Page 1 of 168
A STUDY OF SOME ASPECTS OF THE METABOLISM OF FATTY
ACIDS BY THE HUMAN RED BLOOD CELL IN HEALTH AND DISEASE,
WITH SPECIAL REFERENCE TO DIA3STES MELLITUS.
A dissertation presented in fulfillment of the
requirements for the degree of Doctor of Medicine
in the Faculty of Medicine, University of the
Page 2 of 168
f a . \
6) P | ' ■ '
3 3* <S~
I hereby certify that this dissertation is my own work
and that it has not been presented in fulfillment of the
requirements for other degrees at any University.
> ^ A A &s£LrxJl^
Page 3 of 168
I wish to express my sincere appreciation to:
Professor H. B. Stein for his valuable advice and for making
available the facilities of his laboratory during the course of
To Dr. G. S. Getz who has taken a great deal of interest in
this work; for his stimulating discussions and aid in correcting
To Dr. B. M. Bloomberg for discussions during the earlier
portion of this work.
To Professor G. A. Elliot for allowing me all facilities in
To Dr. R. W. Charlton for making available to me his prima- quine sensitive patients.
To Dr. J. Metz and Mr. D. Hart for the samples of jl31~oleic
Last but not least I would like to record my deep gratitude
to my wife for her invaluable help in the preparation of the
manuscript and for her forbearance.
Page 4 of 168
AD? adenosine diphosphate
A.R. analytical reagent
ATP adenosine triphosphate
ATPase adenosine triphosphatase
BC3 brilliant cresyl blue
b.p. boiling point
cmm cubic millimetre
CoA coenzyme A
cpm / mg counts per minute per milligram
DNA desoxyribonucleic acid
FPA free (non-esterifled) fatty acid
' g gram; gravity
OSH reduced glutathione
Me3 methylene blue
HeBII leuco methylene blue
Me tHb methaemoglobin
mM (mmole) millimole
Page 5 of 168
NAD nicotinamide adenine dinucleotide
form) (coenzyme Ij DPN)
NADH nicotinamide adenine dinucleotide
NAD? nicotinamide adenine dinucleotide phosphate
(oxidised form) (coenzyrae II; 'IPN)
NADPH nicotinamide adenine dinucleotide
Qo2 oxygen quotient
RNA ribonucleic acid
rpm revolutions per minute
S.D. standard deviation
v/v volume by volume
Page 6 of 168
TABLE OF CONTENTS
Chapter I. Certain aspects of the metabolism of the mature
mammalian red blood cell........................ 4
Historical sketch........................... 4
Metabolism of nucleated red cell precursors.. 6
* Carbohydrate metabolism..................... C
Oxidative pathways of carbohydrate
-n' metabolism.............................. 1 1
Respiratory metabolism...................... 15
" Respiratory Supplement ".............. 13
x Methylene blue and related dyes....... 20
, Lipid metabolism............................ 23
Chapter II. Materials and Methods.......................... 28
I. Subjects studied
a) . For In vivo studies.................... 28
b) . For in vitro studies................... 28
II. Preparation of subjects
a) . For in vivo studies.................... 29
I). Carbohydrate tolerance test......... 29
ii). 1^31-oleic acid studies............. 29
b) . For in vitro studies................... 29
Page 7 of 168
III. Preparation of blood samples........ ..... .
A) . Intact erythrocytes................... 30
i). For In vivo studies................ 30
ii). For in vitro studies............... 30
B) . Kaemolysates.......................... 31
IV. Incubation conditions
a) . Employing non-radioactive substrate..... 31
b) . Employing radioactive substrate....... 32
V. Extraction procedures...................... 34
a) . Isolation of 1^31 labelled FFA........ 34
b) . Isolation of red cell total fatty acids. 34
c) . Isolation of red cell total fatty acid
d) . Isolation of red cell lipids for radio- activity assay......................... 35
VI. Analytical procedures
a) . Determination of plasma, whole blood
and erythrocyte FFA.................... 37
b) . Determination of plasma, whole blood
and erythrocyte triglyceride........... 4i
c) . Determination of erythrocyte total oxi- dized pyridine nucleotides.............. 45
d) . Determination of erythrocyte fatty acid
e) . Determination of glucose............... 46
f) . Determination of lactate............... 46
g) . Determination of pyruvate.............. 46
Page 8 of 168
h). Haematocrit determination.............. 46
VII. Assay of radioactivity..................... 46
Chapter III. Some interrelationships of carbohydrate, lipid
and pyridine nucleotide metabolism in the human
erythrocyte in vivo. A comparison betx*een
diabetic patients and normal subjects......... 48
X Chapter IV. Some interrelationships of carbohydrate, lipid
and pyridine nucleotide metabolism in the human
erythrocyte in vitro. A comparison between
diabetic patients and normal subjects.......... 64
A) . Studies in intact cells and haemoly- sates.................................. 64
B) . Studies with inhibitors of glycolysis
and the hexose monophosphate shunt..... 74
C) . Studies in intact cells and haemoly- sates employing radioactive techniques.. 77
Speculation on a possible alternative path- way of carbohydrate metabolism............. 92
\ Chapter V. Certain aspects of carbohydrate and lipid meta- bolism In primaquine sensitive cells............ 100
Page 9 of 168
Chapter VI r a). A brief discussion on some aspects of fatty
b). Some clinical implications of the results
presented in this thesis..................... 110
Appendix I. List of enzymes present in mature mammalian
Appendix II................................................ .
Page 10 of 168
A DESCRIPTION OF THE RED PARTICLES OF THE BLOOD IN THE
HUMAN SUBJECT AND IN OTHER ANIMALS.
"It is a curious and important fact, that these particles
are found so generally through the animal kingdom, that is, they
are found in the human species, in all quadrupeds, in all birds,
in all amphibious animals, and in all fish, in which animals they
are red, and colour the blood.
The blood even of insects contains particles similar in
shape to those of the blood of more perfect animals, but differing
In water insects, as the lobster and shrimp, these parti- cles are white; in some land insects, as the caterpillar and
grasshopper, they appear a faint green when In the vessels, as I
am persuaded from experiments. I have seen them in an insect no
bigger than a pin's head, and suspect they exist almost universal- ly through the animal kingdom.
What is so generally extended through the creation, must be
of great Importance in animal economy, and highly deserving the
attention of every inquirer into the works of nature."
William Hewson, P.R.S. 1777.
Page 11 of 168
The awakening of an appreciation of the importance of bio- chemistry to medical science in recent years has prompted the
clinical chemist to study and explain disease mechanisms in terms
of fundamental cellular metabolism. However, his efforts have
been considerably hampered by obvious difficulties involved in
obtaining suitable tissues for experimental purposes. Most of our
rather meagre understanding of the underlying mechanisms causing
metabolic diseases in the human subject has been obtained by abso- lute chemical analysis of the body fluids of which it is probably
true to say that the blood ( either whole blood or serum ) has
been the most extensively studied. It is becoming increasingly
apparent however, that this type of approach does not offer a
satisfactory explanation for the disturbed metabolic reactions in
In this thesis an attempt has been made to study the bio- chemical problems associated with diseases of metabolism utilizing
a more dynamic approach. For this purpose, the concentration
changes taking place simultaneously in both the plasma and
erythrocyte compartments of various naturally occurring compounds
( e.g, free fatty acids ) were compared at intervals over a period
of a few hours during a glucose tolerance test. Using this tech- nique, relatively large differences from the resting values were
noted in the plasma as well as in the red blood cell within a
short time after administration of the carbohydrate. To obtain
further insight into the nature of the metabolic changes taking
Page 12 of 168
place In the erythrocyte in vivo, an in vitro technique was de- veloped in which the metabolic behaviour of this cell could
apparently be made to simulate that occurring in vivo. The find- ings presented in this study are concerned mainly with certain
interrelationships of carbohydrate and lipid metabolism in the
erythrocyte. A comparison has been made between diabetic patients
and normal subjects. It will be shown that the metabolic be- haviour of normal and diabetic red cells with respect to the bio- synthesis of lipid from carbohydrate is quite different both in
vivo and in vitro. Compared with the normal, the diabetic
erythrocyte is about one third as active in its ability to synthe- size fat from carbohydrate. Furthermore, this defect in lipid
synthesis in the diabetic red cell can be corrected to a large
extent by insulin whether administered intravenously or added to
the cells in vitro. The implications of these findings with
reference to the possible metabolic aberrations in human diabetes
mellitus will be discussed. Employing the same in vitro tech- nique mentioned above, some of the metabolic interrelationships
of carbohydrate and lipid have also been studied in the
glucose-6-phosphate dehydrogenase deficient cell. These results
are discussed from the point of view of their importance to our
understanding of the mechanism of red blood cell haemolysis that
occurs in sensitive individuals following exposure to a wide
variety of chemical compounds. A discussion of the findings in
this study as they relate to certain mechanisms of lipid syn- thesis is also included.
The observation that the human erythrocyte in vivo was
Page 13 of 168
apparently not completely inert from a metabolic standpoint indi- cated that this cell may have important metabolic functions in the
body economy other than the traditional one of a carrier of re- spiratory gases. Thus it was hoped, using the above mentioned
techniques, that a study of the fundamental processes of metabo- lism in the human erythrocyte in health and disease might reveal
abnormalities which could possibly be a reflection of altered
metabolic reactions occurring elsewhere in the body. The results
of this investigation strongly suggest that this is in fact the
case, and furthermore, that a proper study of the human red blood
cell might provide valuable information concerning the basic
cellular defects associated with human metabolic diseases.
Page 14 of 168
CERTAIN ASPECTS OF THE METABOLISM OF THE MAMMALIAN
RED BLOOD CELL
Johannes Swammerdam in 1658 was most likely the
first person to observe that the circulating fluid of the body,
the blood, consisted not only of a liquid part, but also of cellu- lar elements ( "orbicular particles" ) suspended in it. Probably
the earliest report of the red blood corpuscle can be found in
Swammerdam's Biblia Naturae posthumously published by Boerhaave in
1737. The original accurate description of the erythrocyte we owe
to the Dutch microscopist Anthony van Leeuwenhoek. He gave an
account of the red corpuscles of man in the Philosophical Trans- actions of 1674; in various communications he carefully described
red blood cells of different animals, showing that while circular
in man they were oval in birds, frogs and fishes and proving in
all cases that the colour of the blood was due to these red bodies.
The English physician William Hewson (1777) in his book entitled
" An Experimental Inquiry into the Properties of the Blood ",
recognized that the red particles of the blood were in reality flat
discs rather than globules and suggested that they must be of great
value in the body economy. Haemoglobin in crystalline form was
isolated from blood by Hunefeld in 1840, but it was not until 1867
that Hoppe-Seyler demonstrated how haemoglobin possessed the abili- ty to take up and discharge oxygen thus clarifying the functional
significance of haemoglobin and of the cells in which it was
Page 15 of 168
Thirteen years earlier, P. W. Pavy (18 54) noted that blood
which had been allowed to stand for come time outside the body
lost its sugar, stating " Under the changes of the decomposition
of the blood, normal animal glucose is very readily metamorphosed.
The rapidity of the metamorphosis depends on the activity of the
decomposition of the animal substances present, and when the
destruction of the sugar is complete the blood has assumed an
acid reaction." Although Pavy was aware of the existence of lactic
acid it is clear from his communication that he did not attribute
any direct relationship between the loss of sugar and lactic acid
production in shed blood. Claude Bernard (18 5 5 ) first predicted
that the consumption of sugar by mammalian blood in vitro could
be accounted for in very large part by new formation of lactic
acid. Bernard also stressed the fact that the destruction of
glucose by blood was due to a process of fermentation. He wrote
(1877)... " Ce ferment lactlque se rencontre dans le sang, dans
les muscles, dans le foie lui-mSme; car, j'ai constate que les
muscles et divers tissus ne deviennent acides apres la mort
qu'autant qu'ils renferment du sucre ou de la matiere glycogfene
qui subit trfes rapidement une fermentation lactique ..... J'ai vu
egalement la fermentation lactique dans le sang. Quand on prend
sur un animal qu'on vient de sacrifier le sang le plus sucre,
celui des veines sus-hepatiques, on constate que sous 1 'influence
d'une chaleur moder^e le sucre disparal't du sang en lul dormant
une reaction acide, au lieu de la reaction alcaline n o m a l e....
J'admets done, quant a moi, que la destruction du sucre a lieu par
Page 16 of 168
fermentation et non par 1 'Influence directe des alcalis du sang,
qui favorisent seulement cette reaction."
In 1890, Lepine Introduced the term "glycolysis" to describe
the manner in which glucose was "destroyed" in the blood and showed
that this process required the presence of a "ferment". Since
glycolysis was found not to occur in plasma, Lepine stated that it
was in some way connected with the blood cells and further
suggested that the "ferment" was located in the leucocytes. This
prediction was verified twenty three years later when Levine and
Meyer (1913) demonstrated the role of leucocytes in glycolysis.
In the same year, MacLeod (1913) and Kona and Arnheim (1913)
showed that the glycolytic process was also affected by the
erythrocyte thus attributing to this cell another function besides
that of oxygen transport. Since this time numerous Investigators
have been engaged in acquiring knowledge concerning the multiple
metabolic activities which are manifested by the mammalian
erythrocyte, and It is now fully apparent that this cell, far
from being just a carrier of oxygen, maintains a highly organized
state of metabolism in common with other tissue cells of the body.
NUCLEATED RED CELL PRECURSORS
A brief survey of the metabolism of the immature erythrocyte
The nucleated red cell
This cell appears to participate in the
general metabolic reactions characteristic of nucleated cells of
other tissues. Synthesis of desoxyribonucleic acid (DNA) occurs
mainly in the nucleus, and that of ribonucleic acid (RNA) in the
Page 17 of 168
cytoplasm and nucleolus. It Is able to synthesize protein, lipid
and carbohydrate, while synthesis of haemoglobin is a major
function of the nucleated erythrocyte. For reviews see Rimington
(1959) and London (i960).
Reticulocyte maturation is indicated by a pro- gressive loss of RNA, the transition to the mature cell being
marked by the disappearance of this compound. When compared with
the mature cell glycolysis is more active, and since the enzymes
of the tricarboxylic acid cycle are present, respiration occurs
at a much higher rate. The reticulocyte is capable of haemoglobin
synthesis (Walsh et al. 194-9)* The biochemistry of the reticu- locyte has recently been reviewed by Lowenstein (1959).
Page 18 of 168
THE MATURE MAMMALIAN ERYTHROCYTE
SOME ASPECTS OF CARBOHYDRATE METABOLISM IN THE MARgjALIAN
The tern glycolysis was introduced by Lepine to indicate
the disappearance of carbohydrate during the metabolic activities
of a tissue. This rather wide meaning was later restricted by
Warburg to the fission of the carbohydrate molecule by a fermenta- tive reaction to yield acidic products, which usually proved to be
lactic acid in the case of animal tissues. The pioneer investiga- tions of Harden and Young and by Robison on the significance of
phosphoric esters in alcoholic fermentation led directly to the
study of similar phosphorylations in muscle by Embden, Pamas,
Meyerhof, Needham and others. Finally, the contributions of Cori's
laboratory to the phosphorylation mechanism, and the unifying con- cept of the high-energy phosphate bond by Lipmann, Kalckar and
others have revealed the inner energetic basis of the glycolytic
mechanism with remarkable clarity. Numerous excellent reviews on
the subject have appeared (Meyerhof 1941; Barron 1943; Dorfman
1943; Stotz 1945; Dickens 1951; Stumpf 1954).
The mature human erythrocyte is unique among the cells of
the body in that its cellular integrity is preserved in spite of
the absence of a nucleus and of cytoplasmic subcellular particles.
Tills cell has become adapted to a metabolism dependent almost en- tirely upon glycolysis, retaining only atrophied remnants of
tricarboxylic acid cycle activity. However, there is evidence
Page 19 of 168
that the oxidative phase of carbohydrate metabolism functions
adequately in the mature red cell and the possibility exists that
this pathway may be of importance in the regulation of certain
biosynthetic activities by the erythrocyte.
Meyerhof (19 3 2), many years ago established that the
glycolytic system in the erythrocyte was closely similar to that
of muscle and other tissues, and in recent years all the inter- mediary compounds of the glycolytic sequence together with most
of the enzymes necessary for their continued degradation and
synthesis have been isolated from this cell (Bartlett 1959; Altman
1959). For every mole of glucose utilized, 2 moles ATP are
required in the initial phosphorylation stages to fructose 1 ,
6-diphosphate and 4 moles ATP are synthesized during the further
breakdown of this phosphorylated sugar to lactic acid. Thus, in
the overall reaction 2 moles ATP are formed which represents a
net gain of about 19,000 calories. This energy is then made
available for the numerous metabolic functions the cell has to
perform, for example, regulation of the active transport of ions
and molecules across the glycolipoprotein envelope of the
There is however an important feature of difference between
the glycolytic system of the red blood cell in the majority of
mammalian species and that of other tissues. The red cell main- tains a high steady state concentration of 2 ,3-diphosphoglycerate
with special regulatory mechanisms for controlling the level of
this compound within the cell. According to Guest and Rapoport
(1938) 2 ,3-diphosphoglycerate makes up about 50$ of the total acid
Page 20 of 168
soluble phosphorus and constitutes almost the entire "stable
phosphate" fraction of the red blood cell. This ester is a co- enzyme in the conversion of 3-phosphoglycerate to 2-phospho- glycerate (Sutherland et al. 19^9) and is therefore present in
most cells in catalytic amounts only. Guest and Rapoport (1941)
and Prankerd and Altman (1954) demonstrated with the aid of radio- active phosphorus that red cell 2 ,3-diphosphoglycerate had a high
turnover of the labelled atom indicating that it must be taking
an active part in glycolysis or at least be in equilibrium with one
of the metabolic intermediates. Bapoport and Leubering (1950)
obtained evidence that 2 ,3-diphosphoglycerate (in the human
erythrocyte) was in equilibrium with 1 ,3-diphosphoglycerate, the
interconversion being catalyzed by a " diphosphoglycerate mutase ".
Further work by these authors (19 5 1; 19 5 5) proved the existence
in red cells of an enzyme, diphosphoglycerate phosphatase, which
catalyzes the conversion of 2 ,3-diphosphoglycerate to 3-phos- phoglyceric acid and a supplementary cycle at the triose phosphate
level of the glycolytic reaction sequence was postulated. The
functional significance of this pathway has not yet been fully
elucidated, but Altman (1959) considers that the "Rapoport- Leubering Cycle" constitutes a means whereby excess energy rich
phosphate bonds produced during glycolysis can be dissipated since
the reactions in the conversion of 1 ,3 -diphosphoglycerate to 3-
phosphoglyceric acid via this cycle involve a loss of 14 Kcal. in
toto. He argues that since endergonic reactions utilizing adeno- sine triphosphate (ATP) have atrophied in the mature human red
cell, and the ATPase system which regulates ATP and adenosine
Page 21 of 168
diphosphate (ADP) concentrations in other cells of the body is
functionally inactive, in the absence of a system which can dis- pose of excess energy rich phosphate bonds, ATP and 1,3-diphospho- glycerate would accumulate and the concentration of ADP and in- organic phosphorus would decrease, resulting in retardation or
even complete cessation of glycolytic activity. It is perhaps
difficult to understand why the erythrocyte should have evolved
such a complicated mechanism for energy " wastage " especially
since several workers have demonstrated the presence of an active
ATPase in the mature red blood cell (Clarkson and Maizels, 1952;
Garzo et al., 1952). Rather, it would seem more reasonable to
suppose that this energy could well be utilized in many of the
biosynthetic reactions (e.g. synthesis of glutathione, pyridine
nucleotides) which are known to occur in this cell.
Evidence for the existence of an oxidative pathway of carbohydrate
metabolism in the mature mammalian erythrocyte
It was principally due to the studies of Warburg and his
collaborators dealing with the effect of methylene blue on the
metabolism of erythrocytes that the way was paved for the dis- covery of an alternate pathway for the metabolism of glucose In
the red cell as well as in tissue cells generally. Warburg and
Christian (1931a) first proved that glucose-6-phosphate was the
substrate metabolized when methylene blue was added to red blood
cells in catalytic quantities. It was demonstrated in haemolysates
that glucose could no longer be utilized but glucose-6-phosphate
was readily oxidized to 6-phosphogluconic acid. This finding
Page 22 of 168
suggested that the intact red cell contained an enzyme system,
which was destroyed by haemolysis, necessary for the phosphory- lation of glucose, a fact which was soon confirmed by Meyerhof
(1932). The same workers (19 31b) also succeeded in isolating from
horse blood haemolysates an enzyme "Zwischenferment" and a co- enzyme "Coferment", which together in the presence of methylene
blue brought about the oxidation of glucose-6-phosphate. Subse- quent studies have shox/n that "Zwischenferment" is the enzyme
glucose-6-phosphate dehydrogenase while "Coferment" has since been
identified as nicotinamide adenine dinucleotide phosphate (NADP).
Both enzyme and coenzyme are necessary for the first stage of the
oxidative cycle of carbohydrate metabolism. Further investiga- tions by Dickens, Lipmann, Horecker, Racker and Cohen have been
instrumental in elucidating the main sequences of this pathway.
The following overall reactions are known to occur:-
1 . 2Glucose-6-phosphate + Og — > ribose-5-phosphate +
xylulose-5-phosphate + C02
2. Ribose-5-phosphate + xylulose-5-phosphate ------— >
sedoheptulose-7-phosphate + glyceraldehyde-3-phosphate
3. Sedoheptulose-7-phosphate + glyceraldehyde-3-phosphate
fructose-6-phosphate + erythrose-4-phosphate
4. Erythrose-4-phosphate + xylulose-5-phosphate------->
fructose-6-phosphate + glyceraldehyde-3-phosphate.
Although Dickens (1938) observed a slight oxidation of ribose
phosphate by horse blood haemolysates in the presence of methylene
blue, it was Dische (1938) who first drew attention to the possi- bility of this pathway operating in red cells when he reported
Page 23 of 168
that adenosine added to human red cell haemolysates took up inor- ganic phosphorus while at the same time the ribose fraction of the
nucleoside disappeared. The phosphate which disappeared was found
in the haemolysate in the form of fructose-1 ,6-diphosphate in
equilibrium with triose phosphate and an unidentified reducing
ester. This observation suggested that a ribose ester of unknown
structure was formed from adenosine which broke dovm immediately
into a 3-carbon and a 2-carbon compound. Addition of guanosine
and inosine to the haemolysate had the same effect as adenosine.
Extending these studies, Dische (1951) later found that the ribose
moiety of adenosine could be phosphorylated in ATP free haemoly- sates to form what was in all probability ribose-5-phosphate. The
latter compound was further metabolized to a mixture of hexose and
triose phosphates. The mechanism whereby haemolysates of human
erythrocytes rapidly metabolized the ribose portion of adenosine,
guanosine or inosine to hexose phosphate was not clearly defined
until the recent isolation of a purine nucleoside phosphorj^lase
from human red cells (Gabrio and Huennekens 1955; Sandberg et al.
1955; Tsuboi and Hudson 1957). This enzyme, which catalyzes the
phosphorylytic cleavage of guanosine and inosine provides a means
for the introduction of pentose phosphate into the erythrocyte.
Adenosine is utilized in red cells after its conversion to inosine
by adenosine deaminase also present in erythrocytes (Gabrio et al.
195o). The existence of phosphoribomutase, the enzyme catalyzing
the conversion of ribose-1 -phosphate to ribose-5-phosphate, is
inferred from the work of Dische who showed that ribose-1-phosphate
formed from the metabolism of nucleosides in fluoride treated
Page 24 of 168
haemolysates led to the formation of hexose phosphate presumably
via ribose-5-phosphate. Guarino and Sable (1955) have reported
on the presence of both phosphoglucomutase and phosphoribomutase
in human erythrocytes. Bruns et al. (1958), while studying the
metabolism of ribose-5-phosphate in mammalian haemolysates, proved
conclusively the existence of a complete pentose phosphate cycle
in the red cells of man and other mammals. Non-nucleated human
erythrocytes were found to contain a very high activity of
phosphoribose isomerase, which catalyzes the reaction:-
ribose-5-phosphate ----•> ribulose-5-phosphate.
The identification of sedoheptulose-7-phosphate as a product of
ribose-5-phosphate metabolism in human red cells has also been
confirmed by Marinello (19 58).
There is thus considerable evidence to indicate that the
mature human red cell possesses a cyclic oxidative pathway for
glucose metabolism with the production of triose, tetrose, pentose,
hexose and heptulose phosphates as intermediates. Recently,
Murphy (i960) has demonstrated with the aid of differentially
labelled glucose that whereas the relative amount of glucose
metabolized via the pentose phosphate cycle was small (an average
of 1 1 ^ of the total glucose utilized), this pathway contributed
25$ of the available potential energy as reducing pother for the
erythrocyte in the fora of NADPH which it generated. Although the
experiments reported by Murphy wrere performed in vitro, the cells
were treated under conditions resembling as near as possible those
occurring in vivo and it is not unreasonable to infer from these
studies that carbohydrate is able to be metabolized via the oxida
Page 25 of 168
tive cycle by the erythrocyte in vivo.
In confinnation Of the earlier postulates of Barron and
Karrop (1928) that the addition of methylene blue to mature mamma- lian red blood cells caused an oxidation of the glucose molecule
after it had been phosphorylated, various workers using more ele- gant techniques, have recently demonstrated that the dye stimu- lates the glucose oxidative pathway in the human erythrocyte.
IXibovsky and Sonka (1955) showed that the pentose phosphate con- tent of mammalian red cells increased markedly when methylene blue
was added to them, from which they concluded that the pentose
phosphate shunt activity was enhanced by the dye. However, more
convincing evidence of stimulation of the glucose oxidative cycle
in human erythrocytes by methylene blue has been presented by Brin
and Yonemoto (1953). On incubation of red cells with glucose- l-C"^ and glucose-2-C^ in the presence of methylene blue, the
radioactive carbon dioxide recovered accounted for 9 5 of the oxi- dation, and 85/5 of the observed oxygen consumption could be
ascribed to the C ^ O g recovered from radioactive glucose. Glueose- 6- C ^ produced no radioactive carbon dioxide. These workers also
found that the glucose oxidative cycle of thiamine deficient rat
erythrocytes (Brin et al. 1958) and in patients with Wernicke1s
encephalopathy (Wolfe et al. 1958) when stimulated by methylene
blue was markedly retarded at the transketolase stage i.e.
ribose-5-phosphate + xylulose-5-phosphate -------- >
sedoheptulose-7-phosphate + glyceraldehyde-3-phosphate.
In severely thiamine deficient red cells, pentose phosphates were
found to accumulate and the recovery of C-^'Og from glucose-2-C1^
Page 26 of 168
was depressed to one-seventh the normal level. Since transketolase
requires thiamine pyrophosphate as a cofactor (Racker et al. 1953)
these experiments lend further support to the argument that methy- lene blue stimulates the pentose phosphate pathway in red blood
CERTAIN ASPECTS OP RESPIRATORY METABOLISM IN THE MATURE MAMMALIAN
A slow disappearance of oxygen from normal blood main- It
tained at 37°C was reported as long ago as 1867 by Pfluger.
Cohnstein and Zuntz (1884) showed that a similar loss of oxygen
occurred in foetal blood while an absorption of oxygen and a pro- duction of carbon dioxide in the blood of anaemic animals was
demonstrated by Morawitz and Pratt (1908) and confirmed by Morawitz
in 1909. During the same year* Warburg (1903) reported that normal
adult red cells had a negligible oxygen consumption, whereas
nucleated avian erythrocytes possessed a high rate of oxygen up- take. Further experiments by Morawitz and Itami (1910) led to a
conclusion similar to that of Warburg, that the rate of respira- tion of normal rabbit blood was very low. However, oxygen con- sumption in the blood of anaemic animals was appreciable and pro- portional to the degree of reticulocytosis present. For the next
20 years it was generally believed that the normal mature mamma- lian erythrocyte possessed a negligible respiratory rate, and that
such oxygen uptake as occurred in the blood could be accounted for
by the presence of reticulocytes and leucocytes (Harrop 1919;
Dennecke 1926; Derra 1928; Daland and Isaacs 1927; Michaelis and
Page 27 of 168
Salomon 1930; Wright 1930). The earlier work dealing with the
oxygen consumption of the mammalian erythrocyte has been summarized
by Morawitz (1928).
In 1932, Ramsey and Warren, using the respirometer method of
Penn and making due allowance for the number of reticulocytes and
leucocytes present in their blood cell suspension, proved that the
rate of consumption of oxygen by normal mature rabbit red cells
was not quite so small as had been previously assumed. The average
rate of oxygen uptake by these cells was 8 .5 cmm. / g. / hour.
Since this oxygen consumption had earlier been shown by Ramsey and
Warren (1930) and Harrop and Barron (1928) to be accompanied by a
production of heat and of carbon dioxide which yielded a respira- tory quotient within the physiological range, one must consider
the process as a true respiration. The oxygen consumption was not
due to the haemoglobin becoming more saturated during the course
of the experiment. WTien a comparison is made between the Qo2
values of mammalian red cells and other tissue cells of the body,
it is clear that the mature mammalian erythrocyte consumes oxygen
at a slow rate. All the figures reported for mature human red
cells are less that -1.0 (Warburg 1909; Harrop and Barron 1928;
Ramsey and Warren 1930) as compared with -21 for kidney, -11 for
liver and -2.5 for resting muscle (Dixon 1952). However, since
95/3 of the dry weight of the non-nucleated red cell consists of
haemoglobin, Warren (19^8) has argued that the erythrocyte Qo2
values ought therefore to be multiplied by a factor of between 6
and 20 to make them comparable with those of other cells. Berger
(1930) has even suggested that the Qo2 value should be multiplied
Page 28 of 168
by a factor of 100 as the fixed framework of the mammalian
erythrocyte is only about 1% of the dry weight of the cell. When
considered in terms of the mass of its protoplasm, the respiratory
activity of the red cell becomes quite considerable and of importance
to the energy metabolism of this cell.
Two independent experimental findings which may contribute
to our understanding of certain phenomena involved in mammalian
erythrocyte metabolism will now be discussed.
a). " Respiratory Supplement 11
Michaelis and Salomon (1930) discovered that saline
extracts of various rat tissues, in particular the liver, had the
property of greatly stimulating the oxygen consumption of mamma- lian red cells in vitro. They named the active material
" respiratory supplement ". The substance has not yet been puri- fied or fully identified although the possibility of it being one
of the cytochromes, coenzyme I, ascorbic acid or the anti-anaemic
principle of the liver was excluded by the authors. Overbeek
(l939a;b,* 19^0a;b) further studied the Michaelis - Salomon effect
and observed a species difference. The erythrocytes of man and
the rabbit both markedly increased their consumption of oxygen in
the presence of liver extract whereas rat red blood cells did not
react. However, if rat erythrocytes previously poisoned with
sodium fluoride were now treated with liver extract, a considerable
rise in oxygen consumption by these cells was noted; fluoride
poisoned rabbit red cells showed an even greater uptake of oxygen
than normal. The effect of fluoride on human red cell metabolism
Page 29 of 168
will be discussed later. The Michaelis - Salomon reaction was
also demonstrated in rabbit and human erythrocyte haemolysates and
a partial inhibition by potassium cyanide in this system was
observed. Respiratory quotient values of 0.9 suggested that the
material being oxidized was mainly carbohydrate. Furthermore, it
was shown that the active principle of liver extract \*as thermo- labile and could be destroyed by irradiation with ultra violet
light. As a result of his investigations Overbeek could not ex- plain satisfactorily the mechanism of the Michaelis - Salomon
reaction, and although no direct evidence was presented he con- cluded that flavine mononucleotide (the " old yellow enzyme " of
Warburg and Christian) was in part responsible for the effect.
Pirwitz and Haaf (1950) noted that human serum obtained from
reticulocyte rich blood was able to stimulate oxygen consumption
by normal erythrocytes. The authors claimed that this effect was
due to adenosine triphosphate present in the serum.
Various other naturally occurring substances are known to
increase the oxygen consumption of the mature mammalian erythrocyte
in vitro e.g. cysteine and reduced glutathione (GSH) (Bruns and
Rummel (I950a;b). This effect is dependent upon the structural
integrity of the cell since it is not apparent in haemolysates.
Carver et al. (i960) have recently observed that menadione (vita- min K3 ) is also able to stimulate the oxygen consumption of mature
No further work has since appeared on the identification of
" respiratory supplement ", but its existence in a water soluble
fora in mammalian tissues suggests that under an appropriate stimu
Page 30 of 168
lus this substance night be released into the circulation to pro- duce an enhanced respiratory rate together with an increased
metabolic activity of the red blood cell in vivo.
b) Methylene blue and related dyes
Lauth in 1876 reported
that the reaction of hydrogen sulphide with dimethyl-p- phenylenediamine in the presence of an oxidising agent led to the
formation of a deep blue coloured product which he did not identify
or isolate. In the following year a German patent on this com-,
pound* 3 ^7-bis(dimethylamino) phenasathionium chloride (methylene
blue)* was obtained for use as a dye.
The pioneering work of Ehrlich on the use of methylene blue
in vital nerve staining led to the discovery by Gutman and
Ehrlich in 1891 of its specific action in malaria, while in 1901
Ehrlich's student Michaelis first used the dye as a haematological
stain. It was not until 1911 that the effect of methylene blue on
cell respiration was observed when Palladin and coworkers reported
that the addition of methylene blue to plant cells caused a large
increase in their carbonic acid production. The oxygen consump- tion of the cells was not measured. Meyerhof (1 9 1 2 ) then noted
that the respiration of living pyogenic staphylococci could be
inhibited by 0.005^ methylene blue to the extent of 10 - 2Q£>* while
acetone extracted cocci showed a somewhat increased respiratory
activity. However* if the same were heated in vacuo at 100°C a
marked increase in respiration was observed in the presence of
small amounts of the dye. The hypothesis was put forward that
Page 31 of 168
methylene blue functioned as a hydrogen acceptor and could replace
a respiratory enzyme which had been destroyed in the staphylococci
by heating. Sixteen years later Harrop and Barron (1928) dis- covered that the oxygen consumption of mature mammalian red blood
cells could be enormously increased by the addition of catalytic
amounts of methylene blue or of other dyes vThich can be reversibly
oxidized and reduced* e.g. cresyl blue, toluylene blue etc. The
substrate for the reaction was glucose; cyanide was ineffective as
an inhibitor. These authors further studied the effect of
methylene blue upon glycolysis and lactic acid formation in mamma- lian and avian erythrocytes. They found that the glycolytic quo- tient i.e. millimoles lactic acid produced of normal human blood
2 x millimoles glucose degraded
vias for all practical purposes unity, whereas after the addition
of methylene blue the glycolytic quotient always fell well below
1. These findings were interpreted as indicating that glucose
disappeared from blood in two different ways. In the first, the
hexose molecule was split into two molecules of lactic acid with- out coincident utilization of oxygen; in the second, oxidation of
glucose with considerable liberation of energy was thought to have
taken place. It was further suggested that methylene blue acted
by increasing the oxidation of some degradation product of glucose,
most probably a phosphorylatcd derivative. This paper is also
interesting in that it contains perhaps the first allusion to the
existence of an oxidative pathway for the metabolism of glucose in
the mammalian erythrocyte.
These results were communicated by Harrop and Barron to
Otto Warburg who began a series of investigations into the nechan-
Page 32 of 168
ism of action of methylene blue in the erythrocyte. From their
extensive studies with methaemoglobin-containing red cells, proof
of the formation of intracellular methaemoglobin by the dye and
the reduction of intracellular methaemoglobin by glucose, the
Warburg school interpreted the methylene blue effect in erythro- cytes as follows:-
MeB + lib — > MeBH + MetHb
MetHb + metabolite — ? Hb + oxidation product of metabolite
MeBH + C>2 — MeB + H2O2 (destroyed by catalase),
where MeB, MeBH, lib, MetHb are methylene blue, leucomethylene
blue, haemoglobin and methaemoglobin respectively.
Thus, the catalytic role for the oxidation of the metabolite
(glucose or lactic acid) had been placed on methaemoglobin. This
rather unphysiological mechanism was criticized by Wendel (1929;
1930; 1 9 3 1) who proved that lactic acid could be oxidized without
the intermediary participation of methaemoglobin in the methylene
blue-red cell system. It v,ras also shown by Wendel that lactic
acid vias oxidized to pyruvic acid which apparently could not be
further utilized by the cells. The oxidation of lactic acid by
erythrocytes was not prevented by sodium fluoride or cyanide.
Since cyanide inactivates many metal containing enzymes (notably
those containing iron), and fluoride inhibits glycolysis, it was
concluded that neither of these reactions were necessarily con- cerned in the metabolism of lactic acid by red cells In the pres- ence of methylene blue. Wendel now postulated a direct electron
acceptor function between the substrate and atmospheric oxygen for
methylene blue, and stated furthermore that the dye might take the
Page 33 of 168
place of some physiological acceptor molecule which had been lost
during the process of maturation of the cell, a theory identical
with that put forward by Meyerhof in 1912 working with staphylococci.
Warburg later recognized the interpretation given by Wendel and
Harrop and Barron as the more correct one and that the formation
of methaemoglobin was a supplementary rather than an obligatory
step in this reaction. Although it is probable that methylene blue
reacts mainly with HADP dependent systems in the mammalian erythro- cyte, there is evidence from studies of the red cell methaemoglobin
reductase enzymes (Kiese 1944) to indicate that the dye can also
activate triosephosphate and lactic dehjrdrogenase, both enzymes
having nicotinamide adenine dinucleotide (NAB) as a coenzyme
Finally, it should be noted that the oxygen uptake induced
in the mature red blood cell by methylene blue and related dyes
does not represent an entirely catabolic process, for even in
haemolysates, energy made available by this oxidation is utilized
for the synthesis of organic phosphate compounds (Engelhardt and
Ljubimova 1930; Engelhardt 1930; 1932; Runnstrom and Michaelis
THE METABOLISM OF LIPIDS IN THE MAMMALIAN ERYTHROCYTE
Although information regarding the pathways of carbohydrate
metabolism in the mature mammalian erythrocyte is quite consider- able, a paucity of data exists concerning the manner in which fat
is utilized by this cell (see reviews by Prankerd 1955; 1956).
Halm and Hevesey (1939) concluded from in vitro experiments with
Page 34 of 168
radioactive phosphorus that whereas there was a certain amount of
exchange of phospholipids between red cells and plasma, these
compounds were not synthesized to any appreciable extent by
erythrocytes. A rapid interchange of free cholesterol between red
blood corpuscles and plasma was noted by Hagerman and Gould (1951 ),
while London and Schwartz (1953) demonstrated in vivo a dynamic
equilibrium between the cholesterol of plasma and the red cell
stroma using deuterium labelled cholesterol. The latter authors
stated from their in vitro experiments that there was insignifi- cant cholesterol synthesis by the mature human erythrocyte. How- ever, Muir et al. (1951) after injecting radioactive sodium ace- tate into rabbits, took blood for analysis of red cell lipids at
intervals of 6 hours to 52 days and found that the cholesterol of
the red cell membrane was metabolically active having a turnover
rate which reached a maximum within 24 hours. Similar experiments
on the red blood cells of mice, rabbit, dog and man performed by
Altman et al. (1951), Altman (1953) and Altman and Swisher (1954)
both in vivo and in vitro revealed an incorporation of radioactive
acetate into all the lipid fractions of the stroma together with a
rapid turnover of the radioactive constituents. The sphingolipids
had the highest activity while a slow synthesis of cholesterol was
also observed. These findings indicated that the stromal lipid
components were metabolically active and it was also thought that
they may be able to exchange with chemically identical constituents
present in the plasma.
An interchange between red cell and plasma lipids in vitro
has been noted recently by James and coworkers (1959; i960) and
Page 35 of 168
Marks et al. (i960). O'Donnell et al. (1958) investigated the
biosynthesis from actetate-l-C^ of lipids in the formed blood
elements, and concluded that whereas reticulocytes were able to
synthesize fatty acids and cholesterol from labelled acetate in
vitro, this activity was lost during the course of maturation of
the cells and was completely lacking in the mature erythrocyte
of the chicken, rabbit and man. Similar results have been re- ported by Marks et al. (i960). The biosynthesis of lipids by
human blood cells was studies in vitro with the aid of methyl- labelled acetate by Jaumes and coworkers (1959; i960). After in- cubation with radioactive acetate all the common saturated and
unsaturated fatty acids of the erythrocyte including araehidonic,
linoleic and linolenic acids were labelled. The long chain fatty
acids were found to be incorporated into triglycerides as well as
phospholipids but not into cholesterol esters. Triglycerides
rapidly moved out of the cells into the plasma where they were
incorporated into the oc-lipoproteins while the phospholipids
were more readily found in the -lipoprotein fraction of the
plasma. Howe (1959), studied the incorporation of both radioact- ive carbon and radioactive phosphorus into the phospholipids of
human blood cells in vitro. After 6-|- hours all the phospholipid
fractions were labelled with both and the amount of each
label appearing in the following order: cephalins > lecithins
The experimental procedures used by James et al. (1959*
i960) and Rowe (1959) do not really permit of an assessment of the
extent to which leucocytes participated in fat synthesis. It has
Page 36 of 168
been demonstrated by Buchanan (i960) using similar techniques to
James and coworkers that synthesis by leucocytes alone was suffi- cient to account for all the fat synthesized by whole blood.
More recently, Howe et al. (i960) compared the in vitro incorpo- ration of C^-acetate into the lipids of both itfhite blood cells
and erythrocytes of human blood. The leucocytes were responsible
for all the detectable incorporation into the unsaponifiable lipid
fraction and most of that into phospholipid fatty acids. However,
these authors also found quite a considerable incorporation into
the neutral lipid fatty acids of the red cells. It would seem
therefore that under certain in vitro conditions biosynthesis of
fat by the red blood cell is very limited. Yet there are the
findings of Muir et al. (1951) and Altaian et al. (1951) that after
Injection of radioactive acetate in vivo the erythrocyte lipids
exhibited a rapid turnover of the label.
Other evidence attesting to an active metabolism of fat by
the red cell in vivo can be found in the work of Horwitt et al.
(1959) who fed normal human subjects diets rich in linoleic acid
and observed that this fatty acid appeared in the lipids of their
erythrocytes in increased amounts within about a week. Mead and
Pillerup (1957) showed that half an hour after the ingestion of
earboxy-labelled methyl stearate in rats, labelled fatty acids
were present in the erythrocyte lipids, the highest activities
appearing in the triglyceride fraction. Holman (i960) noted that
the polyunsaturated fatty acid content of rat red blood cells
could be altered by manipulating the dietary intake of these fatty
acids. Identical changes in fatty acid composition occurred in
Page 37 of 168
other tissues (e.g. heart). It was concluded that the erythrocyte
polyunsaturated fatty acid pattern reflected the dietary intake of
essential fatty acids and that this might be used to evaluate the
essential fatty acid status of the animal. Whether the changes
described above represent a de novo synthesis of lipid or merely
an exchange of lipid between plasma and red cells cannot be
answered on the available evidence.
In conclusion, there exists as yet little direct evidence
that the mature human red blood cell is capable of synthesizing
lipid in vitro whereas the results of in vivo studies support the
argument for an active fat metabolism in the erythrocyte. These
findings indicate some factor necessary for the maintenance of an
active erythrocyte metabolism within the body which is rapidly
destroyed in vitro.
In this thesis evidence will be presented to indicate that
a) the normal human red blood cell is probably capable of syn- thesizing lipid from glucose in vivo; b) in severely diabetic
patients requiring insulin, this faculty of lipogenesis in the red
cells is significantly impaired; c) the parenteral administration
of insulin to these patients is able to restore in their red cells
the ability to synthesize fat from carbohydrate. An in vitro
technique will also be described in which the above metabolic re- actions of this cell can apparently be made to simulate those
occurring in vivo. These findings will be discussed with particu- lar reference to the possible cause (s) for the abnormality in
lipogenesis from carbohydrate in human diabetes mellitus. A brief
discussion of the results as they relate to certain mechanisms of
lipid synthesis is also included.
Page 38 of 168
MATERIALS AND METHODS
I. Subjects studied
a) . For in vivo studies
The control group consisted of
apparently healthy adults of both sexes ranging in age from 19 to
58 years. The second group were patients suffering from diabetes
mellitus and included both males and females ranging in age from
14 to 69 years. All diabetic patients had a severe form of the
disease and required large doses of soluble insulin for adequate
b ) . For in vitro studies
Venous blood was obtained from
healthy adults and patients with diabetes mellitus. The subjects
with a congenital deficiency of glucose-6-phosphate dehydrogenase
In their erythrocytes were adult male Bantu patients at Johannesburg
hospitals Suffering from a variety of medical and surgical condi- tions (with the exception of diabetes mellitus). The normal
controls for this group were Bantu hospital patients or Bantu
members of staff.
Detection of primaquine sensitivity: The screening test
for glucose-6-phosphate dehydrogenase activity of red cells de- scribed by Motulsky and Cambell was applied. This method proved
reliable and convenient (see Charlton and Bothwell, 1961). All
subjects presented no obvious haeraatological abnormalities.
Page 39 of 168
II. Preparation of subjects
a ) . For In vivo studies
i) . Carbohydrate tolerance test. This was performed In
all subjects after a 12 hour fasting period. In the case of the
diabetic patients Insulin was discontinued for at least 8 to 10
hours before the start of the experiment. Following the removal
of initial blood samples, 50 g glucose was administered orally to
each subject and further blood samples were collected at half hour- ly intervals for 3 hours. During the test all subjects were com- pletely at rest. As far as possible diabetic patients were com- pared with normal subjects of a similar age group.
ii) . For the experiments using I1 31_ labelled fatty acid
the following procedure was employed. Apparently normal fasting
subjects were given 50 uc I1^1- labelled oleic acid by mouth. A
specimen of blood was removed 5 hours later and 50 g glucose was
then taken by the subject. A further sample of blood was withdrawn
after 1 -| hours.
b ) . For in vitro studies
All subjects were in the fasting
state for at least 12 hours prior to the venepuncture. The dia- betic patients were being controlled on soluble insulin which i?as
discontinued from 8 to 10 hours before blood was removed for the
test. As far as possible diabetic patients and primaquine sensi- tive subjects were compared with normal controls of a similar age
group. The age of the subjects investigated ranged from 14 to
Page 40 of 168
III. Preparation of blood samples
a). Intact erythrocytes
i ) . For in vivo studies;
For FFA analyses the blood and
plasma samples were precipitated into the extraction mixture with- in 15 minutes after withdrawal from the vein. This precaution was
taken to minimize the lipolytic effect on triglycerides with subse- quent release of FFA of endogenous lipoprotein lipase which is
known to occur in some persons. Red cell FFA, triglycerides and
pyridine nucleotides were calculated from the whole blood, plasma
and haenatocrit values. In some experiments these parameters were
determined directly on washed erythrocytes.
The red cells from subjects who had taken I1^1- labelled
oleic acid by mouth were washed four times with 5 volumes of 0.85$
sodium chloride to eliminate any radioactive iodine contamination
from the serum.
ii) . For in vitro studies:
The blood was collected into
small conical flasks and defibrinated using a glass rod. Plasma,
buffy coat and the upper 10 to 15$ red cells were removed after
centrifugation at 800 g for 15 minutes. The remaining erythrocytes
were washed twice with cold 0.85$ sodium chloride. After each
washing approximately 10$ of the upper layer of red cells together
with remnants of the buffy coat were aspirated. The washed red
cells were then suspended in cold Krebs-Ringer phosphate buffer
adjusted to pH 7.5 to give a final haematocrit between 40 and 50$.
In some of the earlier experiments the washed erythrocytes were
Page 41 of 168
re-suspended in their own serum. However, this system was dis- continued for reasons which will he mentioned later. The leucocyte
count of the blood prepared as described above was less than O.lfj
of the original white cell count. Erythrocytes obtained from nor- mal subjects, diabetic patients and subjects with a deficiency of
glucose-6-phosphate dehydrogenase in their red cells, will be
referred to as normal red cells, diabetic red cells and primaquine
sensitive (or enzyme deficient) red cells respectively,
The preparation of erythrocytes up to the
stage of suspension in Krebs-Ringer phosphate buffer was the same
as described above. Nicotinamide (■'10 umoles / ml red cell sus- pension) was then added to prevent destruction of the coenzymes
NAD and NADP x-rhich readily occurs during haemolysis of red cells.
Haematocrit values v.rere determined before haemolysis. The
erythrocytes were haemolyzed by rapid freezing and thawing three
times. To remove as much of the stroma as possible the haemolysate
was centrifuges at 2000 g for 30 to 45 minutes. It is not claimed
that this procedure produced a completely stroma free haemolysate
since this was not necessary for the purposes of the present
IV. Incubation conditions
a). Employing non-radioactive substrate
Incubation of the red
cell suspensions was carried out in 50 ml conical flasks in an
atmosphere of air. The flasks were mechanically shaken for 3 - 5
Page 42 of 168
hours in an incubator maintained at 37°C. Haematocrit determina- tions performed before and after incubation showed that no haemo- concentration had occurred. Haemolysis was not detected under
Additions to the red cell suspension during incubation.
Except when otherwise stated, the undermentioned substances were
added to 5 to 10 ml erythrocyte suspension in the following con- centrations :
Glucose:- 10 umoles/ml
Brilliant cresyl blue (BOB):- 0.3 pimoles/ml. 0.16 ml
of a 0.06^ solution of the dye in Krebs-Ringer phosphate buffer
was added per ml red cell suspension. This concentration was not
critical. Methylene blue may be substituted for BCB in the same
concentration but the use of the latter dye was preferred.
Crystalline insulin:- 0.6 units/ml. The insulin was
obtained through the courtesy of Eli Lilly and Company.
Chemical procedures: Samples of the erythrocyte suspension
were removed initially and at various time intervals during incu- bation for determination of FFA, triglycerides and pyridine nucleo- tides as described below. Red cell concentrations were calculated
from the haematocrit value when the erythrocytes were suspended in
phosphate buffer and estimated directly on washed cells when serum
was employed as the suspending medium. All determinations were
performed in duplicate.
b). Employing radioactive substrate
The preparation of the
erythrocytes and additions to the red cell suspensions was the same
Page 43 of 168
as described above. In addition uniformly labelled glucose
(glucose-U-ClZl) (Specific activity 72.9 me per mmole. The Radio- chemical Centre, Amershan, England) was added in a final concen- tration of 0.4 juc per ml red cell suspension. Incubation was
carried out in Warburg-type flasks. The centre well of each
flask contained 0.5 ml of 2 .5N sodium hydroxide to trap liberated
carbon dioxide. The flasks were stoppered and incubation allowed
to proceed at 37°C for 3 hours with gentle shaking. At the end
of the incubation period, the erythrocyte suspension was centri- fuged at 500 g for 5 minutes and the supernatant fluid removed.
The red cells were washed once with 0.85$ sodium chloride before
being submitted to the lipid extraction procedure outlined below.
Haemolysates were added directly to the extraction medium.
Crystalline insulin vrtien added to haemolysates was used in con- centrations ranging from 0.1 to 1.0 units per ml haemolysate.
Because haemolyzed red cells have been shown to lose their
ability to metabolize glucose due to the destruction of hexokinase
and adenosine triphosphate (ATP), these compounds were routinely
added to the haemolyzed erythrocytes. Such haemolysates will be
referred to as reconstituted haemolysates. In certain experi- ments the effect of a NADPH generating system on the biosynthesis
of lipids in human red cell haemolysates was studied. The system
used in this investigation was isocitric acid, isocitric dehydro- genase and NADP. The reagents ATP, NADP, hexokinase (type IV),
isocitric acid (DL + Alio - 42$ "LM; 84$ DL) and isocitric dehy- drogenase (30 units / mg) were all products of the Sigma Chemical
Page 44 of 168
2'-adenylic acid was obtained from Schwarz BioResearch,
Inc., New York.
Y. Extraction procedures
a) . Isolation of labelled FFA in experiments using radio- actlve oleic acid.___________________________________
Labelled FFA was isolated as follows: one volume of
red cells was treated with 5 volumes methylal : methyl alcohol
(4:1) or chloroform : methyl alcohol (l:l), the mixture stirred
with a glass rod for 30 minutes and then centrifuged at 800 g for
5 minutes. The extraction was repeated three times and the com- bined extracts evaporated to dryness. The residue was dissolved
in 10 volumes of petroleum ether (b.p. 40 - 60) which was then
washed 4 times with equal volumes of aqueous 0.5$ potassium carbo- nate. After the alkali washings were acidified to a pH below 4
with 50$ sulphuric acid, they were reextracted 4 times with 2
volumes of petroleum ether. The combined petroleum ether extracts
were taken to dryness, the residue dissolved in 4 ml isopropyl
alcohol and the radioactivity assayed in a scintillation counter.
b ) . Isolation of red cell total fatty acids
Total lipids were extracted from one
volume of red cells with chloroform : methyl alcohol (1 :1 ) (5
volumes). The extraction was repeated three times and the solvent
completely removed. Two millilitres of a 10$ solution of potassium
hydroxide in methyl alcohol was added to the total lipid extract
and saponification carried out at 80°C for 1 hour. Preliminary ex- periments revealed that saponification was complete under these
Page 45 of 168
conditions. The mixture was acidified by the addition of 2 ml
ION sulphuric acid and the total fatty acids were extracted into
low boiling point petroleum ether. The petroleum ether extract
was taken to dryness and the fatty acids dissolved in bensene
(usually 1 ml). An aliquot was then removed for colorimetric
estimation using the rosaniline reagent as described below.
c ) . Isolation of red cell total fatty acid esters.
One volume of red cell suspension was treated with 5
volumes of methylal:methyl alcohol (4:1) and well stirred for 10
minutes. The mixture was centrifuged at S00 g for 5 minutes and
the supernatant fluid decanted. The extraction was repeated twice
and the combined supernatants evaporated to dryness. The residue
was dissolved in 5 volumes of chloroform which was then washed
twice with an equal volume of water to remove small quantities of
dye taken up by the chloroform. After washing with water the
chloroform layer was practically colourless. Total fatty acid
esters were determined on the residue after evaporation of the
chloroform by a modification (Antonis, i960) of the method of
Stern and Shapiro (1953)- In these experiments methylene blue was
employed because brilliant cresyl blue contained a reddish fluo- rescent component which could not be removed from the chloroform
by washing and which interfered with the colour reaction for fatty
d) . Isolation of red cell lipids for assay of radioactivity
All solvents were redistilled before use. Two and a
half to five millilitres of red cells were treated with 5 volumes
chloroform : methyl alcohol (l: 1) in a 50 ml conical flask and
Page 46 of 168
stirred mechanically with a glass rod for 30 minutes after which
time the supernatant fluid was filtered through fat free Whatman
No. 1 filter paper into a 100 ml round bottom flask. The ex- traction was repeated three times and the pooled supernatants taken
to dryness under an atmosphere of nitrogen. Five millilitres of
20$ potassium hydroxide in methyl alcohol v.ras added to the dried
total lipid extract and saponification carried out by boiling
under reflux for 1 to 2 hours. After cooling to room temperature,
the digest was diluted with sufficient xvater to bring the methyl
alcohol concentration to approximately 30$ and the mixture trans- ferred to a separating funnel. The round bottom flask was washed
twice with 20 ml petroleum ether (b.p. 40 - 60°C), the washings in
each case being transferred to the separating funnel. Non- saponifiable material was removed by extracting the aqueous solu- tion three times with 20 ml petroleum ether. These extracts were
combined, washed three times with water and the washings added to
the original aqueous phase which was then acidified to a pH below
2 with 50$ sulphuric acid. Fatty acids were extracted from it
four times with 20 ml petroleum ether and the composite extract
washed free of mineral acid and dried over sodium sulphate. The
pooled extracts containing total fatty acids were evaporated to a
small volume under nitrogen and weighed aliquots were plated on
Since lipid extracts of biological material invariably con- tain non-lipid contaminants such as glucose, amino acids etc., in
each experiment a zero time control i.e. an aliquot of the red
cell suspension taken before incubation was put through the same
Page 47 of 168
lipid extraction procedure described above. The counts in the
zero time control fatty acids were subtracted from those obtained
after incubation for 3 hours. As a check on this method experi- ments were performed in which the lipids were purified by one or
other of the following procedures
a) the liquid washing procedure described by Polch et al.
b) purification of the lipids on a cellulose column accord- ing to Smith (1954).
The results after lipid purification using procedures a) and b)
agreed well with the method used in this study. In some instances
the presence of radioactivity in the fatty acids was confirmed
after separation of the latter by gas-liquid chromatography and
assay of radioactivity in the isolated fatty acids. VI.
VI. Analytical Procedures
a). The determination of free fatty acids (PPA) applicable
to plasma, whole blood and erythrocytes.______________
In connection with certain experiments needed for this in- vestigation it became necessary to devise a method for the esti- mation of micro amounts of higher free fatty acids (hereafter
referred to as FFA) in blood. At the time all the methods avail- able for the determination of blood FPA required a relatively
large amount of material (l - 15 ml). Furthermore, since these
procedures utilized either aqueous or alcoholic alkali to neutra- lize the acid groups, they were liable to the inherent disadvant- ages of a titration method such as the need for the preparation of
Page 48 of 168
fresh standard alkali each day, the performance of a titration in
a carbon dioxide-free atmosphere, and because the concentration of
PPA in blood is so low, the requirement of specialized equipment
(ultra-micro burettes). It was therefore thought worth while to
explore the possibilities of colorimetry.
P r i n c i p l e Plasma, whole blood or erythrocyte suspension is acidi- fied with normal sulphuric acid to a pH below 4.0 and the higher
PFA isolated using a two-phase extraction system. The upper phase
solvent containing the PFA is evaporated to dryness and the resi- due dissolved in isopropyl alcohol. Rosaniline reagent is added,
to the mixture which is heated at 46°C for half-an-hour and the
resulting red colour measured in a photoelectric colorimeter. The
concentration of the PPA is determined from a calibration curve
prepared by using pure stearic acid as a standard. Pure palmitic
or oleic acid may also be used to prepare the standard curve.
All glassware is kept scrupulously clean and free from
even the slightest traces of acid.
a) . Methylal (Merck).
b) . Methyl alcohol (Merck-acetone free).
c) . Methylal : Methyl alcohol (4 : Iv/v)
d) . Petroleum ether (b.p. 40 - 60°C). A.R.
e) . Iso-Propyl alcohol A.R.
f) . Benzene A.R.
g) . Normal sulphuric acid.
Page 49 of 168
h) . Stearic acid (pure).
i) . Rosaniline (base).
Preparation of rosanillne reagent
About 0.5 g rosaniline (base) is
added to 50 ml benzene in a 100 ml round bottom flask equipped
with a reflux condenser. The mixture is heated for one hour under
reflux and then allowed to cool to room temperature. The solution
is decanted from the undissolved dye and centrifuged for five
minutes. The dark-orange fluorescent supernatant is poured into
a 500 ml beaker and benzene added until the diluted reagent shows
25$ transmission in an Evelyn photoelectric colorimeter set to
100$ transmission with distilled water using a 490 mu filter and
the 10 ml aperture.
Blood obtained by venepuncture is collected into tubes
containing 2 mg dried potassium oxalate / ml blood. The plasma is
separated within fifteen minutes after collection. 0.2 ml plasma,
whole blood or erythrocyte suspension is acidified with 0.02 ml
normal sulphuric acid in a 15 ml centrifuge tube; 1.0 ml methylal;
methyl alcohol (4:1 v/v). is added to the tube well shaken. Then
3.0 ml methyl alcohol and 0.8 ml petroleum ether (b.p. 40-60°C)
are added with mixing until the solvents from a single phase. Now
0.8 ml distilled water is added when the system immediately
separates out into two phases. This is followed by a further 2 ml
petroleum ether and the contents of the tube well stirred for one
minute using a glass rod flattened at one end. The tube is centri- fuged at about 3000 rpm for two minutes after which the clear upper
Page 50 of 168
petroleum ether phase Is transferred to a pyrex tube graduated at
10 ml using a Pasteur pipette with a fine bore. Care was taken
not to draw into the Pasteur pipette even the smallest drop of the
highly acid lower phase which would obviously introduce a consider- able error into the results. The extraction is repeated using a
further 2 ml petroleum ether. With the first extraction 95$ of the
FPA is removed and with the second extraction recoveries are for
practical purposes complete.
The two petroleum ether extracts are combined in a pyrex
test tube graduated at 10 ml and evaporated to dryness in a hot
water bath. The residue is dissolved in 1 ml iso-propyl alcohol
and then 1 ml rosanlline reagent is added with mixing. The tube
which is stoppered with a loose fitting glass bulb is placed in a
rack in a constant temperature water bath at 46°C for 30 minutes
after which time it is placed in a beaker of cold water for 5
minutes and the contents made up to 10 ml with benzene. The rea- gent blank consists of 1 ml iso-propyl alcohol + 1 ml rosanlline
reagent treated exactly as the test. The colour of the test solu- tion is read in an Evelyn photoelectric colorimeter using a 520 mu
filter against the reagent blank. The concentration of free fatty
acids in jaeq / litre is calculated from a calibration curve. The
assumption is made that the average molecular weight of the higher
FFA in the blood is 280.
In later experiments a permanent standard of stearic acid in
benzene was used in place of the calibration curve. The last part
of the procedure outlined above was then slightly modified as
follows:- the dried petroleum ether extract of the test was dlssolv-
Page 51 of 168
ed In 0.1 ml benzene and made up to 1 ml with iso-propyl alcohol.
1 ml rocaniline reagent was added and the procedure continued as
before. Standards in benzene containing 0 - 0.352 peq stearic
acid per 0.1 ml were prepared and treated similarly to the test.
The amount of benzene had to be kept small since this solvent
interfered with the development of colour in larger amounts. The
Beer-Lambert lav; was obeyed from 0 - 0.352 pieq using the modified
Calculation:- T x S„ = where,
---------- S C c
T = optical density of test
S = optical density of standard
Sc = concentration of standard
Tc = concentration of test
For further details of the method the reader is referred to
b). The determination of triglycerides applicable to plasma,
whole blood and erythrocytes.__________________________
Current methods available for the estimation of
triglycerides in biological fluids seem to be unsatisfactory.
Indirect procedures based on the measurement of total fatty acid,
cholesterol esters and phospholipids and requiring calculation of
triglyceride by difference, are unsuitable mainly because of the
assumptions which have to be made, particularly of the proportion
of fatty acids in phospholipids. The separation of triglycerides
from other lipid components by column chromatographic methods
tended to be too cumbersome and time-consuming. For these reasons
a relatively simple, accurate, specific and sensitive procedure
Page 52 of 168
Glycerol, liberated from triglycerides by alkaline hydro- lysis, is heated at 140° with o-aminophenol in the presence of
concentrated sulphuric acid and an oxidizing agent to form 8-
hydroxyquinoline. The fluorescence produced by chelation of
8-hydroxyquinoline with a divalent metal ion in alkaline solution
is utilized to provide a quantitative measure of the glycerol, and
therefore of triglyceride concentration.
All reagents and solvents used v.rere analytical reagent
grade, with the exception of o-aminophenol, of which only a tech- nical grade was available. Light petroleum ether (b.p. 30 - 60°)
was redistilled before use. Diethyl ether and isopropyl ether
were freed from peroxides by passage through a column of activated
alumina (heated overnight at 170°) prior to use.
a) . Silicic acid: Silicic acid (Mallinckrodt; 100 mesh,
suitable for chromatography) was size graded by sedimentation with
distilled water. The fine particles, which constituted approxi- mately 50$ of the total, was separated, dried, and activated over- night at 170°.
b) . Arsenic acid solution (0.6$). 50 g of arsenic pent- oxide was dissolved in 100 ml. water, allowed to stand 3 to 4 days
to form H^AsO^, and filtered if necessary. One ml of this stock
solution was diluted to 100 ml with concentrated sulphuric acid.
Page 53 of 168
c) . o-Aminophenol solution (1.6#). Technical grade
o-aminophenol v;as purified by sublimation at 170° in an atmosphere
of nitrogen. The sublimed compound tends to oxidize rapidly at
this temperature, and it was therefore either recrystallized from,
or washed 3 times with, a small volume of isopropyl ether until a
colourless solution was obtained when the pure white crystalline
compound was dissolved in acetone. Immediately before use, 0.16 g
of the pure o-aminophenol was dissolved in 10 ml acetone.
d) . Magnesium solution (120 pg/ml). MgSOi;. 7H20 (0.12 g)
was dissolved in distilled water and made up to 100 ml.
e) . Triolein standards. A stock standard solution was
prepared, containing 100 mg triolein (or equivalent amounts of
other triglycerides) per 100 ml in chloroform. Dilute standards
were prepared, containing 0.2 to 1.0 mg/ml, corresponding to the
range 100 to 500 mg triglyceride per 100 ml serum when carried
through the procedure.
Activated silicic acid (1.2 g) was slurried in a glass- stoppered test tube with 1 ml of isopropyl ether, and 0.3 ml of
serum added dropwise with shaking. A further 6.5 ml of isopropyl
ether was then added, together with a few glass beads, and the
mixture well shaken for half an hour. The silicic acid was allowed
to settle, and a 5 nil aliquot of the supernatant extract (equiva- lent to 0.2 ml serum) was taken off into a glass-stoppered 15 ml
conical centrifuge tube and evaporated to dryness on a hot water
bath using an air blower. One ml aliquots of the triolein stan- dards were similarily evaporated to dryness.
Page 54 of 168
To the dried extract was added 3 to 4 drops ether, 0.5 nil
of methanol, and 3 drops of 2$ methanolic potassium hydroxide
solution, and the mixture saponified for 30 minutes at 60 - 70°.
Two drops of 6$ methanolic acetic acid solution were then added,
and the mixture evaporated just to dryness on a boiling water
bath. Six ml of petroleum ether were added to the hot tube,
followed by 0.5 ml of 10 normal sulphuric acid, the tube was
stoppered, well shaken, and then centrifuged for 1 to 2 minutes.
The petroleum ether layer was carefully pipetted off and discarded.
Duplicate 0.1 ml aliquots of the aqueous glycerol phase were then
treated as below.
One-tenth ml aliquots of the 1 .6$ o-aminophenol solution
were pipetted into a number of test tubes fitted with glass
stoppers, and the solvent evaporated, using an air blower. A 0.1
ml aliquot of the serum glycerol extract prepared as above was
added, followed by 0.4 ml of the 0.6$ arsenic acid solution, and
the mixture heated in a silicone oil bath at 140° for 15 minutes.
The mixture was ccdled in ice water, and 1 ml of the magnesium
solution was cautiously added with mixing. Five ml of 28$ ammonia
solution was carefully added, and the tubes stoppered and well
A blank (0.1 ml of 10 normal sulphuric acid solution) and
glycerol extracts (0.1 ml), derived as before from 1 ml aliquots
of the standard solutions of triolein, were run simultaneously
throughout the procedure. Glycerol standards containing 2 to 20
jig of glycerol per 0.1 ml in 10 normal sulphuric acid, can be
used, since the hydrolysis of the triglyceride is quantitative.
Page 55 of 168
After 5 to 10 minutes, aliquots of the above solutions were poured
into Farrand fluorometer tubes, and the fluorescence produced by
ultraviolet light was measured in a Farrand fluorometer, using
aperture 6, with Coming 5874 (primary) and 2424 (secondary)
filters. The fluorometer was set at 100%, transmission with the
highest standard (i.e. 200 pg corresponding 500 mg / 100 ml serum).
Where low serum triglyceride concentrations were expected, lower
concentration glycerol standards were used for setting the in- strument at 100%, transmission, but in this case the blank reading
Since the use of silicic acid for the separation of tri- glycerides from phospholipids was unsuitable when applied to whole
blood and erythrocytes, triglycerides were extracted into the
upper phase of the tiro phase system as described in Appendix II.
This technique is also effective in separating phospholipids from
other neutral lipid components present in blood. For further
details of the method the reader is referred to Appendix III.
c). Estimation of total oxidized pyridine nucleotides in
Red cell pyridine nucleotides were determined by a
slight modification of the alkali addition procedure as described
by Kaplan et al. (1951). 0.2 ml whole blood or packed cells were
added dropwise with constant shaking to 0.8 ml cold 10% tri- chloracetic acid. After standing for 15 minutes the mixture was
filtered through Whatman No. 40 paper. Four tubes were set up as
follows:- blank - 1 ml water; standard - 0.2 ml and 0.4 ml (con- taining 2 pg and 4 pg NADP respectively) + 0.8 ml and 0.6 ml water
Page 56 of 168
respectively; test - 0.2 ml trichloracetic acid filtrate (0.04 ml
blood or cells) + 0.8 ml water. To each tube was added 1 ml 5
normal sodium hydroxide with mixing after which the tubes were
placed in a boiling bath for 5 minutes. The tubes were then
cooled and the contents made up to 8 ml with water. Fluorescence
was read in a Farrand fluorometer against the standard using pri- mary filter 5874 and secondary filters 4308 + 3389. The blank was
set at 20 on the galvonometer scale.
d) . Estimation of esterified fatty acids
These iirere deter- mined by a modification (Antonis, i960) of the method of Stern and
e) . Estimation of glucose
For the in vivo studies whole
blood glucose was measured according to Asatoor and King (1954).
For the in vitro experiments glucose was measured using glucose
oxidase (Salomon and Johnson, 1959).
f ) . Lactic acid was determined according to the procedure
of Barker and Sumnerson, (1941).
g) . Pyruvic acid \>;as determined according to the procedure
of Friedemann and Haugen, (1943).
h) . All haematocrit values were estimated according to
Wintrobe (1956). VI.
VII. Assa?/ of radioactivity
i). I1 21 _iabelled fatty acids were assayed in a scintilla- tion counter employing a thallium activated sodium iodide crystal.
Page 57 of 168
Radioactivity was estimated with an absolute efficiency of 60c,j .
At least 10,000 counts were measured for each sample. At the
same time FFA was determined quantitatively in the red cells by
the procedure already described.
\ 14 ii). C -labelled fatty acids were plated on to planchettes
and accurately weighed. Radioactivity was assayed at ini'ini te
thinness in a windowless gas flow counter. Counting was con- tinued until the statistical error was less than 5^.
Page 58 of 168
Some interrelationships of carbohydrate, lipid and pyridine
nucleotide metabolism in the human erythrocyte in vivo. A
comparison between diabetic patients and normal subjects.
Figure 1 summarizes the changes in both plasma and erythro- cyte FFA of thirty normal individuals undergoing a glucose toler- ance test. It is evident that within half an hour after the in- gestion of carbohydrate the red cell FFA had begun to rise while
at the same time the plasma FFA concentration was declining. This
reciprocal relationship continued during the active absorption of
glucose and reached a maximum in 1 to l|- hours after which there
was a gradual return to the original resting values in both plasma
and red cell. Under the influence of a carbohydrate load the
average maximum Increase in the normal erythrocyte FFA concentra- tion was 290 ^eq/litre red cells. In all instances the greatest
rise in red cell FFA occurred after the blood glucose had reached
its peak and had already begun to decline.
Erythrocyte and plasma triglyceride concentrations in normal
subjects after the ingestion of 50 g glucose Is indicated in
Figure II. In a similar manner to the behaviour of FFA a rise in
red cell triglyceride values concommitant with a fall in plasma
triglyceride levels was noted. The average increase in erythro- cyte triglyceride concentration was 124 ^ieq/lltre cells. Again,
the maximum rise in red cell triglycerides occurred after the
blood glucose had reached its peak but the time taken for the re- turn to resting values was longer than that noted for erythrocyte
Page 59 of 168
ihe effect of oral ingestion of 50 g glucose on
erythrocyte and plasna FFA in 30 normal subjects.
A - erythrocyte FFA
B - plasma FFA
Curves represent mean values. Vertical lines
indicate the standard deviation.
Page 60 of 168
A - erythrocyte triglyceride
13 - plasma triglyceride
The effect oi* oral Ingestion of 50 g glucose on
erythrocyte and plasma triglyceride levels in
30 normal subjects.
I GLUCOSE mg 7.
Page 61 of 168
The effect of oral ingestion of 50 g glucose on
erythrocyte total oxidized pyridine nucleotides
in 30 normal subjects.
A - erythrocyte pyridine nucleotide
0 1 2 3
GLUCOSE m g
Page 62 of 168
The change in erythrocyte total oxidized pyridine nucleotides
during a glucose tolerance test in normal individuals is shown in
Figure III.. Since the pyridine nucleotides are to be found almost
exclusively within the red cells, only these values are reported.
A small but definite increase in red cell total oxidized pyridine
nucleotides was noted after the ingestion of carbohydrate in normal
individuals. The average rise was 11 yxg / ml cells. The return
to resting values took place somewhat more rapidly than that ob- served for red cell FFA.
Figure IV depicts the manner in which the diabetic red cell
responded to a glucose load in vivo. Both the rise and fall In
red cell and plasma FFA concentrations respectively was much less
and more irregular than in the normal. The largest Increase in
erythrocyte FFA noted in the diabetic group was 98 pieq/litre cells.
Eight cases showed no increase while in two patients a slight fall
in red cell FFA values during the test was actually observed.
No significant change In diabetic red cell triglyceride and
pyridine nucleotide concentrations was noted during a glucose
tolerance test. (Figures V and VT).
To determine whether the lack of response to glucose In the
diabetic erythrocyte could be corrected, the effect of intraven- ous soluble insulin was studied during a carbohydrate tolerance
test in diabetic subjects. As can be seen from Figure VII Insulin
administered 60 to 90 minutes after the oral ingestion of glucose
had within half an hour dramatically brought about the appearence
of a normal picture. Normal individuals receiving insulin at a
Page 63 of 168
The effect of oral ingestion of 50 g glucose on
erythrocyte and plasma FFA in 35 diabetic patients.
A - erythrocyte FFA
3 - plasma FFA
Page 64 of 168
TRIGLYCERIDE pEQ/t- 5^
A - erythrocyte triglyceride
B - plasma triglyceride
The effect of oral ingestion of 50 g glucose on
erythrocyte and plasma triglycerides in 35
-i- ~ •£___ _ _
GLUCOSE m g '/.
Page 65 of 168
A - erythrocyte pyridine nucleotide
The effect of oral ingestion of 50 g glucose on
erythrocyte total oxidised pyridine nucleotides
in 35 diabetic patients.
Page 66 of 168
F ^ A JJEQ/L
The effect of intravenous insulin given during a
glucose tolerance test on erythrocyte and plasma
FFA in 10 diabetic patients.
A - erythrocyte FFA
B - plasma FFA
Page 67 of 168
time when the blood sugar was at a maximum demonstrated very little
further increase in red cell FFA.
The fasting plasma FFA concentration in the diabetic group
(870 jueq/1; S.D. ± 129) was significantly higher than that of the
control group (506 p.eq/1; S.D.* 115)* The red cell FFA varied
widely in both groups.
The object of the experiments using I^l-ia^eiied oleic acid
was to determine whether the observed rise in red cell FFA values
in non-diabetic individuals was due to a transport of FFA from
plasma into erythrocytes. Oleic acid was chosen because it is
known that this acid comprises a considerable portion of the red
cell and plasma total FFA fraction in man (Schrade et al. i960).
After procuring a reasonable level of I1 -*1 activity in the FFA
component of the blood, glucose was administered by mouth and the
radioactivity of erythrocyte FFA measured at a time when the con- centration of FFA in this cell was at its height. It was argued
that if the increased level of FFA in the red cell after a carbo- hydrate load was due merely to a shift from plasma into the red
cell, then a rise in the specific activity of erythrocyte FFA
would be expected to occur. The results of these experiments are
shown in Table 1, where it can be seen that concomitant with the
rise in red cell FFA concentration after glucose in normal sub- jects there occurs a marked decrease in specific activity.
Page 68 of 168
t e x t f o r d e t a ils
CO U ) M CD
O 3 “ (i
U i H*
P cd H*
5 33 3 )
ct p . t-«
3 03 CO 3* M 9 CD
H- to 33
3 3> ct
S ' ct
CD cr CD
P P -
I - ro
d H" CD
CD CT* ro
H* C_j. ct H- O cd 3 3> o P H* P ct ct O o ro H*
• ct O O H* M 3> <
M 3 “
d •d ►d CD
3 ct ct
H 3 }
CD CO 3 CD>
OD —■3 •fr- U1 cd V : O o u> &
O ro U ) 00 • • • •
o uo C \
C t H-<
c t *<
ro ro 4=-
VO! U ) vo
cv ro ro
ro o j
*d 3> 3 H*o I
c t H*<p .
Page 69 of 168
That a close relationship exists between carbohydrate and
fat metabolism in general has been realized for many years. The
studies of Stetten and Boxer (1944) using deuterium-labelled
water, demonstrated that approximately 90^ of the retained glucose
fed to rats must be converted to fat. Masoro and coworkers (1949)
employing radioactive glucose clearly showed lipogenesis from
carbohydrate in the liver of the mouse, while Parvager and Gerlach
(1955) reported that 12 minutes after the administration of
labelled glucose to a rat as much as 3# of the total glucose given
was changed to fat. From the experiments reported here on normal
Individuals, a noteworthy increase In red cell FFA and triglyceride
concentration was observed 1 - 1 -| hours after the ingestion of
glucose. Since there are many factors influencing the turnover of
fatty acids, both esterlfied and non-esterifled, in the body it is
not possible to draw any definite conclusions concerning the meta- bolic fate of these substances based on concentration measurements
The results of the experiments with labelled oleic acid
however, suggest that the observed rise in erythrocyte FFA on in- gestion of carbohydrate was not due to transport from plasma to
red cell. The marked fall in red cell FFA specific activity while
the plasma FFA specific activity was rising is strongly suggestive
of new FFA being formed in this cell. Further evidence of metabo- lic activity in the normal human erythrocyte during a carbohydrate
load v/as the increase in total oxidized pyridine nucleotides (Fig.
III). Since these compounds are exclusively intracellular the
Page 70 of 168
question of a transport from plasma to red cells does not arise.
It has repeatedly been observed that in the diabetic animal
lipogenesis from carbohydrate is seriously impaired. Stetten and
coworkers (19^4,19^6) demonstrated that fatty acid synthesis from
glucose was practically abolished in the alloxan-diabetic rat and
also in diabetic rabbits. The inability of diabetic liver slices
to utilize glucose for the synthesis of fatty acids was reported
by Chemick and Chaikoff (1950). As can be seen from Pigs. IV, V
and VI, the erythrocyte of a diabetic person shows on the average
very little Increase in PPA, triglyceride and pyridine nucleotide
concentration after an oral glucose load. This defect in the case
of FFA could be corrected almost immediately by the intravenous
administration of soluble insulin given during a carbohydrate
tolerance test (Fig. VII). A similar effect has also been noted
by Chaikoff (1953) who reported that the inability of liver slices
from diabetic rats to incorporate glucose into fatty acid3 could
be corrected by prior administration of insulin in vivo for 1 to
3 days. The marked difference In behaviour between the normal
and diabetic erythrocyte with regard to an increase in triglyceride,
pyridine nucleotide and PPA concentration when presented with a
glucose load together with the dramatic response of the latter to
Insulin, suggests that this cell may reflect in its own metabo- lism the metabolic defect (s) which are known to occur in other
tissue cells of diabetic animals and probably also In the diabetic
Recently the non-esterifled fatty acid (FFA) fraction of
the blood has attracted considerable attention (Fredrickson and
Page 71 of 168
Gordon, 1958). The factors controlling the regulation of FFA
metabolism are being actively studied at present and an important
relationship between carbohydrate metabolism and the metabolism of
FFA has already been clearly established. Administration of glu- cose to a normal fasting animal or human subject produces a consi- derable drop in circulating plasma FFA concentration whereas no or
very little effect is observed in patients with severe diabetes
mellitus (Bierman et al. 1957a); (see also Figs. I and IV). The
reason for the fall in plasma FFA after administration of carbo- hydrate has not yet been fully elucidated. Bierman and coworkers
(1957b) and Gordon (1957) concluded from their studies with FFA-C1^
in animals that the drop in plasma FFA after glucose or insulin
was due to a decrease in the rate of flow of fatty acids into the
plasma mainly from adipose tissue. Results which could be inter- preted in a similar manner were obtained by Laurell (19 5 7) and
Fredrickson and Gordon (19 58).
Various tissues (e.g. myocardium, liver, skeletal muscle)
are capable of utilizing FFA directly for metabolic purposes if
carbohydrate Is not available. However, when an excess of glucose
is present they need no longer extract FFA from the plasma as the
cells themselves are able to synthesize fat from carbohydrate.
Consequently fatty acid mobilization from the fat depots to the
plasma is decreased (Gordon and Cherkes 1956; Dole 1956; Gordon
1957). The red cell may represent another example of a tissue
which can form new lipid from carbohydrate. A further point of
interest is the recent observation by Goodman (1958) demonstrating
the existence of specific receptor sites for long chain fatty acids
Page 72 of 168
on the red cell membrane, a finding which suggests that the ery- throcyte may also be able to utilize FPA directly for its metabo- lism. If this is in fact the case, then it is not unreasonable
to assume that the human red cell probably possesses a metabolic
behaviour pattern analogous, in at least some aspects of lipid
metabolism, to that found in other tissue cells.
The rise in red cell FFA reported here could simply be ex- plained as being due to the passage of FFA from plasma to erythro- cyte. However, if we accept the view that the decrease in plasma
FFA when carbohydrate is given is the result of a reduced outflow
of FFA from the fat stores, and if the red cell were unable to
synthesize lipid, then either a fall or no change in erythrocyte
fatty acids would be expected to occur. Furthermore, the marked
drop in red cell FFA I1 ^1 specific activity at a time when the
concentration of FFA is rising in this cell, again indicates that
the high levels of FFA in the erythrocyte after the ingestion of
carbohydrate probably comes from some source other than the plasma.
Although the results presented by these studies cannot as yet be
regarded as conclusive, it would appear that the normal human red
blood cell is capable of synthesizing lipid from glucose in vivo.
Further support for this statement is given by the work of Muir
et al. (1951) and Altman and coworkers (1951; 1953; 195^) already
discussed above. In contrast, the diabetic erythrocyte seems to
be unable to convert carbohydrate into lipid, a defect which can
be rapidly corrected by the administration of insulin in vivo.
Page 73 of 168
Some aspects of the metabolism of carbohydrate, lipid and
pyridine nucleotides by the human red blood cell in vivo have been
studied. Following the oral ingestion of glucose by a normal sub- ject there occurs an increase in erythrocyte triglyceride, pyridine
nucleotides and FFA concentration. In contrast, the red cell of
the diabetic patient shows a very much smaller rise In these para- meters during a glucose tolerance test; in many of the diabetic
cases studied no increase could be detected at all. This defect
in the diabetic erythrocyte in the case of FFA could be corrected
by the intravenous administration of insulin during the perform- ance of a carbohydrate tolerance test.
Although the findings do not permit any definite conclusions
to be drawn, it is suggested that the human red blood cell is able
to synthesize lipid from carbohydrate In vivo. This faculty seems
to be impaired in the human diabetic erythrocyte.
Page 74 of 168
Some interrelationships of carbohydrate, lipid and pyridine nucleo- tide metabolism in the human erythrocyte in vitro. A comparison
between diabetic patients and normal subjects.____________________
A. Studies in intact cells and haemolysates.
The difficulties that would be encountered in trying to
elucidate further the mechanisms responsible for the findings in
the human subject in vivo recorded in Chapter III are obvious. An
attempt was therefore made to determine whether these results
could be reproduced in vitro. It will be shown that the above
mentioned in vivo findings are under certain circumstances apparent- ly reproducable in vitro.
Initially, experiments were performed using washed red cells
in their own serum with only glucose added. In many of the cases
studied (both normal and diabetic) a significant increase In serum
PPA was noted during the incubation period which was probably due
to the presence in the blood of endogenous lipoprotein lipase
(Engelberg 1958) causing lipolysis of triglycerides with subse- quent liberation of FFA. Although no increase in erythrocyte FFA
was ever observed in these experiments, it was thought advisable
to eliminate serum as a suspending medium since it could be argued
that any increase in red cell lipid might be the result of direct
transfer from serum to erythrocyte. Therefore, in all subsequent
experiments the red cells were suspended in Krebs-RInger phosphate
buffer pH 7.5. Red cells incubated at 37°C for 3 to 5 hours in
buffer with only glucose added showed no increase in FFA. This
Page 75 of 168
applied both to cells from normal subjects and to those from
patients with diabetes mellitus, From Table II it is apparent
that in non-diabetic subjects there occurs a significant increase
in erythrocyte FFA concentration when in addition to glucose the
medium contained brilliant cresyl blue (BCB). After 5 hours in- cubation the average rise in red cell FFA was 434 jaeq/1 cells. In
all instances a detectable rise was observed within one hour.
Under the same experimental conditions and in contrast to normal
cells, the average increase in red cell FFA from diabetic patients
was 120 peq/1 cells (Table III). Of the diabetic cases studied
35fa showed no increase In erythrocyte FFA.
The possibility that the observed rise in red cell FFA
values might be due to lipolysis of more complex lipids occurring
within the cell with liberation of FFA was considered. Table IV
indicates that during incubation of normal cells with glucose and
methylene blue there is actually a small rise in total fatty acid
ester concentration. Preliminary experiments not reported here
have shown that a considerable portion of the total red cell ester
increase was made up by the triglyceride fraction, the rest being
due to phospholipid. The contingency that the Increase In FFA
occurred as a result of some lipolytic process within the cell was
thus thought to be unlikely. Fatty acid ester increment in dia- betic red cells under the same experimental conditions was very
The effect of addition of crystalline Insulin to the system
was then studied. Table V demonstrates the rise in FFA production
by diabetic red cells in phosphate buffer containing glucose and
Page 76 of 168
Number Concentration Increase in red cell FFA
of of FFA jieq/1 cells
1 hour 3 hours 5 hours
30 456 85.6 235 434
(120) (13.03) (67.3) (49.6)
The effect of incubating normal red blood cells suspended
in Krebs-Ringer phosphate buffer pH 7 .5 , containing glucose and
BCB on erythrocyte FFA concentration. Each flask contained 5-10
ml red cell buffer suspension; glucose 10 pinoles / ml; BCB 0.3
pioles / ml. Temperature 37°C. The figures represent mean values
with the standard deviation in parenthesis.
Page 77 of 168
Number Concentration Increase In red cell FFA
of of FFA
zero time yeq/1 cells
1 hour 3 hours 5 hours
35 570 ~ 42 120
(67.3) (15.9) (40.6)
The effect of incubating diabetic red cells suspended in
Krebs-Ringer phosphate buffer pH 7-5, containing glucose and
BOB on erythrocyte FPA concentration. Additions to each flask as
in Table II. The increase in red cell FFA during the first hour
was too small to be significant.
Page 78 of 168
Total red cell fatty acid ester
Zero time 5 hours
( 0 . 2 0 )
The effect of incubating normal red cells suspended in
Krebs-Ringer phosphate buffer pH 7.5* containing glucose and
methylene blue on erythrocyte total fatty acid esters. Additions
to each flask as in Table II. Methylene blue was used in the same
concentration as BCB.
Page 79 of 168
Number Increase in red cell FPA
of jueq / 1 cells
1 hour 3 hours 5 hours
15 63 256 412
(37.2) (42.6) (55.1)
The effect of incubating diabetic red cells suspended in
Krebs-Kinger phosphate buffer pH 7.5, with insulin, on erythro- cyte PPA concentration. Additions to each flask and explanation
of the Table as in Table II. Crystalline insulin 0.6 units / ml.
Page 80 of 168
BCB when insulin was added. Over a 5 hour incubation period the
observed increase was of the same order as that seen in normal red
cells without Insulin. Insulin caused no rise in both diabetic
and non-diabetic red cell FFA concentrations when BCB was omitted.
The action of insulin and BCB on normal cells was to produce
slightly higher FFA values after incubation for 5 hours than those
noted when insulin was excluded.
Table VI indicates the effect of BCB on both normal and
diabetic erythrocyte total oxidized pyridine nucleotides when in- cubated as described above. Only in the presence of the dye was
a significant Increase in pyridine nucleotide concentration of
normal cells observed, whereas under the same conditions diabetic
cells showed no significant rise in pyridine nucleotides.
The next stage in this study was to determine whether the
above noted findings required an intact cell or whether they could
be produced in essentially stroma free haemolysates. Using the
concentrations of hexokinase and ATP indicated in Table VII It was
observed that glycolysis proceeded somewhat more rapidly In
haemolysates than in intact cells. An even greater consumption of
glucose was noted in haemolysates on the addition of BCB. No in- crease in FFA occurred In reconstituted haemolysates incubated
without BCB. However, a significant rise in FFA concentration was
noted in normal haemolysates incubated with BCB. This rise was
much lower in diabetic haemolysates incubated under the same condi- tions (Table VII). When insulin was added to both normal and
diabetic haemolysates a significant increase in FFA concentration
occurred above that noted without the hormone (Table VIII).
Page 81 of 168
Number Addition Total oxidized pyridine nucleotides
of of / ml cells
Zero time 1 hour 2 hours 3 hours
Normal - 78 79 79
15 -i- 73 70 ,75 77
(3.4) (3.01) (2.6) (2.8)
The effect of brilliant cresyl blue on the concentration
of normal and diabetic erythrocyte pyridine nucleotides in vitro.
Conditions of incubation as in Table II. The figures represent
mean values with the standard deviation in parenthesis.
Page 82 of 168
ueq/ 1 cells
Increase in FPA ueq/1 cells
1 hr. 3 hrs. 4 hrs.
Normal - - - -
(1 8 .2 )
20 + 198
(1 6 )
A comparison of FPA production in normal and diabetic re- constituted haemolysates. The following concentrations were used:
ATP 1 junole / ml; hexokinase 0.4 mg / ml; glucose 10 umoles / ml;
BCB 0.3 jumoles / ml. Temperature 37°C. Mean values are given
with standard deviation in parenthesis.
Page 83 of 168
Increase in FFA ^eq/1 cells
jueq/1 cells 1 hr. 3 hrs. 4 hrs.
(2 1 )
effect of insulin on FPA production in normal and
diabetic reconstituted haemolysates in the presence of BCB.
Additions to the flasks and explanation of the Table as in Table
VII. Crystalline insulin 0.5 units / ml.
Page 84 of 168
B. Studies with inhibitors of glycolysis and the hexose monophos- phate shunt.__________________________________________________
Previous workers (Huennekens et al. 1957; Brin and Yonemoto
1958) have demonstrated a marked stimulation of the glucose oxida- tive pathway in human red blood cells in the presence of catalytic
quantities of redox dyes like methylene blue or brilliant cresyl
blue (BCB). From the results reported here it is apparently only
when this aspect of carbohydrate metabolism is stimulated in red
blood cells that increased cellular FFA values are observed.
Moreover, since the production of FFA in diabetic cells was: impair- ed, this observation suggested the possibility of a block occurring
Somewhere in the pentose phosphate shunt of diabetic erythrocytes.
For this reason an inhibitor of the glucose oxidative pathway was
sought to determine whether in its presence the increased FFA
values in normal cells incubated with BCB could be prevented.
Neufeld et al. (1955) showed that 2 '-adenylic acid was a potent
inhibitor of yeast glucose-6-phosphate dehydrogenase. The effect
of this compound on FFA production and glycolysis in normal ery- throcytes was thus determined (Table IX). It can be seen that in
the presence of 2*-adenylic acid FFA production in normal red cells
was effectively inhibited while glycolysis was decreased by 25$.
In a further series of experiments the influence of an inhibitor
of the anaerobic glycolytic sequence on FFA production by normal
red cells was determined. Table X shows the effect of sodium fluo- ride on FFA production and glycolysis in normal red cells incubated
with BCB. In the presence of BCB together with fluoride normal
erythrocytes consistently utilized a small (18$) amount of glucose
Page 85 of 168
TABLE D l
Increase in FFA
pteq/ 1 cells
1 hr. 3 hrs.
mmoles / 1
cells / 3 hrs.
- 98 215 6.84
(o - 23)
The effect of 2'-adenylic acid on FFA production and glycolysis
in normal intact red cells Incubated with BCB. Conditions of incu- bation as in Table II. 0.60 pmoles 2'-adenylic acid was added per
ml red cell suspension. Mean values are given. Figures in paren- thesis indicate range of values.
Page 86 of 168
ueq/ 1 cells
Increase in FFA
1 hr. 3 hrs.
mmoles / 1
cells / 3 hrs.
- 84 201 6.01
112 287 1 .1 2
The effect of sodium fluoride on PFA production and glycolysis
in nomal intact cells incubated with BCB. Conditions in incubation
as in Table II. Concentration of fluoride was 0.025M. Mean values
Page 87 of 168
while at the same time FFA production was significantly higher than
with BCB alone.
C. Studies on intact cells and haemolysates employing radioactive
The in vitro findings discussed above, while not presenting
unequivocal proof, strongly suggested that the BCB stimulated nor- mal human erythrocyte was able to synthesize fatty acids from carbo- hydrate, The final stage of this study was undertaken to determine
the validity of this statement. The techniques which were used are
described in Chapter II.
In Table XI Is presented data on the Incorporation of glucose- U-C1^ into erythrocyte lipids of fasting normal subjects when incu- bated in phosphate buffer containing labelled glucose but without
the addition of BCB. Although the specific activity of the isolated
fatty acids was lour, in all the cases studied activity was observed.
When BCB was added to erythrocytes from fasted normal subjects and
incubated as described above, Isotope incorporation into red cell
fatty acids increased considerably (Table XI).
Erythrocytes obtained from fasting, Insulin deprived, dia- betic patients and treated with BCB under similar conditions des- cribed for normal red cells, demonstrated a significantly lower
specific activity in their fatty acids when compared with normal
erythrocytes (Table XII). Included in this Table are the results
on a depancreatized male patient whose red cells also demonstrated
a much lower capacity to synthesize fat from carbohydrate than
Page 88 of 168
Number Addition of Total fatty acids.
of BCB Specific activity
cases cpm / mg
7 - i7
(14 - 24)
10 4- 165
(150 - 175)
The effect of BC3 on the incorporation of glucose-U-C
into normal erythrocyte lipids. Each flask contained 5 - 10 ml
erythrocyte suspension in Krebs-Ringer phosphate buffer pH 7.5;
10 pmoles labelled glucose (specific activity 0.04 uc / umole;
BCB when used was added in a concentration of 0.3 umoles / ml.
Incubation in air at 37°C for 3 hours. Mean values are given
with the range of results in parenthesis.
Page 89 of 168
Total fatty acids.
cpm / mg
(46 - 80 )
The incorporation of glucose-U-C1^ into diabetic red cell
lipids. BCB added to the erythrocyte suspension.
Conditions of incubation a3 in Table XI. ^Results obtained
on erythrocytes from a depancreatized male patient.
Page 90 of 168
Table XIII indicates the effect of insulin added to diabetic
red cells in vitro. The ability of diabetic erythrocytes to syn- thesize fat from glucose was considerably enhanced by Insulin. In
three of the cases studied the specific activity of the isolated
lipids fell within the normal range as defined in this study.
Since NADPIi has been shown to be necessary for some of the
more Important reactions involved In the biosynthesis of fatty
acids in mammalian tissues, It was considered important to deter- mine whether this requirement for NADPII applied to normal human
haemolysates as well. As is indicated in Table XIV reconstituted
haemolysates were able to synthesize fat from glucose only in the
presence of BCB. Furthermore, the addition of a NADPII generating
system such as isocitrate, NADP and Isocitric dehydrogenase brought
about approximately a 6Qfo increase in the incorporation of radio- active glucose into fatty acids in reconstituted haemolysates.
The contribution of leukocytes to the observed biosynthesis
of lipids in erythrocytes was studied as follows: blood was ob- tained from a healthy donor and divided into two portions. One
part was prepared as described above to give an erythrocyte sus- pension poor in leukocytes; the second portion contained the origi- nal quantity of white cells present in the specimen of blood 1J.e.
an erythrocyte suspension rich In leukocytes. Further treatment
of the two samples was the same as that described for a red cell
suspension poor in leukocytes. The results of this experiment are
recorded in Table XV.
Page 91 of 168
Number of Addition of Total fatty acids.
cases Insulin Specific activity
cpm / mg
15 - 60
(46 - 80)
12 4* 136
(120 - 135)
The effect of insulin in vitro on the incorporation of
glucose-U-C1^ into diabetic red cell lipids. BCB added to the
The conditions of Incubation as in Table XI. 0.6 units per
ml red cell suspension was added in each case.
Page 92 of 168
Number of Addition Addition Total fatty acids,
cases of BCB of NADPH Specific activity
cpm / mg
4* •f 293
The effect of BCB and NADPH on the incorporation of
glucose-U-C1^ into normal haemolysate fatty acids. Conditions
of incubation as in Table XI. For the generation of NADPH the
following were added per ml haemolysate:- 10 pmoles Isocitrate;
1 umole NADP; 1 mg isocitric dehydrogenase. Mean values are given.
Page 93 of 168
Total fatty acids.
cpm / mg
Leukocyte poor red cell
suspension. No BOB added 23
Leukocyte rich red cell
suspension. No BOB added 145
Leukocyte poor red cell
suspension. BOB added 157
Leukocyte rich red cell
suspension. BOB added 163
The influence of leukocytes on the incorporation of glucose- U-C1^ into normal red cell lipids. Addition to each flask are
described in Table XI. Incubation time was 2|- hours.
White cell count of leukocyte rich specimen:- 4000 / cmm.
White cell count of leukocyte poor specimen:- 6 / cmm.
Red cell count of each specimen:- 5.2 million / cmm.
Page 94 of 168
The findings observed previously (Chapter III; Mendelsohn
19ola) when studying the metabolism of lipids by the human red
blood cell in vivo, namely, the increase in red cell FFA concen- tration in the normal subject during a carbohydrate tolerance
test, the very much smaller rise in the erythrocyte of the insulin
deprived diabetic patient under the same conditions and the correc- tion of the defect in the diabetic red cell when insulin was ad- ministered parenterally to the patient, have now all apparently
been reproduced in vitro. Under the influence of BCB an increased
concentration of total oxidized pyridine nucleotide has also been
It has been noted above (see Chapter I) that mature mamma- lian erythrocytes suspended either in their own plasma or in an
artificial buffered medium, are able to synthesize lipid only to.: a
limited degree from C1^-acetate (Rowe et al. i960; Altman 1953).
The present investigations using glucose-U-C1^ as substrate, the
results of which are depicted In Table XI, support these observa- tions since the normal erythrocyte fatty acids were only slightly
radioactive. However, the addition of small quantities of BCB to
a suspension of normal red cells caused a considerable increase in
the specific activity of the total cell fatty acids, Indicating
an ability to sjaitheslze lipid from glucose.
In 1928, Barron and Karrop made the interesting observation
that the very small oxygen consumption of mature human erythrocytes
could be Increased many fold in the presence of catalytic amounts
of certain dyes which possess the property of being reversibly
Page 95 of 168
oxidized and reduced e.g. methylene blue, cresyl blue, toluylene
blue etc. The mechanism whereby these dyes increase the respira- tory metabolism of the mammalian erythrocyte has not yet been
fully elucidated. It is known that methylene blue stimulates
pentose phosphate shunt activity in mammalian red blood cells
(Brin and Yonemoto 1958; Huennekens et al. 1957). With the opera- tion of the pentose phosphate cycle under normal conditions a fur- ther source of potential energy would become available to the red
cell in the fora of NADPH. Recent work has revealed the impor- tance of an adequate supply of NADPH which serves as a source of
hydrogen for reductive synthesis of fatty acids (Langdon 1957;
Porter et al. 1957; Titchener et al. 1958). It has also been
suggested by Siperstein (1959) that the activity of glucose meta- bolism over the pentose phosphate pathway with subsequent genera- tion of NADPH may be an Important factor In regulating tissue
lipogenesis. Prom the present studies, it would seem reasonable
to conclude that there exists an intimate relationship between
pentose phosphate shunt activity and lipid metabolism in the
cresyl blue treated normal red cell, since it is only when this
aspect of carbohydrate metabolism is stimulated that a biosyn- thesis of fatty acids from glucose Is observed. However, Huennekens
et al. (1957) have postulated that BCB stimulates the pentose
phosphate cycle in human red cells as follows
-e -e -e -e
substrate-- -> NADPH ---> Plavoprotein -- ? BCB *--> 02
Under these conditions it is unlikely that the H+ from NADPH pro- duced during shunt activity would be utilised directly for fat
synthesis. It would appear therefore, that the relationship
Page 96 of 168
between lipogenesis and pentose phosphate shunt activity in the
3CB stimulated erythrocyte is probably an Indirect one.
The data presented In Table XII demonstrates that red blood
cells from diabetic patients synthesize much less lipid when com- pared with normal erythrocytes under the same experimental condi- tions. A similar finding of a defect In lipogenesis from carbo- hydrate In diabetic animals has been noted by numerous workers.
For example, an Impairment In fat synthesis from carbohydrate
precursors in diabetic rats was observed by Drury (1940) and
Stetten et al. (1944;1946) while a number of investigators (Chemick
and Chaikoff 1950; Brady and Gurin 1950; Hausberger et al. 1954;
Wong and van Bruggen i960) have shown that liver slices and adipose
tissue from diabetic animals were unable to utilize glucose for
the biosynthesis of fatty acids. Concerning the diabetic erythro- cyte, many workers have shown that glycolysis via the Embden- Meyerhof pathway Is normal In this cell (Tolstoi 1924; Cajori and
Crouter 1924; Katayama 19 26). It is of interest therefore to
speculate on the findings observed here which indicate that the
metabolic block in the diabetic erythrocyte may occur in the
pentose phosphate pathway. Support for this theory is given by
the investigations of Sonka et al. (1957) who noted that the red
cells of patients with severe diabetes deprived of insulin showed
a very much reduced pentose phosphate shunt activity. An abnor- mality of the oxidative metabolism of glucose In the erythrocytes
of alloxan diabetic rats has also been reported (Ferrari and
Gastaldi i960). An interesting report has recently appeared
(long and Carson 19 6 1) which demonstrates that human diabetic
Page 97 of 168
erythrocytes have a significantly higher activity of glutathione
reductase when compared with normal red cells. An elevated acti- vity of this enzyme has also been shown in tissues (muscle, liver)
of alloxan diabetic rats (Langdon and Mize 19 6 1). Prom the re- ports of Schrier et al. (1958) and Couri and Packer (1959) it would
appear that alterations of glutathione reductase activity are
associated with lesions in the oxidative breakdown of glucose.
However, in view of the recent findings (Matthes et al. i960;
Gibson et al. 1980) that the llpogenetlc abnormality in diabetic
animals resides in the activity of enzymes directly concerned with
fatty acid synthesis, the possible occurrence in the human diabe- tic red blood cell of a defect in this metabolic system has also
to be considered. 1
It should be emphasized that the erythrocytic aberration in
lipogenesis from glucose was only observed In patients with severe
diabetes deprived of insulin. Red cells of the obese, elderly
subject whose diabetes could be controlled by limiting the dietary
caloric intake, or diabetic patients adequately controlled by
insulin showed no significant difference in lipid synthesis when
compared with erythrocytes of normal subjects in the same age
group. Vallance-Owen et al. (1955) showed that plasma insulin
levels were normal in both the mild obese group of diabetics as
well as in patients satisfactorily controlled by insulin, whereas
the uncontrolled insulin requiring diabetics had no measurable
plasma insulin activity. It would seem therefore that a defi- ciency of Insulin in the blood is necessary before the above men- tioned defect In red cell lipogenesis can be observed in vitro.
Page 98 of 168
This is further borne out by the findings in one depancreatized
patient used in the present study whose erythrocytes also demon- strated a depressed synthesis of fat from glucose in vitro (Table
The findings with reconstituted haemolysates indicate that
the metabolic reactions responsible for the above in vitro results
all occur within the cell and are not due to any differences In
membrane phenomena. The fact that reconstituted, essentially
stroma free haemolysates exhibited a metabolic behaviour pattern
analogus to the intact cell suggests that this cell free system
could be utilized with advantage when studying the cause of meta- bolic abnormalities such as diabetes mellitus in the human subject.
This would be true at least with respect to certain aspects of
carbohydrate and lipid metabolism. The increased lipogenesis from
carbohydrate observed In the erythrocyte of the diabetic patient
as a result of the action of Insulin in vitro, is probably com- parable to Its effect in restoring fatty acid biosynthesis in
certain tissues of diabetic animals (Matthes et al. 1980; Gibson
et al. i960). The present Investigation indicates that human
diabetic red cells and the tissues of diabetic animals behave in
a similar manner with respect to a defective lipogenesis from
carbohydrate and the action of insulin in restoring this abnor- mality. It Is suggested that the diabetic red cell may well re- flect in its own metabolism one of the abnormalities which pro- bably occurs In other tissues of the human diabetic i.e. defective
fat synthesis from carbohydrate. An easily available cell.would
thus seem to be at hand for the study of this problem, and valu
Page 99 of 168
able Information might be obtained concerning the site (s) of the
enzymatic defect (s) present in the patient suffering from diabetes
It has been stated that the human red blood cell is unres- ponsive to the action of insulin both in vivo and in vitro (RenoId
and ¥inegrad i960). The results of the present study do not sup- port this statement since an effect of this hormone on the human
erythrocyte has been demonstrated under both in vivo and in vitro
conditions (see also Mendelsohn, 196ld). A further point of in- terest is the finding that insulin has been shown to exert an
action on lipid metabolism in both normal and diabetic human hae- molysates. While the vast majority of the work on the mechanism
of action of insulin indicates that this hormone produces its
effect primarily on cell permeability to glucose and on the enzyme
hexokinase, a growing body of evidence is accumulating which
demonstrates that these theories of insulin action are inadequate
to explain all the diverse known effects of this hormone (Beloff- Chain and Pocchiari i960). The latter authors have advanced the
hypothesis that insulin must have a more general action resulting
in the stimulation of a number of synthetic processes, not always
involving glucose. They suggested that insulin may act by rais- ing the energy potential of the cell, enabling it to perform a
larger amount of energy requiring metabolic reactions. These
workers have also proposed that the NADP/MADPH system Is specifi- cally affected by insulin when they referred to the function of
NADP/NADPH in fat synthesis and In the oxidative pathway of glu- cose metabolism and discussed the known insulin effects on these
Page 100 of 168
reactions. Further support for the hypothesis that the NADP/MADPH
system is influenced by insulin is given by the work of McLean
(1959) who showed, using rat mammary gland, that the stimulation
by insulin of a reductive synthetic process utilizing NADPH, such
as fatty acid synthesis, could lead to a secondary increase in
glucose oxidation via the pentose phosphate pathway. The same
effect could also be obtained with an artificial electron acceptor
(phenazine methosulphate) and it was suggested that the increased
oxidation of glucose by the pentose phosphate shunt using either
insulin or phenazine methosulphate was mediated by the rate of re- oxidation of NADPH. Essentially the same results have been demon- strated here. It has been shown using the human red blood cell
that only in the presence of an artificial electron acceptor (BCB),
which causes stimulation of the shunt pathway does any biosynthe- sis of fatty acids occur. Furthermore, in normal reconstituted
haemolysates containing BCB, insulin is able to increase both FFA
production and incorporation of labelled glucose into fatty acids.
The present results therefore, would tend to confirm the postu- lates of Beloff-Chain and MeLean that insulin might influence and
perhaps control certain inportant facets of lipid metabolism in
the cell directly.
Because certain investigators (Mark3 et al. i960,* Buchanan
i960; Rowe et al. i960) have shown that the leukocyte possesses a
much greater capacity for fat synthesis than does the mature red
blood cell, it became necessary to consider the influence of the
former type of cell on erythrocyte lipid biosynthesis observed in
this study. The results In Table XV Indicate that while the white
Page 101 of 168
cells contributed very significantly to the lipid synthesized in
whole blood preparations in the absence of BCB, they played no role
in the stimulation of fat synthesis in whole blood on addition of
the dye. In fact, the above results are consistent with the in- terpretation that BCB substantially inhibits lipogenesis from
carbohydrate in human leukocytes, and this system might then offer
a convenient method of differentiating between fat synthesis in
the two types of blood cells. Of interest in this regard are the
findings of Cahill et al. (1958). Certain oxidation reduction
mediators (pyocyanln, methylene blue) were observed to inhibit the
incorporation of either labelled glucose or pyruvate into fatty
acids by normal rat livers. It can therefore be concluded that
under the in vitro conditions employed in the present investigation
the contribution if any of the limited number of leukocytes to fat
synthesis in the BCB stimulated erythrocyte system can be ignored.
The metabolism of pyridine nucleotides in the mammalian erythrocyte.
Earlier reports In the literature have described the synthe- sis of pyridine nucleotides from added nicotinic acid by human red
blood cells (Kohn and Klein 1939i Klein and Kohn 19^0). This syn- thesis, which occurred both In vivo and in vitro, was demonstrated
with nicotinic acid but not when nicotinamide was employed (Handler
and Kohn 19^3). In none of these studies however, was the synthe- sized material properly identified due to Inadequacy of the micro- biological methods used. Some years later Leder and Handler (1951)
incubated human erythrocytes in a medium containing phosphate,
glucose and 20 nicotinamide, and were able to show that 75 - 900
Page 102 of 168
of the synthesized pyridine nucleotide was nicotinamide mononucleo- tide, the remainder being due to NAD. Synthesis of pyridine nucleo- tides after the ingestion of nicotinic acid and tryptophan res- pectively was observed in human red blood cells by Duncan and
Sarett (1951). Hofmann (1955) proved that rabbit blood haemoly- sates were able to synthesize NAD from nicotinamide and ATP; the
presence of a NAD-nucleosidase was also found in the red cells of
this species (see Malkin and Densdedt 1956). Further work by
Preiss and Handler (1958a; b) indicated that human erythrocytes
could convert nicotinamide and nicotinic acid to NAD via separate
The results of the present study indicate that the normal
human red cell is able to synthesize pyridine nucleotides both in
vivo and in vitro when utilizing carbohydrate as substrate. The
defective synthesis of these compounds in diabetic erythrocytes is
probably of Importance in accounting for certain metabolic abnor- malities (e.g. inability to synthesize lipid from carbohydrate)
which occurs in these cells.
Speculation on a possible alternative pathway of carbohydrate
metabolism in the human red blood cell.______________________
The results using inhibitors of both oxidative and non-oxida- tive phases of carbohydrate metabolism in erythrocytes present an
interesting topic for discussion. The addition of 2 '-adenylic
acid to normal human erythrocytes incubated with BOB produced two
noteworthy effects. First, an almost complete inhibition of FFA
production was noted and second, glucose utilization was decreased
Page 103 of 168
by about 25fo. Since 2 ’-adenylic acid has been shown to inhibit
yeast glucose-6-phosphate dehydrogenase it is tempting to assume
that it is acting similarily in the red cell. A point in favour
of this assumption is that the magnitude of the drop in glycolysis
coincides very well with the increased amount of glucose which is
metabolized via the oxidative pathway in the BCB or methylene blue
stimulated erythrocyte (Carson I960; Charlton and Mendelsohn 19 6 1).
The fact that FFA production was almost abolished in the presence
of 2 '-adenylic acid could point to a relationship between pentose
phosphate shunt activity and lipid metabolism in normal human
When fluoride alone is added to mammalian red cells suspended
in buffer with glucose as substrate anaerobic glycolysis is for all
practical purposes completely inhibited. Table X demonstrates that
in the presence of BCB, the fluoride inhibited red cell now ex- hibited a rate of glycolysis of about 15 - 20£. Since BCB stimu- lates the pentose phosphate shunt this quantity of glucose must be
metabolized along this pathway in the fluoride treated red cell.
Under these conditions it can be seen that in normal red cells FFA
production is significantly Increased.
Although the above studies using inhibitors of the two phases
of carbohydrate metabolism in normal human erythrocytes are not
entirely conclusive, they do suggest that an adequately function- ing pentose phosphate shunt is necessaryr for normal fat metabolism
in this cell.
A further point of interest arises out of the findings ob- tained in the BCB stimulated red cell incubated with fluoride.
Page 104 of 168
According to present day concepts the reactions of the pentose
phosphate shunt are considered to be cyclic In nature. The over- all reactions may be summarized as follows :
3glucose-6-phoephate ------ > 2triose phosphate +■
lfructose-6-phosphate 4- 3CC>2 .
The triose and hexose phosphates then enter the glycolytic sequence
or are converted back to glucose-6-phosphate. Since fluoride in- hibits enolase, the enzyme which converts 2-phosphoglycerate to
2-phosphoenolpyruvate, the triose and hexose phosphates formed dur- ing a cycle of the pentose phosphate shunt are unable to be meta- bolized via the glycolytic pathway to pyruvate and are probably
recycled through the oxidative route. However, for the biosynthe- sis of fatty acids to take place the production of pyruvate is
necessary. The latter is oxidatively decarboxylated to acetyl-Co
A from which fatty acids are formed by a series of reductive reac- tions. It therefore became of interest to determine whether the
BCB stimulated erythrocyte in the presence of fluoride was able to
produce pyruvate. The results of six such experiments are pres- ented in Table XVI where it can be seen that under the influence
of BCB, the fluoride inhibited red cell Is In fact able to produce
both lactate and pyruvate. Since the anaerobic glycolytic pathway
to pyruvate and lactate has been blocked in fluoride, it would
seem that a possible explanation of the Results is the fact that
the human erythrocyte possesses another means of forming lactate
and pyruvate from glucose which can bypass the enolase stage. Of
interest in this regard are the experiments recently reported by
Rappoport et al. (1961). Using haemolysates prepared from rat
Page 105 of 168
Number Addition Production after 3 hours incubation
of of mmoles / 1 cells
4“ 3.5 1.07
The effect of sodium fluoride on the production of lactate
and pyruvate in normal erythrocytes. Conditions of incubation as
in Table II. Mean values are given.
Page 106 of 168
erythrocytes, they showed that even in the presence of fluoride,
ribose-5-phosphate could be metabolized to lactate.
It might be of interest here to discuss briefly some possi- bilities for a third pathway of carbohydrate metabolism in mamma- lian red blood cells. Glyoxylic and glycollic acids have been
shown to be precursors of glycine in the rat, and the conversion
of glycolaldehyde to glycine has also been indicated (Weinhouse
and Friedmann (1951); Weissbach and Sprinson (1953). Furthermore,
it is known that "active" glycolaldehyde participates in the meta- bolism of carbohydrate by the pentose phosphate shunt. The latter
compound could be converted via glycolic and glyoxylic acids to
glycine which is then converted to pyruvate via serine:
glucose-6- p h o s p h a t e -----> ribose-5-phosphate — --
ribulose-5-ph o s p h a t e---- > glyceraldehyde-3-phosphate
glycollic acid — glyoxylic acid — > glycine ---^
serine — -> pyruvic acid.
It is interesting to note that Komberg and Sadler (i960) have
demonstrated a dicarboxylic acid cycle In microorganisms accord- ing to the following scheme:
— ,— --- glyoxylate
—> acetyl CoA malate
f I I 1
•~> pyruvate ^
All of the above intermediate compounds i-fith the exception of
glycollate and glyoxalate are knovrn to be present in mammalian red
Page 107 of 168
A further series of reactions could be postulated as follows:
glucose-6-phosphate -------> g l y c i n e ------ > threonine
aminoacetone ------methylglyoxal lactate
The reactions from glycine to aminoacetone have been demonstrated
in micro-organisms (Jleister 1957; Neuberger and Tait i960).
Threonine and aminoacetone are to be found in human erythrocytes,
while the conversion of aminoacetone to methylglyoxal has recently
been demonstrated in ox plasma by Elliot (i960). This metabolic
sequence would thus suggest a reason for the active glyoxalase sys- tem which occurs in the mammalian red blood cell, and for which no
known function has as yet been postulated.
Page 108 of 168
A considerable increase in the concentration of FFA and
total oxidized pyridine nucleotides was observed in normal human
erythrocytes when they were incubated in a phosphate buffer con- taining glucose and BCB in vitro. This effect only occurred when
small quantities of dye were added to the red cell buffer suspen- sion.
A much lower rise in the level of FFA was apparent in the
erythrocyte of the insulin deprived diabetic patient when compared
with the normal.
Addition of insulin to the diabetic red cell suspension
corrected this deficiency.
Ivhen diabetic red cells were incubated with BCB and glucose
no significant increase in total oxidized pyridine nucleotides was
The above effects with respect to FFA were also observed in
reconstituted haemolysates of normal and diabetic erythrocytes.
The biosynthesis of lipids in the human erythrocyte in vitro
was studied with the aid of uniformly labelled glucose.
Normal red cells when incubated in phosphate buffer with
labelled glucose exhibited only a slight tendency to synthesize
The addition of BCB to normal red cells markedly increased
their capacity to synthesize lipid from carbohydrate.
Diabetic red cells when Incubated in phosphate buffer with
labelled glucose demonstrated a significantly lower ability to
synthesize fat from glucose when compared with normal red cells.
Page 109 of 168
In the presence of insulin added in vitro the efficiency of
lipogenesis from carbohydrate In diabetic red cells wa3 restored
to approximately the same degree shown by normal red cells.
The implications of the above findings are discussed with'
reference to the possible aberrant metabolic processes in human
subjects with diabetes mellitus.
Prom the results with inhibitors of various aspects of car- bohydrate metabolism some novel alternate pathways of glucose meta- bolism are suggested in the human erythrocyte.
Page 110 of 168
CERTAIN ASPECTS OF CARBOHYDRATE AND LIPID METABOLISM
IN PRIMAQUINE SENSITIVE ERYTHROCYTES
The genetically transmitted defect known as primaquine sen- sitivity is widely distributed throughout the world, and is found
in the South African Bantu with a frequency of 2-3$ (Charlton and
Bothwell, 1961). Among the metabolic derangements manifested by
these cells is a deficiency of the enzyme glucose-6-phosphate de- hydrogenase, the first enzyme in the pathway of oxidative metabo- lism of glucose (Carson, i960). This hexose monophosphate shunt
pathway is believed to be the only source of NADPH In the erythro- cyte, and primaquine sensitive cells have been shown to be defi- cient in reactions linked to NADPH (Alving et al., i960). The
results already presented in this thesis have indicated that a
relationship exists between the oxidative metabolism of glucose
and lipid biosynthesis in the normal mature cresyl blue stimulated
human erythrocyte. Because of the well-documented deficiency of
glucose-6-phosphate dehydrogenase and increased NADP/NADPH ratio
(Schrier et al., 1958) In primaquine sensitive erythrocytes, it
wa3 thought of interest to assess the performance of these abnor- mal cells In terms of carbohydrate and lipid metabolism using the
in vitro techniques described previously.
A comparison of the alteration in glucose utilization pro- duced by BCB in normal red blood cells and primaquine sensitive
Page 111 of 168
c h a n g e in g lu c o s e u t il iz a t io n
A comparison of normal (closed circles) and primaquine
sensitive (open circles) erythrocytes incubated with
glucose and BOB showing the effect of the dye on glucose
utilization and formation of pyruvate.
PYRUVIC ACID mg
Page 112 of 168
A comparison of the changes in erythrocyte FFA and total
fatty acids (TFA) in normal (N) and primaquine sensitive
(D) cells incubated with glucose and BOB. The figures
represent average TFA specific activity.
Page 113 of 168
erythrocytes is shown in Figure VIII. The effect of the dye was
to Increase the consumption of glucose in normal red cells by an
average of 50/2, ivhereas utilization was decreased in enzyme defi- cient cells by an average of 17/2. Pyruvic acid production after
incubation of normal and abnormal cells with BOB is also indicated
in Figure VIII. In a proportion of cases exhibiting the enzyme
deficiency a significantly greater accumulation of erythrocyte
pyruvic acid was noted when compared with normal cells.
The concentration of erythrocyte FFA increased when normal
cells were incubated with the dye. In contrast, enzyme-deficient
cells produced no significant increase In the level of FFA, and
in fact, in some instances a fall in the concentration was observed.
There was also a marked difference between normal and primaquine
sensitive erythrocytes with regard to total red cell fatty acids.
On Incubation with brilliant cresyl blue an Increase in the con- centration of total fatty acids In normal cells occurred. In the
enzyme-deficient cell on the other hand, a pronounced decrease in
total fatty acids was demonstrated. These results are shown In
When uniformly labelled glucose was employed, radioactivity
was detected in the total fatty acids of both normal and sensitive
cells. The average specific activity of the fatty acids was some- what higher in the case of the sensitive cells (Figure IX).
The stimulation of oxygen consumption and glucose utilization
in normal erythrocytes by dyes such as methylene blue and brilliant
Page 114 of 168
cresyl blue Is well known (Harrop and Barron, 1928). The decrease
in glucose consumption produced by BCB in primaquine sensitive
cells was constant, and represents a marked difference from the
behaviour of normal cells. Similar observations have recently been
reported by Carson (i960).
Murphy (i960) has estimated that only 10 - 20$ of the glucose
metabolized by normal cells at pH 7 .5 passes via the shunt pathway,
and it is reasonable to expect that even less follows this route
in enzyme deficient cells. The degree of inhibition observed in
these studies (averaging 17$) appeared therefore to be too great
to be accounted for only by inhibition of the shunt. Johnson and
Marks (1953) have reported on the consumption of oxygen and the
14 14 formation of C 02 from C -1-glucose by primaquine sensitive cells
in the presence of methylene bluie. Oxygen utilization was found
to be 35$ and C02 production 50$ of normal. In the absence of an
intact tricarboxylic acid cycle In the mature red cell, the only
source of C ^ 0 2 from C^-l-glucose is the hexose monophosphate
shunt, and these observations imply that the shunt does function
in primaquine sensitive cells incubated with the dye. Under the
experimental conditions reported here, therefore, it would seem
that oxidative metabolism does take place, albeit at a lower level
than in normal erythrocytes. The overall diminution in glucose
utilization cannot thus be due solely to the action of the dye on
the shunt pathway. It must therefore be postulated that the dye
produces an inhibition of anaerobic glycolysis. Murphy (i960) has
recently shown that Embden-Meyerhof glycolysis was in fact inhi- bited by methylene blue even in normal red cells. Since the in
Page 115 of 168
crease In metabolism via the shunt pathway produced by the dye in
normal cells was greater than the inhibition of anaerobic glycoly- sis, the overall effect was to increase the consumption of glucose.
In primaquine sensitive cells, however, the shunt pathway is de- fective. Even though oxidative metabolism was stimulated by the
dye, the increase in glucose utilization via this pathway was not
as great as the diminution in anaerobic metabolism, with the result
that the net effect was an overall decrease in the amount of glu- cose metabolized.
In a proportion of cases more pyruvic acid accumulated when
primaquine sensitive cells were incubated with BCB than when normal
cells were used. Although the amount of pyruvic acid formed by
cells from the same individual on different occasions was reason- ably constant, there appeared to be a wide range of values in both
normal and enzyme deficient cells. If the difference is real, its
significance is not obvious, although a possible explanation would
be increased lactic dehydrogenase activity in primaquine sensitive
cells. This has in fact been reported by Larizza et al. (1958),
although Johnson and Marks (1958) found no difference from the nor- mal.
The evidence for a disordered lipid metabolism in the prima- quine sensitive cells as defined in this study seems clear.
Normal erythrocytes incubated with glucose and BCB showed a net
increase in both total fatty acids and in FFA. In addition, the
incorporation of carbon from uniformly labelled glucose into ery- throcyte lipids during incubation with the dye further confirmed
the fact that there had been a net synthesis of fat from glucose.
Page 116 of 168
Although this incorporation of radioactivity into red cell lipid3
also occurred in glucose-6-phosphate dehydrogenase deficient cells,
little or no increase in the FFA fraction together with a pronounced
fall in total fatty acids was noted. The radioactivity observed in
the various lipid fractions of the primaquine sensitive cells was
evidence of some synthesis of fat, hut the decrease in total cell
fatty acids must mean that there had been a greater degree of cata- bolism than of anabolism. Of interest in this regard is the re- cent observation of Krachovec et al. (19 6 1) that rat erythrocytes
were able to oxidize palmitic acid. The ability of human erythro- cytes to metabolize fatty acids is being Investigated at the pres- ent time by the author.
This aspect of aberrant metabolism in primaquine sensitive
cells may be of importance in accounting for their increased sus- ceptibility to haemolysis by a variety of agents in vivo. Although
information has rapidly accumulated in recent years, it is not yet
clear which, if any, of the several abnormal aspects of metabolism
already recognized in these cells is directly related to the sus- ceptibility to haemolysis, 3ince there appear to be valid objec- tions to each (Carson i960). Danon et al. (19 6 1) have reported
morphological differences on electron microscopy between the cell
membranes of primaquine sensitive erythrocytes and normal cells,
and it may be that a lipid component of the membrane is defective.
Lohr and Waller (19 58) found no abnormality in stromal lipid
fractions of erythrocytes from a case of congenital nonspherocytic
haemolytic anaemia in which the deficiency of glucose-6-phosphate
dehydrogenase was even more marked than in primaquine sensitive
Page 117 of 168
cells. However, Alving efc al. (i960) reported that the total red
cell lipids of primaquine sensitive individuals decreased during
the acute haemolytic episode produced by administration of prima- quine, and that the return to previous values was very slow. In
the present investigation a pronounced fall In total fatty acids
of primaquine sensitive cells Incubated in vitro with BCB has been
demonstrated. Of Interest In this connection is the recent find- ing that methylene blue is able to cause haemolysis in sensitive
individuals (Carson i960), while Szeinberg et al. (19 6 1) have
demonstrated that many of the compounds which cause haemolysis In
sensitive patients stimulate pentose phosphate shunt activity in
the red blood cell. In view of the relationship between pentose
phosphate shunt activity and lipid metabolism brought out in this
thesis, it seems possible at least that the disorder of lipid
metabolism observed in the BCB stimulated sensitive cell is of
Importance with regard to the sensitivity to primaquine.
Some aspects of carbohydrate and lipid metabolism in prima- quine sensitive and normal erythrocytes incubated with glucose and
brilliant cresyl blue have been studied in vitro.
Whereas the dye markedly Increased glucose consumption of
normal cells, a decrease In the overall utilization of glucose in
enzyme deficient erythrocytes was observed.
On Incubation of normal cells with uniformly labelled glucose
and dye, a net synthesis of fatty acids was noted.
Although radioactivity from labelled glucose could be detected
Page 118 of 168
in the fatty acids of enzyme deficient cells, total fatty acid
concentration dropped on incubation with BCB, indicating that lipid
catabolism had occurred to a greater extent than synthesis.
The possible significance of this abnormal lipid metabolism
in primaquine sensitive cells has been discussed in relation to
their increased susceptibility to haemolysis by certain agents in
Page 119 of 168
a). A BRIEF DISCUSSION ON CERTAIN ASPECTS OF FATTY ACID SYNTHESIS
Certain points arising out of these studies might be briefly
considered \*hen discussing mechanisms of fatty acid synthesis. A
number of workers (Langdon, 1957; Siperstein, 1959) have suggested
that one of the more important mechanisms controlling lipogenesis
from carbohydrate was the rate of generation of NADPH by the hexose
monophosphate pathway of glucose metabolism. Decently, Ilasoro
(1962), both on the basis of theoretical considerations and avail- able experimental evidence, has attempted to refute the above
statement. One of the puzzling features of this study was the
finding that the red cell could synthesize fat from carbohydrate
only in the presence of a redox dye (Mendelsohn, 196lb;c). These
dyes are known to act by stimulating the shunt as electron accep- tors in the oxidation of glucose. Under those conditions it is
unlikely that any NADP1I would be available for reductive synthesis
of fatty acids. Furthermore, primaquine sensitive cells, which
are known to be deficient in reactions of the pentose phosphate
shunt were also able to incorporate glucose into fatty acids in
the presence of BC3. These findings tend to support the conten- tion of Masoro that NADPII produced by the pentose phosphate shunt
rarely, if ever, appears to be of physiological importance in the
regulation of lipogenesis. However, the present studies do seem
to indicate that pentose phosphate shunt activity is in some manner
related to lipid biosynthesis in the human rod blood cell. It is
of interest that Sass-Kortsak et al. (1962) have obtained evidence
Page 120 of 168
for the existence of a lactic dehydrogenase transhydrogenase system
In human erythrocytes. It is also known that BCB is able to stimu- late red cell lactic dehydrogenase (Kiese, 19^4; Mendelsohn, un- published results). The 11+ from NADH produced during this reaction
could then be transferred to NADP by the transhydrogenase thus
forming NADPH necessary for lipogenesls. Lowensteln (19 6 1) has
postulated a similar mechanism for NADPII production in rat liver.
Also pertinent to the observations made In this study are the find- ings of Stansly and Beinert (1953) and Seubert and Lynen (1953).
Using partially purified enzymes these workers noted independently
that the presence of a redox dye was necessary for fatty acid syn- thesis to occur in their systems. With respect to lipid synthesis
in the erythrocyte it is possible that BCB might be acting in a
b). SOME CLINICAL IMPLICATIONS OP THE RESULTS PRESENTED IN THIS
Our understanding of the mechanisms underlying the numerous
metabolic aberrations which occur In the human body have been
considerably hampered because of the obvious difficulties involved
in obtaining suitable tissues for biochemical analysis. While the
mature erythrocyte perhaps cannot be regarded as a typical body
cell since it does not possess a nucleus or cytoplasmic subcellu- lar particles, there exists a considerable amount of evidence to
show that changes In the biochemistry and metabolism of the red
blood cell do In fact reflect a similar state of events elsewhere
In the body. The following examples may be cited:-
Page 121 of 168
a) Feigelson and Conte (1954) reported that rat red blood
cells could utilise galactose as a substrate. These findings were
confirmed in human erythrocytes by Schwartz et al. (1956) and
l3selbacher et al. (19 56). Galactose is converted to glucose by
the following series of reactions
1) Galactose + ATP ------ > Galactose-1-phosphate + ADP
2) Galactose-1-phosphate + UDPG -7---- ~-- ---->
Glueose-l-phosphate + UD?-Galactose
3) UDP-Galactose --------> UDPG
4) UDPG -4- pyrophosphate ----------- ~> UTP + Glucose-1-phosphate
where UDP, UDPG and UTP are uridine diphosphate, uridine diphos- phate glucose and uridine triphosphate respectively. Haemolysates
from subjects with congenital galactosemia were found to be lacking
in a single enzyme, galactose phosphate uridyl transferase, while
the activities of the other three enzymes were normal (isselbacher
and coworkers 19 56). The same enzymic defect was later discovered
In the liver of congenital galactosemic patients (Anderson et al.
1957a). Because of these findings, the diagnosis of congenital
galactosemia has now become a relatively simple procedure using the
readily available red cell as a means of determining the deficiency
of galactose phosphate uridyl transferase (Anderson et al. 1957b).
b) Although it is generally considered that glycogen does
not take part in the metabolism of the mature human erythrocyte,
Comblath et al. (i960) have recently reported that phosphorylase
is present in human red cells by demonstrating & synthesis of
Page 122 of 168
glycogen In haemolysates. In normal haemolysates small quantities
of glycogen had to be added to prime the reaction, but in two cases
of glycogen storage disease this was not necessary as the erythro- cytes already contained from 2-10 mg glycogen / g haemoglobin. The
reverse reaction i.e. glycogen ---> glucose-1 -phosphate occurred
with equal facility in both instances. Erythrocytes of patients
with Type III glycogen storage disease accumulated up to 500 times
the amount of glycogen found in normal subjects. In common with
liver and muscle tissue the structure of the red blood cell glyco- gen, as determined by degradation with jS -amylase, was also found
to be abnormal. One patient with amylo-1,4 to 1,6-transglucosidase
deficiency (Type IV) had a normal erythrocyte glycogen concentra- tion but the glycogen was shown to be abnormal, being amylopectin- like in structure. A further abnormality in the red cells of
patients with glycogen storage disease has been described by Hsia
(1 9 6 1 ). The activity of an erythrocyte enzyme system which hydro- lyzed glucose-6-phosphate was found to be significantly decreased
among patients and their parents when compared with normal subjects.
These findings suggest that an important aid in the diagnosis of
glycogen storage disease in the human subject might be simply to
estimate the red cell glycogen content or alternatively, to assay
for the glucose-6-phosphate hydrolyzing enzyme.
c) Angelone and coworkers (195^) found a considerable in- crease in erythrocyte oxygen consumption of hyperthyroid patients
when compared with euthyroid subjects. This returned to normal
after the appropriate treatment. Beutler and Necheles (1959) noted
that physiological concentrations of triiodothyronine could stimu
Page 123 of 168
late the glucose oxidative pathway In normal human erythrocytes
in vitro suggesting that thyroid hormone might stimulate this meta- bolic pathway in vivo. In a later communication (1962) they did
in fact obtain evidence for a stimulation of the pentose phosphate
shunt by triiodothyronine in the Intact rat as a whole.
d) A marked retardation in activity at the transketolase
stage of the glucose oxidative pathway has been demonstrated In
the red cells of patients with Wernicke's encephalopathy (Wolfe
et al. 1958). Since transketolase requires thiamine pyrophosphate
as a coenzyme it Is more than likely that an abnormally function- ing pentose phosphate shunt is a generalized cellular phenomenon
in this disease.
e) Rats reared on an essential fatty acid free diet soon
exhibited marked alterations in the polyunsaturated fatty acid con- tent of their red blood cells (Holman i960). Since Identical
changes in fatty acid composition occurred in other tissues, it was
concluded that the erythrocyte polyunsaturated fatty acid pattern
reflected the dietary Intake of essential fatty acids and that this
fact might be employed to evaluate the essential fatty acid status
of the animal. In humans, alterations in dietary fatty acid com- position have also been shown to cause changes in erythrocyte
fatty acid content within a few days (Horwitt et al. 1959).
f) The results of the present investigation also Indicate
that the human red blood cell cannot be regarded solely as a
vehicle for the transport of respiratory gases. After the inges- tion of glucose in normal subjects a rapid rise in the concentra- tion of erythrocyte lipid and pyridine nucleotides occurs. This
Page 124 of 168
strongly suggests that the red cell Is able to utilize ingested
carbohydrate directly for lipogenesis and pyridine nucleotide syn- thesis in common with other tissues. A failure to metabolize glu- cose for the synthesis of fatty acids has been demonstrated on
numerous occasions in the experimentally induced diabetic animal
but only indirect evidence is available that a similar metabolic
defect is present in the human diabetic patient. The data pres- ented here have shown that the insulin deprived diabetic red cell
exhibits an abnormality with respect to the biosynthesis of lipid
from carbohydrate both in vivo and in vitro when compared with the
normal erythrocyte. Furthermore, this defect can be corrected to
a large extent by the administration of insulin indicating that the
metabolic activities of this cell are under hormonal control. The
above findings taken together point to the fact that the red blood
cell might very well reflect in its own metabolism that which
occurs in other tissues. It is suggested that the metabolic ab- normality which Is present in the erythrocyte of the human diabetic
is part of a generalized phenomenon and will also be found else- where in the body.
Concerning the precise nature of the metabolic defect in the
diabetic erythrocyte little can be said at present. The earlier
hypothesis that a lack of NADPK (primarily due to deficient pentose
phosphate shunt activity) caused a failure of lipogenesis in dia- betes must be considered unlikely in the light of more recent ex- perimental evidence. Abraham et al. (1 9 6 2 ) have shown that NADPH
production is definitely not the rate limiting step for lipogene- sis in diabetic tissues. These workers obtained evidence that the
Page 125 of 168
defect In diabetes is probably inherent in certain enzymes direct- ly concerned with fatty acid synthesis. Whether this is the case
in the human diabetic red cell remains to be determined.
It is also suggested from the present investigation, that
since both the Intact red cell and the reconstituted haemolysate
are responsive to the action of Insulin, this system might Increase
our limited knowledge concerning the mechanism of action of insulin
in human tissues.
Finally, the abnormal Interrelationship of carbohydrate and
lipid metabolism in the BCB stimulated primaquine sensitive red
cell indicates that this system could be employed to study haemoly- tic mechanisms In the red blood cell.
Thus it can be seen that a careful study of the metabolism
of the red blood cell could provide valuable information about the
nature of the abnormalities which occur in the numerous diseases
of metabolism afflicting man. Studies are continuing on various
aspects of red cell metabolism In an effort to determine more
accurately the enzymatic defect (s) causing diabetes mellitus in
the human subject as well as to obtain further Information about
the metabolic aberrations in such ill understood syndromes as
familial hypercholesterolaemia and hypertriglyceridaemla.
The interdependence of the fields of biochemistry and medi- cine is becoming increasingly apparent. The rapid strides that
biochemical technique and theory have made in recent years has
greatly extended the approach to pathology that disease can be in- terpretable in terms of disordered biochemistry. Apart from their
significance for a fuller understanding of the pathogenesis of
Page 126 of 168
human metabolic disease, it is hoped that the ideas presented in
this thesis will also be of practical importance in the diagnosis
of these diseases and thus in the management of the patient.
Page 127 of 168
This thesis has been concerned mainly with certain aspects
of carbohydrate and lipid metabolism in the human red blood cell.
Numerous topics of mammalian erythrocyte metabolism have not been
discussed. These include changes in erythrocyte metabolism during
ageing of the cell in vivo and in vitro, preservation of red cells
for blood transfusion, permlabllity of the erythrocyte membrane to
inorganic ions and organic molecules, alterations in erythrocyte
enzyme activity in haematologlcal disorders affecting the cell,
the problem of haemoglobin synthesis, to name but a few.
It is becoming more apparent that the mature mammalian red
blood cell represents a highly organized biochemical unit concerned
with a wide variety of metabolic activities. Phosphorylated inter- mediates of the glycolytic and hexose monophosphate shunt pathways
are continuously being degraded and resynthesized to produce energy
for a multitude of metabolic purposes. These actions are under
hormonal control and any variations noted in the red cell may well
indicate similar metabolic activities occurring In other cells of
the body. Furthermore, numerous studies have shown that the meta- bolism of the erythrocyte Is influenced both by changing nutritional
factors and disease processes affecting the organism as a whole.
In many instances variations in certain red cell components under
a given stimulus are known to reflect analogous modifications in
other tissue cells. In vivo investigations have demonstrated that
the lipid components of the red cell membrane are In an active
metabolic state and suggest that this cell is capable of lipid syn- thesis. Other biosynthetic energies of the erythrocyte are taken
Page 128 of 168
up by glutathione and the pyridine nucleotides. Finally, numerous
isolated reports have recently appeared concerning novel enzyme
systems present in the mammalian red blood cell. The significance
of these enzymes remains obscure at the present time, but their
existence must conceivably attribute to the red cell a much greater
importance in the body economy than solely that of a transport
agent for the respiratory gases.
Page 129 of 168
1 1 9
Page 130 of 168
A list of the enzymes known to be present In the mature mammalian
I. Enzymes of carbohydrate metabolism
a) . Glycolytic enzymes
Phosphogluc omu tas e
Pho s phohexo is omerase
2.3- Diphosphoglycerate muta3e
2.3- Eiphosphoglycerate phosphatase Kashket et al., 1957
Enolase Lohr et al., 1958
Phosphopyruvate kinase "
Lactic acid dehydrogenase Quastel and Wheatley, 1938
b ) . Enzymes of the hexose monophosphate shunt
Glucose-6-phosphate dehydrogenase Lohr et al., 1958
Bartlett and Marlow, 1951
Guarino and Sable, 1955
Comblath et al., i960
Tsuboi et al., 1958
Blanchaer et al., 1955
Lohr et al., 1958
Buffa and Maraini, 1959
Piapoport and Leubering, 1950
Page 131 of 168
6-Phosphogluconic acid lactonase
Buffa and Maraini, 1959
Bruns et al., 1958
Bruns efc al., 1958
Guarino and Sable, 1955
c). Enzymes of galactose metabolism
Galactokinase Isselbacher et al., 1956
II. Enzymes related to protein and non-protein nitrogen metabolism
Glutathione reductase Meldrum and Tarr, 1935
Glutathione peroxidase Mills, 1957
Proteinases I,II,III Morrison and Neurath, 1953
Peptidases Adams et al., 1952
Alanine transaminase Albert and Brin, i960
Glutamic-oxalacetic transaminase Karmen et al., 1955
Glutamic-pyruvic transaminase "
Acetylcholine esterase Alles and Howes, 19^0
Page 132 of 168
Arginase Covolo and West, 1947
Urease Weil, 1944
Glutamlnase Mauri and Torelli, 1959
III. Enzymes concerned with reduction of methaemoglobln
NADP dependent Huennekens et al., 1957
Flavoprotein reductase(dlaphorase)Gibson, 1948
Aldehyde dehydrogenase Matthies, 1959-
IV. Enzymes concerned with the metabolism of nucleosides and
Nucleoside phosphorylase Sandberg et al., 1955
NAD and NADP nucleosidase Hofmann et al., 1959
NAD nucleotidase Alivisatos et al., 1958
NADP nucleotidase tt
Rubinstein and Denstedt, 1955
Preiss and Handler, 1958
Aspartate carbamyltransferase Smith and Baker, i960
Page 133 of 168
Ilehteeojdt t, 1943
Clarkson and Maizels, 1952
Kashket and Denstedt, 1958
Malkin and Denstedt, 1958
VI. Enzymes concerned with steroid metabolism
0estradiol-17©£ dehydrogenase Portius and Repke, i960
0estradiol-17jS dehydrogenase ”
Cholesterol esterase Guardamanga et al., i960
Cholesterol oxidase Danielsson and Horning, 1962
VII. Enzymes of the tricarboxylic
NADP-isocitrlc dehydrogenase VI.
Quastel and Wheatley, 1938
Beutler and Yeh, 1959
Meldrum and Houghton, 1932
Keilin and Wang, 1947
Jones and McCance, 1949
Page 134 of 168
NADU-cytochrome c reductase
Phosphatidic acid phosphatase
c 10 N-\.N -methylene tetrahydrofolate
cf-amino levulinlc acid dehydrase
Aryl amine acetylase
Goldstein and Rieders, 19t>3
Wagenknecht et al., i960
Hokin and Kokin, 1961
Bertino et al., 1962
Rimington and Booij, 1957
Motulsky and Steinmann, 1962
Page 135 of 168
Page 136 of 168
S. Afr. J. med. Sci. (1958), 23, 75-82,
A COLORIMETRIC METHOD FOR THE DETERMINATION OF MICRO- AMOUNTS OF HIGHER UNESTERIFIED FATTY ACIDS (C12—C18) IN BLOOD
Department of Pathology and Microbiology, Division of Chemical Pathology,
University of the Witwatersrand, Johannesburg.
[Received 24th M ay, 1958]
In connection with certain experiments at present in progress in this laboratory,
it became necessary to devise a method for the estimation of micro-amounts of
higher unesterihed fatty acids (UFA) in blood. All the methods [Shore and Nichols,
1953; Grossman et al., 1954; Seldin and Westphal, 1955; Dole, 1956; Gordon
and Cherkes, 1956] at present available for the determination of plasma UFA
require a relatively large amount of material (1—15 ml.). Furthermore, since all the
above procedures utilize either aqueous or alcoholic alkali to neutralise the acid
groups, they are liable to the inherent disadvantages of a titration method such
as the need for the preparation of fresh standard alkali each day, the performance
of a titration in a carbon dioxide-free atmosphere, and because the concentration
of UFA in plasma is so low, the requirement of specialised equipment (ultra-micro
burettes). Therefore it was thought worth-while to explore the possibilities of
It has for long been known [Schmidt, 1955] that the colourless carbinol bases
of the fuchsin dyes, e.g. rosaniline, will unite with one equivalent of an acid with
loss of water to form intensely coloured red salts according to the reactions shown
in Fig. 1.
FIG . 1. Showing the reactions of the colourless carbinol base of rosaniline with acid.
This fact was made use of by Krainick and Muller , who devised a photometric
micro-method for the determination of total fatty acids in 0-1 ml. plasma. In their
method the plasma is first saponified with alkali, acidified to liberate the fatty acids
which are then extracted into benzene. An aliquot of the benzene extract is
evaporated to dryness, the residue dissolved in iso-propyl alcohol which is then
treated with rosaniline reagent and the resulting colour read photometrically. In
this study the procedure described by Delsal  was found to be most suitable
for the extraction of plasma UFA when modified for small volumes of plasma.
No saponification is necessary and as will be shown, the plasma UFA are
quantitatively extracted by this method. The total plasma lipids are first brought
into a one-phase system consisting of methylal: methyl alcohol: petroleum ether
with subsequent conversion into a two-phase system by the addition of water. The
upper petroleum ether layer which contains sterols, glycerides and UFA is taken
to dryness, the residue dissolved in iso-propyl alcohol and the solution then treated
Page 137 of 168
76 The South African Journal o f Medical Sciences
with rosaniline reagent. The effective range of Krainick and Muller’s method was
50—500 fxg. By suitable adjustment of the rosaniline reagent it has been possible
to increase the sensitivity of the procedure still further.
Plasma is acidified with normal sulphuric acid to a pH below 4 ■ 0 and the higher
UFA isolated using a two-phase extraction system. The upper phase solvent
containing the UFA is evaporated to dryness and the residue dissolved in isopropyl
alcohol. Rosaniline reagent is added to the mixture which is heated at 46° C for
half-an-hour and the resulting red colour measured in a photoelectric colorimeter.
The concentration of the UFA is determined from a calibration curve prepared
by using pure stearic acid as a standard. Pure palmitic or oleic acid may also be
used to prepare the standard curve.
All glassware must be kept scrupulously clean and free from even the slightest
traces of acid.
(a) Methylal (Merck).
<(b) Methyl Alcohol (Merck—acetone free).
(c) Methylal: Methyl Alcohol (4:1 v/v).
(d) Petroleum Ether (b.p. 40—60° C). A.R. High blank values are sometimes
obtained if the petroleum ether is allowed to stand too long and consequent- ly this reagent should then be freshly redistilled every 3—4 weeks.
(e) iso-Propyl Alcohol A.R.
(f) Benzene A.R.
(g) Normal Sulphuric Acid. This reagent need not be accurately standardised.
(h) Stearic Acid (pure).
(i) Rosaniline (base) (British Drug Houses, England). Other brands of rosaniline
have been tried and found not to be satisfactory.
Preparation of Rosaniline Reagent
About 0-5 g. rosaniline (base) is added to 50 ml. benzene in a 100 ml. round
bottom flask equipped with a reflux condenser. The mixture is heated for one hour
under reflux and then allowed to cool to room temperature. The solution is
decanted from the undissolved dye and centrifuged for five minutes. The dark-orange
fluorescent supernatant is poured into a 500 ml. beaker and benzene added until the
diluted reagent shows 25% transmission in an Evelyn photoelectric colorimeter
set to 100% transmission with distilled water using a 490 m/x. filter and the 10 ml.
aperture. The reagent is stored in a dark bottle and is stable at room temperature
for one month. If the reagent is to be used after having stood longer than this time,
it is advisable to draw up a new calibration curve in which case a slight variation
in the slope of the curve may be found. The reagent should not be stored in a
refrigerator as the dye will precipitate out of solution.
Blood obtained by venepuncture is collected into tubes containing 2 mg. dried
potassium oxalate/ml. blood. The plasma is separated within fifteen minutes after
collection. 0-2 ml. plasma is acidified with 0-02 ml. normal sulphuric acid in a
15 ml. centrifuge tube; 1-0 ml. methylal: methyl alcohol (4:1 v/v) is added and the
Page 138 of 168
Mendelsohn: Fatty Acids in Blood 77
tube well shaken. Then 3-0 ml. methyl alcohol and 0-8 ml. petroleum ether (b.p.
40—60° C) are added with mixing until the solvents form a single phase. Now
0-8 ml. distilled water is added when the system immediately separates out into two
phases. This is followed by a further 2 ml. petroleum ether and the contents of the
tube well stirred for one minute using a glass rod flattened at one end. The tube is
centrifuged at about 3,000 r.p.m. for two minutes after which the clear upper
petroleum ether phase is transferred to a pyrex tube graduated at 10 ml. using a Pasteur
pipette with a fine bore. Care must be taken not to draw into the Pasteur pipette
even the smallest drop of the highly acid lower phase which will obviously introduce
a considerable error into the results. The extraction is repeated using a further 2 ml.
petroleum ether. When removing the petroleum ether a thin layer of this solvent
is left covering the aqueous phase and it is not necessary to attempt to remove the
petroleum ether completely. With the first extraction 95% of the UFA is removed
and with the second extraction recoveries are for practical purposes complete.
A third extraction produces less than 0-5% of acid.
The two petroleum ether extracts are combined in a pyrex test tube graduated
at 10 ml. and evaporated to dryness in a hot water bath. The residue is dissolved in
1 ml. iso-propyl alcohol and then 1 ml. rosaniline reagent is added with mixing.
The tube which is stoppered with a loose fitting glass bulb is placed in a rack in a
constant temperature water bath at 46° C for thirty minutes after which time it is
placed in a beaker of cold water for five minutes and the contents made up to 10 ml.
with benzene. The reagent blank consists of 1 ml. iso-propyl alcohol + 1 ml.
rosaniline reagent treated exactly as the test. The colour of the test solution is read
in an Evelyn photoelectric colorimeter using a 520 m/x filter against the reagent
blank. The concentration of unesterified fatty acid in meq./l is calculated from a
calibration curve prepared as described below.
If it is suspected that the UFA content of a particular plasma sample will be high
the extraction can satisfactorily be performed on half quantities of plasma and
reagents but still using 2 ml. petroleum ether for the final extractions.
Preparation of the calibration curve
Dissolve 10 mg. stearic acid in 50 ml. iso-propyl alcohol. Since even at room
temperature a solution of stearic acid in iso-propyl alcohol will slowly become
esterified the curve should be prepared immediately all the stearic acid has dissolved.
Pipette 0-1 ml. to 0-5 ml. (i.e. 20—100 p.g. stearic acid) into each of five pyrex test
tubes and the volume in each tube made up to 1 ml. with iso-propyl alcohol. The
reagent blank is 1 ml. iso-propyl alcohol. To each tube add 1 ml. rosaniline reagent
with mixing and continue the procedure as described above. The colours should
be read within five minutes after dilution with benzene.
It is not necessary to prepare a fresh standard curve each time the rosaniline
reagent is made up if it is prepared under the conditions described previously and
provided the same iso-propyl alcohol is used. With each new batch of rosaniline
base and iso-propyl alcohol a fresh standard curve must be drawn up mainly
because of the variability in the acid content of the commercially available alcohols.
The results expressed as millequivalents per litre are calculated on the assumption
that the average molecular weight of the higher unesterified fatty acids in blood
Page 139 of 168
78 The South African Journal o f Medical Sciences
RESULTS AND DISCUSSION
Validity of the Method
In using Delsal’s method it still had to be determined whether the higher
unesterified fatty acids present in blood would be quantitatively extracted into the
upper phase. Table I shows the percentage recoveries of various pure fatty acids
of differing chain length when extracted as described in the method above. It can be
seen that the fatty acids containing twelve to eighteen carbon atoms are for all
practical purposes quantitatively recovered. The lower fatty acids being volatile,
are lost when the petroleum ether is evaporated to dryness and are therefore not
determined by this method.
Recoveries of Fatty Acids of Different Chain Lengths when
extracted as described in the method
Propionic.................................. 3 50 — —
Butyric .................................. 4 50 — —
Caproic .................................. 6 50 — —
C a p r y lic.................................. 8 50 — —
Capric .................................. 10 50 8 16
Laurie .................................. 12 50 47 94
Myristic .................................. 14 50 48 96
Palmitic .................................. 16 40 39-5 99
O leic.......................................... 18 50 51 102
Stearic .................................. 18 80 80 100
Table II shows the recoveries of Palmitic, Stearic and Oleic acids when added
directly to sera of known UFA content; again it is evident that the recoveries of
these fatty acids are complete.
Of the various organic acids present in biological fluids which might contribute
to the acidity of the upper phase the following were subjected to the extraction
procedure: cholic acid, succinic acid, pyruvic acid, citric acid, oxalic acid, ascorbic
acid, uric acid and lactic acid. In each case less than 1 % of the acidity could be
recovered in the petroleum ether phase. Thus there is good evidence that the
procedure described will quantitatively and specifically extract the higher UFA
present in blood.
Page 141 of 168
80 The South African Journal o f Medical Sciences
FIG. 2. Showing the effect of time of heating at 46° C on colour development.
A — 30 minutes at room temperature.
B •— 15 minutes at 46° C.
C — 30 minutes at 46° C.
D — 60 minutes at 46° C.
FIG. 3. Showing the stability of colour with respect to time.
A — colours read within 5 minutes after dilution with benzene.
B — colours read 30 minutes after dilution with benzene.
Page 142 of 168
Mendelsohn: Fatty Acids in Blood 81
curve A figure 2. Although colour development is not complete under these conditions,
this standard curve may be used if a constant temperature water bath is not available,
but it is necesssary to emphasise that in this case the colours are not as stable as they
are after heating at 46° C.
From what has already been said it follows that increasing the temperature will
decrease the time necessary for full colour development. At higher temperatures
(e.g. 70° C), however, it is more difficult to obtain consistent results. The reason
for this is thought to be due to the fact that at 70° C there is considerable
“creeping” of the solvent up the side of the tube with a tendency to loss of solvent
by evaporation, whereas at 46° C this does not occur.
Reproducibility of the Method
Reproducible results can be obtained if the following conditions are carefully
(a) The test tubes in which the colour is to be developed must be made of good
quality glass and should be absolutely acid free.
(b) The test tubes should all be of uniform dimensions. Pyrex tubes 7 X £ inch
were used throughout this study.
(c) The heating conditions should be uniform. This has been achieved by
keeping the temperature of the water bath constant throughout the time of heating
and the water in the bath at a constant level. Four calibration curves were made
separately over a period of four months, each curve being prepared under identical
conditions and using the same reagents. Taking into account small errors introduced
in the weighing of the stearic acid each time, the graphs all fell on the same straight
The amount of acid present in various commercial makes of iso-propyl alcohol
when determined by the above method has been found to vary considerably. For the
best results the reagent blank (i.e. 1 ml. iso-propyl alcohol + 1 ml. rosaniline reagent)
should give a transmission of between 70 and 80% when read against a blank
consisting of 1 ml. iso-propyl alcohol + 9 ml. benzene (set at 100% transmission
and using a 520 m/z. filter). If the acid content of a particular batch of iso-propyl
alcohol is found to be greater than that mentioned above, 0-5 instead of 1-0 ml.
alcohol can be used to dissolve the residue and the procedure is continued as
Fasting blood samples were taken from twenty volunteers aged 15 to 28 years
and the plasma UFA determined in duplicate on each specimen. The plasma UFA
ranged from 0-107 to 0-410 meq./l with a mean value of 0-306 meq./l and
S.D. ± 0-025.
A colorimetric method is described for determination of micro-amounts of
higher unesterified fatty acids in plasma. The method is rapid and convenient.
The conditions for obtaining reproducible results are discussed and a range of
normal values is given.
I am greatly indebted to Professor H. B. Stein for his advice and encouragement. I also wish to
thank Drs. B. M. Bloomberg and A. Antonis for reading the paper and for their helpful criticism.
The expenses incurred in this work were defrayed by a research grant from the University of the
Witwatersrand Research Fund.
Page 143 of 168
82 The South African Journal o f Medical Sciences
D elsal, J. L. (1954). Fractionnement des lipides du serum sanguin par les solvants organiques.
Bull. Soc. Chim. Biol., 36, 1329-1334.
D ole, V. P. (1956). A relation between non-esterified fatty acids in plasma and the metabolism of
glucose. J. Clin. Invest., 35, 150-154.
G ordon, Jr., R. S. and Cherkes, A. (1956). Unesterified fatty acid in human blood plasma. J. Clin.
Invest., 35, 206-212.
G rossman, M. I., Palm, L., Becker, G. H. and Moeller, H. C. (1954). Effect of lipemia and
heparin on free fatty acid content of rat plasma. Proc. Soc. Exper. Biol. & Med., 87, 312-315.
K rainick, H. G. and Muller, F. (1942). Photometrische Mikrobestimmung der Fettsauren.
Mikrochim., 30, 7-14.
Schmidt. (1955). Organic Chemistry. 7th ed. 555-558. Oliver and Boyd (Edinburgh and London).
Seldin, G. L. and Westphal, U. (1955). Non-esterified higher fatty acids in serum of chloroform
treated and normal rats and other species. Proc. Soc. Exper. Biol. & Med., 89, 159-162.
Shore, B., N ichols, A. V. and Freeman, N. K. (1953). Evidence for lipolytic action of human
plasma obtained after intravenous administration of heparin. Proc. Soc. Exper. Biol. & Med.,
Page 144 of 168
Page 145 of 168
C opyright © , 1961 by Lipid R esearch, Inc.
R ep rin ted from J o urn a l of L ip id R e s e a r c h , Vol. 2, No. 1
(Jan u ary , 1961), 45-50
A fluorimetric micro glycerol method
and its application to the determination
of serum triglycerides
D e n n is M en d elso h n an d Arnold A n to n is
Department of Pathology and Microbiology, Division of
Chemical Pathology, University of the Witwatersrand,
Johannesburg, South Africa; and Ernest Oppenheimer
Heart Research Unit, South African Institute for
Medical Research, Johannesburg, South Africa
[Received for publication April 18, 1960]
A fluorimetric method has been developed for the estim ation of glycerol in aqueous solution.
It utilizes a series of reactions in which glycerol is heated with o-aminophenol in the presence
of concentrated sulfuric acid and an oxidizing agent, to form 8-hydroxyquinoline which pro- duces fluorescence on chelation with a divalent m etal ion in alkaline solution. Experim ental
details are given for the estim ation of serum triglycerides on phospholipid-free serum lipid
extracts. The method can also be used for the estim ation of phosphatide glycerol.
f^-ecent interest in fat metabolism has created
a demand for a specific routine method for the direct
determination of serum triglyceride levels. While in- direct methods which are based on the measurement of
serum total fatty acids, cholesterol esters, and phos- pholipids are available, these require calculation of
triglyceride levels by difference, and are unsatisfac- tory mainly because of the assumptions which have
to be made, particularly of the proportion of fatty
acids available in the phospholipids.
The application of more exact techniques has been
rendered possible by the development of column chro- matographic methods for the separation of triglyce- rides from other serum lipid components, but these
tend to be too cumbersome and time-consuming for
routine use. The recently published method of Van
Handel and Zilversmit (1), with minor modifications,
has proved very suitable for routine separation of
phospholipids from other serum lipid components. Sub- sequent alkaline hydrolysis of the triglyceride extract
according to Carlson and Wadstrom (2) allows for
measurement of the liberated glycerol by the chromo- tropic-acid method of Lambert and Neish (3).
The micromethod to be described, which has been
adapted from a spot test described by Feigl (4), is
simple, accurate, and very sensitive. It utilizes a series
of reactions based on the Skraup (5) quinoline syn- thesis, in which glycerol, liberated from triglycerides
by alkaline hydrolysis, is heated at 140° with o-amino- phenol in the presence of concentrated sulfuric acid
and an oxidizing agent to form 8-hydroxyquinoline.
The fluorescence produced by chelation of 8-hydroxy- quinoline with a divalent metal ion in alkaline solu- tion is utilized to provide a quantitative measure of
the glycerol, and therefore of triglyceride concentra- tion.
Reagents. All reagents and solvents used were an- alytical reagent grade, with the exception of o-amino- phenol, of which only a technical grade was available
Light petroleum ether (b.p. 30°-60°) was redistilled
before use. Diethyl ether and isopropyl ether were
freed from peroxides by passage through a column of
activated alumina (heated overnight at 170°) just
prior to use.
Silicic acid: Silicic acid (Mallinckrodt; 100 mesh, suit- able for chromatography) was size graded by sedi- mentation with distilled water. The fine particles,
which constituted approximately 50% of the total,
were separated, dried, and activated overnight at
Arsenic acid solution (0.6c/c ): Fifty g of arsenic pen- toxide are dissolved in 100 ml water, allowed to
Page 146 of 168
46 MENDELSOHN AND ANTONIS J. Lipid Research
stand for 3 to 4 days to form H3As04, and filtered
if necessary. One ml of this stock solution is di- luted to 100 ml with concentrated FI2S04.
o-Aminophenol solution (1.6c/c ): Technical grade
o-aminophenol is purified by sublimation at 170°
in an atmosphere of nitrogen. The sublimed com- pound tends to oxidize rapidly at this tempera- ture, and it is therefore either recrystallized from,
or washed 3 times with, a small volume of iso- propyl ether until a colorless solution is obtained
when the pure white crystalline compound is dis- solved in acetone. This pure product is stable in- definitely if stored in a brown, stoppered bottle.
Immediately before use, 0.16 g of the pure o-ami- nophenol is dissolved in 10 ml acetone.
Mg++ solution (120 fxg/ml): MgS04-7H20 (0.12 g) is
dissolved in distilled water and made up to 100 ml.
Triolein standards: A stock standard solution is pre- pared, containing 100 mg triolein (or equivalent
amounts of other triglycerides) per 100 ml in
chloroform. Dilute standards are prepared, con- taining 0.2 to 1.0 mg/ml, corresponding to the
range 100 to 500 mg triglyceride per 100 ml serum
when carried through the procedure.
Procedure. Activated silicic acid (1.2 g) is slurried
in a glass-stoppered test tube with 1 ml of isopropyl
ether, and 0.3 ml of serum added dropwise with shak- ing. A further 6.5 ml of isopropyl ether is then added,
together with a few glass beads, and the mixture well
shaken for half an hour. The silicic acid is allowed to
settle, and a 5 ml aliquot of the supernatant extract
( = 0.2 ml serum) is taken off into a glass-stoppered
15 ml conical centrifuge tube and evaporated to dry- ness on a hot-water bath using an air blower. One ml
aliquots of the triolein standards are similarly evapo- rated to dryness.
To the dried extract is added 3 to 4 drops ether, 0.5
ml of methanol, and 3 drops of 2% methanolic KOH
solution, and the mixture saponified for 30 minutes at
60° to 70°. Two drops of 6% methanolic acetic acid
solution are then added, and the mixture evaporated
just to dryness on a boiling water bath. Six ml of pe- troleum ether are added to the hot tube, followed by
0.5 ml of 10 N Hl»S0 4, and the tube is stoppered, well
shaken, and then centrifuged for 1 to 2 minutes. The
petroleum ether layer is carefully pipetted off and dis- carded. Duplicate 0.1 ml aliquots of the aqueous glyc- erol phase are then treated as below.
One-tenth ml aliquots of the 1.6% o-aminophenol
solution are pipetted into a number of test tubes fitted
with glass stoppers, and the solvent evaporated, using
an air blower. A 0.1 ml aliquot of the serum glycerol ex- tract prepared above is added, followed by 0.4 ml of
the 0.6% arsenic acid solution, and the mixture heated
in a silicone oil bath at 140° for 15 minutes. The mix- ture is cooled in ice water, and 1 ml of the Mg++ solu- tion is cautiously added with mixing. Five ml of 28%
ammonia solution is carefully added, and the tube
stoppered and well shaken.
A blank (0.1 ml of 10 N H2S04 solution) and glyc- erol extracts (0.1 ml), derived as before from 1 ml
aliquots of the standard solutions of triolein, are run
simultaneously throughout the procedure.1 After 5 to
10 minutes, aliquots of the above solutions are poured
into Farrand fluorometer tubes, and the fluorescence
produced by ultraviolet light is measured in a Farrand
fluorometer, using aperture 6, with Corning 5874 (pri- mary) and 2424 (secondary) filters. The fluorometer
is set at 100%- transmission with the highest standard
(i.e., 200 fj.g corresponding to 500 mg/100 ml serum).
Where low serum triglyceride concentrations are ex- pected, lower concentration glycerol standards may be
used for setting the instrument at 100% transmission,
but in this case the blank reading will be higher, as
shown in the Figures. It is important that all glassware
should be scrupulously clean in order to obtain repro- ducible results.
The Fluorescence Reaction. The Skraup reaction is
commonly used for the synthesis of quinoline deriva- tives from aromatic amines and glycerol in the pres- ence of concentrated sulfuric acid and an oxidizing
agent such as nitrobenzene or arsenic acid. Acrolein,
formed by oxidation of glycerol, is probably an inter- mediate product in the reaction since it will also form
quinoline derivatives. The use of o-aminophenol as the
aromatic amine results in the formation of 8-hydroxy- quinoline; however, a large excess of the reagent is
required to ensure quantitative reaction with regard to
glycerol. This leads to relatively high blanks as well
as marked quenching of the fluorescence subsequently
produced. Optimum conditions have nevertheless been
established in which the influence of these two factors
is minimized, and the graph of fluorescence against
glycerol concentration is linear for the range 0 to 20
Triglycerides such as triolein, tripalmitin, tristearin,
and trilinolein have all been found to give quantitative
recoveries of glycerol after hydrolysis when checked
1 A lternatively, glycerol standards, containing 2 to 20 <ug of
glycerol per 0.1 ml in 10 N H2SO4, can be used, since the
hydrolysis of the triglyceride is quantitative. Satisfactory re- sults are also obtained with aqueous glycerol solutions.
Page 147 of 168
Number 1 FLUORIMETRIC ESTIMATION OF SERUM TRIGLYCERIDES 47
against glycerol standards. These results have also
confirmed, as shown by Carlson and Wadstrom (2),
that glycerol is not lost during the solvent evaporation
after saponification. Triolein has been used as the
standard for serum triglycerides', and its fluorescence
curve is shown in Figure 1.
T R I G L Y C E R I D E JJG
F ig. 1. Relationship between triglyceride concentration and
intensity of fluorescence produced by pure o-aminophenol.
o—:—o----- o; 200 /tg standard set at 100% transmission.
x----- x----- x; 100 /rg standard set at 100% transmission.
The precision of the fluorescence reaction has been
determined on duplicate aqueous glycerol extracts de- rived from 20 different sera. The results indicate an
average error of 3.9 mg for the range 17 to 363 mg/100
ml serum.2 The error will also depend on the sensi- tivity of the fluorometer, since the range 20% to 100%
transmission corresponds to 0 to 200 /rg triglyceride
when the instrument is set at 100%> transmission with
the 200 /eg standard. As indicated in the procedure and
shown in Figure 1, the error can be reduced by ap- proximately one-third by setting at 100%- transmis- sion with the 100 /eg standard. In this case, the range
35% to 100% transmission corresponds to 0 to 100 /eg
triglyceride. Since the method utilizes the glycerol lib- erated from triglycerides, small errors will be intro- duced by the presence of mono- and diglycerides which
can occur up to about 5% to 10% in serum (6).
The influence of a number of serum components on
the fluorescence reaction has been investigated. Inosi- 2 This has been calculated as the standard error of a single
determination, S.D. = ^ -T.V ., where A is the difference be- tween two single tests performed on each sample, and N is the
number of double determinations.
tol, glucose, and other hexose sugars do not affect the
reaction. Serine, and bases such as choline and ethanol- amine, also have no effect. Methanol and ethanol pro- duce interfering fluorescence probably through their
aldehydes; however, they are completely removed by
evaporation after saponification of the triglycerides.
Fatty acids and sterols, when heated with dehydrating
agents such as concentrated sulfuric acid, produce in- terfering fluorescence, and are therefore removed with
petroleum ether after saponification. Glycerophospha- tides and sphingomyelin interfere in the reaction (see
below), and are removed from the serum lipid extracts
by silicic acid, as determined by the sensitive phos- phorus assay procedure of Chen et al. (7). The con- centration of free glycerol in serum is small relative to
that of triglyceride, and its influence on the fluores- cence would therefore be negligible. Moreover, addi- tion of as much as 100 mg of glycerol or glucose to
serum prior to extraction with silicic acid and iso- propyl ether has had no influence whatsoever on the
It has not been possible to assess directly whether
any other serum components alter the serum blank or
influence the amount of fluorescence. However, excel- lent correlation is obtained when the results of the
fluorimetric method are compared with those of other
procedures described later, and this indicates that any
such interference is negligible.
Extraction of Nonphospholipid Serum Lipid Com- ponents. In their preparation of phospholipid-free se- rum extracts, Van Handel and Zilversmit (1) use Dou- cil and chloroform; however, filtration of the extract
is necessary because of the high density of the solvent.
Diethyl ether and isopropyl ether, which also extract
nonphospholipids from silicic acid, do not require fil- tration since the silicic acid settles easily from the
lower density ethers. Diethyl ether proved to be too
volatile and losses were obtained at the shaking stage.
These were prevented by the use of isopropyl ether
(b.p. 67.5° I, which has the additional advantage of
having a very low mutual solubility with water.
Elution of all nonphospholipid serum lipid compo- nents by the isopropyl ether is quantitative. Choles- terol ester and free cholesterol estimations on the ex- tracts by the method of Sperry and Webb (8), were
not significantly different from values obtained on
whole serum extracts.
The extraction procedure for triglycerides has been
tested by recovery experiments. Triolein was added to
serum by two methods: Method 1. Chloroform solu- tions of triolein were added to the stoppered test tubes,
and the solvent blown off. Serum (0.3 ml) was added
Page 148 of 168
48 MENDELSOHN AND ANTONIS J. Lipid Research
to the tubes and the mixture shaken. Silicic acid, pre- mixed with a small amount of isopropyl ether, was
then added and the mixture slurried with a glass rod.
The remainder of the isopropyl ether, together with a
few glass beads, was then added, and the mixture well
shaken. Aliquots of serum were analyzed with and
without the addition of triolein. Method 2. Chloroform
solutions of triolein were added to larger (5 ml) serum
aliquots, the solvent was blown off with No, and the
mixture was well shaken for half an hour. Aliquots of
serum were analyzed before and after the above addi- tion. Recoveries of added triolein varied from 97% to
103% as shown in Table 1.
T A B L E 1. Recovery of T rio lein Added to D if fe r e n t Sera
Serum Added Recovered Recovered
ng ng p er cent
100 100 100
A (M ethod 1) 150 155 103
200 202 101
80 78.5 98
A (M ethod 2) 40 39.6 99
20 19.4 97
50 50.5 101
100 97 97
B (M ethod 2) 150 147 98
200 204 102
Triglycerides were estimated fluorimetrically on du- plicate aliquots of six serum samples. Results are
shown in Table 2. The mean error between duplicate
serum analyses was 4.1 mg/100 ml serum, not signifi- cantly different from the error on duplicate extracts of
the same serum.
Comparison with Other Methods of Triglyceride
Analysis. Triglycerides were estimated on 15 sera by
the following three methods: (a) The proposed fluori- metric procedure. (6) By colorimetric estimation of
ester groups using a modification (9) of the ferric hy- droxamate procedure of Stern and Shapiro (10). This
technique was carried out on the extract derived in (a)
and also on whole sera extracted with Bloor solvent.
Free and ester cholesterol were estimated on both ex- tracts by the method described in the previous section
(8). The mean molecular weight of the cholesterol
ester fatty acids was assumed to be 280. On the whole
serum extracts, triglyceride fatty acids were calculated
according to the method of Thannhauser and Rein- stein (11), serum phospholipids being estimated on the
extracts according to the method of Fiske and Sub- barow (12). (c) By the liberation and electrometric
titration at pH 9.0 of fatty acids obtained from serum
triglyceride fractions after chromatography of whole
scrum lipid extracts on silicic-acid columns according
to Barron and Hanahan (13).
The results obtained by these methods are shown in
Table 3. On the same isopropyl ether extract the fluori- metric method and the ester group method showed an
average error of 4.1 mg/100 ml serum, not significantly
different from the errors present in each method. Com- parison of the fluorimetric and titrimetric methods
produces a slightly higher average error of 5.4 mg/100
ml serum, while comparison against the ester group
method on whole serum extracts shows the highest
average error of 13.0 mg/100 ml serum. The latter
method gives lower values than the other methods,
probably because of an error in the assumption of the
factor (0.69) for the calculation of phospholipid fatty
Elution of free fatty acids by the isopropyl ether is
also quantitative, and provides a simple method for
their extraction free from other acidic serum compo- nents. Using larger serum aliquots the free fatty acids
have been titrated directly on the isopropyl ether ex- T A B L E 2. E stim ation of T r ig ly cerid e C oncentration
on D u plicate Serum Aliq u o t s*
(mg/100 ml serum)
1st Aliquot 2d Aliquot
- S 390
S S TM-5
; ; 69.o
ao 690 69
108 108 0
108 108-° in n “
110.0 !S 1180
6 251 254 o
257 £ 2619
* S.D. of a single determination is 4.1 mg/100 ml serum.
Page 149 of 168
Number 1 FLUORIMETRIC ESTIMATION OF SERUM TRIGLYCERIDES 49
TABLE 3. T rig ly cerid e C oncentration
by D if f e r e n t M eth o d s*
Triglyceride Concentration (mg/100 ml serum)
A B C D
1 82 80 84 78
2 275 265 283 236
3 17 14 20 0
4 160 163 165 150
5 240 233 251 225
6 263 270 273 253
7 122 121 123 118
8 75 73 75 70
9 55 57 63 38
10 363 359 347 323
11 260 264 265 241
12 266 268 277 247
13 327 312 333 315
14 168 166 163 162
15 122 122 125 118
* A: Fluorimetric procedure on isopropyl ether extract.
B: E ster group procedure on same isopropyl ether extract as
C: Titrim etric method.
D: E ster group procedure on Bloor extract of whole serum.
Comparison A /B : S.D. of a single determination = 4.1
mg/100 ml serum.
Comparison A /C : S.D. of a single determination = 5.4
mg/100 ml serum.
Comparison A/D: S.D. of a single determination = 13.0
mg/100 ml serum.
tracts by electrometric titration at pH 9.0 (or with
bromthymol blue) using alcoholic KOH. Recoveries of
added palmitic and stearic acid have been quantita- tive. Results have compared favorably with those ob- tained by other methods.
Normal Values. Normal values obtained by the fluo- rimetric procedure for the White and Bantu popula- tion in South Africa were reported at the First An- nual Congress of the Nutrition Society of Southern
Africa, in November, 1959, and have recently been
published (14). Mean fasting serum triglyceride levels
for young White and Bantu males were 8G and 80
mg/100 ml serum, respectively.
The procedure recommended consists of a number
of independent stages, each of which has been shown
to be quantitative and highly selective. The initial ex- traction with silicic acid and isopropyl ether provides
an efficient procedure for the extraction of the nonphos- pholipid serum lipid components. It has recently been
shown by Cheng and Zilversmit (15), using Doucil in
the analysis of rat plasma triglycerides, that the par- ticle size of the adsorbent is an important factor gov- erning the complete lysis of the protein-lipid bonds
and adsorption of nonlipid serum components. We
have similarly found it necessary to employ silicic acid
having a particle size finer than 100 mesh, since larger
particle sizes lead to low recoveries of the lipid com- ponents. Mallinckrodt silicic acid (100 mesh) has
therefore been size graded before activation at 170°.
We have subsequently used Baker Analyzed reagent
grade silicic acid powder, of which not more than 7%
is retained on a 100-mesh sieve, and found that re- coveries were quantitative, neither the size-grading
nor the activation at 170° being necessary.
The procedure for saponification of the triglycerides
and isolation of the aqueous glycerol extract is almost
identical to that of Carlson and Wadstrom (2) and
the recovery experiments have confirmed that glycerol
is not lost during evaporation of the methanol after
saponification. For the acidification, 10 N sulfuric acid
was used in order to maintain the water content of the
final reaction mixture at a minimum.
As mentioned before, the fluorescence reaction is
subject to a fairly high reagent blank produced by the
large excess of o-aminophenol. It is therefore essential
that the reagent be as pure as possible and absolutely
colorless before use. The amount of oxidizing agent
present is not as important a factor. The production of
fluorescence with Mg++ ions will be affected only by a
very large excess of the latter, which may cause precip- itation of the Mg++-8-hydroxyquinoline complex.
In attempting to purify the technical grade o-ami- nophenol (Kodak) by recrystallization from ether, a
deep-red compound was isolated from the mother
liquor after chromatographic separation on a silicic- acid column according to the following procedure:
50 g of technical grade o-aminophenol was refluxed
for 1 hour with 500 ml of diethyl ether. Then 500 ml of
petroleum ether (b.p. 30°-60°) was added, and the
mixture cooled at —15° for half an hour. The residue
was filtered off, and the clear supernatant was concen- trated to a small volume (± 20 ml), and added to a
silicic-acid column (100 g; previously activated at 170°
and prepared in a 50:50 ether:petroleum ether mix- ture). Elution was carried out with this solvent
mixture (which removed dissolved o-aminophenol) un- til no residue was obtained on evaporation of an ali- quot of the eluatc. Elution was then continued with
pure ether, and a deep-red band began to move down
Page 150 of 168
50 MENDELSOHN AND ANTONIS J. Lipid Research
the column. The eluate containing this fraction was
collected and evaporated to dryness under N2 (yield:
approximately 0.5 g). This impurity has a profound
fluorescence quenching effect when added back to pure
o-aminophenol. On its own, however, it proved to be
a highly sensitive reagent for the quantitative estima- tion of glycerol. Four-tenths ml of a 0.0125% solution
of this compound freshly prepared in the 0.6% arsenic
acid solution, when added to the 0.1 ml glycerol ex- tract and treated according to the procedure outlined
above, produced an intense red fluorescence and a
much lower blank reading (Fig. 2). In addition, the
F ig. 2. Relationship between triglyceride concentration and in- tensity of fluorescence produced by red im purity obtained from
o-----o------o; 200 /ig standard set at 100% transmission.
x-----x------x; 100 fig standard set at 100% transmission.
fluorescence produced in the ammoniacal solution did
not require the presence of a divalent metal ion. The
nature of this compound is not known, but current in- vestigations suggest a highly polymerized product with
the required polar reactive groups, since removal of
either glycerol or the oxidizing agent from the reac- tion mixture did not produce fluorescence.
Application to the Estimation of Phosphatide Glyc- erol. The proposed method is particularly suited to
the determination of glycerol in glycerophosphatides.
Lecithins, cephalins, inositol phosphatides (containing
inositol instead of a base in the glycerophosphatide),
and lysolecithins obtained by fractionation of serum
phospholipids on silicic-acid columns according to
Hanahan et al. (16) have all given 1:1 glycerol:phos- phorus molar ratios when analyzed for glycerol content
according to the above procedure. The mild alkaline
hydrolysis recommended in the procedure readily de- acylates the fatty acid moiety of the glycerophospha- tide resulting in the production of an equilibrium mix- ture of a- and ^-glycerophosphate esters (17). It
would appear, therefore, that both of these esters react
quantitatively in the fluorescence reaction.
The authors express their thanks to F. E. du Toit,
for technical assistance,'and to Professor H. B. Stein,
of the Department of Pathology and Microbiology,
University of the Witwatersrand, Dr. J. H. S. Gear,
Director of the South African Institute for Medical
Research, and Dr. I. Bersohn, Director of the Ernest
Oppenheimer Heart Research Unit of the South Afri- can Institute for Medical Research, for their interest
in the work.
1. Van Handel, E., and D. B. Zilversmit. J. Lab. Clin. Med.
50: 152, 1957.
2. Carlson, L. A., and L. B. Wadstrom. Clin. Chim. Acta
4: 197, 1959.
3. Lambert, M., and A. O. Neish. Can. J. Research 28B:
4. Feigl, F. Spot Tests in Organic Analysis, 5th ed., New
York, Elsevier Publishing Company, Inc., 1956, p. 387.
5. Skraup, H. Mo. 2, 139, 1881.
6. Carlson, L. A., and L. B. Wadstrom. Third International
Conference on Biochemical Problems of Lipids, Brussels,
1956, p. 123.
7. Chen, P. S., T. Y. Toribara and H. Warner. Anal. Chem.
2 8 : 1756, 1956.
S. Sperry, W. M., and M. Webb. J. Biol. Chem. 187: 97,
9. Antonis, A. J. Lipid Research 1: 485, 1960.
10. Stern, I., and B. Shapiro. J. Clin. Pathol. 6: 158, 1953.
11. Thannhauser, S. J., and H. Reinstein. A.M.A. Arch.
Pathol. 3 3 : 646, 1942.
12. Fiske, C. H., and Y. Subbarow. J. Biol. Chem. 66: 375,
13. Barron, E. J., and D. J. Hanahan. J. Biol. Chem. 231:
14. Antonis, A., and I. Bersohn. Lancet 1: 99S, 1960.
15. Cheng, A. L. S., and D. B. Zilversmit. J. Lipid Research
1: 190, 1960.
16. Hanahan, D. J., J. C. Dittmer and E. Warashina./. Biol.
Chem. 228: 685, 1957.
17. Maruo, B., and A. A. Benson. J. Biol. Chem. 234 : 254,
Page 151 of 168
Abraham, S., Migliorini, R. H., Bortz, ¥. and Chaikoff, I. L.,
(196 2). The relation of lipogenesis to reduced triphosphopyridine
nucleotide generation and to certain enzyme activities In the liver
of the " totally " depancreatized rat. Biochim. biophys. Acta,
Adams, E., McFadden, M. and Smith, E. L., (19 5 2). Peptidases of
erythrocytes. 1. Distribution in man and other species. J. biol.
Chem., 198, 603.
Albert, D. J. and Brin, M., (i960). Comparison of serum and ery- throcyte hemolysate transaminase systems. Fed. Proc., 1%, 321.
Alivisatos, S. G. A., Kashket, S. and Den3tedt, 0. F., (1956). The
metabolism of the erythrocyte. IX. Diphosphopyridine nucleotidase
of erythrocytes. Canad. J. Biochem. Physiol., 34, 46.
Alles, G. A. and Hawes, R. C., (1940). Cholinesterases In blood of
man. J. biol. Chem., 133j 375.
Altman, K. I., Watman, R. N. and Salomon, K., (1951). The incor- poration of <sl~ C1^ _ acetate into the stroma of the erythrocyte.
Arch. Biochem. Biophys., 33j 168.
Altman, K. I., (1953). The in vitro incorporation of oc - C1^ -
acetate into the stroma of the erythrocyte. Arch. Biochem. Biophys.,
Altman, K. I. and Swisher, S. N., (1954). Incorporation of acetate-oc
C1^ into human erythrocyte stroma as a function of storage.
Nature, 174, 459.
Altman, K. I., (1959). Some enzymologic aspects of the human
erythrocyte. Amer. J. lied., 27, 936.
Alvlng, A. S., Tarlov, A. R., Brewer, G., Carson, P. E.,
Kellermeyer, R. W. and Long, W. K., (i960). Glucose-6-phosphate
dehydrogenase deficiency. Some biological implications. Trans.
Amer. Ass. Physcns., J3, 8l.
Anderson, E. P., Kalckar, H. M. and Isselbacher, K. J., (1957a).
Defect in uptake of galactose-1-phosphate into liver nucleotides
In congenital galactosemia. ' Science, 125, 113.
Anderson, E. P., Kalckar, H. M., Kurahashi, K. and Isselbacher, K.
J., (1957b). A specific enzymatic assay for the diagnosis congeni- tal galactosemia. I. The consumption test. J. Lab. Clin. Med.,
£ 0, 469.
Page 152 of 168
Angelone, L., Watkins, D. H. and Angerer, C. A., (195^). Oxygen
consumption of erythrocytes from patients with various thyroid
conditions related to their respective serum protein-bound iodine
concentrations. Blood, 9, 953.
Antonis, A., (1959/1960). The colorimetric determination of ester
groups in lipid extracts. J. Lipid Res., JL, 485.
Asatoor, A. M. and King, E. J., (1954). Simplified colourimetric
blood sugar method. Biochera. J., ^6 , xliv.
Barker, S. B. and Sumanerson, W. H., (1941). The colorimetric
determination of lactic acid in biological material. J. biol.
Chem., 13 8 , 535.
Barron, E. S. C-. and Harrop, Jr., 0. A., (19 28). Studies on blood
cell metabolism. II. The effect of methylene blue and other dyes
upon the glycolysis and lactic acid formation of mammalian 'and
avian erythrocytes. J. biol. Chem., 72, 6 5.
Barron, E. S. G., (19^3). Mechanisms of carbohydrate metabolism.
An essay on comparative biochemistry. Adv. Enzymol., 3, 149.
Bartlett, G. R. and Marlow, A. A., (1951). Enzyme systems in the
redblood cell. Bull. Scripps Metabol. Clin., 2, 1.
Bartlett, G. R., (1959). Human red cell glycolytic intermediates.
J. biol. Chem., 2^4, 449.
Behrendt, H., (19^3). Phosphatase of human erythrocytes. Proc.
Soc. exp. Biol., N. Y., j>4, 268.
Beloff-Chaln, A. and Pocchiari, F., (i960). Carbohydrate metabo- lism. Annu. Rev. Biochem., 2£, 295.
Berger, W., (1930). liber Glykolyse der roten Blutkorperchen. Arch,
exp. Path. u. Pharmakol., 150 , 298.
Bernard, C., (18 5 5). Lecons de physiologie experimental appliqule
a la medicine. Paris, Balliere, p. 379.
Bernard, C., (18 7 7). Lecons sur le diab&te. Paris, Balliere, p.
Bertino, J. R., Simmons, B. and Donohue, D. M., (1962). Purifica- tion and properties of the formate-activating enzyme from erythro- cytes. J. biol. Chem., 237, 131^.
Beutler, E. and Yeh, M. K. Y., (1959). Aconitase in human blood.
J. Lab. Clin. Med., 54, 456.
Bierman, E. L., Dole, V. P. and Roberts, T. N., (1957a). An abnor- mality of non esterified fatty acid metabolism in diabetes mellitus.
Diabetes, 6, 475.
Page 153 of 168
Bierman, E. L., Schwartz, I. and Dole, V. P., (1957b). Action of
insulin on release of fatty acids from tissues stores. Amer. J.
Physiol., 191, 359.
Blanchaer, M. C., Brox-mstone, S. and Williams, H. R., (1955).
Decreased phosohofructokinase activity in preserved blood. Amer.
J. Physiol., 18 3 , 95.
Brady, R. 0. and G-urin, S., (1950). Biosynthesis of labeled fatty
acids and cholesterol in experimental diabetes. J. biol. Chem.,
Brin, M., Shohet, S. S. and Davidson, C. S., (1958). The effect
of thiamine deficiency on the glucose oxidative pathway of rat
erythrocytes. J. biol. Chem., 230, 319.
Brin, II. and Yonemoto, R. H., (19 58). Stimulation of the glucose
oxidative pathway in human erythrocytes by methylene blue. J.
biol. Chem., 230, 307.
Bruns, F. H. and Hummel, W., (1950a). Uber der einflus von
Cystein und SII-Glutathion auf den Stoffwechsel der Kemlosen roten
Blutzellen. I. Mitteilung. Biochem. Z., 321, 197.
Bruns, F. H. and Pummel, W., (1950b). liber den einflus von h
Cystein und SII-Gluthion auf den Stoffwechsel der Kemlosen roten
Blutzellen. JI. Mitteilung. Zur Frage der Ferment - und Sub- stratspezifitat. Biochem. Z., 321, 236.
Bruns, F. H., (195^). Uber die Aldolase der Erythrocyten. Biochem.
Z., ^2S, H29.
Brums, F. H., IJoltmann, E. and Vahlaus,„E., (1958). liber den
Stoffwechsel von Ribose-5-phosphat in Ilamolycaten. I.
Aktivitatsmessung und Eigenschaften der Phosphoribose-isomerase.
II. Der Pentosphosphatcyclus in roten Blutzellen. Biochem. Z.,
330, ^8 3.
Buchanan, A. A., (i960). Lipid synthesis by human leucocytes in
vitro. Biochem. J., 75, 315.
Buffa, F. and Maralni, G., (1959). Recenti contributi in campo di
patologia enzymatica del globulo rosso. Minerva Hedica, 50, 3.
Cahill, Jr., G. F., Hastings, A. B., Ashmore, J. and Zoltu, S.,
(1958). Studies on carbohydrate metabolism in rat liver slices.
X. Factors in the regulation of pathways of glucose metabolism.
J. biol. Chem., 2^0, 125.
Cajori, F. A. and Crouter, C. Y., (192^). A comparison of the
rate of glycolysis in different bloods with special reference to
diabetic blood. J. biol. Chem., l60, 765.
Page 154 of 168
Carson, P. E., (i960). Glucose-6-phosphate dehydrogenase defi- ciency in hemolytic anemia. Fed. Proc., 1£, 995-
Carver, M. J. and Ryan, ¥. L., (i960). Stimulation of erythrocyte
metabolism by menadione. Proc. Soc. exp. Biol., N.Y., 104, 710.
Chaikoff, I. L., (1953). Metabolic blocks in carbohydrate metabo- lism in diabetes. Harvey Lectures, 47, 99.
Charlton, R. ¥. and Bothwell, T. H., (19 6 1). Primaquine-sensi- tivity of red cells in various races in Southern Africa. Brit,
med. J., I, 941.
Charlton, R. ¥. and Mendelsohn, D., (19 6 1). Certain aspects of
carbohydrate and lipid metabolism in primaquine sensitive erythro- cytes. S. Afr. J. med. Sci., 26, 109.
Chernick, S. S. and Chaikoff, I. L., (1950). Insulin and hepatic
utilization of glucose for lipogenesis. J. biol. Chem., 186, 5 3 5.
Clarkson, E. K. and Kaizels, M., (1952). Distribution of phos- phatase in human erythrocytes. J. Physiol., 116, 112.
Clutterbuck, P. ¥., (1928). SuccInoxidase; influence of phosphate
and other factors on action of succlndehydrogenase and fumarase of
liver and muscle. Biochem. J., 22, 1193-
Cohnstein, J. and Zuntz, N., (1884). Untgrsuchungqn uber d&s
Blut, den Kreislauf und die Atmung beim Saugtier-Fotus. Pflugers
Arch. ges. Physiol., 34, 173-
Comblath, M., Levin, E. Y., Marquetti, E. and House, E. Y., (i960).
Phosphorylase activity in human erythrocytes. Fed. Proc., 19, 68.
Couri, D. and Racker, E., (1959)- The oxidative pentose phosphate
cycle. V. Complete oxidation of glucose-6-phosphate in a recon- structed system of the oxidative pentose phosphate cycle. Arch.
Biochem. Biophys., 8j>, 195-
Covolo, G. C. and West, R., (1947). The activity of arglnase in
red blood cells. J. clin. Endocrin., 7, 325-
Daland, G. A. and Isaacs, R., (19 2 7). A comparative study of the
oxygen consumption of blood from normal individuals and patients
with increased leucocyte counts (sepsis; chronic myelogenous leuce- mia). J. exp. Med., 46, 53-
DanielS3on, H. and Horning, M. G., (1962). On the oxidation of
cholesterol by blood in vitro. Acta chem. scand., 16, 774.
Danon, D., Sheba, Ch. and Ramot, B., (19 6 1). The morphology of
glucose-6-phosphate dehydrogenase deficient erythrocytes: elec- tron - microscopic studies. Blood, 17, 225.
Page 155 of 168
Dennecke, C., (1926). Uber den Stoffwechsel der Erythrozyten im
anamischen Blute. Deutsch. rned. Wschr., 52, 280.
Derra, £., (19^8). Sauers toff z^hrung,, und Vitalgranulation bei
pemlzlozer Anamie nach Leberdiat. Munich med. Wschr., 75> 1494.
Dickens, F., (1938). Oxidation of phosphohexonate and pentose
phosphoric acids by yeast enzyme. I. Oxidation of phosphohexonate.
II. Oxidation of pentose phosphoric acids. Biochem. J., 32, 1626.
Dickens, F., (1951). Anaerobic glycolysis, respiration and the
Pasteur effect. The Enzymes, vol. 2 (Part I). Eds., J. 3. Sumner
and K. Myrback, Academic Press Inc., pp 624.
Dlsche, 2., (1938). Phosphorylierung der im Adenosln enthaltenen
d-RIbose und nachfolgender zerfall des Esters und Triosephosphat- bildung in Blut. Naturwisschenschaften, 26, 252.
Dische, Z., (1951). Synthesis of hexosemono- and diphosphate from
adenosine and ribose-5-phosphate In human blood, in Phosphorus
Metabolism, vol. 1. Eds. W. D. McElroy and B. Glass, Johns Hopkins
Press, Baltimore, p. 1 7 1 .
Dixon, M., (1952). Manometric Methods. 3rd. Edition, Cambridge
University Press, p. 8 3.
Dole, V. P., (1956). A relation between nonesterifled fatty acids
in plasma and the metabolism of glucose. J. clin. Invest., 35,
Dorfman, A., (1943). Pathways of glycolysis. Physiol. Rev., 23,
Drury, D. R., (1940). The role of insulin in carbohydrate metabo- lism. Amer. J. Physiol., 131, 536.
Dubovsky, J. and Sonka, J., (1955). Oxydative glycolysis of red
blood cells. Cas. 16k. csek., Jg4, 1027.
Duncan, M. and Sarett, H. P., (1951). Effect of nicotinic acid
and tryptophan on pyridine nucleotides of red blood cells In man.
J. biol. Chem., 1Q^, 317.
Elliot, W. H., (i960). Methylglyoxal formation from aminoacetone
by ox plasma. Nature, 18 5, 467.
Engelberg, H., (1958). Human endogenous plasma lipemia clearing
activity. Observations In 482 individuals. J. appl. Physiol.,
Engelhardt, W. A., (1930). Ortho - und Pyrophosphat in aeroben1
und anaeroben Stoffwechsel der Blutzellen. Biochem. Z., 227, 16.
Page 156 of 168
Engelhardt, V/. A. (1932). Die Beziehungen zwischen Atmung und
Pyrophosphaturnsatz in Vogelerythrocyten. Biochem. 2., 251, 343.
Engelhardt, W. A. and Ljubimova, M., (1930). Glykolyse und Phos- phorsaureumsatz in den Blutzellen verschiedener Tlere. Biochem.
Z,, 227, 6 .
Farvager, P. and Gerlach, J., (1955). Recherches sur la synthese
dds graisses a partir d'acetate ou de glucose. II. Les rbles res- pectifs due foie, du tissu adipeux et de certains autres tissus
dans la lipogenese chez le souris. Helvet. physiol, et pharmacol.
acta, 1 3 , 96.
Feigelson, P. and Conte, F. P., (195^). Studies on adaptive enzyme
formation in mammals. I. Galactose raetabolism. J. biol. Chem.,
Ferrari, V. and Gastaldi, F., (i960). Studio del metabolismo oss- idativo del glucoslo in eritrociti di ratti diabeticl. Arch, per
le scienze Medische, 109, 230.
Folch, J., Lees, M. and Stanley, G. H. S., (1957). A simple method
for the isolation of total lipides fromanim&l tissues. J. biol.
Chem., 221, 497.
Fredrickson, D. S. and Gordon, Jr., R. S., (19 58). Transport of
fatty acids. Physiol. Rev., 38, 505.
Friedemann, T. E. and Haugen, G. E., (19^3). Pyruvic acid. II.
The determination of keto acids in blood and urine. J. biol. Chem.,
Gabrlo, B. W. and Huennekens, F. M., (1955). The role of nucleo- side phosphorylase in erythrocyte preservation. Biochim. biophys.
Acta, 18, 585.
Gabrio, B. W., Finch, C. A. and Huennekens, F. M., (1956). Ery- throcyte preservation: A topic In molecular biochemistry. Blood,
n 10 3.
Garz<£ T., U^lmann, A. and Straub, F. B., (1952). Die ATPase der
roten Blutkorperchen. Acta physiol. Hungary, 3, 513.
Gibson, Q. H., (1948). The reduction of methaemoglobin in red
blood cells and studies on the cause of idiopathic methaemoglo- blnaemia. Biochem. J., 42, 13.
Gibson, D. M., Hubbard, D. D. and Love, W. C., (i960). A primary
defect in fatty acid biosynthesis In liver of alloxan diabetic
rats. Fed. Proc., ljj?, 227.
Goldstein, F. and Rieders, F., (1953). Conversion of thiocyanate
to cyanide by an erythrocyte enzyme. Amer. J. Physiol., 173. 287.
Page 157 of 168
Goodman, D. S., (19 58). The Interaction of human erythrocytes with
sodium palmitate. J. clin. Invest., ^7, 1727.
Gordon, Jr., R. S. and Cherkes, A., (1958). Unesterified fatty- acid in human blood plasma. J. clin. Invest., 35, 206.
Gordon, Jr., R. S., (1957). Unesterified fatty acid in human
plasma. II. The transport function of unesterified fatty acid.
J. clin. Invest., 38, 810.
Guardemanga, C., Massari, N. and Santambrogio, C., (1980). Enzyme
activity of human erythrocyte stroma at various ages. II. Choles- terol esterase and cholesterol. Giom. Gerontol., 8 , 161; Chem.
Abstr. j35, 718.
Guarino, A. J. and Sable, H. Z., (1955). Studies on phosphomutases.
II. Phosphorlbomutase and phosphoglucomutase. J. biol. Chem.,
Guest, G. M. and Rapoport, S., (1938). Effects of overdosage of
irradiated ergosterol in rabbits: changes of diphosphoglyceric
acid in the blood cells. J. biol. Chem., 124, 599.
Guest, G. M. and Rapoport, S., (1941). Organic acid-soluble phos- phorus compounds of the blood. Physiol. Rev., 21, 4l0.
Hagerman, J. S. and Gould, R. G., (1951). The in vitro Inter- change of cholesterol between plasma and red cells. Proc. Soc.
exp. Biol., N.Y., 78, 329.
Ilahn, L. and Hevesey, G., (1939). Interaction between phosphatides
of plasma and corpuscles. Nature, 144, 7 2 .
Handler, P. and Kohn, II. 1., (1943). The mechanism of cozymase
synthesis in the human erythrocyte: a comparison of the roles of
nicotinic acid and nicotinamide. J. biol. Chem., 1 5 O , 447.
Harrop, G. A., (19 19 ). The oxygen consumption of human erythro- cytes. Arch. int. Med., 2^, 745.
Harrop, G. A. and Barron, E. S. G., (1920). Studies on blood cell
metabolism. I. The effect of methylene blue and other dyes upon
the oxygen consumption of mammalian and avian erythrocytes. J.
exp. Med., 48, 207.
Hausberger, P. X., Milstein, S. 17. and Butman, R. J., (1954). The
influence of insulin on glucose utilization in adipose and hepatic
tissues in vitro. J. biol. Chem., 208, 431.
Hewson, W ., (1777). An experimental inquiry into the properties
of blood. 2nd. Edition, T. Cadell, London.
Hofmann, E. C. G., (1955). Abbau und Synthese des DPN in den roten
Blutkorperchen Kaninchens. Biochem. Z., 32 7, 273.
Page 158 of 168
Hofmann, E. C. G., Koradshowa, M. and Klecker, J., (1959). Uber
vergleichende Untersuchungen DPN-und TPN- spezifischer Nucleo- sidosen in Erythrocyten und Retllculocyten. Folia Haematol., 76 ,
Hokin, L. E. and Hokin, M. R., (19 6 1). Diglyceride kinase and
hosphatidic phosphatase in erythrocyte membranes. Nature, 189,
Holman, H. T., (i960). The ratio of trienoic:: tetraenoic acids
in tissue lipids as a measure of essential fatty acid requirement.
J. Nutr., 70, 405.
Hoppe-Seyler, F., (£867)). Zur Chemie des Elutes und seiner
Bestandtheile. Ned. chem. Untersuch. a. d. Lab....zu Tubing.,
Berlin, 1, 293.
Horvfitt, M. K., Harvey, C. C. and Century, B., (1959). Effects
of dietary fats on the fatty acid composition of human erythrocytes
and chick cerebella. Science, 130, 917.
Hrachovec, J. P., Lablpxc, M. and Rockstein, M., (19 6 1). Oxida- tion of palmitate-1 -0^4 by red blood cells. Proc. Soc. exp. Biol.,
N.Y., 107, 205.
Hsia, D. Y. Y., (19 6 1). Letter to Nature, 192, 266.
Huennekens, F. M., Liu, L., Myers, H. A. P. and Gabrio, B. W.,
(1957). Erythrocyte metabolism. III. Oxidation of glucose. J.
biol. Chem., 227, 253.
Hunefeld, F. L., (1840). Die Chemismus in der thierischen Organi- sation. Leipzig, p. loO.
Isselbacher, K. J., Anderson, P., Kurahashi, K. and Kalekar, H. M.,
(19 56). Congenital galactosemia, a single enzymatic block in
galactose metabolism. Science, 123, 6 35.
James, A. T., Lovelock, J. S. and Webb, J. P. W., (1959). The
lipids of whole blood. I. Lipid biosynthesis in human blood in
vitro. Biochem. J., 73* 106.
Jones, P. E. H. and McCance, R. A., (1949). Enzyme activities in
the blood of infants and adults. Biochem. J., 45, 464.
Johnson, A. B. and Marks, P. A., (1958). Glucose metabolism and
oxygen consumption in normal and glucose-6-phosphate dehydrogenase
deficient human erythrocytes. Clin. Res., o, 187.
Karmen, A., Wroblewski, F. and la Due, J. S., (1955). Transaminase'
activity in human blood. J. elln. Invest., 34, 126.
Kaplan, Nv 0., Colowick, S. P. and Barnes, C. C., (1951). Effect of
alkali on diphosphopyridine nucleotide. J. biol. Chem., 1 9 1 , 46l.
Page 159 of 168
Kashket, S., Rubinstein, D. and Denstedt, 0. F., (1957). Studies
on the preservation of blood. V. The influence of the hydrogen
ion concentration on certain changes in blood during storage.
Canad. J. Biochem. Physiol., 35, 827.
Kashket, S. and Denstedt, 0. F., (1958). The metabolism of the
erythrocyte. XV. Adenylate kinase of the erythrocyte. Canad. J.
Biochem. Physiol., 36, 1057.
Katayama, I., (1926/7). Studies in blood glycolysis. General con- sideration of glycolysis in relation to the blood colls, and the
production of lactic acid and carbon dioxide. J. Lab. clin. lied.,
Keilln, D. and Wang, Y. L., (1947). Stability of haemoglobin and
of certain endoerythrocytic enzymes in vitro. Biochem. J., 4l,
Kiese, M., (1944). Die Reduktion des Hamiglobin. Biochem. 2.,
Klein, J. R. and Kohn, H. I., (1940). Synthesis of flavine adenine
dinueleotlde from riboflavin by human blood cells. J. biol. Chem.,
Kohn, H. I. and Klein, J. R., (1939). Synthesis of cozymase and
of factor V from nicotinic acid by human erythrocytes in vitro and
in vivo. J. biol. Chem., 130 , 1.
Komberg, H. L. and Sadler, J. R., (i960). Microbial oxidationof
glycollate via a dicarboxylic acid cycle. Nature, 1 8 5 , 153.
Langdon, R. G., (1957). The biosynthesis of fatty acids in rat
liver. J. biol. Chem., 226, 6 1 5 .
Langdon, R. G. and Mize, C. E., (19 6 1). Recent studies on a mech- anism of action of insulin. Bull. Johns Hopkins Hosp., 106, 146.
Larizza, P., Brunetti, P., Grlgnani, E. and Ventura, S., (1958).
L' Individualita bioenzymatica dell-eritrocit.i " fabico " sopra
alcune anomalie biocheraische ed enzymatische delle emazie nei
pazienti affetti da favismo e nei loro familiari. Haematol., 43,
Laurell, S., (1957). Turnover rate of unesterified fatty acids in
human plasma. Acta physiol, scand., _4l, 15 8 .
Lauth, Ch., (18 76). Sur une nuovelle classe de matieres colorantes.
C. R. Acad. Sei., 82, l44l.
Leder, I. G. and Handler, P., (1951). Synthesis of nicotinamide
mononucleotide by human erythrocytes in vitro. J. biol. Chem.,
Page 160 of 168
Leeuwenhoek, Anton van, (1719/33). Of the globules of blood. Phil.
Tr., London, J_, 562.
Lepine, R., (1890). Les sucres du sang. C. R. Acad. Scl., 112, 6 5.
Levene, P. A. and Meyer, G. M., (1912/13). On the combined action
of muscle plasma and pancreas extract on some mono- and disaccha- rides. J. biol. Chem., 1A, 347.
Lohr, G. W. and Waller, H. D., (1958). Hamolytische erythrocyto- pathle durch fehlen von glukose-6-phosphat dehydrogenase in roten
Elutzellen als dominant verer-bliche Xrankheit. Klin. Wschr., 36,
Lohr, G. W., Waller, II. D., Karges, 0., Schlegel, B. and Muller,
A. A., (19 58). Zur Biochemie der Alterung menschlicher Erythrocy- ten. Klin, Wschr., 3 6 , 1008.
London, I. M. and Schwartz, II., (1953). Erythrocyte metabolism.
Metabolic behaviour of the cholesterol of human erythrocytes. J.
clin. Invest., ^2, 1248.
London, I. M., (i960). Metabolism of the mammalian erythrocyte.
Bull. N.Y. Acad. Med., 36, 79.
Long, W. K. and Carson, P. E., (19 6 1). Increased glutathione re- ductase activity in diabetes mellitus. Biochem. Biophys. Res.
Comm., 5, 394.
Lovelock, J. E., James, A. T. and Rowe, C. E., (i960). The lipids
of xtfhole blood. II. The exchange of lipids between the cellular
constituents and the lipoproteins of human blood. Biochem. J.,
Lowenstein, J. M., (19 6 1). The pathway of hydrogen in biosyntheses.
I. Experiments with glucose-l-IP and lactate-2-IP. J. biol. Chem.,
Lowenstein, L. M., (1959). The mammalian reticiilocyte. Int. Rev.
Cytol., 8, 135.
Kacleod, J. J. R., (1913). Blood glycolysis : its extent and sig- nificance in carbohydrate metabolism. The supposed existence of
" sucre virtuel " in freshly drawn blood. J. biol. Chem., 1/5, 497.
Malkin, A. and Denstedt, 0. F., (19 56). The metabolism of the
erythrocyte XII. Diphosphopyrldlne nucleotide nucleosidase of the
rabbit erythrocyte. Canad. J. Biochem. Physiol., J36, 141.
Marlnello, E., (19 58). Identificazione del sedoeptulose-7-fosfato
tra i prodotti del metabolism del ribose-5-fosfato in erltrocitti
umani. Arch. sci. biol. Bologna, 42, 320.
Page 161 of 168
Marks, P. A., (1958). Red cell glucose-6-phosphate and 6-phospho- gluconic dehydrogenase and nucleoside phosphorylase. Science, 127,
Marks, P. A., Gellhorn, A. and Kidson, C., (i960). Lipid synthe- sis in human leukocytes, platelets and erythrocytes. J. biol. Chem.,
Masoro, E. J., Chaikoff, I. L. and Dauben, VJ. G., (1949). LIpo- genesis from glucose in the normal and liverless animal as studied
with C^-4 - labeled glucose. J. biol. Chem., 179, 1117.
Masoro, E. J., (196 2). Biochemical mechanisms related to the
homeostatic regulation of lipogenesis in animals. J. Lipid Res.,
2 , 149.
. " 11
Matthies, H., (1959). Uber eine Aldehyddehydrase in roten Blutkor- perchen. Folia Haematol., 7 6, 439.
Matthes, K. J., Abraham, S. and Chaikoff, I. L., (i960). An enzyme
defect in fatty acid synthesis by alloxan-diabetic rat liver.
Biochim. biophys. Acta, 27, l80.
Mauri, C. and Torelli, y., (1959). Glutamin-desaminierung in
normalen Erythrocyten wahren