312 Am J Clin Nutr 1995;61:312-9. Printed in USA. © 1995 American Society for Clinical Nutrition

Brain tryptophan concentrations and serotonin synthesisremain responsive to food consumption after the ingestionof sequential meals13

Madelyn H Fernstrom and John D Fernstrom

ABSTRACT The response of brain tryptophan concentra-tion and serotonin synthesis to the ingestion of two sequentialmeals was examined in rats. Fasted rats ingested a carbohy-

drate meal followed 2 h later by a protein-containing meal andwere examined 2 or 4 h after the first meal. Other rats ingesteda protein meal first, followed by a carbohydrate meal. Whenthe carbohydrate meal was fed first, brain tryptophan concen-

trations and serotonin synthesis increased at 2 h; these changeswere reversed at 4 h if the second meal contained protein.When the protein meal was fed first, there were no changes inbrain tryptophan or serotonin at 2 h, and a second carbohydratemeal at 2 h did not raise brain tryptophan or serotonin 2 h later.Carbohydrate ingestion 3 h after a protein meal, however, didraise brain tryptophan and serotonin 2 h later. Brain tryptophan

concentrations and serotonin synthesis are thus responsive tothe sequential ingestion of protein and carbohydrate meals ifthere is a sufficient interval between meals. Am J Clin Nutr

1995;61 :312-9

KEY WORDS Tryptophan, serotonin, brain, rat, diet, di-

etary protein, plasma amino acids, large neutral amino acids

Introduction

The synthesis and release of serotonin by brain neurons is

rapidly influenced by the local tryptophan (Trp) concentration

( 1 , 2). Brain Trp concentrations, in turn, appear to reflect Trpuptake from the circulation; uptake occurs via a transportcarrier, located at the blood-brain barrier (the brain capillaryendothelial cells). The carrier is shared between Trp and sev-

eral other large neutral amino acids (LNAAs) and is competi-tive (3). Changes in the blood concentration of Trp or any of

the other LNAAs can thus alter competition at the transport

sites and thereby influence Trp uptake into brain and brain Trp

concentrations (1).

The ingestion of food is one of the most potent physiological

processes to alter the blood concentrations of Trp and its

LNAA competitors (principally leucine, isoleucine, valine, ty-

rosine, and phenylalanine). By this mechanism, food ingestion

influences serotonin synthesis in brain. It has been observedthat the ingestion of a largely carbohydrate, protein-free meal[the fat content is not important (4)] by fasting rats rapidlyraises brain Trp concentrations and stimulates serotonin syn-

thesis, whereas the consumption of a protein-containing meal

(typically 18-40% protein by wt) fails to raise brain Trp

concentrations or to stimulate serotonin synthesis, despite very

large increments in serum Trp concentrations (1). The carbo-hydrate meal is known to produce its effects on brain Trp and

serotonin via insulin secretion (5), which raises serum Trp

concentrations and lowers the serum concentrations of the

other LNAAs, thereby giving Trp a competitive advantage forbrain transport (1). The protein-containing meal fails to raisebrain Trp concentrations because its ingestion causes serum

Trp concentrations and the concentrations of its transport com-petitors to rise by proportionally similar amounts, resulting in

no net change in competition for uptake. Meal-induced changesin brain Trp concentrations thus appear to be predicted by thealterations produced by the meal in the serum concentration ofTrp relative to that of the other LNAAs, a relationship thatcan be simply expressed as a serum ratio of the Trp con-centration to the sum of the concentrations of the other

LNAAs (Trp/ILNAA); this ratio rises when carbohydrates

are ingested and fails to change after the consumption of

protein-containing meals (1).Over the past decade, the recognition that single meals can

influence serum LNAA concentrations, and thus ultimatelybrain serotonin synthesis, has led to the suggestion that thebrain may use this metabolic and neurochemical cascade tomonitor the recent history of carbohydrate and/or protein

consumption. Indeed, in some hypotheses of macronutrientappetite regulation, brain Trp concentrations and serotonin

synthesis are viewed as fluctuating from meal to meal as a

function of the protein and carbohydrate contents of themeal (6, 7). These food-induced neurochemical changes are

then said to be used by the brain to decide the macronutrient

selection at the next meal.

Appealing though such hypotheses are, they involve a majorassumption regarding food intake and serotonin synthesis,namely that brain Trp uptake and serotonin synthesis can vary

1 From the Departments of Psychiatry, Pharmacology, and Behavioral

Neuroscience, University of Pittsburgh School of Medicine.2 Supported by a grant from the National Institutes of Health (HD24730)

and an NIMH Research Scientist Award (MH00254) to JDF.

3 Address reprint requests to MH Fernstrom, Department of Psychiatry,

University of Pittsburgh School of Medicine, Western Psychiatric Institute

and Clinic, 3811 O’Hara Street, Pittsburgh, PA 15213.

Received November 29, 1993.

Accepted for publication September 8, 1994.

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

FOOD INGESTION AND BRAIN SEROTONIN 313

on a meal-to-meal basis. However, the only studies that have

examined the relationship of food ingestion to brain serotonin

synthesis have involved feeding large, single meals to rats after

an overnight fast. No attempts have been made to determine

whether the ingestion of a carbohydrate- or protein-containing

meal by a nonfasting rat would produce the same effects on

brain Trp concentrations and serotonin synthesis as occur after

their ingestion by fasting animals.

The present studies therefore examine the issue of whether

nonfasting rats that ingest a carbohydrate- or protein-contain-

ing meal show changes in the serum LNAA pattern, brain Trp

concentration, and serotonin synthesis similar to those ob-

served in fasting rats. In particular, we conducted studies in rats

fed two sequential meals after an overnight fast. The meal size

(16 kcal, 67 kJ) and intermeal interval (2 h) were high, but in

the physiological range for rats (8, 9), and were chosen to

ensure that biochemical changes would be detected if they

occurred.

Materials and methods

All experimental procedures were approved by the

University of Pittsburgh Animal Care and Use Committee.

Male Sprague-Dawley rats weighing 175-200 g (Hilltop

Laboratories, Scottdale, PA) were acclimated to our animal

quarters for 7 d before experimentation. During this time, water

and food (Purina Rodent Laboratory Chow 5001; Purina Mills,

St Louis) were provided ad libitum. The animals were exposed

to 12 h of light daily (0700-1900) and an ambient temperature

of 22 #{176}C.At 1700 on the day before each experiment, rats were

deprived of food but not water. At 0900 the next morning,

groups of seven rats were given free access to a 4-g meal (dry

wt), or remained fasting. All food was consumed within 15

mm. In some experiments, rats received only one meal while in

others they received a second 4-g meal, presented 2 or 3 h after

the first meal. All rats were decapitated 2 h after their last meal.

Thirty minutes beforehand, they received an injection of m-

hydroxybenzylhydrazine (NSD-1015), an inhibitor of aromatic

L-amino acid decarboxylase, to allow estimation of serotonin

synthesis rate (10). The brains were rapidly removed and

placed on an ice-cold glass plate. The hypothalamus and pieces

of cerebral cortex were removed from each brain (1 1) and

rapidly frozen on dry ice. Blood was collected from the cervi-

cal wound into glass tubes and allowed to clot in an ice bath.

The sera were then harvested after low-speed centrifugation at

1200 x g for 20 mm at 4 #{176}Cand aliquoted into test tubes. All

samples were stored at -70 #{176}Cuntil assayed.

The meals were prepared by using ingredients obtained from

Teklad (Madison, WI). The 0% protein (carbohydrate) diet

contained (in g/kg dry wt) 250 g sucrose, 200 g dextrose, and

460 g dextrin. The 6% protein diet (6% protein by wt, 5.9% of

total energy) contained 60 g casein, 250 g sucrose, 200 gdextrose, and 400 g dextrin. The 12% protein diet (12% by wt,

I 1.7% of total energy) contained 120 g casein, 250 g sucrose,

200 g dextrose, and 340 g dextrin. The 24% protein diet (24%

by wt, 23.5% of total energy) contained 240 g casein, 250 g

sucrose, 200 g dextrose, and 220 g dextrin. The 40% protein

diet (40% by wt, 39. 1% oftotal energy) contained 400 g casein,

250 g sucrose, 200 g dextrose, and 60 g dextrin. In addition to

these ingredients, each diet contained 50 g Mazola Oil (CPC

International, Englewood Cliffs, NJ) and 40 g agar. The agar

was mixed with 1 L water, heated, and then homogenized with

the other ingredients. After cooling, the final food had a firm

consistency. The energy density of the diets was ‘�17.2 kJ/g

dry wt (4. 1 kcal/g dry wt).

Trp concentrations in serum and brain were quantitated

fluorometrically (12-14) with a Hitachi F-2000 (HitachiInstruments, Danbury, CT) spectrophotofluorometer. The

concentrations of other LNAAs in serum (tyrosine, phenyl-

alanine, leucine, isoleucine, and valine) were quantitated by

using a Beckman 6300 amino acid analyzer, equipped for

fluorometric detection of o-phthalaldehyde derivatives of

the amino acids (15). 5-Hydroxytryptophan (5-HTP) con-

centrations in brain were measured by using HPLC and

electrochemical detection (16).

The data were analyzed statistically by using one-way analysis

of variance and the Newman-Keuls test, or Student’s t test (17).

Results

In initial studies, rats were tested by using two diets: 0% and

40% protein. Animals were fasted overnight; the next morning

at 0900 they either remained without food or received a car-

bohydrate (0% protein) meal (4 g dry wt). Ninety minutes later,

one group of fasted rats and one group of rats consuming

carbohydrates received an injection of NSD-1015 and were

killed 30 mm later (2 h after the meal was presented). Other

groups of rats fed the carbohydrate meal at 0 h received a

second meal at 2 h, consisting of either 0% (carbohydrate) or

40% protein. Another group of fasted rats remained without

food. Ninety minutes later, all rats received NSD-l0l5 and

were killed 30 mm later (4 h after the first meal was presented).

The results are presented in Figure 1 . The rats consumed (dry

wt/rat) 3.55 ± 0.1 g (61.1 ± 1.7 kJ, 14.6 ± 0.4 kcal) ofthe first

carbohydrate meal, 2.5 ± 0.45 g (42.7 ± 7.5 kJ, 10.2 ± 1.8

kcal) of the second carbohydrate meal, and 3.45 ± 0.3 g (59 ±

5 Id, 14.1 ± 1.2 kcal) ofthe 40% protein meal. Ingestion of thefirst carbohydrate meal caused the serum Trp/�LNAA to rise

�=60% over fasting values at 2 h. This increase was accompa-

nied by significant increases over fasting values in Trp con-

centrations and in 5-HTP accumulation rates in both cortex and

hypothalamus (Fig 1). After ingestion of the second meal,

presented 2 h after the first meal, the serum Trp/INAA re-

mained high at 4 h if the second meal was carbohydrates, as did

Trp concentrations and 5-HiT synthesis rates in both cortex

and hypothalamus. However, if the second meal contained 40%

protein, the serum Trp/ILNAA, as well as Trp concentrations

and 5-HTP accumulation rates in cortex and hypothalamus, all

declined to or below fasting control values (Figure 1).

The next studies were designed to determine the dietary

protein concentration necessary in the second meal to reverse

the increments in brain Trp concentrations and 5-HTP synthe-

sis produced by an initial meal of carbohydrates. In these

experiments, rats were fasted overnight and presented at 0900

the next morning with either nothing or with a portion of the

carbohydrate diet (4 g dry wt). Two hours later, groups of rats

either remained without food or received a second meal

(4 g dry wt) of either carbohydrate or 6%, 12%, 24%, or 40%

protein. As before, they received NSD-1015 90 mm after the

second meal was presented and were killed 30 mm thereafter.

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

A. . . .B

25

V0� 15

10

200

� 150

serun�TRPn;LNA:a�

In

Coilical TRP �.

CortIcal 5HTP

l� ‘!�

314 FERNSTROM AND FERNSTROM

0.40

______ ‘iid il� I�

10 � 5H1P

9 ** **

�: *

_____ � I�I!TTh�)

FIGURE 1. Effect of ingesting sequential meals of carbohydrates or

carbohydrates and 40% protein on brain tryptophan (Trp) concentrations

and 5-hydroxytryptophan (SHiP) synthesis. A, cerebral cortex; B, hypo-

thalamus. At 2 h: open bar = fasted; shaded bar carbohydrate meal. At

4 h: open bar = fasted; shaded bar two sequential carbohydrate meals;

black bar = carbohydrate meal followed by protein meal. * ,D < 0.05, ** P< 0.01 vs fasted groups at 2 h (t test). * �P < 0.05, ** P < 0.01 vs fastedvalues at 4 h (Newman-Keuls test). .1 ± SEM; n = 7/group.

In this study all animals rapidly consumed all of the food at

both meals. The ingestion of any of the meal pairs raised serum

Trp concentrations significantly over fasting values (Table 1).As in the initial experiments, the ingestion of two sequentialcarbohydrate meals caused the serum Trp/ILNAA to rise

(Table 1). Cortical and hypothalamic Trp concentrations and

TABLE 1

5-Hi? synthesis rates were also significantly elevated above

fasting control values. Also analogous to the first study, inges-tion of an initial carbohydrate meal followed by a 40% protein

meal caused the serum Trp/ILNAA, brain Trp concentrations,

and 5-Hi? synthesis to be at or below fasting control values

(Table 1). The failure of the ratio to exceed fasting values,despite the rise in serum Trp concentrations, was attributable to

the large increase in the serum concentrations of the other

LNAAS. Between these two endpoints, the serum Trp/ILNAAand each of the brain variables could be seen to fall gradually

from the highest values, obtained after two meals of carbohy-

drates, to the lowest values, obtained after ingestion of carbo-

hydrates followed by 40% protein. Of particular note, the

serum Trp/ILNAA and cortical and hypothalamic Trp and5-HTP concentrations were as high when carbohydrates were

followed by a 6% protein meal as the values obtained when two

meals of carbohydrates were ingested. At 12% protein, the

serum Trp/ILNAA and cortical Trp concentrations were sig-

nificantly elevated above fasting concentrations, but hypotha-

lamic Trp and cortical and hypothalamic 5-HTP accumulation

rates were not. At 24% protein, the serum Trp/ILNAA was not

significantly elevated over control values, because the second

meal elevated the serum concentrations of the other LNAAs by

almost as much, proportionally, as the serum Trp concentration

(Table 1). Brain Trp concentrations and 5-HTP synthesis were

also at fasting values. From these studies it is apparent that

when a protein-containing meal is consumed 2 h after a car-

bohydrate meal, the protein content of the second meal must

contain >6% protein to lower Trp concentrations and 5-HTP

synthesis rates in the brain.

Another series of studies examined the ability of a second

meal of carbohydrates to increase the serum Trp/ILNAA and

brain Trp and 5-Hi? concentrations after an initial meal con-

taming protein. As a preliminary step in this study, we evalu-ated the effects of single meals containing different amounts of

protein on the serum Trp/ILNAA, and on cortical and hypo-

thalamic concentrations of Trp and 5-I-ITP synthesis. As antic-

ipated, a single meal of carbohydrates caused all of thesevariables to rise significantly 2 h later (Table 2). Smaller but

nonetheless significant increases were also evident after inges-

Changes in tryptophan (Trp) concentrations and 5-hydroxytryptophan (5-HTP) synthesis rate in cerebral cortex and hypothalamus in rats ingesting a

carbohydrate (CHO) meal followed by a protein-containing meal’

Group Serum lipSerumLNAA

Serum Trp/�LNAA

Trp 5-HTP

Cortex Hypothalamus Cortex Hypothalamus

pjnollL p.snolIL nmol/g �i.molJg protein ng/g p.g/g protein

No food 96 ± 6 482 ± 14 0.19 ± 0.01 28 ± 2 0.34 ± 0.01 113 ± 7 4.1 ± 0.1

CHO-CHO 140 ± 62 376 ± 21 0.38 ± 0.022 36 ± 22 0.41 ± 0.032 159 ± 112 5.5 ± 0.32

CHO-�’6% protein 150 ± 52 437 ± 28 0.35 ± 0.022 36 ± 22 0.38 ± 0.01� 171 ± 112 5.8 ± 0.42

CHO-�12% protein 129 ± 62 385 ± 27 0.33 ± 0.032 35 ± 12 0.35 ± 0.01 136 ± 8 4.4 ± 0.3

CHO-�24% protein 158 ± 10� 675 ± 432 0.24 ± 0.01 28 ± 1 0.30 ± 0.0i3 110 ± 5 4.1 ± 0.2

CHO-+40% protein 159 ± 9� 976 ± 472 0.16 ± 0.01 23 ± 12 0.28 ± 0.012 95 ± 7 3.4 ± 0.2

F 10.74� 45�474 26.67� 17.90� 22.35� 13.22� 13.36�

‘ .1 ± SE. Groups of seven rats, fasted overnight, ingested at 0 h either no food or CHO (4 g dry wt). At 2 h the animals received a second meal: 4 g

dry wt of 0%, 6%, 12%, 24%, or 40% protein (except for the fasted group); 90 mm thereafter, all rats received NSD-1015 and were killed 30 mm later.

LNAA, large neutral amino acids.

2.3 Statistically significant vs no food values (Newman-Keuls test): 2 p < 0.01, -� p < 0.05.

4 Statistically significant vs no food values, P < 0.01 (ANOVA).

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

FOOD INGESTION AND BRAIN SEROTONIN 315

TABLE 2

Changes in tryptophan (Trp) concentrations and 5-hydroxytryptophan (5-HTP) synthesis rate in cerebral cortex and hypothalamus in rats ingesting

single meals containing different amounts of protein’

Group Serum TrpSerumLNAA

Serum Trp/�LNAA

Trp 5-HTP

Cortex Hypothalamus Cortex Hypothalamus

pinoliL pinoliL nmol/g junol/g protein ng/g p.g/g protein

No food 85 ± 8 603 ± 16 0.15 ± 0.01 25 ± 1 0.39 ± 0.01 124 ± 5 4.3 ± 0.2

CHO 128 ± 72 454 ± 28 0.30 ± 0.042 34 ± 22 0.54 ± 0.012 184 ± 82 6.3 ± 0.22

6% Protein 130 ± 42 462 ± 14 0.29 ± 0.012 29 ± 2 0.48 ± 0.012 149 ± 6� 5.2 ± 0.22

12% Protein 123 ± 92 578 ± 48 0.21 ± 0.02� 23 ± 2 0.42 ± 0.01 136 ± 10 4.5 ± 0.2

24% Protein 137 ± 122 728 ± 392 0.17 ± 0.01 27 ± 1 0.38 ± 0.01 133 ± 5 4.1 ± 0.1

40% Protein 170 ± 112 1223 ± 59� 0.14 ± 0.01 20 ± 12 0.35 ± O.Oi� 111 ± 5 3.9 ± 0.1

F 9.21� 49.58� 14.23� 12.37� 35.85� 14.21� 20.82�

‘ .1 ± SE. Groups of seven male rats, fasted overnight, ingested at 0 h either no food or 4 g dry wt of one of the diets indicated in the table (all 4 g was

consumed); 90 mm thereafter, all rats received NSD-1015 and were killed 30 mm later. LNAA, large neutral amino acids; CHO, carbohydrate.

2.3 Statistically significant vs no food values (Newman-Keuls test): 2 p < 0.01, �? p < 0.05.

4 Statistically significant vs no food values, P < 0.01 (ANOVA).

tion of the 6% protein meal (except for cortical Trp, which

increased but not significantly so in this experiment). Con-

sumption of the 12%, 24%, or 40% protein meals did not

elevate any of the brain variables above fasting control values,though ingestion of 12% protein did cause a small rise in theserum Trp/ILNAA. Also, the 40% protein meal caused noincrease over fasting values in the serum Trp/�LNAA despitethe large rise in serum Trp concentrations, because of thesubstantial increments in the serum concentrations of the otherLNAAs. Significant reductions in cortical and hypothalamic

concentrations of Trp also occurred (5-Hi? synthesis also

declined, but not significantly so).Other groups of fasted rats were then given an initial meal of

either carbohydrates or 6%, 12%, 24%, or 40% protein. Two

hours later a second meal was offered that consisted of carbo-hydrates only. The rats were killed 2 h after the second meal,

and had received NSD-1015 30 mm beforehand. As in theother studies, serum Trp concentrations rose after the ingestionof each of the meals (Table 3). The serum Trp/ILNAA and

cortical and hypothalamic concentrations of Tip and 5-HTP

were increased at the 4-h time point in animals consuming two

consecutive carbohydrate meals (Table 3). If the rats consumed

TABLE 3

6% protein as their first meal, the serum Trp/ILNAA and all of

the brain variables measured were almost as high after the

ingestion of the second carbohydrate meal as they were when

both meals had been carbohydrate. When the initial meal

contained 12% or 24% protein, the second (carbohydrate) meal

raised the serum Trp/ILNAA above fasting values, but cortical

and hypothalamic Trp concentrations were not significantly

increased, and 5-Hi? accumulation did not rise significantly

over fasting values (for cortex or hypothalamus). At 40%

protein, neither the serum Trp/ILNAA nor any of the brainvariables was significantly increased over fasting values (Table

3). The principal conclusion from these results is that a second

meal of carbohydrates, 2 h after an initial meal of 12-40%

protein, does not significantly increase brain Trp concentra-tions or serotonin synthesis.

A final series of studies was conducted to determine whether

a longer interval between the first and second meals would

allow a second carbohydrate meal to elevate brain Trp and

5-HTP after an initial meal of high protein content (24%

protein). An intermeal period of 3 h was selected. As a pre-

liminary step, we measured cortical and hypothalamic Trp

concentrations and 5-HTP synthesis 3 h after fasting rats re-

Changes in tryptophan (Tm) concentrations and 5-hydroxytryptophan (5-HiP) synthesis rate in cerebral cortex and hypothalamus in rats ingesting a

protein-containing meal followed by a carbohydrate (CHO) meal’

Group

Serum

Trp

Serum

LNAASerum Trp:

�LNAA

Trp 5-HTP

Cortex Hypothalamus Cortex Hypothalamus

pinoliL p.mol/L nmol/g pinolig protein ng/g pg/g protein

No food 142 ± 3 731 ± 36 0.20 ± 0.01 26 ± 1 0.35 ± 0.01 125 ± 1 4.3 ± 0.1

CHO-CHO 162 ± 82 497 ± 442 0.33 ± 0.012 33 ± 12 0.41 ± 0.0i3 181 ± 102 5.4 ± 0.22

6%-’CHO 165 ± 72 533 ± 132 0.31 ± 0.012 31 ± i� 0.40 ± 0.02� 173 ± 10�� 5.2 ± 0.12

12%-*CHO 190 ± 92 521 ± 272 0.31 ± 0.022 27 ± 1 0.34 ± 0.02 152 ± 3 5.0 ± 0.2

24%-*CHO 184 ± 72 643 ± 122 0.28 ± 0.012 28 ± 1 0.33 ± 0.01 141 ± 6 4.6 ± 0.2

40%-CHO 194 ± 62 814 ± 362 0.24 ± 0.01 25 ± 1 0.35 ± 0.01 141 ± 12 4.6 ± 0.5

F 8.86� 14.96� 11.45� 7.01� 4.00� 8.79� 4.90�

‘ .1 ± SE. Groups of seven male rats, fasted overnight, ingested at 0 h either no food or 4 g dry wt of 0%, 6%, 12%, 24%, or 40% protein. At 2 h the

animals received a second meal of CHO (4 g dry wt, except for the fasted group); 90 mm thereafter all rats received NSD-1015 and were killed 30 mm

later. LNAA, large neutral amino acids.

2.3 Statistically significant vs no food values (Newman-Keuls test): 2 p < 0.01, -� P < 0.05.

4 Statistically significant vs no food values, P < 0.01 (ANOVA).

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

316 FERNSTROM AND FERNSTROM

ceived a 4-g meal (dry wt) of carbohydrates or 24% protein

(Table 4). The results were very similar to those obtained 2 h

after the ingestion of either meal (Table 2). That is, ingestion of

the carbohydrate meal raised Trp concentrations and 5-HTP

synthesis rates in brain, whereas the consumption of the 24%

protein meal did not (Table 4). (Both meals raised serum Trp

concentrations, with the protein-containing meal producing agreater effect, as in the 2-h studies; the serum concentrations of

the other LNAAs were not measured in this experiment.) Other

groups of fasted rats were then studied by using the two-meal

paradigm: one group ingested two meals of carbohydrates (0%

protein), a second group consumed two meals of 24% protein,and a third group consumed an initial meal of 24% protein

followed by a second meal of carbohydrates (Table 5). The

second meal was offered 3 h after the initial meal. As before,

all rats received NSD-1015 30 mm before being killed (at 5 h).

The serum Trp/ILNAA, cortical and hypothalamic Trp con-

centrations, and 5-HTP synthesis rates were all significantly

higher in rats ingesting two sequential carbohydrate meals than

in animals consuming two 24% protein meals. These variables

were all significantly elevated as well in rats ingesting the

protein meal followed by the carbohydrate meal, when com-

pared with values in animals consuming two 24% protein

meals. Thus, an intermeal interval of 3 h appears to be suffi-

cient to allow a meal of carbohydrates subsequent to a meal of

24% protein to raise brain Trp concentrations and 5-HTP

synthesis to values comparable with those obtained when two

meals of carbohydrates are consumed.

Discussion

These results demonstrate that brain Trp concentrations and

the rate of serotonin synthesis [as estimated via 5-HiP accu-mulation rate (10)] respond predictably to the sequential inges-

tion of two meals differing in protein content, though with

some qualifications. An initial meal of carbohydrates (0%

protein) raised brain Trp concentrations and serotonin synthesis

2 h after its ingestion. A second meal containing moderate to

large amounts of protein, ingested 2 h after the first, generally

reversed these effects 2 h later. If the initial meal contained

protein (� 12%), brain Trp concentrations and serotonin syn-

thesis were usually unchanged at 2 h (an anticipated result).

However, a second meal containing carbohydrates (0%

protein) offered at 2 h failed to raise brain Trp concentrations

TABLE 4

or to stimulate serotonin synthesis 2 h thereafter. However, if

the second meal of carbohydrates was offered 3 h after the

protein-containing meal, Trp concentrations and serotonin syn-

thesis rates 2 h later were as high as those seen in animals

consuming two meals of carbohydrates. We conclude that brain

Trp concentrations and the serotonin synthesis rate, at least in

the present context, can respond to the presence or absence of

protein in a meal, even in nonfasted animals, but that this

continued responsivity requires a significant intermeal interval.

Some of the present findings were not unexpected; namely,

that when a protein meal follows a carbohydrate meal, the

serum Trp/ILNAA would fall and brain Trp concentrations

and 5-HTP synthesis would decline. This expectation derived

from some [though not all (18)] earlier reports showing that

protein ingestion can lower the serum Trp/ILNAA and brain

Trp (19, 20). The same effect could be predicted in rats ingest-

ing protein after a carbohydrate meal: the influx of competitors

to Trp into the blood after the protein meal would lower the

already elevated serum Trp/�LNAA (Table 1). The unantici-

pated result was that a carbohydrate meal failed to raise brain

Trp concentrations and 5-HTP synthesis when consumed 2 h

after a protein meal (12-40% by wt; Table 3). This nonrespon-

siveness to carbohydrate might have been due to the lingeringof elevated serum concentrations of Trp’s LNAA competitors

after the first protein meal (Table 2). This would make it

difficult for the second carbohydrate meal (and insulin secre-

tion) to reduce LNAA concentrations sufficiently to raise the

serum Trp/ILNAA (and thus brain Trp concentrations and

serotonin synthesis). This possibility is consistent with the

results of an experiment in which the intermeal interval was

lengthened to 3 h. In this case, the serum Trp/ILNAA, brain

Trp concentrations, and Trp hydroxylation rate were all in-

creased by a carbohydrate meal that followed a 24% protein

meal (Table 5). Perhaps the longer intermeal interval allowed

time for serum LNAA concentrations and insulin concentra-

tions to fall sufficiently toward baseline to permit a subsequent

carbohydrate meal (and insulin secretion) to produce a marked

lowering of serum LNAA concentrations, and thus an increase

in the serum Trp/�LNAA (Table 5).

Is this temporal difference in the ability of a carbohydrate

meal to raise brain Trp concentrations and serotonin synthesis

compatible with current hypotheses regarding the role of meal-induced changes in brain serotonin synthesis in carbohydrate

intake regulation? The notion that carbohydrate intake is reg-

Changes in tryptophan (Tip) concentrations and 5-hydroxytryptophan (5-HTP) synthesis rate in cerebral cortex and hypothalamus in rats 3 h after

ingesting single meals containing different amounts of protein’

GroupSerum

Trp

Trp 5-HTP

Cortex Hypothalamus Cortex Hypothalamus

pnzollL nmol/g p.znol/g protein ng/g p.g/g protein

No food 88 ± 6 20 ± 1 0.45 ± 0.02 113 ± 5 4.0 ± 0.2

CHO 116 ± 62 26 ± 12 0.57 ± 0.022 156 ± 42 5.3 ± 0.22

24% Protein 140 ± 72 19 ± 1 0.47 ± 0.02 115 ± 5 4.2 ± 0.1

F 16.14� 17.50�’ 12.27� 23.59� 20.57�

‘ .t ± SE. Groups of seven male rats, fasted overnight, ingested at 0 h either no food or 4 g dry wt of one of the diets indicated in the table (all 4 g was

consumed); 150 mm thereafter all rats received NSD-1015 and were killed 30 mm later. CHO, carbohydrate.

2 Statistically significant vs no food values, P < 0.01 (Newman-Keuls test).

3 Statistically significant vs no food values, P < 0.01 (ANOVA).

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

FOOD INGESTION AND BRAIN SEROTONIN 317

TABLES

Changes in tryptophan (Tm) concentrations and 5-hydroxytryptophan (5-HTP) synthesis rate in cerebral cortex and hypothalamus in rats ingesting a

24% protein meal or a carbohydrate (CHO) meal followed 3 h later by a second CHO or protein meal’

Group

Serum

Tip

Serum

LNAA

Serum Trp/

�LNAA

Trp 5-HTP

Cortex HypothalamusCortex Hypothalamus

j.unol/L j.unol/L nmol/g pinolig protein ng/g �g/g protein

Protein-+protein 155 ± 7 680 ± 18 0.24 ± 0.01 21 ± 2 0.39 ± 0.02 123 ± 5 3.7 ± 0.2

Protein-*CHO 157 ± 7 475 ± 252 0.34 ± 0.022 27 ± 12 0.49 ± 0.03� 169 ± 72 4�7 ± 0.22

CHO-*CHO 132 ± 33 408 ± 282 0.34 ± 0.03� 29 ± 12 0.52 ± 0.04� 186 ± 72 5.0 ± 0.12

F 5#{149}474 26.16� 5.88� 8.89� 6.10� 25.79� 22.08�

‘ I ± SE. Groups of seven male rats, fasted overnight, ingested at 0 h 4 g dry wt of either 0% (CHO) or 24% protein. At 3 h the animals received a

second 4-g meal of protein or CHO; 90 mm thereafter all rats received NSD-1015 and were killed 30 mm later. Each rat consumed all of the first and second

meals (4 g; 68.6 kJ, 16.4 kcal). LNAA, large neutral amino acids.

2.3 Statistically significant vs protein -� protein values (Newman-Keuls test): 2 p < 0.01, � p < 0.05.

4 Statistically significant vs protein -� protein values, P < 0.01 (ANOVA).

ulated, and that serotonin neurons participate in this regulation,

was assembled primarily from three sets of data: 1) that single

carbohydrate meals consumed by fasted rats would rapidly

elevate the serum Trp/ILNAA, brain Trp concentrations, and

serotonin synthesis, whereas ingestion of a protein-containing

meal would not (21); 2) that the injection of drugs that selec-

tively enhance transmission across serotonin synapses (ie, se-

rotonin agonists and re-uptake blockers) selectively reduces the

ingestion of carbohydrates by rats and humans (22-29); and 3)that rats can change their selection of carbohydrates from meal

to meal in both experimental (29, 30) and free-feeding (31, 32)

contexts. Together, these data have been synthesized into a

negative feedback model for regulating meal-to-meal appetite

for carbohydrates (30, 33). When a rat consumes carbohy-

drates, the serum Trp/ILNAA rises, causing brain Trp uptake

and serotonin synthesis to increase (21), which presumably

leads to an enhancement of neuronal serotonin release (2). The

result is a reduction in carbohydrate (and an increase in protein)

intake at the next meal. At the next meal, more protein (and less

carbohydrate) is ingested, leading to a reduction in the serum

Trp/�LNAA and a decline in brain Trp concentrations and

serotonin synthesis and release. The inhibition of carbohydrate

intake is thus relieved, and carbohydrate ingestion increases at

the succeeding meal. Thus, this feedback mechanism, as envi-

sioned, has both rats and humans proceeding to each meal,

choosing more carbohydrate or protein on the basis of whether

brain Trp concentrations and serotonin synthesis have been

raised or lowered at the preceding meal. Indeed, certain forms

of obesity in humans have been attributed to a breakdown in

this feedback loop, with obesity resulting from a deficiency of

serotonin in the brain and thus an excessive craving for

carbohydrates (23).

The present results probably detract from this hypothesis

because of the findings associated with feeding carbohydrate

meals after protein-containing meals. For the feedback loop to

work, it must be responsive to each meal. However, our data

show that a rat needs 3 h after a protein-containing meal to

exhibit an increase in brain serotonin synthesis when a subse-

quent carbohydrate meal is consumed. When the interval is 2 h,

carbohydrate ingestion does not stimulate serotonin synthesis

(Table 3). In a free-feeding context, rats consume many meals

during the night (the normal feeding period), sometimes sepa-

rated by 2 h but more often by less (8, 9, 32). Given such short

time intervals, it is difficult to envision in most cases that a

carbohydrate meal would be able to raise the serum Trp/

ILNAA when consumed after a protein meal. Such nonrespon-

siveness, if it exists, makes the model untenable. However, this

conclusion cannot be drawn definitively until macronutrient

intake patterns are followed temporally in a free-feeding con-

text, and accompanied by on-line measurements of serotonin

synthesis or release [such as might be possible using in vivo

microdialysis (34)].

The issue of the intermeal interval between protein and

carbohydrate meals is the same for humans. However, because

the rate of metabolism in rats is much greater than that in

humans (35), a minimum intermeal interval of 3 h in rats might

be substantially longer in humans. Normal-weight humans eat

on average 5-7 times/d (though some eat 10 times/d) (36, 37).

Hence, the intermeal-intersnack intervals in humans may be

too short at some times of day to allow a carbohydrate meal or

snack to raise the serum Trp/ILNAA if it follows a protein-

containing meal. In this case, the meal-to-meal feedback hy-

pothesis, as described above, could not reliably work in hu-

mans. A similar argument would apply to individuals (eg,

so-called “carbohydrate cravers”) who consume above normal

numbers of snacks each day (22). To clarify this issue further,

it would be of interest to conduct a study in which the normal

food-intake pattern of human subjects was recorded throughout

the day, in association with frequent measurements of the

serum Trp/ILNAA.

A broader question is whether the serum Trp/ILNAA is

sufficiently responsive to meals and snacks in general to pro-

duce notable changes in serotonin synthesis and attendant brain

functions. If not, then it is of little interest whether this ratio

rises after one carbohydrate meal but not after another. In both

cases there would be no effect on the brain (at least in relation

to serotonin synthesis and function). Indeed, some investigators

suggest as much. Teff et al (38) report that the modest changes

in the serum Trp/ILNAA that follow the ingestion of a car-

bohydrate or protein-containing breakfast are not sufficient to

influence cerebrospinal fluid concentrations of Trp (an index ofbrain Trp concentrations in humans). The ratio changes they

obtained are of similar magnitude to those observed by others

using similar dietary treatments (39, 40). Teff et al (38) con-elude from their studies that protein and carbohydrate meals

fail to alter human brain Trp concentrations sufficiently to

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

318 FERNSTROM AND FERNSTROM

influence serotonin synthesis. Delgado et al (41), however,observed a short-term, substantial lowering of mood in remit-

ted depressed patients after the oral administration of an amino

acid mixture that reduces brain Trp concentrations in rats (42).

Such an effect is consistent with the well-established relation-

ship between affective state and brain serotonin (ie, that en-

hancing serotonin release elevates mood, whereas reducing it

depresses mood) (43). Their amino acid treatment causes a

short-term reduction in the serum Trp/ILNAA (44), the mag-

nitude of which is similar to that observed after the ingestion of

carbohydrate or protein meals by humans (39, 40, 45). This

recognition suggests that changes in the serum Trp/ILNAA of

this size might well produce changes in brain function that

signal significant alterations in brain Trp uptake and serotonin

synthesis. Regardless of the ultimate resolution of this issue,

note that this issue remains unresolved.

Three technical points deserve comment. First, we have

focused on Trp concentrations and serotonin synthesis in the

cerebral cortex and the hypothalamus. The hypothalamus was

selected for study because of its role in appetite regulation (46).

An examination of the ability of meals to influence biochem-

ically serotonin neurons projecting into this region was thus of

great interest, particularly because these serotonin projections

are likely to be involved in food-intake regulation (47). Cere-

bral cortex was examined because it is a major recipient of

serotonin axons and nerve terminals (48), but is not known tohave a key role in food-intake regulation. Using cortex, we

could thus explore whether serotonin synthesis in nerve termi-

nals not linked to appetite control is likely to be influenced by

meal-induced changes in Trp concentrations. Clearly, the re-

sults support the view that all serotonin projections are likely to

be influenced by meals, a finding consistent with an earlier

report that carbohydrate ingestion increases Trp concentrations

and serotonin synthesis in several brain regions (49). Though it

is not presently clear why all serotonin neurons should be

susceptible to food-induced changes in transmitter synthesis,

the effects on those projecting into the hypothalamus might be

to influence local appetite circuits.

Second, the present studies have used meals of large, but

normal size (8, 9). The issue of meal size is important because,

in previous single meal studies using fasting rats, animals

consumed a high proportion of their daily energy intake as asingle, large meal. The neurochemical responses, it could thus

be argued, might not be predictive of effects obtained after a

meal of normal size. The present results suggest that the

responses may be the same in the two dietary contexts. And

third, the present results indicate that a meal containing �6%

but < 12% protein raises the serum Trp/ILNAA, brain Trp

concentrations, and serotonin synthesis (Table 2). This findingdiffers from that of Yokogoshi and Wurtman (50), who re-

ported that a meal containing 5% casein was associated with nopostingestion increase in the serum Trp/ILNAA (neither brain

Trp concentrations nor serotonin synthesis were measured in

their study). We have no simple explanation for this difference,

except to note that several days before experimentation, they

entrained their animals to a feeding schedule that included an

extended period of fasting each day (20 h). Metabolic effects

may therefore have been present during their test meal that do

not occur in animals that are fasted only once, the night before

the test meal is presented.

Finally, although the present results focus on serotonin’s role

in food-intake sensing and regulation by the brain, serotonin

neurons are but one of many neurochemically specific cells in

the brain that participate in appetite regulation. Ultimately,

information regarding the participation of serotonin neurons in

brain circuits that monitor nutrient intake and influence appe-

tite must be integrated with knowledge regarding other neuro-

chemically specific neurons that function in the same brain

circuitry before a full understanding of the brain’s role infood-intake regulation can be achieved. El

We gratefully acknowledge the expert technical assistance of Alissa

Ebaugh, Karl Kovalkovich, and Terre Constantine.

References

1. Fernstrom JD. Aromatic amino acids and monoamine synthesis in the

central nervous system: influence of the diet. J Nutr Biochem 1990;

1:508-17.

2. Sharp 1, Bramwell SR, Grahame-Smith DG. Effect of acute adminis-

tration of L-tryptophan on the release of 5-HT in rat hippocampus in

relation to serotoninergic neuronal activity: an in vivo microdialysis

study. Life Sci 1992;50:1215-23.

3. Pardridge WM, Oldendorf WH. Kinetic analysis of blood brain barrier

transport of amino acids. Biochim Biophys Acta 1975;401:128-36.

4. Fernstrom MH, Fernstrom JD. Large changes in serum free tryptophan

levels do not alter brain tryptophan levels: studies in streptozotocin-

diabetic rats. Life Sci 1993;52:907-16.

5. Crandall EA, Fernstrom JD. Acute changes in brain tryptophan and

serotonin after carbohydrate or protein ingestion by diabetic rats.

Diabetes 1980;29:460-6.

6. Wurtman Ri. Dietary treatments that affect brain neurotransmitters:

effects on calorie and nutrient intake. Ann NY Acad Sci 1987;499:

179-90.

7. Anderson GH, Li ETS, Glanville NT. Brain mechanisms and the

quantitative and qualitative aspects of food intake. Brain Res Bull

1984;12: 167-73.

8. Kissileff HR. Free feeding in normal and “recovered lateral” rats

monitored by a pellet-detecting eatometer. Physiol Behav 1970;5:

163-73.

9. LeMagnen J. Advances in studies on the physiological control andregulation of food intake. Prog Physiol Psychol 1971;4:203-61.

10. Carlsson A, Lindqvist M. Dependence of 5-HT and catecholaminesynthesis on concentrations of precursor amino acids in rat brain.

Naunyn Schmiedebergs Arch Pharmacol 1978;303: 157-64.

1 1. Glowinski J, Iverson LL. Regional studies of catecholamines in the ratbrain. I. The disposition of [3Hjnorepinephrmne, [3H]dopamine and

[3H]dopa in various regions of the brain. J Neurochem l966;13:

655-69.

12. Denckla WD, Dewey HK. Determination of tryptophan in plasma,

liver and urine. J Lab Clin Med 1967;69:160-9.13. Lehmann J. Light-A source of error in the fluorometric determination

of tryptophan. Scand J Clin Lab Invest 1971;28:49-55.

14. Bloxam DL, Warren WH. Error in the determination of tryptophan by

the method of Denckla and Dewey. A revised procedure. Anal

Biochem 1974;60:621-5.15. Fernstrom JD, Fernstrom MH, Grubb PE. Twenty-four-hour variations

in rat blood and brain levels of the aromatic and branched-chair, amino

acids: chronic effects of dietary protein content. Metabolism 1987;36:

643-50.

16. Fuller RW, Perry KW. Effects of buspirone and its metabolite,1-(2-pyrimidinyl)piperazine, on brain monoamines and their

metabolites in rats. J Pharmacol Exp Ther 1989;248:50-6.

17. Zivin JA, Bartko JJ. Statistics for disinterested scientists. Life Sci

1976;18: 15-26.

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from

FOOD INGESTION AND BRAIN SEROTONIN 319

18. Fernstrom JD, Faller DV. Neutral amino acids in the brain: changes in

response to food ingestion. J Neurochem 1978;30:1531-8.

19. Glaeser BS, Maher TJ, Wurtman Ri. Changes in brain levels of acidic,basic, and neutral amino acids after consumption of single meals

containing various proportions of protein. J Neurochem 1983;4l:

1016-21.

20. Teff KL, Young SN. Effects of carbohydrate and protein administra-

tion on rat tryptophan and 5-hydroxytryptamine: differential effects on

the brain, intestine, pineal, and pancreas. Can J Physiol Pharmacol

1988;66:683-8.

2i. Fernstrom JD, Wurtman Ri. Brain serotonin content: physiological

regulation by plasma neutral amino acids. Science 1972;178:414-6.

22. Wurtman JJ, Wurtman Ri, Mark 5, Tsay R, Gilbert W, Growdon J.d-Fenfluramine selectively suppresses carbohydrate snacking by obesesubjects. Int J Obes 1985;4:89-99.

23. Wurtman JJ, Wurtman Ri, Growdon JH, Henry P. Lipscomb A, ZeiselSH. Carbohydrate craving in obese people: suppression by treatments

affecting serotoninergic transmission. mt i Eating Disord 1981;

1:2-11.

24. Wurtman JJ, Wurtman Ri, Reynolds 5, Tsay R, Chew B. Fenfluramine

suppresses snack intake among carbohydrate cravers but not among

noncarbohydrate cravers. Int J Eating Disord 1987;6:687-99.

25. Wurtman JJ, Wurtman Ri. Drugs that enhance serotoninergic trans-mission diminish elective carbohydrate consumption by rats. Life Sci

1979;24:895-904.

26. Leibowitz SF, Weiss GF, Walsh UA, Viswanath D. Medial hypotha-

lamic serotonin: role in circadian patterns of feeding and macronutrient

selection. Brain Res 1989;503:132-40.

27. Shor-Posner G, Grinker JA, Marinescu C, Brown 0, Leibowitz SF.

Hypothalamic serotonin in the control of meal patterns and macronu-

trient selection. Brain Res Bull 1986;17:663-71.

28. Weiss GF, Rogacki N, Fueg A, et al. Effect of hypothalamic and

peripheral fluoxetine injection on natural patterns of macronutrient

intake in the rat. Psychopharmacology 1991;105:467-76.

29. Li ETh, Anderson GH. 5-Hydroxytryptamine: a modulator of food

composition but not quantity? Life Sci 1984;34:2453-60.

30. Wurtman JJ, Moses PL, Wurtman Ri. Prior carbohydrate consumption

affects the amount of carbohydrate that rats choose to eat. J Nutr

1983;l 13:70-8.

31. Johnson Di, Li ETh, Coscina DV, Anderson GH. Different diurnal

rhythms of protein and non-protein energy intake by rats. Physiol

Behav 1978;21 :777-80.

32. Tempel DL, Shor-Posner G, Dwyer D, Leibowitz SF. Nocturnal pat-

terns of macronutrient intake in freely feeding and food-deprived rats.

Am J Physiol 1989;256:R541-8.33. Wurtman JJ. Neurotransmitter control of carbohydrate consumption.

Ann NY Acad Sci 1985;443:145-51.

34. Schwartz DH, McClane 5, Hernandez L, Hoebel BG. Feeding in-

creases extracellular serotonin in the lateral hypothalamus of the rat as

measured by microdialysis. Brain Res 1989;479:349-54.

35. Munro HN. Evolution of protein metabolism in mammals. In: Munro

HN, ed. Mammalian protein metabolism. Volume 3. New York:

Academic Press, 1969:133-82.

36. Weltzin TE, Hsu LKG, Pollice C, Kaye WH. Feeding patterns in

bulimia nervosa. Biol Psychiatry 1991;30:1093-110.

37. Fernstrom MH, Weltzin TE, Neuberger 5, Srinivasagam N, Kaye

WH. Twenty-four hour food intake in patients with anorexia

nervosa and in healthy control subjects. Biol Psychiatry 1994;36:

696-702.

38. Teff KL, Young SN, Marchand L, Botenz MI. Acute effect of protein

or carbohydrate breakfasts on human cerebrospinal fluid monoamine

precursor and metabolite levels. J Neurochem 1989;52:235-41.

39. Ashley DVM, Liardon R, Leathwood PD. Breakfast meal composition

influences plasma tryptophan to large neutral amino acid ratios of

healthy lean young men. J Neural Transm 1985;63:271-83.

40. Pijl H, Koppeschaar HPF, Cohen AF, et al. Evidence for brain sero-

tonin-mediated control of carbohydrate consumption in normal weight

and obese humans. Int J Obes 1993;17:513-20.41. Delgado PL, Chamey DS, Price LH, Aghajanian GK, Landis H,

Heninger GR. Serotonin function and the mechanism of antidepressant

action. Arch Gen Psychiatry 1990;47:41 1-8.42. Gessa GL, Biggio G, Fadda F, Corsini GU, Tagliamonte A. Effect of

the oral administration of tryptophan-free amino acid mixtures on

serum tryptophan, brain tryptophan and serotonin metabolism. J

Neurochem 1974;22:869-70.

43. Meltzer HY, Lowy MT. The serotonin hypothesis of depression. In:

Meltzer HY, ed. Psychopharmacology: the third generation of

progress. New York: Raven Press, 1987:513-26.

44. Weltzin TE, Fernstrom JD, McConaha C, Kaye W. Acute tryptophan

depletion in bulimia: effects on large neutral amino acids. Biol

Psychiatry 1994;35:388-97.

45. Fernstrom JD, Wurtman RJ, Hammarstrom Wiklund B, Rand WM,

Munro HN, Davidson CS. Diurnal variations in plasma concentrationsof tryptophan, tyrosine, and other neutral amino acids: effect of dietary

protein intake. Am J Clin Nutr 1979;32:1912-22.

46. Hoebel BG, Leibowitz SF. Brain monoamines in the modulation of

self-stimulation, feeding, and body weight. In: Weiner H, Hofer MA,

Stunkard AJ, eds. Brain, behavior, and bodily disease. New York:

Raven Press, 1981:103-42.47. Leibowitz SF, Weiss GF, Shor Posner G. Hypothalamic serotonin:

pharmacological, biochemical, and behavioral analyses of its feeding-

suppressive action. Clin Neuropharmacol 1988;ll(suppl 1):S51-71.

48. Cooper JR. Bloom FE, Roth RH. The biochemical basis of neurop-

harmacology. 6th ed. New York: Oxford University Press, 1991.

49. Colmenares JL, Wurtman Ri, Fernstrom JD. Effects of ingestion of a

carbohydrate-fat meal on the levels and synthesis of 5-hydroxyindoles

in various regions of the rat central nervous system. J Neurochem

1975;25:825-9.

50. Yokogoshi H, Wurtman RI. Meal composition and plasma amino acid

ratios: effect of various proteins or carbohydrates, and of various

protein concentrations. Metabolism 1986;35:837-42.

by guest on October 25, 2015

ajcn.nutrition.orgD

ownloaded from