Journal of Clinical InvestigationVol. 45, No. 10, 1966

Correlation of Kynurenine Excretion with Liver TryptophanPyrrolase Levels in Disease and after Hydrocortisone

Induction *

KURTALTMANt ANDOLGAGREENGARD(From the Department of Internal Medicine, New York Medical College, and the Institute for

Muscle Disease, Inc., New York, N. Y., and the Department of Biological Chemistry,Harvard Medical School, and Cancer Research Institute, NewEngland

Deaconess Hospital, Boston, Mass.)

In a large variety of apparently unrelated dis-eases, the major pathway of tryptophan metabo-lism, that which proceeds through kynurenine, hasbeen observed to be abnormal. The deviation is aquantitative one: the urinary excretion of one ormore normal metabolites is high. Attempts to ex-plain this phenomenon have largely been restrictedto hypotheses that assume a "metabolic block" inthe catabolism of these metabolites, as exemplifiedby pyridoxine deficiency. However, it seems un-likely that the variety of conditions associated withthe increased excretion of these tryptophan me-tabolites, e.g., Hodgkin's disease, rheumatoid ar-thritis, schizophrenia, porphyria, renal tuberculo-sis, and aplastic anemia, to mention only a fewthat are included in a recent review (1), would allbe coupled specifically with pyridoxine deficiency.Furthermore, the term "metabolic block" is ofdoubtful meaning in conditions not involving theinborn absence of any one enzyme. The accumula-tion of an intermediary metabolite like kynureninemust be the result of an enzymic imbalance (2):a relative increase in its rate of formation or a rel-ative decrease in its rate of degradation (of whichabsence of an enzyme is the extreme case). Thepresent study is chiefly concerned with the formerpossibility.

The oxidation of tryptophan is catalyzed bytryptophan pyrrolase, which was first characterizedin mammalian liver in 1950 by Knox and Mehler

* Submitted for publication April 4, 1966; acceptedJune 16, 1966.

This work was supported by the Arthritis Foundationand by U. S. Public Health Service grant CA 08676-01from the National Cancer Institute.

t Address requests for reprints to Dr. Kurt Altman,New York Medical College, Flower and Fifth Ave. Hos-pitals, Fifth Ave. at 106th St., New York, N. Y. 10029.

(3). It is the first enzyme on the kynureninepathway and so determines how much tryptophanenters the pathway. The regulatory role of tryp-tophan pyrrolase is also indicated by the fact thatits level in rat liver varies significantly and ap-propriately under different physiological condi-tions (4). The observed rises in activity undersuch conditions are called inductions because theyreflect increases in the immunochemically mea-surable amount of the specific enzyme protein (5).Being inducible by its specific substrate (6) and,through a different mechanism involving RNAsynthesis (7), by glucocorticoids (8, 9), it hasserved as the model of mammalian adaptive en-zymes and has drawn attention to the importantphysiological regulation of metabolism mediatedby changes in the amounts of enzymes (10, 4).The quantitative study of tryptophan pyrrolase inhuman liver appeared to be necessary for ap-proaching the causes that underlie the abnormaltryptophan metabolism in various diseases.

MethodsThe hospitalized subjects with various diseases were

all ambulatory and in no acute distress, with the indicatedexception of those who were undergoing abdominal sur-gery. They were kept on a regular hospital diet and re-ceived no medications for at least 1 week before, andduring, these studies. Unless otherwise indicated, no pa--tient was included who had been taking corticosteroids,sedatives, tranquilizers, or vitamin preparations for aprolonged period of time. The urinary excretion of kyn-urenine during the 24 hours after an oral dose of 2 gL-tryptophan was determined by the method of Tomp-sett (11).

Specimens of human liver were obtained by Menghinineedle biopsy by the standard percutaneous aspirationtechnique. When indicated, larger liver samples wereobtained by excision from surgical patients during lapa-rotomy. With these larger samples, a microassay suitable

1527

KURT ALTMANAND OLGAGREENGARD

for the estimation of tryptophan pyrrolase in the needlebiopsy samples of human liver was developed. This as-say (as explained in the Results) requires fresh mito-chondria that were obtained from rat liver at 0 to 40 Cby the following procedure. One g of rat liver was ho-mogenized in 7 ml 0.14 M potassium chloride and cen-trifuged for 10 minutes at 600 X g. The supernatant wascentrifuged for 10 minutes at 10,000 X g. The sedimentedmitochondria were washed once by resuspension in freshpotassium chloride solution and recentrifuged. The sedi-ment, suspended in 35 ml water, will be referred to as thestandard mitochondrial preparation. It was used im-mediately.

Each needle biopsy sample of human liver (approxi-mately 10 to 20 mg wet weight) was immediately placedin a tissue grinder (surrounded by ice) with 0.45 ml of0.14 M potassium chloride solution containing L-trypto-phan (6 mM) and sodium phosphate buffer (pH 7.2,0.02 M). It was homogenized by hand within 30 minutesand centrifuged in a Spinco model L ultracentrifuge, at100,000 X g for 45 minutes. (The above preparation ofrat liver mitochondria was carried out during this time.)The assay of tryptophan pyrrolase activity involvedmeasurement of the increase of absorbance at 360 mu(absorption peak of kynurenine) in microcuvettes (1 cmlight path) placed in a Beckman DU spectrophotometerequipped with thermospacers maintained at 370 C and aGilford attachment for continuous recordings. Eachcuvette contained 0.1 ml of the human liver supernatant(the particle-free, soluble fraction, or "cell sap"),L-tryptophan (3 mM), sodium phosphate buffer (pH 7.2,0.02 M), 0.05 ml of the standard mitochondrial prepara-tion, and 0.05 mg methemoglobin, in a total volume of0.5 ml. The principles underlying this method and someproperties of the enzyme will be described in theResults.

The enzyme activity was calculated from the opticaldensity change, which, after the lag period of 1 to 5 min-utes, was linear for at least 30 minutes. Each assay in-cluded a control cuvette without liver supernatant thatshowed no significant change in optical density. Theprotein content of the liver supernatant was estimatedby a biuret method (12). The activity of tryptophanpyrrolase was expressed as micromoles of kynurenineformed per hour per 100 mg soluble protein. Using thesoluble protein as the basis eliminates errors arising fromdifferent degrees of completeness of homogenizations.

Results

A system for the assay of rat liver tryptophanpyrrolase which ensures that the enzyme concen-tration alone is limiting (13) was adapted to mi-croscale by the previously described "continuousmethod" (14) in which the formation of kynuren-ine is measured directly during incubation withtryptophan. Biopsy samples of liver obtained bysurgical excision provided enough material to test

the suitability of the assay system and to investi-gate some of the properties of human tryptophanpyrrolase. As in other mammalian livers (3, 15),the tryptophan pyrrolase of human liver is localizedin the soluble, particulate-free phase (cell sap),and upon incubation with tryptophan, kynureninerather than formylkynurenine accumulates owingto the high formylase activity. In normal rat liverless than one-third of the tryptophan pyrrolase isin the active holoenzyme form (15). Full activa-tion requires the addition of its heme coenzyme inthe form of hematin or methemoglobin (15, 16),as well as small amounts of microsomes and mito-chondria (13). Figure 1 illustrates that, like ratliver cell sap, the tryptophan pyrrolase activity ofhuman liver cell sap is doubled by the addition ofmethemoglobin and further increased by mito-chondria. Steady rates of product formation areusually preceded by a short lag period. Through-out this study tryptophan pyrrolase activities werecalculated from curves such as those in Figure 1.The cofactors contributed by rat or human liver

0.075r

E

to0

0.025

7

2I

10 15 20 25MINUTES

0

FIG. 1. ILLUSTRATION OF THE MEASUREMENTOF TRYP-TOPHANPYRROLASEACTIVITY. One g of surgical excisionbiopsy sample was homogenized in 20 ml medium butotherwise treated as described in the Methods. Curvesare reproduced from the spectrophotometric records ob-tained with the Gilford automatic equipment that re-cords the optical density for each of four cuvettes, fourtimes per minute. They were transposed arbitrarily alongthe ordinate to avoid superimposition of curves. Thecontent of each of the seven cuvettes was described un-der Methods, with the following modifications: curve 1,no enzyme; curve 2, no methemoglobin or mitochondria;curve 3, no mitochondria; curves 4, 5, 6, and 7, mito-chondria diluted 1 to 8, 1 to 4, 1 to 2, and undiluted,respectively.

1528

0.050[

HUMANLIVER TRYPTOPHANPYRROLASEAND KYNURENINEEXCRETION

mitochondria are equally effective in increasing theactivity of tryptophan pyrrolase in human livercell sap. Owing to the small amount of humanbiopsy material available, the mitochondria wereroutinely prepared from rat liver. They were freeof tryptophan pyrrolase activity (13) as seen bythe lack of optical density change in cuvettes(curve 1, Figure 1) containing all the assay con-stituents including mitochondria, but no solublefraction of the liver. Such control cuvettes wererun with each sample of human liver. A cuvettecontaining the soluble liver preparation but nomitochondria or methemoglobin (curve 2, Fig-ure 1) was also included in each test. This wasdone because in rat liver after the administrationof large doses of tryptophan [but not of hydrocor-tisone (9)] and in experimental porphyria (17)a qualitative change in the enzyme occurs: it is inthe activated form and does not require the addi-tion of methemoglobin and mitochondria. None ofthe enzymes of subjects in the present studyshowed such a change, and therefore only the ac-tivities obtained with optimal supplementationwith the necessary activators will be reportedhere.

Table I shows that the tryptophan pyrrolase ac-tivity of the samples taken by needle biopsy is thesame as the activity of the excision biopsy sam-ples (1 g homogenized in 20 ml medium) fromthe same liver. The last four subjects in Table Iillustrate the reproducibility of the assay. Thetryptophan pyrrolase activities of different biopsysamples of the same individual are, in agreement.The activity (per 100 mg protein) was indepen-

TABLE I

Tryptophan pyrrolase activities of differentsamples tf human liver*

Tryptophan pyrrolase activity

ExcisionSubject Assay biopsy Needle biopsy

.umoles/hour/100 mg soluble protein4 No mhb, no mt 0.31 0.394 No mt 0.64 0.584 Complete mixture 1.18 1.085 Complete mixture 1.08 (1.0),t 1.04 (1.1)7 Complete mixture 1.32 (0.8), 1.14 (1.3)

12 Complete mixture 0.68 (0.6), 0.65 (1.2)17 Complete mixture 0.35 (1.2), 0.32 (0.55)

* Needle biopsy samples (and the additional surgical excision biopsysample from Subject 4) were assayed according to the standard pro-cedure (complete mixture) or with the indicated omissions of methemo-globin (mhb) or mitochrondria (mt). From Subjects 5, 7, 12, and 17,needle biopsy samples were obtained and assayed on two differentoccasions 7 to 10 days apart. For description of subjects see Table II.

t The milligrams of soluble protein in each assay cuvette is shown inparentheses.

dent of the amount of protein within the range of0.5 to 1.5 mgper cuvette.

Table II shows the tryptophan pyrrolase levelsand kynurenine excretions of a group of hos-pitalized subjects. Patients with rheumatoid ar-thritis formed one-third of the group, although wedid not intend to contrast subjects with this dis-ease with others or with the normal population; ourpurpose was merely to investigate a group heter-ogeneous with respect to the amount of urinarykynurenine after a tryptophan load. Rheumatoidarthritis is one of the conditions in which increasedkynurenine excretion is frequently found (18),and this is seen in some of the subjects in thepresent study. Table II indicates that irrespectiveof the diagnosis, those subjects who excrete muchkynurenine also have a high tryptophan pyrrolaseactivity in the liver. A further illustration of thiscorrelation will be seen in Figure 3.

As mentioned in the introduction, the level oftryptophan pyrrolase increases in vivo upon theadministration of tryptophan (3) or hydrocorti-sone to rats (8). Substrate induction was not at-tempted in humans, owing to the toxicity of thehigh dose (100 mg per 100 g body weight of rats)of tryptophan required. The smaller dose (2 gtryptophan) given as a test load in the presentstudies did not affect the tryptophan pyrrolaselevel. Induction of human liver tryptophan pyr-rolase by hydrocortisone, already briefly reported(19), is illustrated in Table III. Increases inthe enzyme level of two-, three-, and fourfoldwere seen in the three subjects studied. Kynuren-ine excretions were measured in the same sub-jects. Before receiving hydrocortisone, these werein the normal range (30 to 80 jumoles). Two to6 days later the same patients received hydrocorti-sone. The tryptophan load was given 5 hours later(immediately after taking a liver biopsy sample)because, judging from observations in other spe-cies, it takes at least 5 hours for the enzyme to at-tain a maximal level, and then it stays elevatedfor another 6 to 8 hours. Most of the adminis-tered tryptophan is metabolized during this pe-riod (20). Table III shows that hydrocortisoneadministration raised the excretion of kynureninein these patients to 187, 135, and 238 Fumoles, re-spectively. The values are in the range of theelevated excretions of kynurenine found in variousdisease states (21). The results in Tables II and

1529

KURT ALTMANAND OLGAGREENGARD

TABLE II

Tryptophan pyrrolase level and excretion of kynurenine in hospitalized subjects

Age

years

Sex Diagnosis

58 M Cancer of head of pancreas29 M Fracture of ankle18 M Hepatitis45 F Rheumatoid arthritis22 M Erosive gastritis23 M Fracture of humerus38 M Cirrhosis of liver33 M Pancreatic cyst49 M Rheumatoid arthritis37 F Rheumatoid arthritis66 F Rheumatoid arthritis47 M Primary amyloidosis52 M Malnutrition with pyridoxine

deficiency59 MI Gout, tophaceous53 F Rhuematoid arthritis46 F Lupus, discoid64 M Choledocholithiasis52 F Rheumatoid arthritis62 M Rheumatoid arthritis

Tryptophanpyrrolaseactivity*

jumoleslhour/

100 mgsolubleprotein

0.260.280.200.400.340.380.530.320.500.560.680.70

0.321.021.201.101.081.041.32

Urinary kynurenineafter tryptophan load

jsmoks/24 hours

30 (30)t40 (55)45 (40)61 (66, 66, 71)65 (75)71 (76)80 (81)92 (81)

111 (136)136 (141)141 (150)156 (175)

212 (247)240 (232, 247)262 (260, 265, 262)266 (262)292 (272)303 (313, 288)328 (317, 328, 313)

* The tryptophan pyrrolase activity of Subjects 1, 5, and 8 was measured in surgical excision biopsy samples and ofthe other subjects in needle biopsy samples of liver.

t Repeat measurements of kynurenine of the same patient 1 to 2 weeks before biopsy are in parentheses.

III suggest that subjects with high tryptophanpyrrolase activity form more kynurenine and ex-

crete that which accumulates because it exceedsthe normal capacity of the enzyme to catabolizekynurenine.

Figure 2 illustrates the effect of pyridoxine on

tryptophan pyrrolase level and kynurenine excre-

tion. When a nutritionally normal subject whohad a somewhat elevated tryptophan pyrrolaselevel and increased urinary kynurenine (Subject11, Figure 2) was treated with pyridoxine, the

tryptophan pyrrolase level and hence the capacityto form kynurenine remained high, but the excre-

tion of kynurenine decreased. Presumably thepyridoxal phosphate enzymes that catabolize kyn-

urenine became more active. Thus, the effective-ness of pyridoxine does not necessarily indicatethat there was an "abnormal block" but, in thiscase, a "block" relative -to increased demand.Even in healthy control subjects given a test loadof tryptophan, the excretion of kynurenine was

reported to decrease as a result of administra-tion of pyridoxine (22). A patient was alsostudied who had chronic malnutrition with pyri-doxine deficiency (Subject 13, Figure 2). As istypical for pyridoxine deficiency, this patient ex-

TABLE III

The effect of hydrocortisone on tryptophan pyrrolase and urinary kynurenine*

Before hydrocortisone After hydrocortisone

Subject 4 3 2 4 3 2Urinary kynurenine, jsmoles/

24 hours 61 45 40 187 135 238Tryptophan pyrrolase activity,

,gmoles/hour/100 mg solubleprotein 0.40 0.20 0.28 0.85 0.75 1.18

* Needle biopsy samples were obtained from Subjects 4, 3, and 2 (cf. Table II). Oral tryptophan was then given andthe 24-hour urine collection begun. Seven to 10 days later the same patients were injected intramuscularly with 250 mghydrocortisone hemisuccinate (Upjohn, Kalamazoo, Mich.). Five hours later, needle biopsy samples were taken, thetryptophan load was given, and urine collection was started.

Patient

1 B.B.2 C.M.3 G.M.4 M.F.5 N.F.6 B.C.7 F.H.8 R.O.9 J.D.

10 B.T.11 L.W.12 L.C.13 L.B.

14 M.K.15 M.B.16 R.S.17 T.S.18 M.M.19 J.M.

1530

HUMANLIVER TRYPTOPHANPYRROLASEANDKYNURENINEEXCRETION

creted large amounts of kynurenine, which re-turned to normal levels after pyridoxine treatment.The purpose of illustrating these two subjects inthe same Figure is to contrast the two kinds ofenzymic imbalance. In the latter case the im-balance is due to abnormally low levels of theenzymes that catabolize kynurenine compared tothe normal level of tryptophan pyrrolase. In theformer case the enzymes that catabolize kynurenineare normal, but their activity is low compared tothe abnormally high tryptophan pyrrolase activity.Pyridoxine removes the symptom, the increasedexcretion of kynurenine, in both cases.

Figure 3 illustrates the spectrum of tryptophanpyrrolase levels and kynurenine excretions col-lected in this study Only one point, that referringto the patient with pyridoxine deficiency describedin Figure 2, deviates significantly from the directcorrelation between these two parameters. Thenutritionally normal group as a whole, includingpatients with rheumatoid arthritis or other con-ditions, exhibits a direct correlation between tryp-tophan pyrrolase levels and kynurenine excretionthroughout the range of values. Furthermore,

FIG. 2. THE EFFECT OF PYRIDOXINE TREATMENTIN A

PYRIDOXINE-DEFICIENT AND A NUTRITIONALLY NORMALSUBJECT. The tryptophan pyrrolase level (dark bars, leftordinate) and, after the tryptophan load, the excretionof kynurenine (light bars, right ordinate), were deter-mined in two subjects, one suffering from chronic mal-nutrition with pyridoxine deficiency (Subject 13) andanother from rheumatoid arthritis (Subject 11). Thesame experiments were repeated on each subject after2 weeks of daily administration of 50 mg pyridoxine-HCL.

1.4

1.2-

< 1.0Lu(,-j 0.8

06za-

I-

cr

0.2

0

0 00.

- 0

09*-0

0

@0

0

A

50 100 150 200 250 300URINARY KYNURENINE pmoles/24 hours

350

FIG. 3. THE CORRELATIONBETWEENLIVER TRYPTOPHANPYRROLASE ACTIVITY AND KYNURENINE EXCRETION INHOSPITALIZED SUBJECTS. The values for tryptophan pyrro-lase activity (ordinate) and urinary kynurenine per 24hours (abscissa) are derived from Tables II and III.Circles represent nutritionally normal subjects (closedcircles, untreated; open circles, after hydrocortisone).Triangle represents the untreated. pyridoxine-deficientsubject (see also Figure 2).

subjects whose tryptophan metabolism has beenchanged by hydrocortisone administration also fitthis same correlation between the level of trypto-phan pyrrolase and the amount of kynurenineexcreted.

Discussion

The present study contributes two simple factsnecessary for approaching the mechanism of in-creased excretion of kynurenine in subjects notsuffering from pyridoxine deficiency. a) The levelof tryptophan pyrrolase correlates with the amountof kynurenine excreted, and b) hydrocortisonecan raise the level of liver tryptophan pyrrolaseand simultaneously raise the excretion of kynuren-ine in man. The relevance of these two obser-vations will be discussed separately.

a) It is reasonable to assume that the correla-tion between kynurenine excretion and liver tryp-tophan pyrrolase activity in man, shown in Fig-ure 3, is a causal one. There is evidence -in ratsthat the activity of tryptophan pyrrolase measuredin freshly excised livers provides a good approxi-mation of its physiological effectiveness in situ

1531

KURT ALTMANAND-OLGA GREENGARD

and that this enzyme regulates the rate of degra-dation of tryptophan at low or high levels of cir-culating tryptophan. For example, rats whoseliver tryptophan pyrrolase level has been artificiallyraised by the administration of the nonmetaboliz-able analogue, a-methyltryptophan, oxidize moretryptophan from their body stores (23). Therate of clearance from the blood oi high doses ofadministered tryptophan also correlates with thelevel of tryptophan pyrrolase as measured in liverextracts (20).

b) In intact rats, exposure to cold (24) and avariety of administered substances including as-pirin (25), chlorpromazine, and reserpine (9, 26)cause as much rise in tryptophan pyrrolase levels asdo high doses of glucocorticoids. In adrenalec-tomized rats, only glucocorticoids and tryptophanare good inducers because the other agents act bystimulating the pituitary-adrenal system (6). Asshown in the present study, the enzyme in humanliver is also inducible by hydrocortisone in vivo.Thus, any form of stress or treatment with a va-riety of drugs that act directly or indirectly tostimulate the pituitary-adrenal system may alsoincrease the level of tryptophan pyrrolase. Thepossibility that factors altering adrenocortical func-tion underlie the increased excretion of trypto-phan metabolites in a number of disparate diseasesreported in the literature must be seriously con-sidered.

The steady state concentration of kynurenineand hence its excretion can, of course, rise as a re-sult of either a relative increase in its rate of forma-tion or a relative decrease in its rate of degrada-tion. In both cases, pyridoxine can lower itsexcretion because even in nutritionally normalsubjects pyridoxal phosphate is not in excess, andby providing an excess, the appropriate enzymescan be activated to catabolize high levels of kyn-urenine (cf. Figure 2). Therefore, the so-called"correction" of the abnormally elevated kynuren-ine excretion by treatment with pyridoxine doesnot necessarily mean that the underlying enzymicdisproportion arose from a subnormal activity ofthe enzymes that degrade kynurenine. This ab-normality, as well as its correction by pyridoxine(which raises the activity of the degrading en-zymes to above normal levels), is equally con-sistent with the alternative cause, an enzymic dis-proportion due to higher than normal activity of

the first enzyme on the tryptophan-kynureninepathway.

The excretion of metabolites of kynurenine isalso governed by their steady state concentration,which depends on both the kynurenine input andthe activity of the enzymes that catabolize them.Beyond a certain input of kynurenine, the capacityof some enzymes may be exceeded, and conse-quently the relative amounts of the various me-tabolites excreted will be different from those atlower inputs of kynurenine. Thus, such differ-ences in pattern, reported to exist among differentdiseases (21), do not necessitate the assumptionof more than one "enzymic abnormality," or infact, any deviation from the normal activity of en-zymes that catabolize kynurenine. A single cause,the (more or less) elevated level of tryptophanpyrrolase, may be responsible for the large kyn-urenine excretion and also for the different relativepatterns of subsequent metabolites in differentdiseases.

When additional measurements of the relativerates of formation and removal of kynurenine be-come available, it will be possible to classify thediseases characterized by elevated excretions ofkynurenine into at least two broad categories: 1)those in which the tryptophan pyrrolase level andthe rate of production of kynurenine are normalbut the degradation of kynurenine is inhibited,and 2) those in which the tryptophan pyrrolaselevel is increased whereas the degradation of kyn-urenine remains normal. Pyridoxine deficiencybelongs to the first category and up to now wasthe only condition in which the increased excre-tion of kynurenine was rationally explained.There may be other diseases that exemplify thefirst category, but there is no clear evidence forthis yet. In view of the present results, most ofthe conditions in which there is elevation of kyn-urenine excretion probably belong to the secondcategory. Causes in addition to the hydrocortisoneeffect we have demonstrated can be visualized forthe high tryptophan pyrrolase levels in this di-verse group of diseases. The nonhormonal induc-tion of tryptophan pyrrolase in animals is regu-lated by the concentration of its specific substrateand its heme coenzyme (27), and these mecha-nisms offer ways in which this enzyme could beelevated in appropriate disease states. For ex-ample, human cases of porphyria show increased

1532

HUMANLIVER TRYPTOPHANPYRROLASEAND KYNURENINEEXCRETION

excretion of tryptophan metabolites (21), and inrats with experimental porphyria the tryptophanpyrrolase level is high as a consequence of induc-tion by the heme coenzyme that is formed in excess(17). Other possibilities are genetic variants thatmight cause the excessive production of this en-zyme. However, increased activity (due to a stress-ful component of the disease or to the therapeuticmeasures taken) would be expected to accountfrequently for the observed increases in the excre-tion of kynurenine or its metabolites. This mayrepresent the most common cause for the commonsign of increased kynurenine excretion in a groupof diseases otherwise disparate in symptomatologyand origin.

Summary

An assay system- suitable for measuring the ac-tivity of tryptophan pyrrolase in needle biopsysamples of human liver was developed. In a groupof hospitalized subjects, the tryptophan pyrrolaseactivity varied between 0.26 and 1.32 U, and in thesame patients, the excretion of kynurenine during24 hours after a 2-g oral dose of tryptophan rangedbetween 30 and 328 tnmoles. The administrationof hydrocortisone caused two- to fourfold increasesin the level of tryptophan pyrrolase and in theamount of urinary kynurenine. In all the nutri-tionally normal subjects studied, including thosegiven hydrocortisone, there was a direct correla-tion between the activity of tryptophan pyrrolaseand the amount of urinary kynurenine. The re-sults suggest that in a variety of diverse diseases(excluding pyridoxine deficiency) the high rateof excretion of kynurenine and its metabolitesmight be due to an increased rate of oxidation oftryptophan. The cause of this, the high level ofliver tryptophan pyrrolase, may frequently be aresult of increased adrenocortical secretion.

Acknowledgments

Weare grateful to Dr. G. Acs for valuable suggestions,to Naomi Mendelsohn for expert technical assistance,and to Dr. W. McKinnon of the Department of Surgery,New York Medical College, for the excision biopsysamples.

References1. Musajo, L., and C. A. Benassi. Aspects of disorders

of the kynurenine pathway of tryptophan metabo-lism in man. Advance. clin. Chem. 1964, 7, 63.

2. Knox, W. E. Adaptive enzymes in the regulation ofanimal metabolism in Physiological Adaptation,C. L. Prosser, Ed. Washington, D. C., AmericanPhysiological Society, 1958, p. 107.

3. Knox, W. E., and A. H. Mehler. The conversion oftryptophan to kynurenine in liver. I. The coupledtryptophan peroxidase-oxidase system formingformylkynurenine. J. biol. Chem. 1950, 187, 419.

4. Knox, W. E., and 0. Greengard. The regulation ofsome enzymes of nitrogen metabolism-an intro-duction to enzyme physiology in Advances in En-zyme Regulation, G. Weber, Ed. New York,Pergamon, 1964, vol. 3, p. 247.

5. Feigelson, P., and 0. Greengard. Immunochemicalevidence for increased titers of liver tryptophanpyrrolase during substrate and hormonal enzymeinduction. J. biol. Chem. 1962, 237, 3714.

6. Knox, W. E. Two mechanisms which increase invivo the liver tryptophan peroxidase activity: spe-cific enzyme adaptation and stimulation of thepituitary-adrenal system. Brit. J. exp. Path. 1951,32, 462.

7. Greengard, O., M. A. Smith, and G. Acs. Rela-tion of cortisone and synthesis of ribonucleic acidto induced and developmental enzyme formation.J. biol. Chem. 1963, 238, 1548.

8. Knox, W. E., and V. H. Auerbach. The hormonalcontrol of tryptophan peroxidase in the rat. J.biol. Chem. 1955, 214, 307.

9. Greengard, O., and P. Feigelson. A difference be-tween the modes of action of substrate and hor-monal inducers of rat liver tryptophan pyrrolase.Nature (Lond.) 1961, 190, 446.

10. Knox, W. E., V. H. Auerbach, and E. C. C. Lin.Enzymatic and metabolic adaptations in animals.Physiol. Rev. 1956, 36, 164.

11. Tompsett, S. L. The determination in urine of somemetabolites of tryptophan-kynurenine, anthra-nilic acid and 3-hydroxyanthranilic acid-and ref-erence to the presence of o-aminophenol in urine.Clin. chim. Acta 1959, 4, 411.

12. Layne, E. Spectrophotometric and turbidimetricmethods for measuring proteins in Methods in En-zymology, S. P. Colowick and N. 0. Kaplan, Eds.New York, Academic Press, 1957, vol. 3, p. 450.

13. Greengard, O., N. Mendelsohn, and G. Acs. Effectof cytoplasmic particles on tryptophan pyrrolaseactivity of rat liver. J. biol. Chem. 1966, 241, 304.

14. Greengard, O., and P. Feigelson. The purificationand properties of liver tryptophan pyrrolase. J.biol. Chem. 1962, 237, 1903.

15. Feigelson, P., and 0. Greengard. A microsomal iron-porphyrin activator of rat liver tryptophan pyrro-lase. J. biol. Chem. 1961, 236, 153.

16. Knox, W. E., M. M. Piras, and K. Tokuyama. Tryp-tophan pyrrolase of liver. I. Activation and assayin soluble extracts of rat liver. J. biol. Chem.1966, 241, 297.

17. Feigelson, P., and 0. Greengard. The activation andinduction of tryptophan pyrrolase during experi-

1533

KURTALTMANAND OLGAGREENGARD

mental porphyria and by amino-triazole. Biochim.biophys. Acta (Amst.) 1961, 52, 509.

18. Flinn, J. H., J. M. Price, N. Yess, and R. R. Brown.The excretion of tryptophan metabolites by pa-tients with rheumatoid arthritis. Arth. and Rheum.1964, 7, 201.

19. Altman, K., and 0. Greengard. Tryptophan pyrro-lase induced in human liver by hydrocortisone: ef-fect on excretion of kynurenine. Science 1966,151, 332.

20. Knox, W. E. The regulation of tryptophan pyrrolaseactivity by tryptophan in Advances in EnzymeRegulation, G. Weber, Ed. New York, Pergamon,1965, vol. 4, in press.

21. Price, J. Some effects of chelating agents on trypto-phan metabolism in man. Fed. Proc. 1961, 20, 223.

22. Crepaldi, G., and A. Parpajola. Excretion of trypto-phan metabolites in different forms of haemoblas-tosis. Clin. chim. Acta 1964, 9, 106.

23. Moran, J. F., and T. L. Sourkes. Induction oftryptophan pyrrolase by a-methyltryptophan andits metabolic significance in vivo. J. biol. Chem.1963, 238, 3006.

24. Maickel, R. P., and B. B. Brodie. Increase in ratliver tryptophan peroxidase as a response to pitui-tary-adrenal stimulation by various centrally actingdrugs. Fed. Proc. 1960, 19, 267.

25. Chiancone, F. M. Enzyme activities of the trypto-phan-nicotinic acid path in physiopathology. Ital.J. Biochem. 1964, 13, 1.

26. Kusch, T., and J. Heinrich. Die Aktivititsverander-ungen der Leber-Tryptophan-Pyrrolase und Ty-rosin-a-Ketoglutarat-Transaminase durch Chlor-promazin und Reserpin unter Berucksichtigung derFunktionen. Acta biol. med. germ. 1963, 10, 554.

27. Greengard, O., and P. Feigelson. The activation andinduction of rat liver tryptophan pyrrolase in vivoby its substrate. J. biol. Chem. 1961, 236, 158.

SPECIAL NOTICE TO SUBSCRIBERS

Post Offices will no longer forward the Journal when you move.

Please notify The Journal of Clinical Investigation, BusinessOffice, 10 Stoughton Street, Boston, Mass. 02118, at once whenyou have a change of address, and do not omit the Zip Codenumber.

1534