| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
The Department of Clinical Chemistry, Leiden University Medical Center, 2300 RC Leiden, The Netherlands
Address all correspondence and requests for reprints to: E. G . W .M. Lentjes, Department of Clinical Chemistry, Leiden University Medical Center, P.O. Box 9600, 2300 RC Leiden, The Netherlands.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We conclude that the erythrocyte-associated cortisol fraction is relatively undervalued but can serve as an important transport vehiculum and storage compartment for cortisol. This fraction can have a considerable effect on the total plasma and free cortisol concentration when strict temperature control during sample handling is not considered.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The free fraction and the albumin-bound cortisol, because of its low binding affinity, are considered responsible for the biological activities of the hormone. Most studies have focused on the distribution of cortisol between plasma water, the free fraction, and the binding proteins.
In contrast to the vast amount of literature on the subject of hormone concentrations in plasma (the free hormone fraction), only a few studies consider the erythrocyte-bound fraction of hormones, although it has been known for decennia that the erythrocytes can bind cortisol (1). One reason is possibly that samples of erythrocytes are always contaminated with plasma (2); and removal of the plasma, by washing the cells, can result in partial loss of an erythrocyte-bound hormone. Another reason is that erythrocyte-bound cortisol was thought not to redistribute to other compartments and, therefore, not to be of interest, from the pharmacokinetic point of few (3).
The first study of cortisol associated with erythrocytes was reported in 1953 (4). Subsequently, Peterson et al. (5) demonstrated that 20–30% of the cortisol in human blood was associated with erythrocytes. This was confirmed by Migeon et al. (6), who found, on average, 25% of the 17-hydroxycorticosteroids to be associated with the erythrocytes, and by Vermeulen (7), who reported that, after infusion of 14C-cortisol in man, 16–37% of the radioactivity was associated with the erythrocytes. Farese and Plager (8) found that the erythrocyte uptake of cortisol ranged from 3–31% and was dependent upon the protein binding of cortisol. Many years later, it was shown that the free cortisol concentration determines the extent of erythrocyte binding and not the CBG concentration. This was reported in two detailed studies on the erythrocyte-associated cortisol (9, 10). The ratio of erythrocyte-cortisol to plasma free cortisol was 2.62 ± 0.16 in six healthy men, measured at 36.8 C (9), or 2.4 at 20 C (10).
We showed previously that the free cortisol concentration in blood was greatly influenced by the temperature (11). The aim of the present study was to investigate whether the binding of cortisol to the erythrocytes is also temperature dependent and to explore the effect of temperature changes on the distribution of cortisol in whole blood.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Venous blood from healthy Caucasian men (age, 19–40 yr) was collected in EDTA-containing tubes (Vacutainer, Becton Dickinson and Co., Rutherford, NJ). The subjects were between 18 and 40 yr of age and did not take any medication. The blood samples were centrifuged at the indicated temperature: 20, 30, or 37 and 40 C. The erythrocytes and the plasma were used within 2 h after blood sampling.
Chemicals
Cortisol was from Sigma Chemical Co. (St. Louis, MO). Other reagents were from Merck (Amsterdam, the Netherlands). [14C]Sucrose (SA, 442 mCi/mmol) and [1,2,6,7-3H]-cortisol (SA, 85.5 Ci/mmol) were obtained from NEN research products (DuPont de Nemours, Boston, MA). This tracer was purified, before use, on a solid-phase (C18) extraction column from J. T. Baker (Phillipsburg, NJ).
Cortisol-free plasma
Cortisol-free plasma was obtained by charcoal treatment (1). To 10 mL plasma, 500 mg charcoal (Merck, Darmstadt, Germany) was added, and the sample was mixed for at least 1 h. The charcoal was removed by centrifugation for 30 min at 2,000 x g and then for 45 min at 130,000 x g. This procedure was repeated, and the plasma was then filtered through a 0.22-µm filter (Costar, Cambridge, MA). Aliquots of this plasma were stored frozen at -20 C.
Cortisol assay
Cortisol measurements in plasma and filtrate samples were done with an RIA. Cortisol was extracted from the samples with 2 x 5 mL ethyl acetate. The organic layers were combined and evaporated to dryness. The residue was dissolved in 500 µL Tris buffer (0.05 mol/L, pH 8.0), and the cortisol in these residues was determined in the assay. A cortisol standard concentration curve, from 0–550 nmol/L, was constructed in phosphate-buffered saline (1.4 mmol/L phosphate and 0.15 mol/L NaCl, pH 7.5) or in cortisol-free (charcoal treated, see above) plasma. These standard curve samples were also extracted with 2 x 5 mL ethyl acetate. In this way, recoveries of cortisol from samples and corresponding standard samples were identical.
The within-run precision was less than 5.5%; the interassay precision, determined over a period of 1 yr, was 10%. The detection limit was less than 1 nmol/L.
Free cortisol
For the determination of the plasma free cortisol concentration, samples were filtrated at 20, 30, 37, and 40 C, using the Amicon MPS-1 ultrafiltration device, according to the method described previously (11). The temperature was monitored during the centrifugation process and was constant at all four temperatures, with a maximal variation of ±0.1 C. The cortisol concentration in the plasma and filtrate specimens were determined with an RIA, as described above.
Cortisol bound to erythrocytes
Erythrocyte-associated cortisol was determined according to Hiramatsu et al. (9), with some modifications. Briefly, an aliquot of blood was centrifuged, and the erythrocyte pellet and plasma were separated. The total amount of cortisol extracted from the erythrocyte pellet (which also contained some entrapped plasma) was determined. The amount of cortisol in the entrapped plasma was then subtracted from the total amount of cortisol. The plasma volume was estimated by using 14C-sucrose (description below).
Several blood-collecting tubes of the same volunteer were combined, and 3.0-mL aliquots were subsequently taken for the experiments. We sometimes increased the cortisol concentration in the samples by adding a small amount of the cortisol standard. A 500-µL aliquot was used for the determination of the hematocrit (Ht) using 14C-sucrose (12). The 3.0-mL aliquots were incubated for 1 h at 20, 30, 37, or 40 C in a water bath and then were centrifuged for 15 min at 1,500 x g at the same temperature in a temperature-controlled centrifuge. In the plasma, separated from the erythrocyte pellet, the total ([Cortpl-1]) and free cortisol ([Cortfree]) concentrations were determined. The erythrocyte pellet was allowed to reach room temperature (20 C). Cortisol that bound to erythrocytes (Erycort) in this pellet, which also contains a small volume of entrapped plasma of volume VR, was then extracted for 20 min at 20 C by resuspending the erythrocytes with 600 µL cortisol-free plasma (see Ref. 13), to which 14C-sucrose was added (Suc1; 14,000 dpm per 100 µL). The samples were then centrifuged at room temperature for 15 min at 2,500 x g. The extraction plasma was removed and stored at 4 C for the determination of the cortisol concentration ([Cortpl-2]) and for counting the radioactive sucrose (Suc2).
Calculation of the volume of the entrapped plasma
The volume of the entrapped plasma (VR) in the erythrocyte pellet was calculated from the dilution of the radioactive sucrose tracer, which is not absorbed by the erythrocytes, according to the formula: VR (µL) = 600 (Suc1 - Suc2)/Suc2. Suc1 and Suc2 are the concentrations of the radioactive sucrose tracer (in dpm per 100 µL); 600 is the volume in µL of the added cortisol-free plasma to which 14C-sucrose was added.
The space associated with erythrocytes that is accessible to small nonprotein molecules (such as sucrose or free cortisol, but not protein-bound cortisol) is 2.1% of the erythrocyte volume, as reported by Hiramatsu et al. (9). We made the appropriate corrections for the amount of cortisol in the entrapped plasma (see below).
Calculation of cortisol bound to erythrocytes
The total amount of cortisol in the extraction plasma (A-Cortpl-2) was calculated from the volume (600 µL plus the entrapped plasma volume, indicated by VR) and the cortisol concentration in this extraction plasma: A-Cortpl-2 = (600 + VR)·[Cortpl-2]. The amount of cortisol bound to the erythrocytes was then obtained by subtracting the amount of cortisol in the entrapped plasma from the total amount of cortisol in the extraction plasma (A-Cortpl-2): Ery-cort = A-Cortpl-2 - (VR·[Cortpl-1]).
The total amount of cortisol in the plasma of the original 3-mL blood sample (A-Cortpl-1, in pmol) was calculated from the Ht and the plasma cortisol concentration according to: A-Cortpl-1 = 3·(1 - Ht)·[Cortpl-1].
Cortisol binding to plasma proteins
For the binding studies, we used plasma from healthy volunteers. The plasma was untreated or steroids were removed by charcoal adsorption (stripped), as mentioned above.
For Scatchard analysis, a series of 6–10 tubes was used, each containing plasma to which an increasing amount of cortisol was added. Incubations were done in borosilicate tubes. When using the radioactive cortisol tracer, a volume of the tracer (100,000 dpm) was pipetted in each tube, and the solution was evaporated to dryness. Then, 400 µL of stripped plasma and 40 µL of a cortisol standard in phosphate-buffered saline (10–12 cortisol working standards were used, giving a final concentration of 50–1,200 nmol/L) were added and mixed. The tubes were incubated for 60 min at 20, 30, 37, or 40 C. Then, the samples were filtered in the Amicon ultrafiltration apparatus at the indicated temperature. The radioactivity in the filtrates was measured in a scintillation counter, or the cortisol concentration was measured in the RIA.
In addition, the nonspecific binding was estimated after heating the plasma for 30 min to 60 C to destroy the CBG binding (14).
The association constant (Ka) of CBG for cortisol was estimated from the slope of the Scatchard plot, and the maximum number of binding sites was calculated from the intercept on the abscissa, when the bound/free ratio equals zero (15).
CBG was estimated by an RIA from Medgenix (Amersfoort, The Netherlands).
Mathematical approach of the binding of steroids to proteins
Consider a system of steroid hormones and binding proteins in binding equilibrium.
The relation for a particular steroid hormone (i) between the plasma
concentration of bound ([HiP]) and free
([Hi]) hormone, in a system of several steroid hormones
(H) and one binding protein (P) can be described by the equation (16, 17):
 | (1) |
In the situation of one steroid hormone (H) and several binding
proteins (Pj), the relation is described by the
equation:
 | (2) |
When there are more steroids that compete for the same binding places
on the proteins, the bound-to-free ratio for steroid i is written
as:
 | (3) |
These calculations can be automated in a spreadsheet computer program. The calculations were verified by comparing the results with those of Dunn et al. (18), using their affinity constants and free hormone concentrations for the 21 steroids and three-protein (CBG, SHBG, and albumin) system at 37 C. However, little information is available about affinity constants of most steroids, or free steroid concentrations at other temperatures. Therefore, we used a system of two steroids (cortisol and cortisone) and two binding proteins (albumin and CBG) for the computer simulations. The differences in the fractional distribution of cortisol in this system, compared with that of Dunn et al. (21 steroids and three proteins) were insignificant: 0.8–3.3% lower values at 0.1 µmol/L plasma cortisol and 0.5–0.7% lower at 1 µmol/L plasma cortisol concentration.
We determined the binding affinity for cortisol to CBG at 20, 30, 37, and 40 C and calculated the binding affinities at other temperatures by interpolation, using a van ‘t Hoff plot (plot of log(K) vs. 1/temperature in degrees Kelvin) (see Ref. 19). The binding affinities of cortisol for human albumin at different temperatures were determined similarly, using the values reported by Westphal (table VI-9) (16).
Statistics were performed with the statistical package SPSS version 6.1.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Another point of concern was the extraction of the erythrocyte-bound cortisol by using the high-affinity binding places of CBG. This proved to be highly efficient: in several experiments (data not shown here), we could extract the cortisol added to cortisol-free blood, for more than 95%. The extraction with this cortisol-free CBG was done at room temperature because the binding affinity of cortisol for CBG is about 10-fold higher at 20 C, compared with 37 or 40 C.
Measurements of the different cortisol fractions were performed at least in triplicate. The within-run coefficient of variation (CV) for the erythrocyte-bound concentration ranged from 2.5–11% (for values >50 pmol cortisol/mL erythrocytes, corresponding to free cortisol concentrations of 20 nmol/L). At lower concentrations, the within-run CV increased to 30% at 25 pmol cortisol/mL erythrocytes. For the free cortisol measurements, the within-run CV’s were less than 20% for the range 0–10 nmol/L, 2–7% for 10–70 nmol/L, and 0.9–3% at 70–140 nmol/L.
Total and free cortisol concentrations in plasma and the amount of
cortisol bound to the erythrocytes were determined at 20, 30, 37, and
40 C in the blood of six healthy persons. The correlation between the
free cortisol concentration and the amount of cortisol bound to
erythrocytes at all four temperatures is shown in Fig. 1
. The data points were fitted with a
linear regression line with a slope of 2.38 ± 0.06 and intercept
of -9.0 ± 4.3 pmol/mL erythrocytes (r2 = 0.995).
|
shows the fractional distribution of
cortisol for a broad range of plasma cortisol concentrations, at four
different temperatures: room temperature (20 C), hypothermia (30 C),
normal temperature (37 C), and hyperthermia (40 C). In this figure, it
can be clearly seen that the free and loosely bound fractions (albumin
and erythrocytes) show an exponential increase with increasing
temperature. In the hypothermic situation, this fraction is much lower,
compared with 37 C (5–15% vs. 25–40% respectively, in
the plasma cortisol reference range); and at 40 C, this fraction ranges
from 55–70%.
|
.
|
and 3 are valid for the situation of an Ht
of 0.4 and normal CBG concentration (0.65 µmol/L). Changing the CBG
concentration can have a major effect on the fractional distribution of
cortisol among the blood compartments. This effect is demonstrated in
Fig. 4, where five curves show the CBG
dependence of the percentage cortisol bound to the erythrocytes.
|
) and total cortisol
6–9.5%. At a lower temperature (e.g. a refrigerated
centrifuge of 4 C), the bias can be as high as 25%.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We therefore used a method that did not disturb the equilibrium distribution of cortisol between plasma and erythrocytes. Instead, the erythrocyte-associated cortisol was determined in an erythrocyte pellet by measuring the total amount of cortisol extracted from this pellet and subtracting the amount of cortisol in the entrapped plasma. Critical steps in these experiments are the measurements of the exact volumes of plasma and erythrocytes, the concentrations of cortisol in the various fractions, and the complete extraction of the erythrocyte-bound cortisol. The volume of the entrapped plasma was accurately measured with 14C-sucrose, which will not enter the cells (12). For the calculation of the total amount of cortisol in the entrapped plasma, a small correction for the protein-inaccessible space (2.1% of the erythrocyte volume) has to be made (9). This volume, which is not accessible to proteins (albumin and CBG), but only to small molecules (like unbound cortisol or sucrose) is probably formed by the aggregation of erythrocytes caused by bridging by plasma proteins, which can easily be disrupted by mechanical force (20).
The cortisol assay needed some modifications because, at the low end of the calibration curve, differences were seen in recoveries from different matrices, even after extraction of the cortisol from the sample. This problem was solved by using similar matrices for standards and samples.
Measurement of the erythrocyte-bound cortisol was reproducible, despite the many steps in the assay. The average CV, determined in the blood of normal men over a period of months, was 7%, which is about the same as the performance obtained by Hiramatsu et al. (9).
After the centrifugation of blood and the removal of about 90% of the plasma, the cortisol bound to the erythrocytes was extracted by steroid-free CBG and albumin in stripped plasma. This procedure was first described by Driessen et al. (13). We found an extraction recovery of more than 95%, which is very similar to the data reported by Driessen et al. (90–95%).
Binding of cortisol to the erythrocytes, estimated in the blood of
several volunteers, is directly related to the free cortisol
concentration and is independent from the temperature (Fig. 1). It
seems that the free cortisol concentration is the most important
determinant of the amount of the erythrocyte-associated cortisol,
because this linear relation is not observed with total plasma
cortisol, or CBG concentration. It was already shown by Hiramatsu
et al. (9) that the cortisol partitioning coefficient
(erythrocyte-associated cortisol concentration divided by plasma
unbound cortisol) was 2.62 ± 0.16 (determined at 36.5 C) and was
constant over a wide range of cortisol concentrations. In our study,
the average ratio of the erythrocyte-associated and free cortisol
(slope of the regression line in Fig. 1) was 2.38 ± 0.06,
slightly lower than the data reported by Hiramatsu et al.,
but consistent with the ratio of 2.4 (determined at 20 C), as
calculated from the original data of Driessen et al.
(21). It is tempting to speculate that, because the relation of
the erythrocyte-associated and free cortisol is linear over a wide
range of free cortisol concentrations, the cortisol in the erythrocytes
is bound by low-affinity protein(s): e.g. a protein with a
binding affinity (Ka) of 1.106 L/mol and
a concentration of 22.5 µmol/L will result in an
erythrocyte-associated/free cortisol distribution ratio of 2.4 that is
constant over a free cortisol range from 0–200 nmol/L. Other possible
combinations are: Ka = 1.105 L/mol and 43
µmol/L concentration, or Ka = 1.104 L/mol and
0.26 mmol/L concentration. A combination of high affinity but low
capacity and low-affinity/high-capacity binding proteins cannot be
excluded, as was suggested by Kornel et al. (22), but these
findings have not yet been confirmed.
After we had established the cortisol distribution ratio between
erythrocytes and plasma water, we could simulate the distribution of
cortisol among the different blood compartments in several conditions.
One of the problems, however, is that the information about binding
affinities of cortisol to CBG and albumin at the different temperatures
is scattered in the literature and, in addition, shows much variation
or even is missing. We therefore measured the cortisol binding affinity
for CBG at 20, 30, 37, and 40 C and calculated the other binding
affinities by interpolation. For albumin, a collection of cortisol
binding affinities was reported by Westphal (16). Figure 2 shows four
plots of the cortisol distribution. The plot of 37 C also shows the
results, indicated by dots, of the direct measurement of the
individual fractions. An excellent correlation with the calculated
fractions was observed. We also did measurements at the other
temperatures, but the blood was from other persons, with other Ht, CBG,
and albumin concentrations. To be able to compare the four situations,
we simulated the other three plots and used the Ht, CBG, and albumin
concentration of the 37-C situation and the appropriate binding
affinities. Both temperature and plasma cortisol concentration have a
strong effect on the percentage erythrocyte-bound cortisol, which is
shown more explicitly in Fig. 3. It is clearly demonstrated that in
situations of, for instance, hyperthermia, the free and loosely
protein- and erythrocyte-bound cortisol fraction in blood increase
exponentially. In addition, reduction of the CBG concentration, which
can be the result of an increase of the free cortisol concentration
(23), will further increase the pool of the easily available cortisol
(see Fig. 4). In situations of hypothermia, e.g. during
operations, the reverse can be expected: more protein binding of
cortisol; thus, less free and loosely-bound cortisol, which implies
that the hypothermic patients at surgery will be more cortisol
deficient than is to be expected from the total cortisol concentration.
This could negatively influence the patient’s recovery from
surgery.
It has been known, for a long time, that hyperthermia or fever, induced from either exposure to high temperatures (sauna) or injection of pyrogens, has produced approximately a 2- to 3-fold rise from baseline in the plasma cortisol concentration (24, 25, 26, 27, 28, 29, 30, 31, 32). Increased cortisol concentrations have also been reported in patients with fever, as the result of infections (33).
In addition to this 2- to 3-fold increase in total plasma cortisol concentration, which is the result of activation of the hypothalamus-pituitary-adrenal axis, the amount of cortisol that is biologically active (the free and loosely-bound cortisol) will be much higher because of the effects described in this manuscript.
The binding of cortisol to erythrocytes also has practical consequences for the preanalytical handling of the blood samples. Separation of the plasma (CBG concentration of 0.65 µmol/L) and erythrocytes, in a refrigerated centrifuge (4 C), can eventually result in approximately 25% higher free cortisol and 11% higher total plasma cortisol concentrations because of the fact that, at lower temperatures, almost no cortisol is associated with the erythrocytes, a fraction that will be separated from the plasma. The effect will be stronger at relatively high cortisol levels (stress situations, Cushing syndrome) and can be misleading in a diagnostic setting. In situations of high CBG concentrations (women using estrogen containing oral contraceptives, or pregnant women), refrigerated separation of the erythrocytes will increase the free cortisol concentration in the plasma by 15–20% compared with the situation in which the erythrocytes were separated at 37 C.
We conclude that strict temperature control during sample handling is very important, to determine the free cortisol concentration.
Received August 5, 1998.
Revised October 15, 1998.
Accepted October 19, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
