Plasma Amino Acids and Insulin Luels in Obesity: Response to CarbohydrateIntake and Tryptophan Supplements

Benjamin Caballero, Nicholas Finer, and Richard J. Wurtman

We assessed the plasma amino acids, glucose, and insulin responses of obese and lean control subjects to midafternooncarbohydrate snacks. After a standard 400 kcai lunch, eight lean and nine obese subjects received, at 2PM, a 30 g sucro.esnack; blood samples were obtained at hourly intervals until 6 PM.Each subject participated in four similar studies in whichthe carbohydrate snack was consumed alone or with 250, 500, or 1,000 mg of L-tryptophan (Trp), offered as a capsule. Theobese group exhibited elevated plasma levels of the branched-chain amino acids, phenylalanine and tyrosine, and the level.of these amino acids declined much less in response to carbohydrate intake than in lean controls. As a consequence, theplasma ratio of Trp:large neutral amino acids (Trp/lNAA ratio). which normally rises after carbohydrate consumption,showed virtually no change in the obese group. The plasma Trp/lNAA response of this group did not reach control value.even when carbohydrate intake was increased to 50 or 75 g. Peak plasma Trp concentrations and Trp/LNAA ratio. after250. 500, and 1.000 mg Trp doses were also significantly lower in the obese. Since brain Trp uptake is strongly correlatedwith the plasma Trp/lNAA ratio. which in turn determines the rate of brain serotonin synthesis, the blunted Trp/LNAAresponse to carbohydrate intake in the obese could contribute to alterations in the serotonin-mediated regulation of foodintake.

~ 1988 by GrunB & Stratton,lnc.

P LASMA AMINO ACID concentrations normallyexhibit marked variations throughout the day in

response to food intake and to food-induced hormonerelease'; the branched chain amino acids can be up to sixfoldhigher after a high-protein meal than after a carbohydratemeal. Studies measuring plasma amino acid levels in obesityreported increased concentrations of the branched-chainamino acids in the postabsorptive state and a blunted fall inthese amino acids in response to a glucose challenge.2.3Sinceplasma tryptophan (Trp) levels are reportedly normal oreven below normal in the obese,3 the result is an abnormallylow plasma tryptophan:large neutral amino acids (Trp/LNAA) ratio, as well as a decrease in the rise in this ratiothat normally occurs after ingestion of carbohydrates.. Sincethis ratio determines how much of the precursor amino acidTrp becomes available to brain neurons synthesizing serota-nin,5.6the alterations in plasma amino acid levels associatedwith obesity could affect serotonin-mediated brain neura-transmission.

An impairment in central serotoninergic function couldcompromise nutrient choice and the regulation of foodintake. When rats7.. or obese people9.lo are treated withserotonin agonists and offered a choice of foods of differentcarbohydrate content, they diminish the proportion of theirtotal caloric intake represented by carbohydrates, whilesustaining protein intake. It has been proposed that thisresponse could be part of a physiological feedback system, bywhich mammals regulate: macronutrient intake.II.'2 Carbo-hydrate meals, by increasing the plasma Trp/LNAA ratio (aphenomenon mediated by insulin), increase brain tryptophan

From lilt DtpaTlmtnt of Brain and Cogni/i"t SC'itncts and liltClinical Rt5tarcll Ctnltr, Massacllustlu Inslitult of TtCilnology.Cambridgt. MA.

Addrtss rtprint rtqutSU 10 Btnjamin Cabal/tro. MD. MITClinical Rt5tarch [t1lltr. 50 Amrs SI. Cambridgt. MA 02/42.

~ 1988 by Grunt & Strollon. Inc.

oo26-0495/88/3707...()O 14$03.00/0

672

uptake, and consequently, serotonin release. Serotonin actsthen physiologically as the serotoninergic agonists do, inhib-iting carbohydrate intake relative to that of -protein insubsequent meals. A decreased rise in the Trp/LNAA ratioafter eating carbohydrates could, consequently, contribute tothe specific carbohydrate hunger reported by many obesepersons characterized as carbohydrate cravers.

The present study explored the response of plasma aminoacids, insulin, and glucose to carbohydrate intake in a groupof obese subjects. We also assessed the effect ofTrp adminis-tration on the plasma amino acid levels and Trp/LNAAratios.

MATERIALS AND METHODS

Nine obese and eight lean subjects were studied. The obese group

was drawn from a larger set of patients who participated in aprevious study to assess patterns of food choice. In that study, all had

been categorized as carbohydrate cravers,' and on-line records of

food intake had shown that they preferentially consumed carbohy-drate-rich snacks in the afternoon. between 2 PM and 4 PM. Thus. in

the present study we used that time period for exploring the plasma

hormone and substrate levels after carbohydrate consumption. Bothlean and obese volunteers were healthy and free of medication, andhad normal blood cell counts, transaminases, and bilirubin values.

The study protocol was approved by the Committee on the Use of

Humans as Experimental Subjects of Massachusetts Institute ofTechnology, and written consent was obtained from all volunteers.

The study consisted of 4 one-day eJlperiments. On each of the

study days, after an overnight fast. subjects received a 195 kcal

breakfast at 8 AM. and a 400 k::al (55% carbohydrate, 15% protein,

and 30% fat) lunch at noon. At 2 PM, a 30 g-sucrose, lemon-ftavoredsnack was administered along with 0, 250. 500, or 1,000 mg of

L-tryptophan. administered by capsule. Plasma glucose. amino

acids. and insulin levels were measured at noon. and hourly between

2 PM and 6 PM. The order of each test was randomly assigned. and aI-week period elapsed between tests. LNAA levels in deproteinized

plasma samples were determined by high-performance liquid chr~

matography with postcolumn dcrivatiz.ation and fluorometric detec-

tion. Total tryptophan was measured b)' the method of Denckla and

Dewey.1J Plasma insulin was measured using a commercial RIA kit(lmmuno Nuclear Co. Stillwater, MN). This assay uses a human

M.,M>oI,sm. Vol 37. No 7IJu1y). 1988: pp 672-676

-- -_...- - - - - - . -...- u __

PLASMA AMINO ACIDS IN OBESITY

Tabla 1. Data of tha Groupi Studied

Age (yr)

Weight (kgl

Body Mass IndexObesity Index

L...Cn-81

30 ~ 4

60.8 ~ 9.9

21.3 ~ 2.3

0.98 ~ 0.1

0beMIn -91

39 ~ 8

107.4 ~ 26.08

38.7 ~ 10.781.7 ~ 0.48

Values given are the mean ~ SD.

8GrOUpSdiffer. P < .01. Student'. ttnt.

insulin antibody. and includes a preliminary PEG step to removeendogenous antibodies; its sensitivity in our laboratory is around 2#iU/mL. and the variability bet~een assays is <7%.

To determine the amount of carbohydrate that obese subjects hadto consume to attain Trp/LNAA ratios comparable to controls. theobese volunteers participated in two additional studies. Using thesame schedule as described above. they received 50 g or 75 gcarbohydrate snacks on two separate days.

Baseline plasma amino acid concentrations for lean and obesewere compared by Student's t test for unpaired samples. The plasmasubstrate response curve was evaluated .by a two-way ANOV A forunbalanced designs. with group and time as the main effects. TheCLINFO and SAS statistical packages were used.

RESULTS

The anthropometric characteristics of the two groupsstudied are summarized in Table 1. Noon levels of plasmainsulin (Fig I) were higher in the obese, but this differencedid not attain statistical significance. Noon values ofbranched-chain amino acids, tyrosine, and phenylalanineand the Trp/LNAA ratio (Table 2) were, on the other hand,significantly higher in the obese group. Two hours afterlunch (2 PM), mean branched-chain amino acid concentra-tions were similar in both groups: 56'>.5and 546.6 mmol/Lin lean and obese, respectively (F'g 2). The day-to-dayvariability of noon amino acid and insulin values was approx-imately 15%. For example, the mean values of branched-chain amino acid levels at noon in each of the four test days

;30"E-;~ 20

/) 1.\\00I .... ...

I. .. ,. ,. ,

Il/ T L ~

o Lean

10

oNoon 2PM 3PM 4

T .me 01 cloy

6

Fig 1. PI..m. insulinI.vels .ftar c.rbohy t. Intak. in laan.nd ob..a Subjects received a 400 keel lunch at 1\00<I and a 30 glucrose snack at 2 PM.V.luea .r. m.an :! SEM. level. in the obese

group et 2 PM, 3 PM. .nd 4 PM .re lignific.ntly highar th.n in leancontrols (P < .051.

673

Tabla 2. Plum. Amino Acid Levall .t Noon

L... 0beM

225.0 % 38.18135.7 ~ 23.08

70.7 % 21.3.

71.1 ~ 12.2.

60. 7 ~ 9.0.

56.0 % 12.1

0.097 % 0.012.

Valine

Leucine

Isoleucine

Tyrosine

Phenylalanine

Tryptophan

Trp/lNAA ratio

185.2 % 47.0

117.2 % 11.255.8 % 6.1

50.2 % 9.8

48.3 % 6.4

62.7 % 8.90.122 % 0.017

Values represent #imol/l, mean % SD of feu' measur_1I on

different days.

8GrOUpSdiffer, P < .05, Student's nnt.

were 358, 372, 345, and 381 Ilmol/L in lean controls and424,442,454, and 444 Ilmol/L in the obese. Likewise, theresponse curve to the 30 g carbohydrate snack was consistentin each individual in all test days, and concurrent administra-tion of tryptophan had no significant effect on plasma aminoacid levels other than tryptophan. .

The 30 g carbohydrate snack produced similar bloodglucose response curves in lean and obese. Plasma insulinconcentration was significantly higher in the obese group at 2PM,3 PM,and 4 PM(Fig I). With the 30 g carbohydrate dose,peak hormone level was 41 :1:20 and 13 :1:4 IlV /mL in obeseand lean, respectively. At the end of the tcst at 6 PM,plasmainsulin leyel was 5 :1:1 in lean and 12 :1:4 ~V /mL in obese.

Administration of 30 g carbohydrate at 2 PM caused asignificant fall in the plasma concentration of branched-chain amino acids in the control group, down to 282.1 :1:73mmol/L at 5 PM.In contrast, the plasma BCAA level of theobese group at this same time point was 422.9 :1:73 mmol/L(P < .05) (Fig 2). Peak plasma Trp/LNAA ratio wasattained one hour after carbohydrate intake in both groups.reaching 0.150 :1: 0.031 in controls and 0.108 :1: 0.017 inobese (P < .05) (Fig 3).

Response to Tryptophan Doses

Administration of 250 mg of tryptophan along with thecarbohydrate snack caused an elevation of mean plasma Trpto reach, at 4 PM,101.3 mmol/L in lean and 92.2 mmol/L inobese. Peak plasma Trp/LNAA ratio was 0.286 :1:0.078 in

200Noon 2PW 3Pu ..~

Time 01 day

.....

Fig 2. Plum. br.nchad-chain .mino acid. (v.lina + leuclna +iIoleucinel aftar ingestion of . 400 kc.llunch .t noon .nd a 30 glucrose Inack at 2 PM. Except for tha 2 PM value., all differ~nca'batwaen le.n .nd obesa ara lignificant (P < .051. V.lua. arama.n % SEM.

600, ,

I 500I r\-..

8-. Obese

i

0-0 Leon

u'" .000E"0C. 300

674

0.160

OJ5Q

OJ"OCICI

3 OJ30.....

~a:... OJ20

0.110 1 1 1

V / ,.. T ..1..,..~#.......................0.100

Noon 2PM 3"'" 4"'"T,me 01 do,

6"'"

Fig 3. Plasma Trp/LNAA ratio after ingestion of a 400 kcallunch at noon and a 30 gsucrose snack et 2 PM.Values are meen :tSEM. All differences between lean end obese ere stetisticallysignificant IP < .051.

controls and 0.159 ~ 0.032 in obese. This peak response foreach tryptophan dose is presented in Fig 4. None of thetryptophan doses used affected the branched-chain aminoacid response to carbohydrate in either group.

Peak plasma tryptophan levels after ingestion of 500 mgand 1,000 mg tryptophan were significantly higher in leanthan in obese. The plasma tryptophan response curve to thehighest tryptophan dose used (I g) is shown in Fig 5.

Response 10Carbohydrate Intake in the Obese Group

Besides the 30 g carbohydrate test, two additional carbo-hydrate tests were performed in the obese group, at levels of50 g and 75 g of intake. After ingestion of 50 g carbohydrateat 2 PM, obese subjects exhibited a fall in branched-chainamino acids from 500.5 ~ 48 to 447.7 ~ lIS mmoljL at 4PM. The TrpjLNAA ratio increased from 0.090 to 0.100.After ingestion of 75 g carbohydrate, the peak TrpjLNAAratio rose to 0.103 ~ 0.015, a level still significantlylowerthan that attained by lean people after ingestion of 30 gcarbohydrate (Fig 6).

Fig 4. Peak pluma Trp/LNAA retios efter orel tryptophenedmini.tration. Subjects recaived the emount of L-tryptophen incap.ula form. elong with a 30 gsucrose sneck. at 2 PM.two hourseftar the standerd 400 keel lunch. Values ere mean :t SEM. Alldifferences between leen end obese are statistically significantIP < .051.

CABAlLERO. FINER. AND WURTMAN

40Noon 2PI,I 3PI,I 4PI,I

Time of day

6 PI,I

Fig 5. Dose-response curve of plasma tryptophan after inges-tion of 19 of L-tryptophan administered in capsule form. elong withthe 30 gsucro.e snack at 2 PM.Values are in I£mollL, mean :1 SEM.Differences between lean and obe.e in values at 3 PM,4 PM, end 5

PM ars statistically significant IP < ,051.

The peak blood glucose level reached after the 75 g dosewas 145 ~ 13 mgjdL, one hour after ingestion.

DISCUSSION

These data show that the plasma amino acid response to a30 g carbohydrate snack differs significantly in obese andlean persons. It also shows that moderately obese subjectswith normal glucose tolerance and plasma insulin levels thatare not significantly elevated exhibit high plasma levels ofthe branched-chain amino acids and of phenylalanine andtyrosine. Our data also indicate that the normal rise in theplasma TrpjLNAA ratio caused by carbohydrate consum~tion is markedly blunted in the obese (Fig 3). The lack of asignificant increase in that ratio is a consequence of persis-tently high levels of branched-chain amino acids, even afterthe meal-induced insulin release (Fig 2). For example, afteringestion of 30 g carbohydrate the amino acid valine fell onlyby 18%, as compared with 46% in the control group. As aresult, the rise in the plasma TrpjLNAA ratio induced bycarbohydrate ingestion was significantly less in the obese.

The abnormal amino acid response to carbohydrate intakeis a manifestation of the insulin resistance associated with

obesity. Early studies reporting elevated plasma levels ofseveral amino acids in obese subjects2.'4 already suggested a

~!?~ Ol~Z..J"-Il.a:...oE..o

if

0100

O.O~O8L ,)0 ~ 7~

CorbohyOrot. 00.. (9)

Fig 6. Peak pluma Trp/LNAA ratios in lean end obese efterInge.tion of e 30 g sucrose sneck. end in obe.e efter ingestion of50 g end 76 gsucrose snacks. all edministered at 2 PMin differ8fltdeys. BL. bueline 12 PMI levels. Velue. are mean :t SEM. Alldifference. between leen end obese are statistically significantIP < .061.

06O 1o Leon

. C Obese

"§:1 o .Oz-'......II.

nnJJna O.2OE.."ir

ooo.10 2:.t o 1000

TRP oo.e m9)

2401

8-8 Obe..0-0 Lean

=i 190

&..." 140E.."ir gO

o L.onObe..

-- -

PLASMA AMINO ACIDS IN OBESITY

link between this hyperaminoacidemia and decreased insulinsensitivity. The amino acids usually elevated are valine,leucine, isoleucine. phenylalanine, and tyrosine.1.3,14.1Salthough high alanine concentrations in the postabsorptivestate have also been reported.14 In addition. abnormally lowplasma tryptophan lev'els have been described in obesesubjects with hyperinsulinemia.' In spite of the associationbetween insulin resistance and elevated amino acid levels.there is no clear correlation between plasma insulin andamino acid concentrations;,.16 and one study found normalvaline disposal in obese persons with impaired glucose dispos-al." The relationship between insulin resistance and aminoacid metabolism is complex, and is affected by factors suchas duration and severity of obesity and lean body mass."Furthermore. it is possible that carbohydrate cravers consti-tute a subgroup of obese persons with higher branched-chainamino acid: insulin ratios than the general obese population.and that this is the basis for their impaired appetite behav-ior.

Although the plasma insulin levels in our obese group werenot significantly different from controls (mainly because ofthe large variability). mean values were 60% higher in theobese. This difference in insulin concentration is similar to

that reported by Forlani et al. IS The mean noon plasmainsulin level of our obese group (15 ~U /mL) is also identicalto the mean fasting value reported by Glass et al in 17obese;' and very similar to that of Pennetti et al in 74 obesewomen and 42 obese men.16

Our study found significant elevations in all the branched-chain amino acids, phenylalanine, and tyrosine at noon(Table 2). Plasma tryptophan levels tended to be lower in theobese. but the difference did not reach the statistical signifi.cance reported by Ashley et al.' These investigators foundthat the only amino acid significantly elevated in nondiabeticobe5e was tyrosine. Other neutral amino acids were elevatedonly in diabetic obese, who had impaired glucose tolerance ormarked hypoinsulinemia or hyperinsulinemia. The authorsalso reported no differences between lean and nondiabeticobese in the 120-minute amino acid response to a 100 g oralglucose challenge. The contrast with findings presented herecould be due to differences in the population studied, butthey may also suggest tt,at the plasma amino acid response tocarbohydrate intake is different in the morning, after ano.ernight fast, than in midafternoon. Studies in normalindividuals indicate that the time of day affects the plasmaamino acid response to meal composition.I9.20We chose thetime of our study based on the pattern of snack consumptionobserved in this population of obese persons, to explore thepossible biochemical correlates of their craving for carbohy-drate frods.

Several lines of evidence support the role of central

675

serotoninergic transmission in the regulation of appetite andfood choice. The serotonin reuptake inhibitor fenfluramineselectively decreases carbohydrate intake in obese persons;oas does the serotonin precursor tryptophan in some obeseindividuals.' A feedback mechanism between diet composi-tion and the serotoninergic neurons, mediated by the plasmaTrp/LNAA ratio, has been proposed. In this model. high-carbohydrate diets, by causing a marked rise in the plasmaTrp/LNAA ratio. would increase brain tryptophan uptakeand serotonin synthesis. which in turn acts by inhibitingspecifichunger for carbohydrates.II Since the plasma Trp/LNAA ratio is correlated with brain tryptophan concentra-tion and serotonin synthesis in animals6.11and presumably inhumans. the blunted Trp/LNAA response to carbohydrateingestion may contribute to the decreased central serotonin-ergic transmission and carbohydrate craving described insome groups of obese persons. Our data showed that. at thetime of day in which they tended to consume more starchysnacks. obese subjects were unable to raise their Trp/LNAAratio to control levels. even when consuming twice as muchcarbohydrate.

Increases in brain tyrosine leveJs have been shown toenhance the synthesis of catecholamines in and the releasefrom physiologically active neurons.22 In contrast. increasesin brain phenylalanine levels can have the opposite effect.ZJHence. the possibility should be considered that the eleva-tions in plasma levels of these two amino acids, as found inour study, might contribute to the behavioral and regulatorydisturbances associated with obesity.

An unexpected finding in our study was the bluntedresponse to oral tryptophan administration in the obese. Notonly were peak plamsa Trp/LNAA ratios significantly lowerthan in controls at all dose levels, but the absolute peakplasma tryptophan concentrations and the areas under thecurve were also lower in the obese, suggesting that theplasma appearance rate of tryptophan after oral administra-tion is different than in lean subjects. This differencepersisted even when the tryptophan dose was corrected perbody weight. Although we measured plasma levels for onlyfour hours after administration, it seems unlikely that thedifferent tryptophan response curve could be explained onlyby differences in gastric emptying and rate of absorption.since the plasma glucose response curves were very similar inlean and obese. Other possible differences may be related tothe rate of first-pass removal of the amino acids by the liver.and in the subsequent oxidation in this organ. In any case.these results indicate that obese subjects may require largerdoses of tryptophan in order to reach plasma levels compara-ble to lean controls. This different response should be consid-ered in studies assessing the effects of tryptophan supple-ments on food intake or behavior in obese.

REFERENCES

J. Wurtman RJ. RoseCM. Chou C. et al: Dailyrhythms in the 3. Ashley DVM, Fleury MO. Golay A. et al: Evidence forconcentrationor variousamino acids in human plasma. N Engl J diminished brain 5-hydroxytryptaminebiosynthesisin obese dia.Med 279:171-175.1968 betic and non-diabetichumans. Am J Clin Nutr 42:1240-1245.

2. Felig P. Marliss E. Cahill GF: Plasma aminoacid levelsand 1985insulinsecretionin obesity.N EnglJ Med 281:811-816.1969 4. Martin-DuPan R. Mauron C. Glaeser B. et al: Effect of

676

various oral glucose doses on plasma neutral amino acid levels.Metabolism 31:937.943,1982

5. Fernstrom JD, Wurlman RJ: Brain serotonin content: Physie>-

logical regulation by plasma neutral amino acids. Science 178:414.416.1972

6. Wurtman RJ, Hefti F. Melamed E: Precursor control of

neurotransmitter synthesis. Pharmacol Rev 3:315.345. 19807. Wurtman JJ. Wurtman RJ: FenRuramine and Ruoxetine

spare protein consumption.. hile suppressing caloric intake by rats.Science 198:1178.1180. 1977

8. Wurlman JJ, Wurlman RJ: Drugs that enhance central sere>-

toninergic transmission diminish elective carbohydrate consumption

by rats Life Sci 24.895.904. 1979

9. WUrlman JJ. Wurlman RJ, Growdon JH, et al: Carbohydrate

craving in obese people Suppression by treatments affecting sere>-toninergic neurotransmission Int J Eating Disord 1:2-15. 1981

10 Wurtman JJ, Wurtman RJ' D.fenfluramine selectively

decreases carbohydrate but not protein intake in obese subjects. Int JObes 8:79.84,1984

II. Wlirlman JJ' The invohement of brain serotonin in excessive

carbohydrate snacking b) obese carbohydrate cravers. J Am DietAssoc 84: 1004.1 007, 1984

] 2 Wurtman RJ, Wurlman JJ Carbohydratc craving. obesityand brain serotonin. Appetite 7:99.]03.1986

13. Denckla WD. Dewey-HK The determination of tryptophanin plasma.li\er and unne. J Lab Clin \fed 69:160-169.1967

]4. Felig P. Wahren J. Hendler R. et al Splanchnic glucose andamino acid metabolism in obesity. J Clin Invest 53.582.590. 1974

CABALLERO. FINER. ANO WURTMAN

15. Forlani G. Vannini p, Marchesini G. et al: Insulin-dependentmetabolism of branched.chain amino acids in obesity. Metabolism33:147-150.1984

16. Penneti V. Galante A. Zonta-SgarameJlat. et al: Relationbet..een obesity. insulinemia. and serum amino acid concentration in

a sample of Italian adults. Clin Chem 28:2219-2224, 1982

17. Glass AR, Bongiovanni R, Smith CEo et al: Normal valine

disposal in obese subjects with impaired glucose disposal: Evidencefor selective insulin resistance. Metabolism 30:578-582. 1981

] 8. Caballero B. Insulin resistance and amino acid metabolism in

obesity. Ann NY Acad Sci 499:84-93.1987

19. Ashley' DVM, Coscina D\'. Anderson GR Selective decrease

in protein intake follo,,'ing brain serotonin depletion. life Sci24:973-984.1979

20. Lieberman H. Caballero B. Finer N: The compositionoflunch determines afternoon plasma tryptophan ratios in humans. JI'eural Transm 65:2] ).217.1986

21. Fernstrom JD. Faller DV. l'eutral amino acids in the brain:

Changes in response to food ingestion J Neurochem 30:1531-1538.1978

22. Milner JD. Wurtman RJ: Tyrosine availability determinesstimulus-evoked dopamine release from rat striatal slices. NeurosciLett 59:215-220.1985

23 During MJ. Acworth IN. Wurtman RJ. An in 'ivo study ofdopi1mine release in striatum. The effects of phenylalanine. inWurtman RJ. Ritter-Walter E (eds) Dietar) Phenylalanine andBrain Function Boslon. Blrkhauser (in press)