1 s2.0 0009898183903704 Main
1 s2.0 0009898183903704 Main
1 s2.0 0009898183903704 Main
Elsevier
CCA 02682
Summary
Introduction
The glucosylation of human haemoglobin A to the form HbA,, has been shown
to proceed via an intermediate labile haemoglobin glucose adduct transformed in an
Amadori rearrangement to HbA,, [l]. The formation of the labile glucosylated
haemoglobin fraction, termed anodal glucohaemoglobin A [2,3] has been recognised
Fig. 1. Haemoglobin isoelectric focusing pattern in whole blood from a diabetic patient. The numbers
designate the following haemoglobin fractions: 1, HbA,; 2, anodal glucohaemoglobin A; 3, HbA,,; 4,
HbA: 5, met HbA; 6, glucosylated HbA,; 7, HbA,.
319
rearrangement to HbA,, (the ketoamine), all reactions are reversible [2]. The
rate-determining step of the Amadori rearrangement is considered to be loss of the
C-2 hydrogen atom. Based on these mechanisms a reversal of the ketoamine into
haemoglobin A and a mixture of glucose and mannose would be anticipated. The
emergence of both aldohexoses would be expected since the asymmetry of C-2 is
destroyed in the reaction [ 111. The purpose of the present study is to determine the
kinetic and equilibrium conditions governing the formation of HbA,,. The clinical
relevance of this investigation is the enhanced understanding of glucosylated
haemoglobin levels during normal metabolic states as well as during periods of
improved or deteriorating glycaemic control in diabetic patients.
In order to study the steady state conditions of HbA,,, 24 children with newly
diagnosed diabetes mellitus (age range 1 to 15 years) were selected. The length of the
period since symptoms were first noticed until diagnosis was on average 4.6 weeks.
In each of the first 24 h the blood glucose was determined. During this period the
mean blood glucose was considered to be representative of the integrated blood
glucose level for the period from symptoms first occurring until admission. High
glucose levels have previously been shown to affect the HbA,, determination when
carried out by IEF [2,12], and this may lead to a systematic overestimation of the
measured HbA,, on the day of admission. Consequently in these patients the HbA,,
values determined on days 2 to 3 were used to reflect the blood glucose level for the
previous period.
Haemolysates with a haemoglobin concentration of 2 to 3 mmol/l were prepared
from 0.5 ml washed and packed red cells lysed in 3.5 ml distilled water.
The samples were stabilised with carbon monoxide [ 131 and stored at - 30°C for
1 to 3 days until analysis. In the experiments in Fig. 2 the haemoglobin solution was
stabilised with 0.03 mmol/l potassium cyanide [ 141 (pH 8.5) to prevent formation of
extra bands due to the presence of fully and partially oxidised haemoglobins [ 151.
Incubation experiments were performed on 2.0 ml washed red cells in 120 ml 0.15
mol/l saline from 0 to 25 mmol/l D-glucose (analytical grade, Sigma). The glucose-
rich saline solution was changed twice within 24 h. The incubation period ranged
from 0 to 60 h. The samples were kept in a thermostatically controlled water bath at
4, 15 and 37°C respectively. The total amount of haemoglobin as measured by Van
Kampen’s modification of Drabkin’s method [ 161 was unchanged during the incuba-
tion at 4 and 15’C. Erythrocytes convert glucose to lactate resulting in cell
haemolysis, and at 37’C a decrease in total haemoglobin occurred when the
incubation period exceeded 24 h. At the lower temperatures the samples were stable
for 60 h.
The separation of haemoglobins was carried out by isoelectric focusing on a
polyacrylamide gel slab (Ampholine-PAG LKB 1804- 13 1, LKB-Produkter AB,
S-161,26 Bromma, Sweden). The HbA,, band was cut out and eluted and the HbA,,
concentration related to total haemoglobin determined spectrophotometrically at
417 nm. Using this technique the minor haemoglobins HbA,, + A,, focused for
320
Fig. 2. Haemoglobin isoelectric focusing pattern in: A, whole blood from a diabetic patient incubated in
saline for 3 h at 37’C; B, purified HbA,, incubated in saline for 20 min at 37°C; C, purified HbA,,
incubated in saline for 24 h at 37°C. The numbers designate the following haemoglobin fractions: 1,
HbA,,; 2, HbA; 3, met HbA; 4, HbA,; 5, HbA,.
towards the anode, well separated from the HbA,, and anodal glucohaemoglobin A
fractions. For further details of the full assignment of the haemoglobin bands see [2].
The intra- and interassay variations for the method were 2 and 4.2%, respectively.
slab stained with Coomassie Brilliant Blue [14]. The relative concentrations of
HbA,, and HbA were determined from a densitometer record performed on a Laser
scanner SL-504 from Biomedical Instruments (Chicago, IL, USA).
Results
Formation of HbA,,
The rate of synthesis of HbA,, was measured by incubation of red cells in 10, 15,
20 and 25 mmol/l o-glucose from 0 to 24 h at 37°C. The results and the rate
constant k, for the forward reaction are shown in Table I.
Dissociation of HbA,,
Purified HbA,, incubated in 0.15 mol/l saline from 0 to 24 h followed by
isoelectric focusing gave rise to a progressively increasing HbA concentration while a
corresponding decrease in the HbA,, concentration occurred (Fig. 2). The results of
the densitometer record of the two haemoglobin fractions are given in Table II. Red
cells were incubated in 0.15 mol/l saline at 4, 15 and 37°C from 0 to 60 h. The rate
of dissociation of HbA,, at the different temperatures is shown in Fig. 3. HbA,,
gave T; 113 h (37°C) while a reduction of 1.5% and 0.7% was found at 15 and 4°C
during 60 h incubation. The rate constant k_, (mean f SE) (n = 24) was 0.45 f 0.01
x 10-6.s-’ (4”(J), 0.85 + 0.02 x 10-6.s-’ (15’C) and 1.7 f 0.05 X 1O-6. s-i
(37°C). The apparent Arrhenius activation energy for the reverse k_, reaction was
calculated to be 28.1 kJ per degree. Table III summarises the results of measured
and predicted HbA,, levels. The predicted values of HbA,, were calculated on the
basis of the equilibrium constants K, and K, and the glucose levels in question (5%
HbA,, = 100 X K, x K, X glc)while the measured HbA,, was determined in children
with newly diagnosed diabetes mellitus. The children were divided into three
subgroups according to the HbA,, level. A close correlation was found between
observed and predicted HbA,, values.
TABLE I
Percentage HbA,, formed after incubation of red cells for 24 h at 37’C at glucose concentrations of 10,
15, 20 and 25 mmol/l
Basal HbA,, value in the red cells before incubation was 5.2 f0.061 (n = 10). (The values are given as
mean *SE)
1200 IO 130
86400 40 100
1 I I *
I I I I
10 20 30 40 50 60 hours
4
13
1
‘;;370c
I , 1 I I I I I 1 *
2 6 10 14 18 22 26 30 hours
Incubation time
Fig. 3. Dissociation of isoelectric purified HbA,, in saline at 4, 15 and 37°C. Note that the time interval
on the X-axis is different in the upper and lower panels.
323
TABLE III
Comparisons of measured and predicted HbA ,c values in children with newly diagnosed diabetes mellitus
(The values are given as mean f SD)
HbA,,9-11%
n = 14 12.7k2.1 10.1~0.5 10.2
HbA,, < 9%
?I=5 8.6k2.1 7.7* 1.1 7.0
The first order rate constant, k_,, for the dissociation of HbA,, was calculated
from the initial slope of the regress curve determined during incubation of HbA ,c in
saline (Fig. 3).
The rate constant, k,, for the formation of HbA,, could now be calculated since
d(HbA,,)/dt = k, X (HbA aldimine) - k_, X (HbA,, ketoamine). Consequently
A(HbA,, ketoamine)
k _ 2 (HbA I= ketoamine) + = k, (HbA aldimine),,
At
where % HbA,, and A ‘5%HbA,, are the relative concentration and the increase in
relative concentration of ketoamine, respectively, at time t. The equilibrium con-
centration of aldimine is designated % aldimine,,.
From these data the equilibrium constant K, = k,/k_, for the formation of
HbA,, was calculated to be 8.4 k 0.3 at 37°C (mean &SE).
324
Discussion
From the results of this study it is evident that the glucosylation of human
haemoglobin to HbA ,c is a reversible reaction since HbA is regenerated during
prolonged incubation of HbA,, (Fig. 2). Owing to these findings there is now no
need to invoke a saturable glucosylation system in man as proposed by Graf et al
[171.
Bunn has previously considered the rearrangement reaction to be irreversible and
using this assumption determined the first order rate constant k, (I .5 x 10ph ’ sp ’ )
for the forward reaction from in vivo [I] and in vitro studies [3]. Taking into account
the reversibility of the Amadori rearrangement, recalculation of the data of Bunn [3]
give a mean value of k, of 4.0 x 10m6. SC’. We have determined the value of k, to
be 14.2 x 10m6. s-‘. However, recalculation of the literature data show a pro-
nounced variation of values e.g. Boden et al [lo] reported data for a group of
insulin-independent diabetic patients giving k, = 28.7 X 10e6. sm.‘. From the data of
Svendsen et al [ 181 k, may be calculated as 7.1 X lo-” . SC’ for a group of
insulin-dependent diabetic patients, while the data of Vialettes et al [9] on diabetic
rats give a value of k, = 14.8 x 10e6. SC’. In their study Weykamp and Penders [ 121,
using “C-glucose incorporation, determined k, = 1.17 X 10m6. sp ‘. This value was
based on the assumption of irreversibility of the HbA,, formation and of a value of
K, = 2.12 l/mol. However, the latter value may be overestimated since it has been
shown that “C-glucose contains reactive impurities [ 191.
The reason for the discrepancy between Bunn’s original value of k2 and the one
determined in this study is at least partly due to the negligible contribution of the
backward reaction. The expression of Bunn is still valid if initial rates are considered
since the backward reaction may be neglected at the start of the reaction. However,
experimentally initial rates are difficult to handle since the equilibrium between
haemoglobin A plus glucose and the aldimine must be established before the steady
state expression describing the initial rate is valid.
Determination of the rate constant of the first order backward reaction, k_,
(1.7 x 10 -6. s-‘) is somewhat hampered by the instability of the haemoglobin A
system. As evident from the densitometer record of Table II this obstacle could be
overcome by using incubation periods not exceeding 24 h at 37°C. Within this
period the densitometer record verifies that the haemoglobin lost in the dissociation
of HbA,, appears as haemoglobin A within the experimental error. The temperature
dependence of the reverse reaction gives rise to an apparent Arrhenius activation
energy of 28.1 kJ per degree, a value much too low to allow any deductions on
activation parameters. This finding is parallel to the situation observed in the
pre-equilibrium (k, and k_ ‘) [2]. The values determined for k, and k_, allow the
equilibrium constant for the formation of HbA,, to be calculated: K = k,/kp, = 8.4.
Since the equilibrium constant for pre-equilibrium has been determined [2] K, = 0.96
l/mol, the system is now kinetically and thermodynamically defined. Therefore,
given a glucose and a HbA concentration, all other concentrations may be calculated
as a function of time, e.g. the equilibrium concentration of HbA,, may be predicted
to be 5.0% with a mean blood glucose of 6 mmol/l. This value compares favourably
32.5
with the experimentally determined HbA,, value of S.5% for normal controls, even
though the calculation does not take into account the removal of HbA,, by decay of
erythrocytes. Furthermore from the results in Table III it is seen that the value for
HbA,, calculated on basis of the equilibrium constants was in close agreement with
the HbA,, value observed at various glucose concentrations in diabetic children.
Clearly further calculations taking into account erythrocyte decay are now war-
ranted. However, the knowledge of the system presented here is qu~itatively in
accord with the clinical observation that glucohaemoglobin levels correlate well with
blood glucose and glucosuria for the preceding few weeks [7,8]. Also, it offers an
explanation of the fact that HbA,, decreased more rapidly than expected owing to
erythrocyte decay during improvement of metabolic control in diabetic patients
[8,9,101.
The instability of HbA,, during saline incubation must be taken into considera-
tion when aldimine is elimitaed by this process [20]. This stresses the need for a
method such as isoelectric focusing, capable of separating the ketoamine from the
aldimine fraction without any pretreatment of erythrocytes.
Acknowledgements
H. Olesen, MD, is thanked for critial discussion of the manuscript. Lene Nielsen
is thanked for her skilful technical assistance and Kirsten Rosby for her able
secretarial assistance. We are indebted to Aage Vgund, MSc, Novo Research
Institute, for carrying out the statistical evaluation.
References
I Buna HF. Haney DN, Kamin S, Gabbay KH, Gallop PM. The biosynthesis of human hemoglobin
A,,. J Clin Invest 1976; 57: 1652-1659.
2 Mortensen HB, Christophersen C. Glycosylation of human hemoglobin A. Kinetics and mechanisms
studied by isoelectric focusing. Biochim Biophys Acta 1982; 707: 154-163.
3 Higgins PJ, Bunn HF. Kinetic analysis of the nonenzymatic glycosylation of hemoglobin. J Biol Chem
1981; 256: 52045208.
4 Koenig RJ, Peterson CM, Jones RL, Saudek C, Lehrman M, Cerami A. Correlation of glucose
regulation and hemoglobin A,, in diabetes mellitus. N Engl J Med 1976; 295: 417-420.
5 Gabbay KH, Hasty K, Breslow JL, Ellison RC, Bunn HF, Gallop PM. Giycosylated hemoglobins and
bng term blood glucose control in diabetes mellitus, J Clin Endocrinol Metab 1977; 44: 859-864.
6 Blanc MH, Barnell DM, Gleason RE, DUM PJ, Soeldner JS. Hemo~obin A,, compared with three
conventional measures of diabetes control. Diabetes Care 1981; 4: 349-353.
7 Dunn PJ, Cole RA, Soeldner JS et al. Temporal relationship of glycosylated haemoglobin concentra-
tion to glucose control in diabetics. Diabetologia 1979; 17: 213-220.
8 Mortensen HB, Vestermark S, Kastrup KW. Metabolic control in children with insulin dependent
diabetes mellitus assessed by hemoglobin A ,=. Acta Paediatr &and 1982; 71: 217-222.
9 Vialettes B, Vovan L, Simon MC, Lassmann V, Altomare E, Vague SL. Kinetics of fast haemoglobin
in diabetic rats. Diabetologia 1982; 22: 264-268.
10 Boden G, Master RW, Gordon SS, Shuman CR, Owen OW. Monitoring metabolic control in diabetic
outpatients with glycosylated hemoglobin. Ann Intern Med 1980; 92: 257-260.
11 Bunn HF, Haney DN, Gabbay KH, Gallop PM. Further identification of the nature and linkage of
the carbohydrate in hemoglobin A,,. B&hem Biophys Res Commun 1975; 67: 103-109.
12 Weykamp CW, Penders TJ. Mechanism and speed of reactions between haemo~obin and glucose;
326
consequenses for the measurement of glycosylated haemoglobins in patient maternal. Clin (‘him Acta
1982; 125: 341-350.
13 Jeppsson JO, Franzen B, Gaal AB. Simplified determination of hemoglobin A ,c m diabetic patrents by
use of electrofocusing. In: Electrophoresis. Vol. 79. Radola BJ ed. New York: decruyter, 1980:
655-661.
14 Mortensen HB. Quantitative determination of hemoglobin A,, by thrn-layer isoelectrrc focusrng. J
Chromatogr 1980; 182: 325-333.
15 Bunn HF. Drysdale JW. The separation of partially oxidized hemoglobins. Biochim Biophys Acta
1970; 229: 51-57.
16 Van Kampen EJ, Zijlstra WC. Standardization of hemoglobinometry. II. The hemiglobincyanide
method. Clin Chim Acta 1961; 6: 538-544.
17 Graf RJ, Halter JB, Porte D. Glycosylated hemoglobin in normal subjects and subjects with
maturity-onset diabetes. Diabetes 1978; 27: 834-839.
18 Svendsen PAa, Christiansen JS, Soegaard U, Nerup J. Synthesis of glycosylated haemoglobin in viva.
Diabetologia 1981; 21: 549-553.
19 Triieb B, Holenstein CC, Fischer RW. Winterhalter KH. Nonenzymatic glycosylation of proteins. J
Biol Chem 1980; 255: 6717-6720.
20 Mortensen HB, Marshall MO. Effect of saline incubation on red cell content of glucosylated
haemoglobins studied by iso-electric focusing. Clin Chim Acta 1983; 132: 213-217.