Journal of Clinical InvestigationVol. 43, No. 6, 1964

Increased Unbound Cortisol in the Plasma of Estrogen-treatedSubjects *

JOHNE. PLAGER, KURTG. SCHMIDT, ANDWILLIAM J. STAUBITZ

(From the Medical Foundation of Buffalo, the Department of Medicine, The State Universityof New York at Buffalo, and the Division of Urology, Edward J. Meyer

Memorial Hospital, Buffalo, N. Y.)

The plasma cortisol 1 level is the sum of the cor-tisol bound to the plasma proteins and that whichis unbound. During pregnancy or estrogen treat-ment the plasma cortisol and 17-hydroxycorticoidlevel rises several-fold (1-8). This rise is sec-ondary to an increase in the plasma content oftranscortin (6, 9), which is a normal a-globulincomponent of plasma (10) with high affinity forbinding cortisol. Since the subject maintainedon long term estrogen therapy does not developthe clinical stigmata of hyperadrenocorticism, ithas generally been considered that the boundplasma cortisol is physiologically inactive (7, 11)and that the unbound plasma cortisol remains atnormal concentration (12) in these subjects.

In the present investigation the plasma levels ofbound and unbound cortisol were determined in agroup of estrogen-treated subjects at differentadrenal cortisol production rates and comparedto those of a control group. At the beginning ofthis work, we postulated that if the unboundplasma cortisol level is the only factor which de-termines the extent of the plasma cortisol effecton cellular functions, then the unbound cortisollevel should be the same in both groups of subjectswhen studied at rest. In addition, if only the un-bound cortisol level controls pituitary ACTH in-hibition via the negative feedback mechanism, thenthis level should be the same in both groups atequivalent adrenal cortisol production rates. Ourresults indicate that the unbound cortisol levelcannot be the only factor which determines periph-eral physiological activity. Estrogen-treated sub-jects were found to have relatively elevated levelsof plasma unbound cortisol.

* Submitted for publication November 27, 1963; ac-cepted January 23, 1964.

These studies were supported by U. S. Public HealthService grants AM-06085 and RG-9914.

1 Cortisol: 1lp,17a,21-trihydroxy-4-pregnene-3,20-dione.

Methods

Radioactive cortisol. Cortisol-1,2-H' (2.08 ,uc per ug)2was purified on two different paper chromatographicsystems. This labeled cortisol did not show significantchange in specific activity when added to carrier cortisoland recrystallized from three different solvent systems.

Ultrafiltrationi procedure. Details of this procedurehave been published (13). Ten ml of heparinized plasmawas adjusted to pH 7.4, and 0.04 ,tg of tritiated cortisolwas then added. The plasma was incubated at 37.5° Cfor 0.5 hour while being slowly shaken and was thentransferred to a cellophane casing and placed in a Tori-bara ultrafiltration apparatus (14). The sample was cen-trifuged initially for 15 minutes at 37.5° C. The 0.2-mlultrafiltrate formed during this period was removed anddiscarded and the sample further centrifuged for 45 min-utes. During this period 0.55 to 0.65 ml of ultrafiltrateformed. One-half ml of the ultrafiltrate was trans-ferred to a liquid scintillator (15) and counted in a liquidscintillation spectrometer. The number of counts in the0.5 ml of ultrafiltrate was considered representative ofthe proportion of unbound cortisol in the sample. Theunbound cortisol in micrograms per 100 ml plasma wascalculated from the total plasma cortisol value and thepercentage of this cortisol which was found to be ultra-filterable by the above technique.

In preliminary experiments the validity of this tech-nique, which was described by other investigators (12,16), was confirmed. The cellophane membrane used inthese experiments was not found to be a selective bar-rier to the free passage of either water or cortisol. Theproportion of cortisol found to be ultrafilterable was notaffected by the volume of plasma studied so long as theplasma was not allowed to become too concentrated dur-ing the ultrafiltration procedure. In addition, the half-hour preliminary incubation afforded equilibrium betweenthe added tracer cortisol and the plasma cortisol; nosignificant alteration of the proportion ultrafilterablecould be detected in plasma from normal or estrogen-treated subjects by prolonging the incubation time to 5hours.

Clinical subjects. Thirty-four male subjects, hospital-ized on a urological service, were studied. Their meanage was 52 years with a range of 16 to 81 years. As

2 Purchased from New England Nuclear Corp., Boston,Mass.

1066

UNBOUNDCORTISOL IN PLASMA

a group, they were individuals in good general healthwho were well into their convalescence from an acuteurological problem. Six subj ects, ages 67 to 78, hadprostatic carcinoma treated by orchiectomy and had beenon stilbestrol (diethylstilbestrol) for at least 1 year atthe time they were studied. None of these subjects hadevidence of metastatic disease. The remaining estrogen-treated subjects were given 5 mg of stilbestrol daily for1 to 3 weeks for the purposes of this study.

Experimental procedure. Intravenous infusions of 5%glucose in water were started at 9:00 a.m. If adreno-cortical stimulation was desired, from 0.3 to 30 U ofACTH3 was infused during the 7-hour experimentalperiod. The amount of ACTHwas varied from patientto patient in an attempt to provide a range of adrenalcortisol production rates. At 1:00 p.m. a 45-ml bloodsample was drawn into a heparinized syringe. Ten mlof a standard solution of tritiated cortisol in 5% ethanol-saline was then administered intravenously within a1-minute period. A total of 5.5 jug of cortisol was givenwhich contained, at 16% counting efficiency, 4 millioncpm. One-half hour after the cortisol was given a 20-ml heparinized blood sample was taken. Five additionalblood samples were taken at one-half hour intervals inthe estrogen-treated subj ects and at 20-minute intervalsin the other subjects. The volumes of these sampleswere also 20 ml except for the final one, which was 40ml. Blood samples were immediately centrifuged, andthe plasma was separated and frozen until analyzed.

Ten ml of plasma from the first sample was analyzedby the ultrafiltration procedure. The last plasma wasdivided into two equal aliquots, and the eight plasmasamples were carried through the extraction procedureand paper chromatographic separation of cortisol previ-ously described (17). The cortisol area from the paperchromatogram of the first and of the last samples waseluted and analyzed with the Porter-Silber reagent (18).The cortisol areas from the six other paper chromato-grams were eluted, transferred to vials, and countedwith an error of less than 5%go in a Packard Tri-Carbliquid scintillation spectrometer.

Calculation of the adrenal cortisol production rate wascarried out by using established isotopic techniques (19-22). In those subjects in whom the cortisol level at thebeginning and the end of the experiment fell within5%So of their mean level, the plasma cortisol was con-sidered to be constant during the experimental period, andin these cases the adrenal cortisol production rate andthe peripheral removal rate were equal. Cortisol radio-activity in counts per minute per milliliter of plasma wasplotted on semilogarithmic co-ordinates against time and,with the occasional exception of the initial point, rep-resented a single exponential curve. From the ordinateintercept, slope of this curve, knowledge of the dose ofadministered radioactivity, and the plasma cortisol level,

3 Supplied as corticotropin solution by the WilsonLaboratories, Chicago, Ill.

4Packard Instrument Co., La Grange, Ill.

the miscible pool, apparent distribution volume, andadrenal production rates were calculated.

In many of the patients, the plasma cortisol level didnot remain constant during the experimental period. Thiswas particularly true in those subjects who did not re-ceive ACTH and whose level fell with the diurnalrhythm. Although these subjects were not, strictlyspeaking, in a steady state, their cortisol production anddisappearance rates may be calculated independently ifboth of these rates are constant, although not equal,throughout the experimental period. To estimate theplasma cortisol level at intermediate experimental times,the initial and final plasma cortisol levels were notedon a semilogarithmic plot, and the cortisol level at in-termediate times was obtained by interpolation. On threeoccasions, additional plasma samples were taken at in-termediate times and analyzed for cortisol, and in theseinstances their analyzed levels fell within experimentalerror of the anticipated interpolated values. From theplasma cortisol levels and the plasma radioactivity levelsthe specific activity (counts per minute per microgramcortisol) could be calculated and plotted on semiloga-rithmic co-ordinates against time. Analysis of the spe-cific activity curve provided the miscible pool size, distri-bution volume, and production rate values; the disap-pearance rate was calculated from the fractional turn-over rate obtained from the plasma radioactivity (countsper minute per milliliter) curve. In all the data to bepresented, it appeared justified to assume that the rateof cortisol production and the rate of disappearancewere constant throughout the experimental period, sinceall the data used showed a single exponential fall.

Results

The data for the 21 control subjects and 13stilbestrol-treated subjects are tabulated in TablesI and II, respectively. The mean experimentalvalues have been calculated for those control andstilbestrol-treated subjects who were not givenACTH.

In Figure 1 the adrenal cortisol production ratesare plotted against the total plasma cortisol con-centration for the control subjects. The plasmacortisol level in this group of subjects provides afair index of their adrenal cortisol productionrate. The values for the stilbestrol-treated sub-jects are not plotted here, but they showed plasmacortisol levels that were markedly elevated relativeto their cortisol production rates when comparedto the control group.

In Figure 2 the adrenal cortisol production ratesare plotted against the unbound plasma cortisolconcentration for the two groups of subjects.The unbound cortisol concentration also appears

1067

J. E. PLAGER, K. G. SCHMIDT, AND W. J. STAUBITZ

TABLE I

Isotopic cortisol studies in control subjects*

InitialInitial distri- Disap- Disap- Produc-

Initial Unbound Final miscible bution pearance pearance tionNo. cortisolt cortisol cortisol pool volume tq rate rate

JAg/100 ml ;100mlmg100m / ml ye L min Jg/min pg/min1 9.6 0.85 5.7 2,340 24.4 68 23.8 112 8.9 0.98 5.3 1,900 21.4 70 19 123 10.5 1.68 7.9 2,760 26.2 84 23 144 12.3 2.1 8.3 2,520 20.4 85 20.5 155 8.6 0.94 2,580 30.0 100 186 8.0 0.96 1,730 20.2 57 21Mean 9.6 1.25 2,305 23.8 15

7 14 2.1 9.3 4,980 36 85 40.6 218 18 2.9 3,830 21.3 90 309 21 3.8 27 3,810 18.3 80 33 40

10 16 2.3 2,940 17.7 50 4011 30 6.6 10,900 35 86 8812 19 3.2 7,450 39 58 8913 44 8.8 11,900 37 93 8914 40 9.9 11,400 29 72 10915 36 7.0 10,450 29 65 11116 36 9.0 10,000 27 56 12317 43 11.6 18,000 42 92 13518 35 7.7 16,900 48 82 14219 42 10.5 20,540 49 71 20020 45 13.5 34,300 76 98 24021 50 16.7 64 35,400 71 77 318 378

* Subjects 7 to 21 were studied during iv infusion of supplemental ACTH.t In those cases in which the initial and final plasma cortisol values fell within 5%of their mean value, the plasma

cortisol concentration is considered to have stayed constant throughout the experimental period, and this mean value isgiven in this column.

to provide a fair index of adrenal cortisol produc-tion but does not appear to be superior to theplasma cortisol level in this respect. In the stil-bestrol-treated subjects the relationship betweenthe unbound cortisol level and the adrenal cortisolproduction rate was not the same as that seen inthe control subjects. Equivalent adrenal cortisolproduction rates produced much higher plasmalevels of unbound cortisol in the stilbestrol-treatedsubjects.

In this same respect, in Tables I and II themean values may be compared for those controland stilbestrol-treated subjects who were not givenACTH. Both groups were producing equivalentquantities of cortisol, i.e., 15 and 15.2 jug per min-ute during the study period. Despite this, the un-bound cortisol level in the stilbestrol-treated sub-jects was several times higher than that seen inthe control subjects. These stilbestrol-treatedsubjects, while apparently maintaining a eucorti-coid status, had appreciable elevation of theirplasma unbound cortisol levels.

The initial miscible pool in the two groups ofnon-ACTH-treated subjects was also different.

The mean pool size in the control subjects was 2.3mg, whereas that in the stilbestrol-treated sub-jects was 3.8 mg, a difference of 1.5 mg. Thedifference in pool size apparently cannot be ac-counted for solely by the larger amount of corti-sol present in the plasma of the stilbestrol-treatedsubjects. If one estimates a mean plasma volumeof 3 L, the plasma cortisol content of the controlgroup will be 3 L x 96 pg per L or about 0.3 mg,and the extraplasma miscible pool will be 2.3 mgminus 0.3 mg or 2.0 mg. By similar calculation,the extraplasma miscible pool for the stilbestrol-treated subjects will be 2.9 mg. Although thedifference between the two extraplasma misciblepools is smaller than the difference between thetotal pools, it is still probably significant.

In the stilbestrol-treated subjects studied with-out supplemental ACTH infusion, it was unclearwhether the elevation of plasma unbound cortisolmight represent some degree of excitement sec-ondary to the experimental procedure. For thisreason, a single blood sample was taken from tensubjects who had been on long-term stilbestroltherapy and from five control subjects. These

1068

UNBOUNDCORTISOL IN PLASMA 1069

TABLE II

Isotopic cortisol studies in stilbesterol-treated subjects*

InitialInitial distri- Disap- Disap- Produc-

Initial Unbound Final miscible bution pearance pearance tionNo. cortisolt cortisol cortisol pool volume tj rate rate

pg/100 ml Jsg/lOO ml esg/lOO ml Jsg L min jsg/min 4g/min1 22 4.2 2,600 11.6 155 11.62 32 5.7 4,700 14.6 235 123 35 2.5 1,690 4.8 90 134 37 5.2 3,300 8.9 160 145 29 3.8 21 3,350 11.5 88 26 156 27 3.6 20 4,350 15.8 110 27 207 37 5.6 24 6,900 18.4 125 38 21Mean 31 4.4 3,841 12.2 15.2

8 47 8.5 10,600 22.5 170 439 83 14.2 13,900 16.8 135 71

10 71 16.3 8,950 12.6 76 8111 75 17 16,200 21.6 120 9412 74 18.3 20,400 27.5 110 12813 127 34.4 55,500 44.0 135 285

* Subjects 8 to 13 were studied during iv infusion of supplemental ACTH.t In those cases in which the initial and final plasma cortisol values fell within 5%of their mean value, the plasma

cortisol concentration is considered to have stayed constant throughout the experimental period, and this mean value isgiven in this column.

samples were taken at a local county home between2 and 3 p.m. The subjects were ambulatory, andthe procedure appeared to be accepted with equa-nimity. The data from these fifteen subjects arepresented in Table III and are consistent with val-ues presented in Tables I and II. The mean un-bound cortisol value for the control subjects was1.1 jug per 100 ml and for the stilbestrol-treatedsubjects 4.0 pg per 100 ml.

The previous data show that neither the levelof bound nor of unbound cortisol similarly re-

.d 2500'k 200

0

1150`0

00L 100to% 5000

0

* 00

0 0 0

* S

10 20 30Total Cortisol

40 50,ug/0oo ml.

FIG. 1. TOTAL CORTISOL IN THE PERIPHERAL PLASMAOF CONTROLSUBJECTS AS A FUNCTION OF THE ADRENAL

CORTISOL PRODUCTIONRATE.

flects the adrenal cortisol production rate in boththe estrogen-treated and control subjects. In-spection of the percentages of unbound cortisolin Table III suggested that 'hese values were simi-lar in the two groups of subjects. The data fromTables I and II were therefore plotted as the per-centage of cortisc.l unbound (ultrafilterable)against the adrenal cortisol production rate (Fig-ure 3). This plot, and the additional data in

c 250

x 200C0

._

0 150la

0a- lo.0t 5000

*0 0

0 0

0

e

0 0

0

e ee eee

2 4 6 8 10 12 14 16 18Unbound Cortisol ,ug/ 00ml.

FIG. 2. UNBOUND CORTISOL LEVEL IN PERIPHERALPLASMA AS A FUNCTION OF THE ADRENALCORTISOL PRO-DUCTION RATE. The control subjects are designated byclosed circles and the stilbesterol-treated subjects by theletter e.

J. E. PLAGER, K. G. SCHMIDT, AND W. J. STAUBITZ

TABLE III

The total cortisol and unbound cortisol in the plasma of agroup of estrogen-treated and control subjects

PlasmaNo. cortisol Unbound cortisol

Control )Ag/IO0 ml % lg/100 ml1 4.4 9.8 0.432 12.0 10.3 1.243 8 13.9 1.114 7.9 14.4 1.145 1 7.0 9.3 1.6

Estrogen-treated1 30 13.2 3.82 34 9.8 3.43 44 7.6 3.44 39 14.5 5.75 30 16.7 5.06 54 11 5.97 59 7.6 4.58 30 8.6 2.69 27 8.8 2.4

10 31 9.7 3.0

Table III, suggest that this proportion is similarlyrelated to adrenal cortisol production in both theestrogen-treated and control subjects. Otherworkers have reported that both the percentage ofultrafilterable cortisol in the plasma and the 24-hour urinary corticoid excretion tend to decreaseduring estrogen administration (12). This rela-tionship would be consistent with the present ob-servation that the percentage of cortisol that isultrafilterable appears to decrease with a decreas-ing adrenal cortisol production rate.

e

0

0 .,*:1 -

e*ee*e w

e * ?eee Wee10 20 30

Unbound Cortisolx 100Total Cortisol

Discussion

Cortisol circulates in human plasma in at leastthree forms: 1) bound to albumin, 2) bound totranscortin, and 3) in unbound form. Duringpregnancy or estrogen administration the totalplasma cortisol rises (5, 7, 8) owing primarily toa marked increase in transcortin-bound cortisol(8). Since the pregnant woman or estrogen-treated man does not develop the clinical stigmataassociated with increased adrenal production ofglucocorticoids (7), it has been suggested that thetranscortin-bound cortisol is physiologically in-active (7, 11) and that the plasma unbound corti-sol is the active plasma cortisol form and is innormal concentration in the estrogen-treated sub-ject (12). If the plasma unbound cortisol is theonly moiety controlling cortisol activity in thetissues, for homeostasis to be maintained thismoiety should also control pituitary ACTH inhi-bition and the hepatic reduction and removal ofcortisol. One would therefore anticipate thatwhile the estrogen-treated subject is secreting theequivalent quantity of cortisol as the normal per-son, the unbound cortisol levels in their plasmaswill be similar.

It is clear from the presented data that at equiva-lent adrenal cortisol production rates the plasmaunbound cortisol level in the estrogen-treated sub-ject is appreciably higher than that of the controlsubject. Since the estrogen-treated subject doesnot develop stigmata of Cushing's syndrome, andsince the urinary corticoid excretion of the estro-gen-treated subject stays within normal range ordecreases (4, 7), the plasma unbound cortisol levelper se is probably not necessarily directly propor-tional to its evident peripheral physiological ac-tivity or to the rate of cortisol removal by theliver. In addition, since the stilbestrol-treatedsubjects had a plasma unbound cortisol level threeto four times higher than the control group dur-ing a time in which they were secreting equivalentquantities of adrenal cortisol, the plasma un-bound cortisol level per se was probably not con-trolling pituitary ACTH release in both groups.The only parameter which appeared to be similarlyrelated to the adrenal cortisol production rate inboth estrogen-treated and control subjects was thepercentage of the total cortisol represented by theunbound fraction. This proportion of unbound

400

300

200

1 00

0

0

._-

U-0

-a

I-0

FIG. 3. PERCENTAGEOF UNBOUNDPLASMACORTISOL ASA FUNCTION OF THE ADRENALCORTISOL PRODUCTIONRATE.The control subj ects are designated by closed circles andthe stilbestrol-treated subjects by the letter e.

1070

UNBOUNDCORTISOL IN PLASMA

to total cortisol in the plasma appears to most Iclosely approximate a physiologically significantindex. Why this should be so is not completely Iclear at the present time. At low adrenal corti-sol production rates the plasma transcortin in thecontrol and estrogen-treated subjects is prob-ably only partially saturated. Since the associa-tion constant for transcortin will be the same inboth groups of subjects, their percentages of un-bound cortisol will also be approximately thesame. This would probably account for the flatpart of the curve that is evident in Figure 3 atlow cortisol production rates. At elevated adrenalcortisol production rates, transcortin is saturated,and there is a steep rise in the percentage of un-bound cortisol as seen in Figure 3. In both theestrogen-treated and control subjects, this riseshould represent a predominant binding equi-librium with plasma proteins of weaker bindingaffinity. The cortisol capacity and affinity of theweaker binding plasma proteins probably do notdiffer in the two groups of subjects.

The mean cortisol miscible pool was 2.3 mg and3.8 mg in the control subjects and in the estrogen-treated subjects, respectively. As calculatedabove, the extraplasma miscible pool was approxi-mately 2.0 mg and 2.9 mg in these two groups,respectively. Appreciable quantities of transcor-tin-like binding have been reported present inthoracic duct lymph, and its binding capacity hasbeen stated to be slightly below that of the periph-eral plasma (23). Although this observation doesnot necessarily mean that transcortin is presentin appreciable concentration in the extracellularfluid, it does suggest the presence of transcortinoutside the vascular compartment and that itsconcentration in estrogen-treated subjects may behigher than in normal subjects. In addition, re-cent experiments in which I125-labeled transcortinwas administered intravenously to normal and es-trogen-treated subjects indicated that 50%o of thetranscortin miscible pool was outside of the plasma(24). The difference in total miscible pool be-tween the estrogen-treated and control subjectswas 1.5 mg, but this difference became 0.9 mgwhen the plasma cortisol content was subtractedfrom each group. This difference in cortisol poolmay become negligible at the tissue level if thereis another area of increased transcortin binding,

besides the plasma, in the estrogen-treated sub-jects. In other words, despite elevated levels ofbound and unbound cortisol in the plasma of theestrogen-treated subject, the total amount of cor-tisol in the tissues may be equivalent to that foundin the control subject. The form in which cor-tisol exerts its physiological activity in the tis-sues is not known.

Failure to observe evidence of hyperadreno-corticism in estrogen-treated subjects may be re-lated to a direct effect of estrogens on cellular en-zyme activity, on cellular and vascular perme-ability for substances besides transcortin and corti-sol, on the rate of cortisol catabolism, or on othermetabolic parameters.

Summary

Data have been presented which test the postu-late that the unbound level of plasma cortisol, ir-respective of what the total level may be, is thedeciding factor in determining whether the hu-man subject will show clinical evidence of normalor abnormal tissue cortisol concentration. Thispostulate was studied in control and estrogen-treated human subjects and was not found to beconsistent with the following observations: 1)Both the bound and unbound plasma cortisol lev-els were found to be appreciably elevated in agroup of clinically eucorticoid, estrogen-treatedsubjects, when compared to the control group.2) Both the plasma bound and unbound cortisollevels are a fair index of the adrenal cortisol pro-duction rate in the nonestrogen-treated subject,but these two parameters are relatively elevatedin the estrogen-treated subject at equivalent ratesof adrenal cortisol production. 3) The data indi-cate that the proportion of unbound to total plasmacortisol bears a closer relationship to adrenal cor-tisol production in the combined group of estro-gen-and nonestrogen-treated subjects than doeseither of these parameters independently.

Acknowledgments

The authors gratefully acknowledge the assistance ofDrs. A. A. Sandberg and W. R. Slaunwhite, Jr., in theircritical review of the manuscript.

Miss Margaret Reszel provided excellent technicalassistance.

1071

J. E. PLAGER, K. G. SCHMIDT, AND W. J. STAUBITZ

References1. Gemzell, C. A. Blood levels of 17-hydroxycorti-

costeroids in normal pregnancy. J. clin. Endocr.1953, 13, 898.

2. Taliaferro, I., F. Cobey, and L. Leone. Effect ofdiethylstilbestrol on plasma 17-hydroxycorticos-teroid levels in humans. Proc. Soc. exp. Biol.(N. Y.) 1956, 92, 742.

3. Wallace, E. Z., H. I. Silverberg, and A. C. Carter.Effect of ethinyl estradiol on plasma 17-hydroxy-corticosteroids, ACTH responsiveness and hydro-cortisone clearance in man. Proc. Soc. exp. Biol.(N. Y.) 1957, 95, 805.

4. Robertson, M. E., N. Stiefel, and J. Laidlaw. Theinfluence of estrogen on the secretion, dispositionand biologic activity of cortisol. J. clin. Endocr.1959, 19, 1381.

5. Migeon, C., J. Bertrand, and P. E. Wall. Physio-logical disposition of 4-C4-cortisol during latepregnancy. J. clin. Invest. 1957, 36, 1350.

6. Slaunwhite, W. R., Jr., and A. A. Sandberg. Trans-cortin: a corticosteroid-binding protein of plasma.J. clin. Invest. 1959, 38, 384.

7. Peterson, R. E., G. Nokes, P. S. Chen, Jr., and R. L.Black. Estrogens and adrenocortical function inman. J. clin. Endocr. 1960, 20, 495.

8. Sandberg, A. A., and W. R. Slaunwhite, Jr. Trans-cortin: a corticosteroid-binding protein of plasma.II. Levels in various conditions and the effects ofestrogens. J. clin. Invest. 1959, 38, 1290.

9. Doe, R. P., and U. S. Seal. Protein binding of cor-tisol in man (abstract). J. clin. Invest. 1963, 42,929.

10. Daughaday, W. H. Binding of corticosteroids byplasma proteins. IV. The electrophoretic demon-stration of corticosteroid binding globulin. J. clin.Invest. 1958, 37, 519.

11. Sandberg, A. A., and W. R. Slaunwhite, Jr. Trans-cortin: a corticosteroid-binding protein of plasma.V. In vitro inhibition of cortisol metabolism. J.clin. Invest. 1963, 42, 51.

12. Mills, I. H., H. P. Schedl, P. S. Chen, Jr., and F. C.Bartter. The effect of estrogen administration on

the metabolism and protein binding of hydrocorti-sone. J. clin. Endocr. 1960, 20, 515.

13. Plager, J. E., R. Knopp, W. R. Slaunwhite, Jr., andA. A. Sandberg. Cortisol binding by dog plasma.Endocrinology 1963, 73, 353.

14. Toribara, T. Y., R. Terepka, and P. A. Dewey.The ultrafilterable calcium of human serum. I.Ultrafiltration methods and normal values. J.clin. Invest. 1957, 36, 738.

15. Bruno, G. A., and J. E. Christian. Determination ofcarbon-14 in aqueous bicarbonate solutions byliquid scintillation counting techniques. Applica-tion to biological fluids. Analyt. Chem. 1961, 33,1216.

16. Chen, P. S., Jr., I. H. iIills, and F. C. Bartter.Ultrafiltration studies of steroid-protein binding.J. Endocr. 1961, 23, 129.

17. Plager, J. E., P. Cushman, Jr., and A. E. Chase.The plasma cortisol response to ACTH in "idio-pathic hirsutism." J. clin. Invest. 1961, 40, 1315.

18. Porter, C. C., and R. H. Silber. A quantitative colorreaction for cortisone and related 17,21-dihydroxy-20-ketosteroids. J. biol. Chem. 1950, 185, 201.

19. Peterson, R. E., and J. B. Wyngaarden. The mis-cible pool and turnover rate of hydrocortisone inman. J. clin. Invest. 1956, 35, 552.

20. Peterson, R. E. The miscible pool and turnover rateof adrenocortical steroids in man in Recent Prog-ress in Hormone Research, G. Pincus, Ed. NewYork, Academic Press, 1959, vol. 15, p. 231.

21. Robertson, J. S. Theory and use of tracers in de-termining transfer rates in biological systems.Physiol. Rev. 1957, 37, 133.

22. Zilversmit, D. B. The design and analysis of iso-tope experiments. Amer. J. Med. 1960, 29, 832.

23. Sandberg, A. A., W. R. Slaunwhite, Jr., and A. C.Carter. Transcortin: a corticosteroid-bindingprotein of plasma. III. The effects of varioussteroids. J. clin. Invest. 1960, 39, 1914.

24. Sandberg, A. A., M. Woodruff, H. Rosenthal, N.Nienhouse, and W. R. Slaunwhite, Jr. Transcor-tin: a corticosteroid-binding protein of plasma.VII. Half-life in normal and estrogen-treatedsubjects. J. clin. Invest. 1964, 43, 461.

1072