nancy than in non-pregnant women, in agreement with Booth eta! ..... Tait, J. F., and Burstein, S., In vivo studies of steroid dynamics in man. InThe Hormones: ...
CLIN. CHEM. 26/12, 1734-1737 (1980)
Improved Ultrafiltration Method for Determining Unbound Cortisol in Plasma lvanka Jerkunica, Judy Sophianopoulos,1 and Demetrios Sgoutas2 We describe a modified ultrafiltration method for measuring unbound cortisol in diluted or undiluted plasma or serum. After equilibration at 37 #{176}C with purified [3H]cortisol, plasma or serum with or without buffer was placed in the
ultrafiltration cell and two successive 0.2-mL fractions of protein-free ultrafiltrate were obtained. Under our conditions, free ligand concentration was independent of flow rate. The second fraction (the first is discarded) was used for determining the proportion of unbound cortisol. The assay is rapid (less than 2 h), practical (no more than 1.5 mL of plasma or serum is necessary), and reproducible (CV: 4.5% within assay and 5.2% in different assays).
Samples from normal men and women (blood taken at 0800 and 1600 hours), from pregnant women, and from patients with Gushing’s disease and adrenal insufficiency gave results that agreed with those obtained by equilibrium dialysis.
AdditIonal Keyphrases:
steroids . adrenal function equilibrium dialysis . binding constants for cortisol binding globulin . reference intervals . circadian differences . changes during pregnancy Measurement of unbound cortisol in plasma is becoming a frequently ordered test in many clinical laboratories in conditions involving abnormalities of cortisol-binding plasma proteins (i.e., corticosteroidbinding globulin and albumin) and where measurements of totalplasma cortisolmay lead to erroneous diagnosis(1,2). The most common methods so far described for measuring unbound cortisolin plasma involve the conjunction of two independent measurements: total plasma cortisoland of the dialysable(3-5) or ultrafiltrable (6-8) cortisol fraction as revealed by [3H]cortisol added to the testplasma. The distribution of [3H]cortisol between unbound and bound plasma cortisol has also been determined by steady-state gel filtration (9) and treatment of the sample with dextran-coated charcoal (10). Recent techniques rely on measurement of the dialyzableor ultrafiltrable cortisolby sensitive radioimmunoassay (RIA) (11, 12). We describean improved ultrafiltration method based on the use of ultrafiltration cellsand membranes similarto those we used forthe measurement of unbound thyroid hormones (13).
Materials and Methods Samples Blood samples were collected, unless otherwise specified, between 0800 and 1000 hours and centrifuged as soon as possible. Serum was stored at -20 #{176}C until assayed. Heparinized blood samples were also collected and treated simi-
larly.
Reagents [1,2,6,7-3H]Cortisol (spec. acty.80-100 Ci/mmol) from New England Nuclear Corp.,Boston, MA 02118, was examined for purity by thin-layerchromatography on silicagelG with the developing system benzene/ethyl acetate (2/1 by vol).To prepare a working solution,dilute0.25 mCi of [3H]cortisol 25-fold with absolute ethanol (14).
Total Cortisol Assay We determined (14),
cortisol directly in serum or plasma by RIA with the use of dextran-coated charcoal to separate
antibody-bound
Apparatus We use a stirredultrafiltration cellof 3-mL capacity(Model 3;Amicon Corp, Lexington, MA 02173) with a 25-mm Diaflo YM-lO membrane (Amicon), which has a molecular mass cut-offof 10 000 daltons.The membrane should be washed before itsfirstuse by soaking for at least1 h in de-ionized water. Change the water four times, allowing the membrane to soak for 15 mm each time. After the lastsoak period,blot the membrane fora few seconds, and locatethe shinierside. This sideisthe one that should be touching the plasma solution.Place the membrane in itscompartment and assemble the ultrafiltration cellas described in the manufacturer’s in-
structions. After ultrafiltration, withdraw the remaining plasma solution from the cell, and wash the cell and membrane four times while the cell is still assembled. For the last wash, add de-ionized water, stirfor 30 s,then ultrafilter for 3 mm and withdraw the wash water from the cell.This washing proce-
dure reduces to background and in the ultrafiltrate.
Received
1734
April
2, 1980; accepted
May 13, 1980.
CLINICAL CHEMISTRY,Vol. 26, No. 12. 1980
the radioactivity
inside the cell
We have on occasionused a singlemembrane formore than a dozen assays,but we recommend changing the membrane after about three or four assays to minimize the possibility of
tearing it.
Procedure The whole experiment was carried out in a room with regulated temperature (37 #{176}C). Add 0.1 mL of [3Hlcortisol (9 X 105 dpm, 4.2 pmol) ethanolic solution to a glass tube and evaporate in a stream of nitrogen.Add 1.5 mL of plasma sample and 1.0 mL of 4-(2-hydroxyethyl)-1-piperazineethanesulfonate (HEPES) buffer(80 mmol/L, pH 7.4)and mix gently and thoroughly. Incubate the resulting solution for 30 mm at 37 #{176}C. Pipet duplicate 50-sL aliquots for “total counts” intotwo scintillation vialsand pipet 2.0mL intothe ultrafiltration cell. Place the cellon a magnetic stirrer(Model PC-353; Corning Glass Works, Corning, NY 14830). Begin sufficient stirring for thorough mixing without frothing. After 30s, ultrafilter the solution under nitrogen pressure of 124 Wa (18 psig),
Department of Pathology and Laboratory Medicine, Emory University School of Medicine, and ‘Department of Chemistry, Emory University, Atlanta, GA 30322. 2 Address communications to this author.
from free cortisol.
with
continuous
stirring.
Under
these
conditions,
the flow rate isabout 60 tL/min. Collect two successive ultrafiltrates into the tubes, about 0.2 mL into the first tube and 0.2-0.3 mL into a previously weighed second tube.Discard the firstultrafiltrate. Determine the volume of the ultrafiltrate by weighing the second tube and its contents. Count the ra-
dioactivityin 0.05-mL aliquots(14). The volume range stated forthe second ultrafiltrate issufficient to contain radioactive hormone with several times background cpm for patients with very low cortisolconcentrations.Between the magnetic stirrer
and the cell, we inserted an asbestos plate (3 mm thick) and a Styrofoam plate (1 cm thick), which kept the temperature inside the cellat 37 #{176}C throughout the experiment. Temperatures were measured inside the cell with a digital thermometer. The ultrafiltration is complete in about 7 mm, 3 mm being sufficient for each tube.
Calculations The concentration
of unbound
cortisol
in undiluted plasma
was calculatedas follows: % unbound
in diluted
(cpm/mL (cpm/mL
plasma
[3H]corticol
[3H]cortisol
in second ultrafiltrate) fraction
X dilution
factor)
X100 “corrected” % unbound (% unbound
in undiluted plasma
=
“corrected”
A more realistic
correction
for dilution
total cortisol (3) was made for each
that in deriving the unbound cortisol fraction in undiluted plasma from the unbound cortisol fraction measured in diluted plasma, we assume that all the constants of the binding proteins remain unchanged.
0 1.40 1.66 2 5 10 20
7.49 7.28 7.34 6.57 5.07
(0.22) (0.42) (0.42) (O.46b)
(O.26b)
7.47 7.51 7.82 7.59 7.87
(0.34) (0.36) (0.48) (0.55) (0.32)
8.7 8.4 8.4 7.7 5.9
(0.21) (0.42) (0.42) (0.51 b)
8.7 (0.42) 8.8 (0.50) 9.1 (0.61) 8.8 (0.62) 9.2(0.41)
(0.38b)
389(010b)
8.08(0.58)
4.6(#{216}.l#{216}b)
2.73
8.74
3.2
(0.08b)
Methods. b Significantly (p 0.05) in unbound cortisol between diluted and undiluted serum if the dilution did not exceed 1.67-fold. Using the
computer program, on the other hand, we could not prove significant (p > 0.05) difference in results between undiluted and diluted up to 10-fold, but there was a significant (p 0.05) higher than the mean unbound cortisol of non-pregnant women, in agreement with previous reports (8). However, the percentage of unbound cortisol in plasma was lower (p
