIS IN WOMEN 1015

Table 2. Linear regression analyses of the relationships between milk prolactin and thedegree of breast fullness, within each breast and within each mother

Right breast Left breast Both breasts

Mother Probability Correlation Probability Correlation Probability Correlation

A 0.0005 0-62 0.95 0-0028 0.12 0.073B 0 0009 0.65 0 0059 0.37 0.0001 0 44C 0.0032 0.45 00004 0.56 00001 049D 0.65 0013 050 0.036 0.012 0.54E 0.0006 0-67 0 032 0 21 0 0001 0 37F 0.0059 0.37 00049 0-38 0.0002 0.33G 00039 0.46 0011 0.36 0.0001 040C* 0.0033 0.25 0.010 0.39 0.0007 0.22

Asterisk indicates analysis of samples from mother C obtained from a previous study (Daly et al. 1993 b), in which thedegree of breast emptying was determined using the CBM system. For this analysis, degree of breast emptying wasconverted into degree of breast fullness, by subtraction of the degree of emptying from one.

between milk prolactin and degree of fullness (Table 2). Subject C also participated in aprevious study (Daly et al. 1993 b) when she was feeding a 5-month-old infant. We havemeasured the prolactin concentrations in milk samples collected during that study andcompared them with the degree of fullness of the breasts determined directly using the CBMsystem. In this case, the degree of fullness explained 25 and 39 % of the variation in theconcentration of prolactin in milk, on the right and left sides, respectively (Table 2).Furthermore, the decrease in the concentration of prolactin from the fore- to hind-milk wassignificantly related to the volume of milk removed from the breast during a feed (Student'spaired t test; P = 0.014).

DISCUSSION

Plasma prolactin and milk synthesisWe found a significant, progressive decrease, from 1 to 6 months of lactation, in the

concentration of prolactin in blood plasma, both before (basal) and 45 min after (suckling-stimulated) the commencement of breastfeeding (Fig. 1). Despite this decline, basal prolactinat 6 months postpartum was still higher than the concentration reported for non-lactatingwomen at 6 months postpartum (Gross & Eastman, 1983). Similar decreases in basalprolactin during lactation have been found for women in the USA during the first month oflactation and from 1 to 7 months of lactation (Noel et al. 1974; Tyson et al. 1975; Battin,Marrs, Fleiss & Mishell, 1985), in Zaire from approximately I to 30 months of lactation(Hennart et al. 1981) and in the Philippines from 1 to 20 months of lactation (Gross, Haynes,Eastman, Balderrama-Guzman & del Castillo, 1980). Basal prolactin, at 1 month of lactation,ranged from as low as 15 + 1.4 ug/l (n = 3) (Noel et al. 1974) to 119 + 19.1 ,ug/l (n = 24)measured in the present study (Fig. 1). This variation may be explained by differences in thefrequency of breastfeeding, which ranged from four to six times per 24 h up to 2 months oflactation and from two to four times per 24 h up to 7 months of lactation for Americanmothers feeding to a strict routine (Noel et al. 1974), to 9 6 + 0-65 feeds per 24 h (n = 11) for




infants breastfed on demand from 1 to 6 months of lactation in the present study (Table 1) andnine to ten feeds per 24 h for at least the first 12 months of lactation in Zairean women(Hennart et al. 1981). Furthermore, Gross et al. (1980) found a significant correlationbetween the frequency of breastfeeding and the basal concentration of prolactin in bothamenorrhoeic and menstruating lactating Filipina women. Since the clearance of prolactinfrom the plasma can take up to 180 min (Noel et al. 1974), more than eight feeds per 24 hmay not allow the concentration of prolactin to fully decline before the next breastfeed and thiscould explain the elevated basal concentration of prolactin associated with more frequentbreastfeeding.We found no change in milk production from 1 month (708 + 54.7 ml/24 h (n = 11)) to

6 months (742 + 79.4 ml/24 h (n = 9)) of lactation (Fig. 1). Similar milk intakes have beenreported for longitudinal studies in the USA by Neville, Keller, Seacat, Lutes, Neifert, Casey,Allen & Archer (1988) (range, 739 + 47-3 ml/24 h (n = 12) to 787 + 24 4 ml/24 h (n = 13))and Dewey & Lonnerdal (1983) (range, 673 + 48.0 ml/24 h (n = 16) to 896 + 36.8 ml/24 h(n = 1)). However, concurrent prolactin measurements were not made in the earlier studies.In the present study, while milk production remained relatively constant until 6 months oflactation, the concentration of prolactin in plasma declined (Fig. 1). This is consistent withHuang et al. (1987), who reported that between 40 and 60 days postpartum there was no

change in milk production (1160 + 102 and 1125 + 108 g/24 h, respectively), while there was

a decline in the concentration of prolactin in the plasma (from 3975 + 533 mU/I in 13 womento 2695 + 334 mU/l in 15 women; P < 0 05) in mothers selected from groups of women

(n = 24 and 26, respectively) on the basis of 'adequate' milk volume. Therefore, 24 h milkproduction was not controlled by either the basal or suckling-stimulated concentration ofprolactin in the blood.

Practical experience suggests that, within women, there is considerable variation in milkproduction between breasts (Mobbs, 1990; Phillips, 1991). We have found that milkproduction from right breasts was significantly higher than that from left breasts (mixed modelanalysis; P= 0021; Table 1), which is consistent with Daly et al. (1993 b), who reportedasymmetric milk production. Whereas almost all of the literature presents only the combinedmilk production for both breasts, the above findings stress the importance of assessing eachbreast independently for investigation of the mechanisms controlling milk production inwomen.We have investigated the possibility of a relationship between the concentration of prolactin

in the plasma and the short-term rate of milk synthesis (i.e. the rate of milk synthesis betweenbreastfeeds). We found no significant change in the mean short-term rate of milk synthesisfrom 1 to 6 months of lactation (mixed model analysis; P = 0.95) and there was norelationship between the mean short-term rate of milk synthesis and either the basal orsuckling-stimulated concentration of prolactin in the plasma. In agreement with Daly et al.(1993b), there was considerable variation in the short-term rate of milk synthesis bothbetween breasts (c.v. = 72 %) and within the same breast (c.v. = 85 %0) (see examples in Fig. 2).Nevertheless, these variations in the short-term r.ate of milk synthesis immediately following abreastfeed were not related to variations in either the basal or suckling-stimulatedconcentration of prolactin in the plasma, controlled for variation due to the individual breasts,the mothers and the stage of lactation (mixed model analysis; basal concentration, P = 0.55;suckling-stimulated concentration, P = 0 93). Since prolactin reaches a peak concentration inblood approximately 45 min after the beginning of a breastfeed (Noel et al. 1974), it was

1016




possible that the effect of prolactin was delayed until after the subsequent breastfeed. However,there were no relationships between the short-term rate of milk synthesis following thesubsequent breastfeed and either the basal or suckling-stimulated prolactin from the previousbreastfeed, controlled for variation due to the individual breasts and the stage of lactation(mixed model analysis; basal concentration, P = 0.23; suckling-stimulated concentration,P = 0.48).The bilateral variation in the short-term rates of milk synthesis following successive

breastfeeds (Fig. 2) indicates that the rate of milk synthesis within one breast was independentof the rate of milk synthesis in the other breast. For example, whereas the rate of milksynthesis in one breast was higher than in the other breast after one breastfeed, the reverse wasoften observed after the next breastfeed (see Fig. 2, mother B). Overall, we found that ten outof thirteen sets of bilaterally paired rates of milk synthesis exhibited significant reversals in therate of milk synthesis occurring following consecutive breastfeeds. Similarly, Daly et al.(1993 b) found significant variations in the short-term rate of milk synthesis between breastswithin women. These findings clearly demonstrate that milk synthesis is controlledindependently in each breast.

Since each breast would be expected to be exposed to the same concentration of bloodprolactin, asymmetry in milk production and bilateral variations in the rate of milk synthesisare inconsistent with changes in the concentration of prolactin in the blood directly regulatingeither the amount of milk produced or the short-term rate of milk synthesis. Nevertheless,threshold concentrations of prolactin in blood are essential for both the initiation (Kulski et al.1978) and maintenance of lactation (Cowie et al. 1980). Thus, for blood prolactin to have aneffect on the regulation of milk synthesis, it would have to act through a locally selectivemechanism within individual breasts.

Milk prolactin and milk synthesisThe concentration of prolactin in milk (Fig. 3) was significantly lower (P = 0.049) than its

concentration in blood plasma (Fig. 1), in agreement with previous reports for milk prolactinat 2 weeks (Healy et al. 1980) and 4 weeks (Yuen, 1988) of lactation. The concentration ofprolactin in the defatted milk varied during suckling, with fore-milk having a significantlyhigher concentration of prolactin than the hind-milk (Fig. 3). Yuen (1988) compared twenty-one paired samples of fore- and hind-milk, collected from women between 7 and 88 days oflactation, and found a similar difference between fore- and hind-milk. Noilin (1979) reportedthat there was cyclic movement of prolactin into the lactocytes (mammary secretory epithelialcells) of the rats, with the highest rate of prolactin entry into the lactocyte occurring when thealveolus was empty, and decreasing as the alveolus filled with newly secreted milk. This leadYuen (1988) to speculate that the variation in the concentration of prolactin between fore- andhind-milk of women may be a result of the cyclic entry of prolactin into the lactocyte inresponse to alveolar distension and contraction. This hypothesis is consistent with therelationship between the concentration of prolactin in the milk and the degree of fullness ofthe breast, providing that there is limited mixing of milk in the ducts, permitting the formationof a concentration gradient for milk prolactin from the alveoli to the nipple. Since prolactinmRNA has been detected in rat lactocytes (Steinmetz, Grant & Malven, 1993), the possibilitythat de novo synthesis of prolactin contributed to the differences in the concentration ofprolactin in fore- and hind-milk cannot be dismissed.

1017




Autocrine control of milk synthesisRecent research suggests that autocrine mechanisms appear to control the rate of milk

synthesis in rats (Heesom, Souza, Ilic & Williamson, 1992), goats (Peaker & Wilde, 1987;Wilde, Calvert, Daly & Peaker, 1987), marsupials (Sharman, 1970) and women (Prentice,Addey & Wilde, 1989; Daly et al. 1993 b). A 7-6 kDa component of the whey fraction ofgoats' milk (feedback inhibitor of lactation, FIL) has been shown to inhibit the synthesis andsecretion of milk by stopping membrane trafficking between the endoplasmic reticulum andthe Golgi body (Wilde, Addey, Boddy-Finch & Peaker, 1995). Proteins similar to FIL havebeen identified in the milk of other species, including cows (Wilde et al. 1995), tammarwallabies (Nicholas, Wilde, Bird, Hendry, Tregenza & Warner, 1995) and women (Prentice etal. 1989). Bennett, Knight & Wilde (1990) found that, when applied to mouse lactocytes, a10-30 kDa whey fraction from goats' milk, containing the autocrine inhibitor (2 timesconcentrated relative to milk), reduced the number of receptors for prolactin on the basalmembrane of the lactocyte, thereby reducing prolactin binding. A reduction in the number ofprolactin receptors is also consistent with the cyclic movement of prolactin into the lactocytesreported by Noilin (1979) and the higher concentration of prolactin in fore-milk (providingfore-milk represents milk secreted immediately following a breastfeed). Daly et al. (1993 b)found that the short-term rate of milk synthesis was higher when most of the available milkhad been removed from the breast. While we did not find a relationship between the short-termrate of milk synthesis and the concentration of prolactin in the milk, it is of interest that we didfind a relationship between the degree of fullness of the breast and the concentration ofprolactin in milk. Indeed, the only mother (mother D) who consistently showed no relationshipbetween milk prolactin and the degree of fullness experienced problems with an over supply ofmilk early in lactation, suggesting a lack of response to local feedback inhibition of milksynthesis. Consequently, as milk accumulates within the mammary gland, the binding ofprolactin to its membrane receptors, signal transduction and subsequent movement ofprolactin into the lactocyte during receptor-ligand uncoupling are likely to be reduced by theaction of FIL. Such a mechanism would permit asynchronous milk synthesis between breastsboth in the short and long term.

In summary, our investigations show that changes in the concentration of prolactin in theblood could not directly regulate milk synthesis in women. They do not, however, precludeprolactin from having a permissive role in milk synthesis. The variations in the rate of milksynthesis, both unilaterally and bilaterally, are indicative of autocrine control mechanismsregulating milk synthesis. Furthermore, the gradient for prolactin in the milk and the knownrequirements of prolactin for the synthesis of milk (Cowie et al. 1980) suggest that the controlof milk synthesis may be by an interplay of autocrine inhibitors and endocrine effectors withinindividual breasts. Since infants consume irregular quantities of milk at irregular times duringthe day (Hartmann, Sherriff & Kent, 1995), this interplay would allow the breast to regulatemilk synthesis such that the variable demands of the infant can be met, with minimal wastageof maternal resources.

We offer special thanks to the volunteer subjects, their families and the Nursing Mothers' Association ofAustralia. We thank Drs R. Parsons and H. Vu of The University of Western Australia, Department ofPublic Health, Biostatistics Consulting Unit, for their valuable statistical assistance; Ms Y. Ge formodifications to the Shape(© Measurement System and Ms B. Lovelock and Ms J. C. Kent for theirtechnical assistance. The project was supported by the National Health and Medical Research Council.

1018




REFERENCES

AONO, T., SHIOJI, T., AKI, T., HIROTA, K., NOMURA, A. & KURACHI, K. (1979). Augmentation ofpuerperal lactation by oral administration of sulpiride. Journal of Clinical Endocrinology andMetabolism 48, 478-482.

ARTHUR, P. G., HARTMANN, P. E. & SMITH, M. (1987). Measurement of milk intake of breast-fed infants.Journal of Pediatric Gastroenterology and Nutrition 6, 758-763.

ATWOOD, C. S. & HARTMANN, P. E. (1992). Collection of fore- and hind-milk from the sow. Journal ofDairy Research 59, 287-298.

BATTIN, D. A., MARRS, R. P., FLEISS, P. M. & MISHELL, D. R. (1985). Effect of suckling on serumprolactin, luteinizing hormone, folicle-stimulating hormone, and estradiol during prolonged lactation.Obstetrics and Gynecology 65, 785-788.

BENNE'TT, C. N., KNIGHT, C. H. & WILDE, C. J. (1990). Regulation of mammary prolactin binding bysecreted milk proteins. Journal ofEndocrinology 127, (suppl.), 141.

COWIE, A. T., FORSYTH, I. A. & HART, I. C. (1980). Hormonal Control of Lactation. Springer-Verlag,Berlin.

DALY, S. E. J., Di Rosso, A., OWENS, R. A. & HARTMANN, P. E. (1993a). Degree of breast emptyingexplains changes in the fat content, but not fatty acid composition, of human milk. ExperimentalPhysiology 78, 741-755.

DALY, S. E. J., KENT, J. C., HUYNH, D. Q., OWENS, R. A., ALEXANDER, B. F., NG, K. C. & HARTMANN,P. E. (1992). The determination of the short-term breast volume changes and the rate of synthesis ofhuman milk using computerized breast measurement. Experimental Physiology 77, 79-87.

DALY, S. E. J., OWENS, R. A. & HARTMANN, P. E. (1993 b). The short-term synthesis and infant-regulatedremoval of milk in lactating women. Experimental Physiology 78, 209-220.

DE CARVALHO, M., ROBERTSON, S., FRIEDMAN, A. & KLAUS, M. (1983). Effect of frequent breast-feedingon early milk production and infant weight gain. Pediatrics 72, 307-31 1.

DELVOYE, P., DEMAEGD, M., DELOGNE-DESNOECK, J. & ROBYN, C. (1977). The influence of the frequencyof nursing and of previous lactation experience on serum prolactin in lactating mothers. Journal ofBiosocial Science 9, 447-451.

DEWEY, K. G. & LONNERDAL, B. (1983). Milk and nutrient intake of breast-fed infants from 1 to6 months: Relation to growth and fatness. Journal of Pediatric Gastroenterology and Nutrition 2,497-506.

GROSS, B. A. & EASTMAN, C. J. (1983). Effect of breast-feeding status on prolactin secretion andresumption of menstruation. Medical Journal of Australia 1, 313-317.

GROSS, B. A., HAYNES, S. P., EASTMAN, C. J., BALDERRAMA-GUZMAN, V. & DEL CASTILLO, L. V. (1980).A cross-cultural comparison of prolactin secretion in long-term lactation. Progress in ReproductiveBiology 6, 179-186.

HARTMANN, P., SHERRIFF, J. & KENT, J. (1995). Maternal nutrition and the regulation of milk synthesis.Proceedings of the Nutrition Society 54, 379-389.

HEALY, D. L., RAYIIGAN, S., HARTMANN, P. E., HERINGTON, A. C. & BURGER, H. G. (1980). Prolactin inhuman milk: Correlation with lactose, total protein, and a-lactalbumin levels. American Journal ofPhysiology 238, E83-86.

HEESOM, K. J., SOUZA, P. F. A., ILIC, V. & WILLIAMSON, D. H. (1992). Chain-length dependency ofinteractions of medium-chain fatty acids with glucose metabolism in acini isolated from lactating ratmammary glands. A putative feed-back to control milk lipid synthesis from glucose. BiochemicalJourncal 281, 273-278.

HENNART, P., DELOGNE-DESNOECK, J., VIS, H. & ROBYN, C. (1981). Serum levels of prolactin and milkproduction in women during a lactation period of thirty months. Clinical Endocrinology 14, 349-353.

HUANG, G., ZHENG, D. & QIAN, Y. (1987). Serum prolactin level of lactating mothers and its relation tomilk volume. Journal of the Western China University Medical School 18, 265-268.

JENSEN, R. G., BITMAN, J., CARLSON, S. E., COUCH, S. C., HAMOSH, M. & NEWBURG, D. S. (1995). Milklipids; A. Human milk lipids. In Handbook of Milk Composition, ed. JENSEN, R. G., pp. 495-542.Academic Press, San Diego.

KULSKI, J. K., HARTMANN, P. E., MARTIN, J. D. & SMITH. M. (1978). Effects of bromocriptine mesylate onthe composition of the mammary secretion in non breastfeeding women. Obstetrics and Gynecology52, 38-42.

1019




LAWRENCE, R. A. (1989). Breastfeeding; A Guide for the Medical Profession, 3rd edn. Mosby-YearBook, St Louis, MO, USA.

MOBBS, E. J. (1990). Suckling and milk production. Medical Journal ofAustralia 152, 616.NEVILLE, M. C., KELLER, R., SEACAT, J., LUTES, V., NEIFERT, M., CASEY, C., ALLEN, J. & ARCHER, P.

(1988). Studies in human lactation: Milk volumes in lactating women during the onset of lactation andfull lactation. American Journal of Clinical Nutrition 48, 1375-1386.

NICHOLAS, K. R., WILDE, C. J., BIRD, P. H., HENDRY, K. A. K., TREGENZA, K. & WARNER, B. (1995).Asynchronous concurrent secretion of milk proteins in the tammar wallaby (Macropus eugenii). InIntercellular Signalling in the Mammary Gland, ed. WILDE, C. J., PEAKER, M. & KNIGHT, C. H.,pp. 153-170. Plenum Press, New York.

NOEL, G. L., SUH, H. K. & FRANTZ, A. G. (1974). Prolactin release during nursing and breast stimulationin post partum and non-post partum subjects. Journal of Clinical Endocrinology and Metabolism 38,413-423.

NOILIN, J. M. (1979). The prolactin incorporation cycle of the milk secretory cell. An integral componentof the prolactin response cycle. Journal of Histochemistry and Cytochemistry 27, 1203-1204.

PEAKER, M. & WILDE, C. J. (1987). Milk secretion: Autocrine control. News in Physiological Sciences2, 124-126.

PHILLIPS, V. (1991). Successful Breastfeeding. Nursing Mothers' Association of Australia, Nunawading,Victoria, Australia.

PRENTICE, A., ADDEY, C. V. P. & WILDE, C. J. (1989). Evidence for local feedback control of human milksecretion. Biochemical Society Transactions 15, 122.

RATrIGAN, S., GHISALBERTI, A. V. & HARTMANN, P. E. (1981). Breast-milk production in Australianwomen. British Journal of Nutrition 45, 243-249.

RENFREW, M., FISHER, C. & ARMS, S. (1990). Bestfeeding: Getting Breastfeeding Right for You.Celestial Arts, Berkley, CA, USA.

SHARMAN, G. B. (1970). Reproductive physiology of mammals. Science 167, 1221-1228.STEINMETZ, R. W., GRANT, A. L. & MALVEN, P. V. (1993). Transcription of prolactin gene in milk

secretory cells of the rat mammary gland. Journal ofEndocrinology 136, 271-276.STERN, I. & SHAPIRO, B. (1953). A rapid and simple method for the determination of esterified fatty acidsand for total fatty acids in blood. Journal of Clinical Pathology 17, 184-187.

TYSON, J. E., KHOJANDI, M., HUTH, J. & ANDREASSEN, B. (1975). The influence of prolactin secretion onhuman lactation. Journal of Clinical Endocrinology and Metabolism 40, 764-773.

WILDE, C. J., ADDEY, C. V. P., BODDY-FINCH, L. M. & PEAKER, M. (1995). Autocrine control of milksecretion: From concept to application. In Intercellular Signalling in the Mammary Gland, ed. WILDE,C. J., PEAKER, M. & KNIGHT, C. H., pp. 227-237. Plenum Press, New York.

WILDE, C. J., CALVERT, D. T., DALY, A. & PEAKER, M. (1987). The effect of goat milk fractions onsynthesis of milk constituents by rabbit mammary explants and on milk yield in vivo. Evidence forautocrine control of milk secretion. Biochemical Journal 242, 285-288.

YUEN, B. H. (1988). Prolactin in human milk: The influence of the nursing and duration of post partumlactation. American Journal of Obstetrics and Gynecology 158, 583-586.

1 020



blood and milk prolactin and the rate of milk synthesis in women

Documents

milk prolactin

h milk production

concentration of prolactin

g24 h n

women david

summaryin women

months of lactation

constant means