Vitamin B6 Metabolism in Chronic Alcohol Abuse

PYRIDOXALPHOSPHATELEVELS IN PLASMAANDTHE EFFECTS

OF ACETALDEHYDEONPYRIDOXALPHOSPHATESYNTHESIS AND

DEGRADATIONIN HUMANERYTHROCYTES

LAWRENCELUMENGand TING-KAI Li

From the Departments of Medicine and Biochemistry, Indiana UniversitySchool of Medicine and Veterans Administration Hospital,Indianapolis, Indiana 46202

A B S T R A CT The plasma pyridoxal-5'-phosphate(PLP) level of alcoholic subjects has been comparedwith that of non-alcoholic individuals in order to as-certain the incidence of abnormal vitamin Be metabolismin chronic alcohol abuse. 66 alcoholic subjects were se-lected on the basis that they did not exhibit abnormalliver function tests and hematologic findings. 35 of themhad plasma PLP concentrations less than 5 ng/ml, thelowest value encountered in 94 control subjects, indi-cating a high incidence of deranged PLP metabolism inalcoholic patients even when hepatic and hematologicabnormalities are absent. The biochemical basis for thealtered PLP metabolism in chronic alcohol abuse wasexamined. Low plasma PLP levels in alcoholics werenot accompanied by decreased pyridoxal kinase andpyridoxine phosphate oxidase activities in erythro-cytes. Further studies with erythrocytes demonstratedthat the cellular content of PLP is determined not onlyby the activities of these PLP-synthesizing enzymes butalso by the activity of a phosphate-sensitive, membrane-associated, neutral phosphatase, which hydrolyzes phos-phorylated B. compounds.

Acetaldehyde, but not ethanol, impaired the net forma-tion of PLP from pyridoxal, pyridoxine, and pyridoxineplhosphate by erythrocytes. However, when the B.-phos-phate phosphatase activity was inhibited by 80 mMphosphate, this effect of acetaldehyde was abolished.By the use of broken cell preparations, it was possible

This work was presented in parts at the 57th AnnualMeeting of the Federation of American Societies for Ex-perimental Biology, Atlantic City, N. J. (1973) and the1st International Symposium on Alcohol and Aldehyde Me-tabolizing Systems, Stockholm, Sweden (1973).

Received for publication 25 Jutne 1973 and in revised formn26 October 1973.

to demonstrate directly that the action of acetaldehyde ismediated by the phosphatase, resulting in an accelera-tion of the degradation of the phosphorylated Be com-pounds in erythrocytes.

INTRODUCTIONPvridoxal-5'-phosphate (PLP)' is the biologically ac-tive, coenzymic form of the vitamin Be compounds. Al-though the importance of this cofactor in metabolism iswell established (1), relatively little is known about itsinvolvement in disease processes. Deficiency of PLP hasbeen causally implicated in several clinical disorders inthe alcoholic patient, including peripheral neuropathy(2), convulsions (3), sideroblastic anemia (4), and liverdisease (5). Of particular interest is that deprivationof vitamin B. produces fatty liver in experimental ani-mals and deficiency of this vitamin in the monkey forprolonged periods results in florid cirrhosis (6, 7). Inthe rat, chronic alcohol feeding has been shown to actsynergistically with vitamin B. deprivation to producemarked fatty infiltration of the liver (5). No humancounterparts of these experimental lesions have beenreported, but it is known that patients with severe al-coholic fatty liver and cirrhosis have decreased plasmaPLP and hepatic vitamin B. content (8, 9).

Hines and Cowan recently studied three alcoholicpatients, during and as long as 3 wk after the chronicadministration of large amounts of alcohol and foundthat alcohol consumption caused a depression of theplasma concentration of PLP and impaired the con-version of pyridoxine to PLP (4). Hines has further

1 Abbreviations used in this paper: Pi, inorganic phos-phate; PLP, pyridoxal-5'-phosphate; TEA, triethanolaminehydrochloride.

The Journal of Clinical Investigation Volume 53 March 1974@693-704 693

reported in a preliminary comiimunication that the pyri-dloxal kinase (ATP: pyridoxal 5'-phosphotransferase,EC 2.7.2.35) activities of red cells and hepatic tissueare reduced in some subjects during alcohol administra-tion, and that kinase activity returns to normal afterthe cessation of alcohol intake (10). If these observa-tions, in fact, represent a direct effect of ethanol onPLP synthesis, the incidence of deranged vitamin Bemetabolism in alcoholic patients should be high, al-though this has not been specifically determined. Thesefindings would further suggest that pyridoxal kinaseis the rate-limlaiting enzyme in the synthesis of PLP inman and a major determinant of the concentration ofPLP in plasma and tissues.

In this conmmunication, we report the prevalence ofabnormally low plasma PLP concentrations in patientsNwith chronic alcohol abuse but without clinically de-tectable hepatic and hematologic disease. The metabo-lism of vitamin B6 in normal hulnman erythrocytes hasbeen studied in order to define the optimal conditionsfor assay of the PLP synthesizing and degradative en-zvmes and to delineate their relative imlportance in theregulation of the iitracellular content of PLP. Theerythrocytic pyridoxal kiniase and pyridoxine phos-lhate oxidase (pyridoxine phosphate: oxygen oxido-

reductase, EC 1.4.3.5) activities in patients with chronicalcoholism who exhibit decreased plasnma PLP levelshave been measured. Lastly, the effects, in vitro, ofethanol aind its oxidative product, acetaldehyde, on thenet svnthesis of PLP in lhulmnan ervthrocvtes have beenexamined.

METHODSSuibjects. 66 male alcoholic patients, 18-68 years of age,

comprised the experimental group. All were in-patients ofthe Alcoholism Treatment Unit of the Indianapolis V. A.Hospital, chosen because they had normal physical examina-tions and laboratory tests (including complete blood count,bromsulfalein excretion, serum albumin, bilirubin, alkalinephosphatase, glutamic-oxaloacetic transaminase, and pro-thrombin time). 18 had transiently elevated serum glutamic-oxaloacetic transaminase activities and 9, transiently in-creased serum alkaline phosphatase activities, both of whichhad returned to normal by the time of the study. Liverbiopsies were not performed, since in no instance werethey clinically indicated. All of the subjects had long-standing histories of chronic alcohol abuse (11), but hadabstained from alcohol 1-5 days at the time of the study.94 healthy male volunteers from the medical center per-sonnel pool and individuals free of chronic and acute ill-nesses undergoing routine annual physical examinations con-stituted the control group. The dietary intakes of the con-trol subjects were considered to be adequate by history, butthose of the alcoholic patients could not be reliably assessed.Individuals taking vitamin supplements were excluded.

Preparation of plasma samples and erythrocytes. Venousblood was collected with 2 mMEDTA as anticoagulant.All specimens were protected from exposure to light inorder to avoid photolysis of PLP. Plasma was separated

within 30 min of collection and stored at 4°C if assayedimmediately or at - 10°C if assayed within 2 days ofcollection. Stored in this manner, no significant loss of PLPfrom the plasma was observed for as long as 4 days. Afterseparation from plasma and buffy coat, the erythrocyteswere washed three times with 0.15 M NaCl at 4°C bycentrifugation and resuspension. The cells were finally sus-

pended in appropriate incubation media to packed cell vol-umes of 15%. Aliquots of the suspended cells were thenadded to the final incubation mixtures.

Preparation of superutatanit antd ghost fractions fromiizwashed erythrocytes. Previously washed and packed ery-throcytes were hemolyzed in 10 vol of a 12 mMtriethanol-amine hydrochloride (TEA)-NaOH buffer (pH 7.4) at4°C and allowed to stand for 10 mim with intermittentmixing. The hemolysate was clarified by centrifugation at30,000 g for 30 min and the supernatant fraction separatedfrom the sedimented erythrocyte ghosts. For studies of theintracellular distribution of enzymes, the ghost fractionwas washed six or more times with TEA buffer until thesupernate became colorless. For other studies the red cellmembrane fraction was washed twice. Where indicated,the erythrocyte ghosts were sonicated, employing a model1000 Insonator (Savant Instruments, Inc., Hicksville,N.Y.) (20 kHz) supplied with a half-inch sonohorn. Thereference meter setting was 75 and sonication was appliedtwice in 10-s pulses at 4°C.

Antalyses. PLP in plasma samples and erythrocytic sus-

pension were assayed enzymatically with tyrosine decar-boxylase apoenzyme. The method described by Chabnerand Livingston (12) was modified as follows: (a) Com-mercially available apoenzyme isolated f rom Streptococcu(sfccalis was further purified by (NH4)2SO4 fractionation.Contaminating pyridoxal kinase activity was completelyremoved by precipitation in a 60%o saturated (NH4) 2S04solution. The apoenzyme was then concentrated by pre-cipitation with (NH4)2SO4 at 85%c saturation, and resolu-bilized by dialysis against 0.3 M sodium citrate buffer (pH6.0) containing 24% (vol/vol) glycerol and 2 mMmer-captoethanol. This partially purified preparation was stablefor as long as 6 wk at - 10°C without significant loss ofPLP-reconstitutable enzyme activity. (b) Deproteinizedplasma samples were prepared by adding 1.0-ml aliquotsto 0.10 ml of 75% (wt/vol) trichloroacetic acid. Aftercentrifugation, the trichloracetic acid was extracted bywater-saturated and peroxide-free ether, thereby obviatingthe problems of sample neutralization. (c) The assay mix-ture contained 0.15 M sodium citrate buffer (pH 6.0), 5nmM L- [1-_4C]tyrosine (sp act 20 uCi/mmol), apoenzyme,and 0-4 ng of PLP standard or unknown. Citrate pro-vided a greater buffering capacity than did acetate at pH6.0. A higher concentration of tyrosine gave higher andmore reproducible reconstituted enzyme reaction rates. \Viththese modifications, 90±8.5% of the P1-P could be recov-

ered and reproducibility of the method was +4.7% (1 SD).All assays were performed in duplicate. In the experimentsin which ethanol or acetaldehyde was employed, thesecompounds in the concentrations used did not interfere withthis assay method for PLP.

Pyridoxal was assayed fluorometrically, after separationby ion-exchange chromatography and oxidation in thepresence of cyanide (13). Hemoglobin wras determined as

cyanmethemoglobin (14).Incubation conditions. All experiments were performed

in either 10-ml or 25-ml Erlenmeyer flasks incubated ina shaking waterbath at 320 C. The reaction flasks andsamples were protected from ordinary room light by the

694 L. Lumeng and T.-K. Li

use of darkened rooms and yellow fluorescent lights inorder to avoid photolysis of PLP. The incubation flaskswere sealed by rubber caps when acetaldehyde or ethanolwas employed. Reactions were initiated by adding the ap-propriate substrates and terminated at different time inter-vals by transferring 1.0-ml aliquots into tubes which con-tained 0.10 ml of 75% trichloracetic acid. Pyridoxal kinaseactivity was determined by measuring the formation ofPLP from pyridoxal and ATP. Pyridoxine phosphate oxi-dase activity was measured by following the formation ofPLP with pyridoxine phosphate as substrate. The activityfrom the coupled kinase and oxidase reactions was deter-mined by measuring the formation of PLP with pyridoxineand ATP as substrates. The Be-phosphate phosphatase ac-tivity was determined by the formation of pyridoxal fromPLP. The optimal conditions for the assay of these en-zymes in human erythrocytes have not been reported previ-ously and hence have been specifically determined for boththe intact cells and hemolysate supernatant fractions aspart of this study.

Reagents. Pyridoxal, pyridoxine, PLP, and tyrosine de-carboxylase apoenzyme were purchased from Sigma Chem-ical Co., St. Louis, Mo. Pyridoxine phosphate was obtainedfrom Schwarz/Mann Div., Becton, r)ickinson & Co.,Orangeburg, N. Y. The concentrations of the vitamin Bfcompounds were determined by their molar absorptivity in0.1 N NaOH (15). Acetaldehyde was redistilled immedi-ately before use.

RESULTS

Plasma PLP concentrations of conWrol sutbjects andalcoholic patients. The plasma PLP levels of controlsubjects were found to be remarkably constant as a func-tion of time. Repeated measurements in two subjectsover periods of as long as 6 mo did not show coefficientsof variation beyond 10% (Fig. 1). A similar findingwas obtained in four other subjects measured at weeklyintervals for 2 mo. Meals did not alter these values sig-nificantly. However, person-to-person variation of theconcentration of plasma PLP was considerable, andthere was an age-related decline of the mean plasma

25 -

r-bD

0

20-o

15-

10-

5-

o SUBJECTT. K. L.X± 1 SD = 19.5 ± 1.5 ng/ml

0

0

& 000*0 O

0

0 *0 0 0 .00

* SUBJECTL. L.x ± 1 SD = 13.6 ± 1.2 ng/ml

OCTI NOV I DEC I JAN I FEB I MAR I1971 1972

FIGURE 1 Serial measurements of the plasma PLP con-centration of two control subjects.

-4

P4

Pi

30-

25

20-

15-

10-

5-

o ALCOHOLIC

0I

CONTROL

0* 0

o00 *

0 0

0~ ~~~0

0

80 8 o

0 0

I I I I I I I10 20 30 40 50 60 70

AGE, YEARS

FIGURE 2 Plasma PLP concentration of alcoholic and non-

alcoholic males, plotted as a function of age. The upperline is the regression line of the control group and thelower, of the alcoholic group.

PLP concentration (Fig. 2). In the 94 male controlsubjects, plasma PLP concentrations ranged from 5 to26.3 ng/ml. This range of values for a normal popula-tion is in agreement with that reported previously (12,16). In contrast, the plasma PLP concentrations of 66alcoholic male individuals free of hematologic and liverdisorders ranged from 1.9 to 13.1 ng/ml. 35 of the alco-holic subjects exhibited values below 5 ng/ml. The slopesof the regression lines for the control and alcohol groupswere not significantly different, indicating that the age-related decline of mean plasma PLP was comparable inthe two groups. However, the age-adjusted mean plasmaPLP concentration of the alcoholic group was 4.7±0.6ng/ml below that of the control group, a highly signifi-cant difference (P < 0.001). These data clearly estab-lish that alcoholic subjects, even when free of hemato-logic or hepatic disease, very commonly have loweredlevels of plasma PLP.

Studies on the metabolismn of PLP in normal humanerythrocytes. Pyridoxal kinase, pyridoxine phosphateoxidase, and Be-phosphate phosphatase activities havebeen described in human erythrocytes (17, 18). Thefirst two enzymes catalyze the synthesis of PLP whereasthe third enzyme catalyzes the hydrolysis of phosphory-lated vitamin B. compounds. When intact erythrocyteswere incubated in media containing either 1.5 or 80 mMinorganic phosphate (Pi), they converted pyridoxal,pyridoxine, and pyridoxine phosphate to PLP. How-ever, in the presence of 80 mMPi, the rates of netsynthesis of PLP from all three substrates were con-

siderably higher and more constant than those in thepresence of 1.5 mMPi (Fig. 3). These findings are in

Deranged Vitamin B6 Metabolism in Alcohol Abuse 695

16

12-

B

430 iM PL 30 jiM PN 10 jM PNP

0 30 60 90 120 0 30 60 90 120 0 30 60 90 120

MINUTES

FIGURE 3 Effect of 1.5 and 80 mMPi on the conversion of vitamin BG compounds to PLPin intact erythrocytes. The low Pi medium contained 50 mM TEA (pH 7.4), 1.5 mMsodium phosphate (pH 7.4), 10 mMKCl, 102 mMNaCl, and 5 mMglucose. The high Pimedium contained 50 mMTEA (pH 7.4), 80 mMsodium phosphate (pH 7.4), 10 mMKCl,7 mMNaCl, and 5 mMglucose. Washed erythrocytes (RBC) equivalent to 215 mg hemo-globin were used. Reactions were started by adding the B6 substrates, 30 /AM pyridoxal (PL),30 ,uM pyridoxine (PN) and 10 ,AM pyridoxine phosphate (PNP). Each point representsthe mean of duplicate determinations.

agreement witlh those reported by Anderson, Fulford- phosphatase, erythrocytes were hypotonically lysed andJones, Child, Beard, and Bateman (17). the ghosts separated by centrifugation. Whereas the

To discern the intracellular location of pyridoxal kinase and oxidase activities were located exclusively inkinase, pyridoxine phosphate oxidase, and B6-phosphate the supernatant fraction of the hemolysate, Be-phosphate

TABLE IDistribution of Pyridoxal Kinase, Pyridoxine Phosphate Oxidase, and B6

Phosphate Phosphatase Actizities in Human Erythrocytes

Preparation Kinase Oxidase Pliosphatase

pg PLP/l/g Hb pg PLP/h/g Hb Ag PLIh/g Hb

Intact erythrocytes 10.540.5 6.540.2Hemolysate supernate 8.640.3 4.2±0.2 0Sonicated ghosts 0 0 4.4±-1.0

The kinase and oxidase activities in washed erythrocytes were assayed in the80 mnMPi medium as described for Fig. 3. The reaction mixtures for assayingthe kinase and oxidase activities in hemolysate supernate and sonicated ghostscontained 10 mMTEA, 1 mMMgC92, 80 mMpotassium phosphate, andhemolysate supernate or sonicated ghosts derived from 1 ml of packed cellsin final volumes of 10 ml. For the kinase reaction, 2.5 mMATP and 30 ,AMpyridoxal were added as substrate and pH was 8.0. For the oxidase reaction,10 jAM pyridoxine phosphate was substrate and pH was 7.4. The phosphataseactivities were assayed in a reaction mixture which contained: 10 ml of 20 mMTEA buffer (pH 7.1); 2 ml of 10 mMMgCI2; 1.0 Og/0.5 ml of PLP; 3.2 ml of6 MKCI; 1.0 ml of sonicated ghosts prepared from 0.25 ml of packed erythro-cytes; and 3.3 ml of water. 1 ml of packed erythrocytes contained 280 mghemoglobin (Hb). Values are means±SE of four experiments.

696 L. Lumeng and T.-K. Li

4 -

2 -

0 .

12 -

10 -

6 -

4

2 -

0

O INTACT RBC, 80 mMPi

* SUPERNATANTFRACTION

I II I IT -l" I0 0.02 0.04 0.06 0.08 0,10 0.20

PNP, mM

0 INTACT RBC, 80 mMPi

* SUPERNATANTFRACTI[if

I I I I I0 0.02 0.04 0.06 .0.08 0.10

PN, mM

0 INTAC'

0

T RBC, 80 mMPi

O SUPERNATANT

I I I I I

0 0.02 0.04 0.06 0.08 0.10

PL, mM

FIGURE 4 Effect of substrate concentratiorconversion of pyridoxal (PL), pyridoxinedoxine phosphate (PNP) to PLP. The 80as described in Fig. 3 was used for intactbation medium for studying the phosphordoxal in hemolysate supernate contained(pH 8.0), 80 mMpotassium phosphate (IMgC92, 2.4 mM ATP and hemolysatefinal hemoglobin concentration of 1.30 g/1dium for studying the conversion of pyricto PLP was as described above exceptomitted and the final hemoglobin concent

phosphatase activity was found to be associated prin-cipally with the erythrocyte membrane (Table I). Al-though erythrocyte membranes are known to containan alkaline phosphatase of broad substrate specificity,the pH optimum for the hydrolysis of PLP was foundto be 7.1, suggesting that the reaction is catalyzed bya separate enzyme. This activity was increased bysonication and inhibited by high concentrations of Pi,

s e.g., 80 mMPi. In contrast, Pi did not influence theactivities of pyridoxal kinase and pyridoxine phosphateoxidase in the hemolysate supernate. The higher ratesof the net PLP synthesis observed in intact erythro-cytes when exposed to a high concentration of Pi canthus be explained by the inhibitory action of Pi upon

.30 0.40 0.50 the neutral Be-phosphate phosphatase activity of theerythrocyte membranes.

Studies on the optimal conditions for the assay ofpyridoxal kinase and pyridoxine phosphate oxidase inintact human erythrocytes. Since 80 mMPi was ef-fective in inhibiting Be-phosphate phosphatase, the ac-tivities of pyridoxal kinase, pyridoxine phosphate oxi-dase, and their coupled reactions could be assayed inintact red cells by incubation in the presence of 80 mMPi. It was found that all reactions in intact red cellswere subject to substrate inhibition (Fig. 4). The op-timal substrate concentrations for pyridoxal, pyridoxine

[ON phosphate, and pyridoxine were 10-30 AM.Studies on the optimal assay conditions and other

1.00 2.00 3.00 enzymatic properties of the pyridoxal kinase and pyri-doxine phosphate oxidase in hemolysate superntatantfractions. In assays of the membrane-free, supernatantfractions of hemolysates, patterns of substrate inhibitionsimilar to those for the intact cells were found for py-

gf* ridoxine and pyridoxine phosphate, but substrate in-hibition was not observed with pyridoxal as the sub-strate (Fig. 4). Several other enzymatic properties ofpyridoxal kinase and pyridoxine phosphate oxidase stud-

I ied are of interest, particularly in regard to the assayof their activities in hemolysate supernates. Zn2+ couldreplace Mg2' as the divalent cation in the kinase reac-tion. Mn2' and Co"+ were substantially less effective thatZn2` or Mg2+, whereas Ni' and Fe2' were inactive.

1.00 2.00 When added to the assay for pyridoxine phosphate oxi-dase, FMNinvariably produced a 10-30% stimulation ofthe measured activity of this enzyme, indicating that

n on the rate of this enzyme is normally not saturated by its coenzyme(PN) and pyri- in erythrocytes.mMPi medium The pH-rate profiles of the enzymes for the conver-

-ylation of pyri- sion of pyridoxal, pyridoxine phosphate, and pyridoxine10 mM TEA to PLP were also examined. The pH optima for pyri-

pH 8.0), 1 mMsupernate to a

00 ml. The me-

loxine phosphatethat ATP was

tration was 1.67

g/100 ml. The medium used for investigating the conver-sion of pyridoxine to PLP was modified so that the pHwas 7.4 and the final hemoglobin concentration was 1.30g/100 ml.

Deranged Vitamin B6 Metabolism in Alcohol Abuse 697

7 -

6-

5

4-

3-

.0

bD.-

be

aOD

1 -

0 -

12 -

10 -

.0

be

beO

a

a~

.0

bD

baaa04

Ir

2

30; AMPL 10aMPN 2M PNP

0 30 60 90 120 0 30 60 90 120 0 30 60 90 120

MINUTES

FIGURE 5 Effect of acetaldehyde (Acd) on the net formation of PLP by intact erythro-cytes (RB,C) in the presence of 1.5 mMPi. Washed cells containing 185 mg hemoglobinwere incubated in the 1.5 mMPi medium as described for Fig. 3. Where indicated, 1 mMAcd was also present. Reactions were started by the addition of B, substrates: pyridoxal(PL), pyridoxine (PN), and pyridoxine phosphate (PNP). The final volumes were 10 ml,and incubation temperature was 32°C. Each point is the mean of duplicate determinations.

doxal kinase, pyridoxine phosphate oxidase, and thecoupled reactions were 8.0, 7.4, and 8.0, respectively.

A previous report had postulate(d that pyridoxalserves as a regulator molecule of PLP metabolism inEscherichia coli since the pyridoxal kinase from thisbacterial species is inactivated by pyridoxal on a time-dependent basis (19). Hence, experiments were per-formed to examine whether a similar control mechanismobtains in human erythrocytes. The supernatant frac-tions of red cell hemolysate were preincubated withpyridoxal, in concentrations as high as 1 mMand fordurations as long as 1 h, in the presence of 0.5 mnM

TABLE II

Erythrocyte Pyridoxal Kinase, and Pyridoxine PhosphateOxidase Activities in Alcoholic and Non-

alcoholic Subjects

Numberof Plasma Oxidase Kinase

subjects PLP activity activity

ng/ml ug/h/g Hb pg/h/g Hb

Alcoholic 7 3.1-5.0 5.6-12.2 4.5-14.9Nonalcoholic 8 8.2-25.1 3.0-7.9 3.7-12.4

The procedure was as described in Fig. 3, using the 80 mMPi medium. The ranges of measured paiameters are shown.

ZnSO4 and 10 mMpotassium phosphate, pH 6.0 (19).In no instance was inactivationi of the kinase detecte(dwhen it was assayed after the addition of ATP.

Erythrocytic pyridoxal kinase and pyridoxinc phos-phate oxidase activities of alcoholic subjects with lo-Wplasmjta PLP concenttrations. The pyridoxal kinase anidpyridoxine phosphate oxidase activities of intact eryth-rocytes from seven alcoholic subjects with abnormallylow plasma PLP concentrations were compared withthose from eight individuals from the control group(Table II). No significant differences were observedfor either enzyme. Moreover, the total and specificactivities of the two enzymes were comparable in mag-nitude in both normal and alcoholic subjects. Theseresults are in contrast to those of Hines, reportedl in apreliminary communication, that the erythrocyte pyri-doxal kinase activity was decreased in alcoholics dur-ing and up to 2 wk after alcohol administration (20).

Effect of ethanol and acetaldehyde on the net synthe-sis of PLP by intact hunan erythrocytes. The effectsof ethanol and acetaldehyde on the net formation ofPLP by human red cells were studied in the presenceof either 1.5 mMPi, a condition in which the activityof the Be-phosphate phosphatase was preserved, or 80mMPi, where the activity of the erythrocytic phos-phatase was inhibited. Ethanol, in concentrations ashigh as 70 mM, did not alter the net synthesis of PLP

698 L. Lumeng and T.-K. Li

by washed erythrocytes under either condition. In con-trast, acetaldehyde (<1 mM) decreased the net con-version of pyridoxal, pyridoxine, and pyridoxine phos-phate to PLP by erythrocytes incubated in 1.5 mMPi(Fig. 5). After 2 h of incubation with 1 mMacetalde-hyde, the amount of PLP synthesized from pyridoxineand pyridoxine phosphate was 30-50% less than thatof the controls. Impairment of net PLP synthesis wasless pronounced when pyridoxal was the substrate.The effect of varying the concentration of acetaldehydeon the net synthesis of PLP from pyridoxine is shownin Fig. 6. Significant impairment was evident even atconcentrations as low as 0.05 mM. However, this ef-fect of acetaldehyde was completely abolished when redcells were incubated in 80 mMPi (Fig. 7). These re-sults therefore imply that the action of acetaldehyde re-quires the presence of an active Be-phosphate phospha-tase.

Effect of acetaldehyde on PLP synthesis by the su-pernatant fraction of erythrocytic hemolysate. Sincethe hemolysate supernate of erythrocytes contains onlythe pyridoxal kinase and pyridoxine phosphate oxidaseactivities but no B6-phosphate phosphatase activity, theeffect of acetaldehyde on PLP synthesis by this fractionwas examined. Acetaldehyde, 1 mM, did not affect theconversion of pyridoxal and pyridoxine phosphate toPLP by the hemolysate supernate (Fig. 8), thus fur-ther implicating the Be-phosphate phosphatase as the

80 mM

28

24

20

X16-ba

P4 12

",:, +0 g/S +1.00 mMAcd

-4

10l,M PN, 1.5 mMPi2 - INTACT RBC

0 1

0 30 60 90 120

MINUTES

FIGuRE 6 Effect of acetaldehyde (Acd) concentration onthe net synthesis of PLP from pyridoxine (PN) by intactred cells (R)3C) in media containing 1.5 mMPi. Whereindicated, 0.05, 0.10, or 1.00 mMAcd was added. WashedRBC containing 225 mg hemoglobin were used. Conditionswere otherwise identical to Fig. 5. The values are meansof duplicate assays.

mediator of the action of acetaldehyde on net PLP syn-thesis by intact erythrocytes.

0 30 60 90 120 0 30 60 90 120 0 30 60 90 120

MINUTES

FIGURE 7 Net synthesis of PLP by intact erythrocytes (RBC) incubated in the 80 mMPimedium, plus or minus 1 mMacetaldehyde (Acd). The 80 mMPi medium was as describedin Fig. 3. Washed RBC containing 188 mg hemoglobin were added. Reactions were initiatedby addition of either pyridoxal (PL), pyridoxine (PN) or ,pyridoxine phosphate (PNP).Final volume, 10 ml; temperature, 32°C. Each value is the average of duplicate assays.

Deranged Vitamin B6 Metabolism in Alcohol Abuse 699

2.5

2.0

bD)=L

A4

1.5

1.0

0.5 -

0-

0 15 30 45

MINUTES

FIGURE 8 Effect of acetaldehyde (Acd) on the net forma-tion of PLP by the supernatant f raction of hemolyzedhuman erythrocytes. Pyridoxal kinase activity was assayedin 0.1 M TEA (pH 8.0), 2.5 mMATP, 1.0 mMMgCl2, 20,uM pyridoxal (PL), and hemolysate supernate containing170 mg hemoglobin. Pyridoxine phosphate oxidase activitywas determined in 0.1 M TEA (pH 7.4), 30 /iM pyri-doxine phosphate (PNP) and hemolysate supernate con-taining 132 mg hemoglobin. Where indicated, 1 mMAcdwas employed. Final volume, 10 ml; incubation tempera-ture, 320C.

Effect of acetaldehyde on PLP synthesis by a recon-stituted system of hemolysate supernatant and sonicatedmembrane fractions. With the knowledge of the in-tracellular locations of pyridoxal kinase, pyridoxinephosphate oxidase, and the Be-phosphate phosphatase,the entire PLP-metabolizing capacity of the erythro-cyte can be restored by combining the erythrocyte mem-branes with the hemolysate supernate. Incubation ofpyridoxine phosphate with the supernatant fraction ineither 1.5 or 80 mMPi, (Fig. 9 A) or with the com-bined mixture of supernate and sonicated ghosts in 80mMPi (Fig. 9 B) resulted in the same amount of PLPsynthesized and the reactions were linear as a functionof time. These results were expected since the super-natant fraction alone did nlot contain the Be-phosphatephosphatase while phosphatase activity in the sonicatedghosts was inhibited by high concentrations of Pi.Tn the presence of 80 mMPi, the addition of acetalde-lhyde (4 mM), with and without preincubation, resultedin little or no inhibitory effect on the net production ofPLP by the recombined system of supernate plus soni-cated ghosts (Fig. 9 B). In 1.5 mMPi, the Be-phos-phate phosphatase activity of the sonicated ghosts wasnot inhibited and as a result, the net formation of PLP

from pyridoxine phosphate was drastically diminished(Fig. 9 C). Under these conditions, acetaldehyde, es-pecially when preceded by preincubation, further cur-tailed the net formation of PLP. The action of acetalde-hyde is thus apparent as an enhancement of the ac-tivity of the B-phosphate phosphatase.

DISCUSSIONPyridoxal phosphate constitutes a substantial portionof the total vitamin Bo compounds present in plasmaand 'tissues and its concentration in plasma is consid-ered to be a reliable indicator of the state of vitamin Benutrition (21). It correlates directly with the PLP con-tent of erythrocytes and appropriately with several func-tional criteria of the adequacy of vitamin Be nutrition,such as tryptophan load testing and the PLP-saturabilitvof erythrocyte transaminases (22, 23). The relativeconstancy of plasma PLP values observed in this studyin the six control subjects for durations as long as 6mo further confirms the usefulness of this measurementfor the assessment of vitamin B nutritional stattus in

A B C

1.5 or 80 mMl Pi 80 mMPi 1.5 mMPi2 mMPNP 2 mMPNP 2 mMPNP

6 66 / 0.8 6

0.604 44

0.4-

0.2

0 000 60 0 60 0 60

MINUTES

FIGURE 9 Effect of acetaldehyde (Acd) on the net syn-thesis of PLP of reconstituted broken cell preparations.The incubation media contained 10 mMTEA (pH 7.4),1 mMMgC12, 80 mMpotassium phosphate (pH 7.4) or1.5 mMpotassium phosphate (pH 7.4) plus 102 mMKCl.Where indicated, hemolysate supernate containing 170mg hemoglobin, sonicated ghosts derived from 0.9 ml ofpacked erythrocytes, or 4 mMAcd was employed. Reactionswere initiated by adding the substrate, 2 mM pyridoxinephosphate (PNP). When preincubation was done, it wascarried out for 30 min in the absence of the substrate.(A) Supernatant fraction alone in either 1.5 or 80 mMPi.* *, control; A A, plus Acd. (B) Supernatantfraction plus sonicated ghosts in 80 mMPi. * 0, con-trol, with or without preincubation; V V, plus Acd;O O, with preincubation in the presence of Acd. (C)Supernatant fraction plus sonicated ghosts in 1.5 mMPi.0-*, control, with or without preincubation; V V,

plus Acd; 0 O, plus Acd and preceded by preincul-bation.

700 L. Lumeng and T.-K. Li

man. The plasma levels for the control population in thepresent report are similar to those determined by Ham-felt (16), Chabner and Livingston (12), and Walsh(24), who used tyrosine apodecarboxylase and glutamic-oxaloacetic apotransaminase for assay, but are lowerthan those reported by Hines and Love (8) who em-ployed apophosphorylase b. The decrease in the plasmaPLP levels with age in the control population has alsobeen observed by others (16, 24). Although no ex-planation is yet available for this phenomenon, it isinteresting that this age-related decline in plasma PLPis preserved in the alcoholic population.

The present findings demonstrate that there is a highincidence of deranged PLP metabolism in chronic al-cohol abuse and that it occurs before the appearance ofliver disease. 53% of the alcoholic subjects had plasmaPLP values less than 5 ng/ml, the lowest level en-countered in the control group (Fig. 2). Since plasmaPLP concentrations have been shown to correlate ap-propriately with other biochemical indicators of thestate of vitamin Be nutrition, the data strongly suggestthat a true state of PLP deficiency, albeit subclinical,prevails in many of the alcoholic patients.

The occurrence of altered vitamin Be metabolism inalcoholic individuals before the development of overtliver disease has important clinical implications. Pyri-doxal phosphate functions as the essential coenzyme atnumerous stages in the metabolism of proteins, carbo-hydrates, and fats and in the synthesis of many biologi-cally important compounds, including biogenic aminesand heme (25). Apart from the possibility that alteredmetabolism of PLP may have a causative role in thegenesis of alcoholic liver disease (5-7), the repair ofcellular damage from alcohol requires an increasedneed for nutrients; the failure to correct the disturbancein the metabolism of an essential vitamin may con-tribute to chronic tissue injury.

Hines has previously reported in preliminary com-munications that the pyridoxal kinase activities oferythrocytes and hepatic tissue are diminished in somealcoholic patients, and has suggested that this mightbe a cause of deranged PLP metabolism (10, 20). Ourmeasurements of the pyridoxal kinase and pyridoxinephosphate oxidase activities in the erythrocytes of al-coholics with abnormally low plasma PLP levels (Ta-ble II) do not show impaired enzyme activities. Itshould be pointed out that Hines' subjects were chronicalcoholic volunteers who were ingesting the equivalentof a fifth of 86 proof whiskey per day and had developedabnormal liver function tests and sideroblastic changesin bone marrows at the time that the pyridoxal kinaseactivities were assayed. In contrast, the subjects in thisstudy were chronic alcoholics devoid of hematologic orliver disease and who had abstained from alcohol forseveral days. Furthermore, Hines employed hemolyzed

cells for his assays whereas the assays in this study wereperformed with intact erythrocyte. The erythrocyteswere incubated in the presence of 80 mMPi, which pre-cluded any interference by the phosphatase, and optimalconcentrations of pyridoxal and pyridoxine phosphatewere employed, thereby avoiding substrate inhibition(Fig. 4).

The studies with the intact erythrocytes demonstratethat the rate of net PLP synthesis and the cellularcontent of PLP are regulated not only by the activitiesof the PLP synthesizing enzymes, pyridoxal kinase, andpyridoxine phosphate oxidase, but also by a Be-phos-phate phosphatase (Figs. 3 and 9). The importanceof the phosphatase as a regulator of PLP concentra-tions in tissues has not been fully appreciated in thepast and considerations of regulatory mechanisms havecentered primarily on modes of inhibition or inactiva-tion of the kinase enzyme (19, 26). The control mecha-nism based on pyridoxal as a regulatory molecule (19)probably does not apply to man since our findings indi-cated that pyridoxal does not inactivate the pyridoxalkinase of erythrocytes. The action of the Be-phosphatephosphatase is relevant to red cell metabolism sincehigh concentrations of PLP have been shown to inter-fere with the activities of many glycolytic and pentosephosphate shunt enzymes (27). Anderson and asso-ciates have also suggested that the phosphatase enzymein erythrocytes may play an important role in convert-ing pyridoxine to pyridoxal which is then transportedfrom erythrocytes to other tissues (17).

The pronounced influence of Pi on the net amountof PLP formed in intact cells can he explained by itsinhibitory action upon B6-phosphate phosphatase (Fig.3), an effect which was reproduced in the reconstitutedsystem of hemolysate supernate and sonicated mem-branes (Fig. 9). In the intact cell, high concentrationsof Pi in the medium will also accelerate the glycolyticpathway (28) and the transport of vitamin Be com-pounds (18). These effects, however, do not obtain inthe reconstituted system.

The oxidation product of alcohol, acetaldehyde, de-creased the net synthesis of PLP in intact human redcells in vitro (Figs. 5 and 6). The experiments usingintact cells and broken cell preparations show clearlythat this phenomenon is the result of an accelerateddegradation of the phosphorylated Be compounds, medi-ated by the activity of the neutral Be-phosphate phos-phatase (Figs. 6, 7, and 9). Since ethanol itself had nodemonstrable effect, the observed phenomenon andmechanism therefore represent a unique action of acet-aldehyde.

The molecular basis for the apparent enhancementof the activity of the Be-phosphate phosphatase is un-known. Acetaldehyde may alter the activity of thisenzyme directly by interaction with amino groups

Deranged Vitamin B6 Metabolism in Alcohol Abuse 701

through Schiff base formation or indirectly throughperoxidation of the lipids of the plasma membrane.Wehave observed that acetaldehyde will also alter theactivity of other membrane associated enzymes in ery-throcytes: at 0.05 mM, acetaldehyde effectively inhibitedthe activities of acetylcholinesterase and p-nitrophenyl-phosphate phosphatase of erythrocytic membrane prepa-rations.2 An alternative hypothesis for the enhancementof PLP hydrolysis is that acetaldehyde competes withPLP for binding to proteins in the erythrocytes anddisplaces bound PLP. Preliminary studies indicate thatwhile free PLP is readily hydrolyzed by the membrane-associated Bo-phosphate phosphatase, PLP bound to in-tracellular proteins of the erythrocyte or to bovineserum albumin is not.' Studies to delineate these pos-sible mechanisms of acetaldehyde action are in progress.

Whether or not the effect of acetaldehvde as hereshown with human red cells plays a significant role inthe genesis of deranged PLP metabolism in chronicalcohol abuse cannot be ascertained at this time forseveral reasons. There is presently little informationregarding the importance of the Bo-phosphate phospha-tase in the control of intracellular PLP levels of tissuesother than in the erythrocyte (17, 29). Furthermore,little is known of the role of the different organs in theinterconversion of Be compounds in vivo (30-32).Human red cells were used in these studies as a firstmodel because of their easy accessibility and the pres-ence of easily measurable PLP-synthesizing and degra-dative enzyme activities. Similar studies need to beperformed with cells from other sources, particularlythose possessing high PLP synthesizing capacities (e.g.,liver, kidney, spleen, and brain). In this regard, theliver is of particular interest, since this organ appearsto be the principal source of circulating PLP whichpresumably is then transferred to tissues with low PLPsynthesizing capacity, e.g. skeletal muscle. Moreover,since the liver is the major site of ethanol oxidation, itappears that the concentration of acetaldehyde in situtmay be higher in this tissue than in other organs. An-derson, Fulford-Jones, Child, Beard, and Bateman (17)have recently suggested that the role of erythrocytesin vitamin Be metabolism in vivo is to convert pyridox-ine to pyridoxal. They speculate that pyridoxal may bethe physiologically important vitamer, more readilyutilized by the cells of other tissues. However, such aconsideration would pertain only if pyridoxal is prefer-entially transported or if pyridoxine phosphate oxidaseactivity is low or absent in other tissues. These aspectsof vitamin Be metabolism also remain to be explored.

2Scott, P., L. Lumeng, and T. K. Li. Unpublished data.sLumeng, L., and T. K. Li. Unpublished data.'Lumeng, L., R. E. Brashear, and T. K. Li. Manuscript

in p)reparation.

Although the majority of the metabolic consequencesof alcohol are considered to be mediated by its effecton the redox state of the nicotinamide adenine di-nucleotide coenzyme in liver (33), several biochemicalactions of ethanol have now been ascribed to acetalde-hyde (34-39). However, there is, unfortunately, stillsome uncertainty with regard to the concentration ofacetaldehyde attaiinable in circulation and in tissuesduring the imbibition of alcoholic beverages. Bloodvalues following alcohol ingestion have been reportedto range from 1.0 to 7.0 Ag/ml (0.023-0.160 mM) asdetermined by a variety of methods (34). AMore re-cent studies by Truitt (40) and by Majchrowicz andMendelson (41) indicate that the concentration of"free" acetaldehyde in blood is usually less than 1.0/g/ml during the consumption of alcoholic solutionslow in or devoid of congeners and 1.1-1.5 ig/ml whenbeverages high in congener content are given. How-ever, there is also an indeterminate quantity of acetalde-hyde in erythrocytes and, perhaps, other tissues whichis chemically bound and which is released by proteinprecipitants (42).

The range of acetaldehyde concentrations employedin these studies is comparable to those employed todemonstrate an inhibitory effect of acetaldehyde onbrain metabolism (36-38, 43), but are much lowerthan those used to discern the metabolic effects ofacetaldehyde on cardiac and hepatic tissues (38, 39).A significant decrease in the net synthesis of PLP ofhuman erythrocytes was evident at concentrations ofacetaldehyde as low as 0.05 ml\I (Fig. 6). This con-centration of total acetaldehyde added to incubations invitro is of the same order of magnitude as the "free'acetaldehyde blood levels reported by Truitt (40) andMajchrowicz and Mendelson (41) in their humanstudies in vivo. Such quantitative comparisons arenoteworthy in evaluating the pathophysiologic signifi-cance of the metabolic effects of acetaldehyde.

Multiple vitamin deficiencies are frequently encoun-tered in chronic alcohol abuse. Folate and vitamin B0deficiencies as etiologic factors in the anemia of alco-holism have been well characterized (44). In additionto primary vitamin deficiency due to inadequate intake,ethanol per se or its metabolic products may produceconditioned vitamin deficiency states (45) due to altera-tions in one or more of several processes which affectvitamin metabolism; namely, absorption, conversion.storage, tissue utilization, and excretion. Eichner aindHillman (46) recently presented evidence that alcoholmay interfere with either the storage of folate or theconversion of folic acid to N-5-methyltetrahydrofolicacid. The studies of Hines and Cowan have indicatedthat during alcohol ingestion, the net formation in vivoof PLP from pyridoxine is decreased (4). The in vitrostudies reported here suggest that this action of alcohol

702 L. Lumeng andl T.-K. Li

may be mediated by acetaldehyde through an accelera-tion of the hydrolysis of phosphorylated B. compounds.

ACKNOWLEDGMENTS

WVe thank Jane Yu and Squire Heard for their excellenttechnical assistance and Dr. Pao-Lo Yu, Department ofMedical Genetics, for the statistical analyses.

This work was supported by U. S. Army Medical Re-search and Development Command grant DADA 17-72-C-2132, U. S. Public Health Service grant RR00750, and aGrant-in-aid from Licensed Beverage Industries. We alsogratefully acknowledge a Veterans Administration Re-search and Education Associateship awarded to L. Lu-meng.

REFERENCES

1. Rodwell, V. W. 1970. Metabolism of vitamin B,. InMetabolic Pathways. D. M. Greenberg, editor. Aca-demic Press, Inc., New York. 3rd edition. 4: 411.

2. Leevy, C. M., E. Valdellon, and F. Smith. 1971. Nu-tritional factors in alcoholism and its complications. InBiological Basis of Alcoholism. Y. Israel and J. Mar-dones, editors. Wiley-Interscience, New York. 365.

3. Lerner, A. M., L. M. DeCarli, and C. S. Davidson.1958. Association of pyridoxine deficiency and convul-sions in alcoholics. Proc. Soc. Exp. Biol. Med. 98:841.

4. Hines, J. D., and D. H. Cowan. 1970. Studies on thepathogenesis of alcohol-induced sideroblastic bone-mar-row abnormalities. N. Engl. J. Med. 283: 441.

5. French, S. W., and J. Castagna. 1967. Some effects ofchronic ethanol feeding on vitamin B, deficiency in therat. Lab. Invest. 16: 526.

6. Okada, M., and A. Ochi. 1971. The effect of dietaryprotein level on transaminase activities and fat deposi-tion in vitamin Be-depleted rat liver. J. Biochem.(Tokyo). 70: 581.

7. Greenberg, L. D. 1964. Arteriosclerotic, dental, andhepatic lesions in pyridoxine-deficient monkeys. Vitam.Horm. 22: 677.

8. Hines, J. D., and D. S. Love. 1969. Determination ofserum and blood pyridoxal phosphate concentrationswith purified rabbit skeletal muscle apophosphorylaseb. J. Lab. Clin. Med. 73: 343.

9. Frank, O., A. Luisada-Opper, M. F. Sorrell, A. D.Thomson, and H. Baker. 1971. Vitamin deficits insevere alcoholic fatty liver of man calculated frommultiple reference units. Exp. Mol. Pathol. 15: 191.

10. Hines, J. D. 1970. Correspondence. N. Eng. J. Med.283: 1173.

11. Criteria Committee, National Council on Alcoholism.1972. Criteria for the diagnosis of alcoholism. Anin.Inztern. Med. 77: 249.

12. Chabner, B., and D. Livingston. 1970. A simple enzymicassay for pyridoxal phosphate. Anial. Biochern. 34: 413.

13. Tamura, Z., and S. Takanashi. 1970. Fluorometricdetermination of pyridoxal and pyridoxal 5'-phosphatein biological materials by the reaction with cyanide.MIfethods Enzymol. 18: 471.

14. Beutler, E. 1971. Red Cell Metabolism-A Manual ofBiochemical Methods. Grune and Stratton, New York.13.

15. Peterson, E. A., and H. A. Sober. 1954. Preparationof crystalline phosphorylated derivatives of vitamin Bs.J. Am. Chem. Soc. 76: 169.

16. Hamfelt, A. Age variation of vitamin Be metabolismin man. Clin. Chim. Acta. 10: 48.

17. Anderson, B. B., C. E. Fulford-Jones, J. A. Child,M. E. J. Beard, and C. J. T. Bateman. 1971. Con-version of vitamin B. compounds to active forms inthe red blood cell. J. Clin. Invest. 50: 1901.

18. Yamada, K., and M. Tsuji. 1968. Transport of vitaminBe in human erythrocytes. J. Vitaminol. (Osaka). 14:282.

19. White, R. S., and W. B. Dempsey. 1970. Purificationand properties of vitamin B. kinase from Escherichiacoli B. Biochemistry. 9: 4057.

20. Hines, J. D. 1969. Altered phosphorylation of vitaminBo in alcoholic patients induced by oral administrationof alcohol. J. Lab. Clin. Med. 74: 882. (Abstr.)

21. Sauberlich, H. E., J. E. Canham, E. M. Baker, N.Raica, Jr., and Y. F. Herman. 1972. Biochemical as-sessment of the nutritional status of vitamin B. in thehuman. Am. J. Clin. Nutr. 25: 629.

22. Hamfelt, A. 1967. Enzymatic determination of pyri-doxal phosphate in plasma by decarboxylation of L-tyrosine-"C (U) and a comparison with the trypto-phan load test. Scatnd. J. Clin. Lab. Invest. 20: 1.

23. Hamfelt, A. 1967. Pyridoxal phosphate concentrationand aminotransferase activity in human blood cells.Clin. Chim. Acta. 16: 19.

24. Walsh, M. P. 1966. Determination of plasma pyridoxalphosphate with wheat germ glutamic-aspartic apotrans-aminase. Am. J. Clin.. Pathol. 46: 282.

25. Sauberlich, H. E. 1968. Biochemical systems and bio-chemical detection of deficiency. In The Vitamins. W.H. Sebrell, Jr., and R. E. Harris. editors. AcademicPress, Inc., New York. 2nd edition. 2: 44.

26. Ebadi, M. S., E. E. McCoy, and R. B. Kugel. 1970.Interrelationships between pyridoxal phosphate and py-ridoxal kinase in rabbit brain. J. Nentrochem. 17: 941.

27. Srivastava, S. K., and E. Beutler. 1972. The effect ofnormal red cell constituents on the activities of red cellenzymes. Arch. Biochemii. Biophys. 148: 249.

28. Rose, I. A., and J. V. B. Warms. 1966. Control ofglycolysis in the human red blood cells. J. Biol. Chem.241: 4848.

29. Turner, J. M., and F. C. Happold. 1961. Pyridoxaminephosphate-oxidase and pyridoxal phosphate-phosphataseactivities in Escherichia coli. Biochem. J. 78: 364.

30. Johansson, S., S. Lindstedt, and H. G. Tiselius. 1968.Metabolism of [FH,] pyridoxine in mice. Biochemistry.7: 2327.

31. Colombini, C. E., and E. E. McCoy. 1970. Vitamin B.metabolism. The utilization of ["Cl pyridoxine by thenormal mouse. Biochemistry. 9: 533.

32. Snell, E. E., and B. E. Haskell. 1971. The metabolismof vitamin B6. Comnpr. Biochem. 21: 47.

33. Hawkins, R. D., and H. Kalant. 1972. The metabolismof ethanol and its metabolic effects. Pharmacol. Rev.24: 67.

34. Truitt, E. B., Jr., and M. J. Walsh. 1971. The role ofacetaldehyde in the actions of ethanol. Itt The Biologyof Alcoholism. B. Kissin and H. Begleiter, editors.Plenum Press, New York. 1: 161.

35. Duritz, G., and Truitt, E. B., Jr. 1966. Importance ofacetaldehyde in the action of ethanol on brain norepi-nephrine and 5-hydroxy-tryptamine. Biochem. Phartna-col. 15: 711.

36. Lahti, R. A., and E. Majchrowicz. 1969. Acetaldehyde-

Deranged Vitamin B6 Metabolism in Alcohol Abuse 703

An inhibitor of the enzymatic oxidation of 5-hydroxy-indoleacetaldehyde. Biochem. Pharmacol. 18: 535.

37. Kiessling, K. H. 1962. The effect of acetaldehyde onrat brain mitochondria and its occurrence in brain afteralcohol injection. Exp. Cell Res. 26: 432.

38. Ammon, H. P. T., C. J. Estler, and F. Heim. 1969.Inactivation of coenzyme A by ethanol-I. Acetaldehydeas mediator of the inactivation of coenzyme A follow-ing the administration of ethanol int vivo. Biochem.Pharmacol. 18: 29.

39. Schreiber, S. S., K. Briden, M. Oratz, and M. A.Rothschild. 1972. Ethanol, acetaldehyde, and myocardialprotein synthesis. J. Clin. Invest. 51: 2820.

40. Truitt, E. B., Jr. 1971. Blood acetaldehyde levels afteralcohol consumption by alcoholic and nonalcoholic sub-jects. In Advances in Mental Science-Biological As-pects of Alcohol. M. K. Roach, W. M. Mclssac, and

P. J. Creaven, editors. University of Texas Press,Austin. 3: 212.

41. Majchrowicz, E., and J. H. Mendelson. 1970. Bloodconcentrations of acetaldehyde and ethanol in chronicalcoholics. Scientce (Wash. D. C.). 168: 1100.

42. Truitt, E. B., Jr. 1970. Ethanol-induced release ofacetaldehyde from blood and its effect on the determi-nation of acetaldehyde. Quart. J. Stud. Alcohol. 31: 1.

43. Lahti, R. A., and E. Majchrowicz. 1967. The effects ofacetaldehyde on serotonin metabolism. Life Sci. 6: 1399.

44. Eichner, E. R., and R. S. Hillman. 1971. The evolu-tion of anemia in alcoholic patients. AmO. J. Med. 50:218.

45. Ershoff, B. H. 1948. Conditioning factors in nutritionaldisease. Physiol. Rev. 28: 107.

46. Eichner, E. R., and R. S. Hillman. 1973. Effect ofalcohol on serum folate level. J. Clin. Invest. 52: 584.

704 L. Lumneng and T.-K. Li