The Journal of Clinical InvestigationVol. 46, No. 9, 1967

Thyroxine Transport in Thyrotoxicosis and Hypothyroidism *MITSUO INADA t ANDKENNETHSTERLING §

(From the Medical Service, Bronx Veterans Administration Hospital and the Department ofPathology, Columbia University College of Physicians and Surgeons, New York)

Summary. (a) The thyroxine-binding proteins were investigated in 23 casesof untreated thyrotoxicosis and 16 cases of untreated hypothyroidism, em-ploying reverse flow paper electrophoresis with the glycine acetate system atpH 8.6 in the Durrum type cell.

(b) In active thyrotoxicosis, all sera exhibited diminished thyroxine-binding prealbumin (TBPA) capacities. However, 17 of the 23 sera alsohad diminished thryroxine-binding alpha globulin (TBG) capacities, as wellas markedly elevated free thyroxine fractions. In contrast, six thyrotoxic serahad normal TBG capacities and normal or slightly elevated free thyroxinefractions.

(c) In hypothyroidism, the TBPA capacities showed no consistent devia-tion from the normal range. 11 of the 16 sera had elevated TBG capacitiesas well as markedly diminished free thyroxine fractions. In contrast, fivehypothyroid sera had normal TBG capacities and normal or nearly normalfree thyroxine fractions.

(d) In thyrotoxicosis and hypothyroidism, the inverse correlation betweenfree thyroxine fraction and TBG was much closer than that with TBPA.When diagnostic categories were considered separately, only TBG bore asignificant inverse relation to the free thyroxine fraction. It is thereforesuggested that in thyroid diseases TBG may sometimes play a more im-portant role than TBPAin determining the free thyroxine fraction.

(e) The demonstrated variations in the binding proteins were consideredsufficient to explain the abnormalities of the free thyroxine fractions in thy-roid disease.

Introduction

Previous studies (1-4) have suggested dimin-ished thyroxine-binding prealbumin (TBPA) insera from patients with thyrotoxicosis without sig-nificant alteration in thyroxine-binding alpha glo-bulin (TBG); slight but significant elevation ofTBG was reported in hypothyroidism (2-4).

* Submitted for publication 9 February 1967; accepted25 May 1967.

Supported in part by grant AM-10739-02 from the U. S.Public Health Service.

: Visiting investigator from the Department of In-ternal Medicine, Kyoto University School of Medicine,Kyoto, Japan.

§ Address requests for reprints to Dr. Kenneth Ster-ling, Veterans Administration Hospital, 130 West Kings-bridge Road, Bronx, N. Y. 10468.

In more recent works (4-7) the free thyroxinefraction in serum has been measured; it appearsto be determined in large part by the thyroxine-binding proteins. Thus, the elevated free thy-roxine fractions in sera from patients with thyro-toxicosis have been ascribed to their diminishedTBPA capacities, as well as to the elevated thy-roxine concentrations (4, 5). However, pub-lished findings do not appear to have excludeddefinitely the possibility that free thyroxine frac-tions might also be related to TBG capacities.

In the present report, we have shown that ap-proximately three-fourths of 23 sera from thyro-toxic patients had diminished TBGcapacities and,moreover, the free thyroxine fractions were corre-lated inversely with TBG capacities in thyrotoxi-

1442

THYROXINETRANSPORTIN THYROTOXICOSISAND HYPOTHYROIDISM

cosis and in hypothyroidism. These findings sug-gested that in thyroid disease, TBG may play amore important role than TBPA in determiningthe free thyroxine fraction.

MethodsClinical material

Thyrotoxicosis. Sera from 23 untreated hyperthyroidpatients were obtained from the Veterans Administra-tion Hospital and its Radioisotope Service, the ThyroidClinic, and the Radiotherapy Department of the Presby-terian Hospital. The diagnoses were verified by thy-roidal uptake of n"I, as well as thyroidal clearance (8),in the patients from the Veterans Administration Hos-pital Radioisotope Service. Protein-bound iodine (PBI)1or thyroxine iodine by column2 was also available, aswell as T3 resin uptake and other diagnostic tests. Allwere considered cases of toxic diffuse goiter (Graves'disease).

Hypothyroidism. Sera from 16 untreated hypothyroid-subj ects were obtained from the same sources as thepatients with thyrotoxicosis. The diagnoses were veri-fied by thyroidal uptake, PBI,l or thyroxine iodine bycolumn,2 and usually by other tests as well.

Cases C.K., R.C., I.H., J. E., and E.C. were hypothy-roid after 1"I therapy, and W.P. after total thyroidectomyfor thyroid carcinoma. Case E.W. became hypothyroid

1J. R. Leonards Medical Laboratory, Cleveland, Ohio,or Bio-Science Laboratories, Van Nuys, Calif.

2 Bio-Science Laboratories, Van Nuys, Calif.

M-MARKER - - - K ___-- -

+_1 A-:1- l

ANODE*| -- j --

<-| ----6 gTBPA %-oALBfTBG

....................a .

... .............~~~~~..................

after subacute thyroiditis. M.C. had unexplained goitroushypothyroidism. A.T., M.S., J.M., E.G., M.O., J.V., F.R.,and O.L. had spontaneous idiopathic myxedema or hy-pothyroidism.

All patients had moderate or severe clinical pictures,and patients with mild thyrotoxicosis or hypothyroidismwere not included in this study.

Normal group. Sera were obtained from 19 healthyvolunteers (12 men, 7 women) with ages ranging from22-45 yr.

Determination of free thyroxine in serum

The free thyroxine fraction in serum was determinedin duplicate or triplicate by the magnesium precipitationmethod of Sterling and Brenner (6). The free thyroxineiodine concentration in mzg/100 ml represents the prod-uct of free thyroxine percentage and PBI value or thy-roxine iodine by column.

'I-labeled L-thyroxine 3 in 50% propylene glycol solu-tion was tested for purity by descending paper chromatog-raphy with the n-butanof-dioxane-ammonia system. Theradiothyroxine was free of appreciable radioactive con-taminants other than 1-5%o iodide. The precise per centof radioactivity due to iodide-'~I was determined bypaper electrophoresis as previously described (6).

Paper electrophoretic study

Reverse flow electrophoresis (9) with a glycine acetatesystem, pH 8.6 (10), was performed to determine the

3 Abbott Laboratories, North Chicago, Ill. Cysteine(0.2%o) was added by the supplier for stabilization.

4 s I ____3T1MARKER --

- ---: -- -I

i I..T'It'_enI.

Iv--- -I .. 1

-

"--,--1 -, ._-- -, _-. :- _- -' -7-

.-- -..I-.--.-.. ,---i l-....----------..- f

FIG. 1. REVERSEFLOW PAPER ELECTROPHORESIS, GLYCINE ACETATE SYSTEM, PH 8.6. Radioactive scan aligned with

-stained paper strip below. The above normal serum had been enriched with a "tracer" amount of thyroxine-'I, 3,ug/100 ml. The scan is typical of sera with "tracer" as opposed to "loading" or saturating concentrations of the hor-mone.

I

-1

11.all-

1443

MITSUO INADA AND KENNETHSTERLING

maximal binding capacities of thyroxine-binding alphaglobulin (TBG) and thyroxine-binding prealbumin (TB-PA), using the Spinco-Durrum apparatus. 580 ml gly-cine acetate solution was poured into the anodal celland 500 ml solution into the cathodal cell. Fresh solutionwas always employed. Serum samples of 8 Al were applied3.5 inches from the anodal end of each Whatman filterpaper strip (grade 1, SP Pattern). Samples were ap-plied on the paper strips in the cold room at 4VC tominimize the evaporation of the solution from the paperduring application. Electrophoresis was performed for19 hr at a constant current reading of 12 mA (0.5 mA/cmof paper strip) at room temperature. After completion

of electrophoresis, the paper strips were dried at 120'Cfor 30 min and then scanned with the Nuclear-ChicagoActigraph II (model C-1OB) equipped with an auto-matic recorder, before they were stained with brom-phenol blue. From both the tracings and the stainedpaper strips, the position of each thyroxine-binding pro-tein was determined, and the maximal binding capacitiesof TBG and TBPA were calculated by planimetry ofthe tracing. In subsequent work planimetry was obviatedby automatic integration (Disc Integrator in Nuclear-Chicago Actigraph III System, model 1004) which yieldedthe same values.

TABLE I

Findings in sera from 23 cases of untreated thyrotoxicosis

Free thyroxine Free thyroxineSubjects Age and sex PBI fraction iodine TBG TBPA

j&g/100 ml % mUg/100 ml jig/100 ml 1lg/1OO mlGrouLp I (diminished TBG)

H.P. 50FE.H. 63FS.B. 45FW.H. 46FB.P. 25FB.S. 42FA.R. 36MH.F. 38FJ.P. 55MI.A. 62FB.L. 40FH.A. 45ML.T. 45FR.R. soMM.L. 30FG.B. 35FL.P. 35M

10.913.1*10.3

9.18.1i9.5

12.6*15.8*13.911.714.612.212.0

8.910.011.6

8.1

0.0910.1110.1460.0880.0760.1250.0730.0940.0650.1120.1130.0720.0920.1040.0660.1840.091

9.9214.5415.04

8.016.16

11.889.20

14.859.04

13.1016.50

8.7811.04

9.266.60

21.347.37

12.717.5

8.914.815.0

9.411.516.015.617.117.115.515.215.313.6

8.814.3

172135132157113156165145199204169147167148194193168

Mean ± SD 11.3 :1: 2.3 0.100 i 0.031

Group II (normal TBG)P.W. 12FE.S. 45FP.R. 45MA.B. 55FB.L. 22FA.N. 45F

18.512.617.316.210.313.6

0.0670.0650.0710.0570.0480.053

Mean 1 SD 14.8 ± 3.1 0.060 i 0.008

11.33 i 4.0 14.0 i 2.8 163 =1 25

12.408.19

12.289.234.947.21

22.320.820.923.521.021.8

120160141176168159

9.04 i 2.9 21.7 i 1.1 154 i1 21

Pt

Mean i SD in all casesof thyrotoxicosis

<0.01 <0.01

12.2 4 2.9 0.090 ± 0.032

<0.001

Normal mean i SD(12 volunteers)

<0.001

6.2 ± 1.0 0.042 4 0.006

NS <0.001 NS

10.73 i 3.88 16.0 :h 4.2 160 i 24

<0.001 <0.001 <0.001

2.56 i 0.49 21.0 i 2.111 259 i 2111

PBI, protein-bound iodine; TBG, maximal binding capacity of thyroxine-binding alpha globulin; TBPA, maximalbinding capacity of thyroxine-binding prealbumin; NS, not significant (P > 0.05).

* Thyroxine iodine by column.$ Probability that the value in group I is identical to the corresponding value in group II.§ Probability that the value in the whole diagnostic category is identical to the corresponding value in the normal

group.11 Mean ± SD of 19 normal volunteers, rather than 12.

1444

THYROXINETRANSPORTIN THYROTOXICOSISAND HYPOTHYROIDISM

As shown in Fig. 1, quite satisfactory separation ofthyroxine-binding proteins was obtained with this method.

It was found by studies with various thyroxine en-richments from 50-1000,gg/100 ml that TBG and TBPAwere saturated by the addition of 75 and 600 lsg/100 ml,nonradioactive L-thyroxine to serum, respectively. There-

fore, samples of sera, enriched with the above amountsof nonradioactive L-thyroxine were used to determine themaximal binding capacities of TBG and TBPA, re-spectively. The appropriate carrier was prepared froma stock solution of 2 mg/ml nonradioactive L-thyrox-ine4 dissolved in 0.067 N NaOH. This was diluted1: 3 with 0.125 g/100 ml human serum albumin in iso-tonic saline solution and stored in a brown plastic bottleat 40C. Although this stock solution was stable for atleast 2 wk, it was ordinarily depleted within 5 days.Appropriate dilutions of the stock solution were madeimmediately before addition to serum.

All measurements of maximal binding capacities weremade at least in duplicate or triplicate and were run withconcomitant analysis of a serum pool and, occasionally, anormal serum as standards. Replicates were in goodagreement, and mean values were determined for eachserum sample. The mean values and standard deviationof maximal binding capacities of TBG and TBPA inthe serum pool were 21.5 ± 0.3 and 186 + 3 ,g/100 ml,respectively. The mean value for TBPA in the serumpool (which included sick patients) was appreciablylower than values in healthy volunteers (259 ± 21 fig/100 ml).

4 Mann Research Laboratories, Inc., New York, N. Y.

30

25

20

l5

5

GROUP I 11

TBG(pg /100 ml)

t

300

250

200

'150

50

TBPA FREE THYROXINE(,ug /100 ml) FRACTION (%)I I I

,*1.0a

0

I I

0.175

0J5c

L125

o.icz

0.07

D.025

0

I I

FIG. 2. TBG, TBPA, AND FREE THYROXINE FRACTION

IN THYROTOXICOSIS. The cross-hatched areas indicate the-normal ranges. The horizontal lines indicate the mean

values in each group of thyrotoxicosis. Group I has di--minished TBG capacity; group II has normal TBG.

TBG TB PA FREETHYROXINE(,Ug/KO ml) (Aug/IOOml) FRACTION (%)

GROUP I ]I I I I I

FIG. 3. TBG, TBPA, AND FREE THYROXINE FRACTIONIN HYPOTHYROIDISM. The cross-hatched areas indicatethe normal ranges. The horizontal lines indicate themean values in each group of hypothyroidism. Group Ihas elevated TBG capacity; group II has normal TBG.

Results

In 23 cases of thyrotoxicosis, the TBG capaci-ties ranged from 8.8-23.5 jug/100 ml (Table I andFig. 2). Despite the wide scatter, the mean valuefor TBG capacities in thyrotoxicosis was signifi-cantly reduced (mean + SD = 16.0 + 4.2 jug/100ml in thyrotoxicosis vs. normal of 21.0 + 2.1, P< 0.001). The TBPA capacities were diminishedin all instances of thyrotoxicosis (Table I andFig. 2). The mean TBPA capacity in thyrotoxi-cosis (mean + SD = 160 + 24 ug/100 ml) wassignificantly below the normal TBPA capacity(mean+ SD= 259±21 jug/100 ml, P< 0.001).

In 16 cases of hypothyroidism, the TBGcapaci-ties were normal or slightly elevated (Table II andFig. 3). The mean value (mean + SD = 25.4 ±3.5 ug/I100 ml) for TBG capacities in hypothy-roidism was significantly elevated (P < 0.001).The TBPA capacities in hypothyroidism showedsome scatter (mean ± SD = 251 ± 36 u.g/100ml), but there was no significant difference fromnormal (Table II).

The scatter diagram of the relation betweenfree thyroxine fraction and PBI value in thyroiddisease revealed the expected direct relation (Fig.4). The correlation coefficient of + 0.63 washighly significant statistically (P < 0.001). Theresults resembled the previous findings from this

1445

1446 MITSUO INADA ANDKENNETHSTERLING

TABLE II

Findings in sera from 16 cases of untreated hypothyroidism

Free thyroxine Free thyroxineSubjects Age and sex PBI fraction iodine TBG TBPA

sg/I100 ml % mAg/100 ml pg/100 ml Jsg/100 mlGroup I (elevated TBG)

C.K. 23F 1.4 0.019 0.27 25.8 256M.C. 36F 2.8* 0.028 0.78 27.8 194A.T. 40F 1.5 0.019 0.29 27.7 267R.C. 42M 2.2 0.026 0.57 27.6 238W.P. 45M 3.3 0.024 0.79 32.2 287M.S. 5OF 1.1 0.025 0.28 28.9 213E.W. 52F 2.2 0.029 0.64 27.0 238J.M. 25M 2.0 0.024 0.48 26.8 160E.G. SOF 1.0 0.026 0.26 24.4 255I.H. 72M 2.0 0.028 0.56 25.3 308M.O. 35F 2.6 0.024 0.62 26.9 275

Mean i= SD 2.0 :4: 0.7 0.025 i 0.003 0.50 ± 0.20 27.3 ± 2.1 245 i 43

Group II (normal TBG)J.V. 45F 0.8 0.031 0.25 21.9 266J.E. 46F 1.3 0.037 0.48 20.6 264F.R. 72M 1.1 0.047 0.52 21.0 255E.C. 38M 3.1 0.033 1.02 18.5 269O.L. 60F 1.1 0.037 0.41 23.5 275

Mean i SD 1.5 4_ 0.9 0.037 i 0.006 0.54 i 0.3 21.1 i 1.8 266 i 7

Pt NS <0.001 NS <0.001 NSMean i SD in all cases 1.8 i 0.8 0.029 i 0.007 0.51 4 0.22 25.4 :1 3.5 251 i 36

of hypothyroidism

P§ <<0.001 <0.001 <0.001 <0.001 NSNormal mean :i- SD 6.2 ±t 1.0 0.042 4t 0.006 2.56 i 0.49 21.0 i 2.1j 259 ± 21

(12 volunteers)

PBI, protein-bound iodine; TBG, maximal binding capacity of thyroxine-binding alpha globulin; TBPA, maximalbinding capacity of thyroxine-binding prealbumin; NS, not significant (P > 0.05).* Thyroxine iodine by column.

t Probability that the value in group I is identical to the corresponding value in group II.§ Probability that the value in the whole diagnostic category is identical to the corresponding value in the normal

group.11 Mean :1 SD of 19 normal volunteers, rather than 12.

0 0*

.0

.

0 * *

5 gl 15PBI ( ug/100 ml)

laboratory (6), which showed elevation of per-centage free thyroxine in thyrotoxicosis and dimi-nution in hypothyroidism. However, the PBIvalue is not considered to be the main factor reg-ulating the fraction of unbound hormone. Sinceloading normal sera with nonradioactive thyroxinein the clinical range produces much less elevationof the free thyroxine percentage than is usuallyobserved in sera from patients with thyrotoxicosis(5, 6), the relation illustrated does not imply thatthe thyroxine concentration of serum per se is amajor determinant. The effects of the bindingproteins are discussed below.

The scatter diagram of free thyroxine fractionagainst TBPA capacity (Fig. 5) revealed theexpected highly significant inverse correlation

;-e WM--- 0.200-z0

<t 0.150-

LLLLUZ 0.100-

0To_I 0.050-

LUJlLULL

FIG. 4. RELATION BETWEENFREE THYROXINEFRACTIONAND PBI IN THYROTOXICOSISAND HYPOTHYROIDISM. Thecross-hatched areas indicate the normal ranges.

THYROXINETRANSPORTIN THYROTOXICOSISAND HYPOTHYROIDISM

(- 0.59, P < 0.001). However, an even morestriking inverse correlation was evident in theplot of free thyroxine fraction against TBG (Fig.6), which had a correlation coefficient of - 0.84(P < 0.001). The data were subjected to addi-tional statistical treatment, including computationof partial and multiple correlation coefficients, aswell as simple correlation coefficients within thediagnostic categories.

As may be anticipated by inspection of Fig. 4,no significant correlation existed between freethyroxine fraction and PBI value if either diag-nostic category was considered separately (P >0.1). Similarly, within either the thyrotoxic orhypothyroid group, the free thyroxine fractionbore no significant relation to the TBPA capacity(Fig. 5). On the other hand, TBGcapacity borea statistically significant inverse relation to thefree thyroxine fraction- both in thyrotoxicosis (P< 0.001) and in hypothyroidism (0.025 < P <0.05).

The partial and multiple correlation coefficientsof free thyroxine fraction against PBI and/orTBPA and/or TBG revealed highly significantcorrelations only when TBG was included, re-gardless of inclusion or exclusion of PBI andTBPA. For thyrotoxicosis, P < 0.001; and forhypothyroidism, 0.01 < P < 0.025. All the sta-tistical data were compatible with a more signifi-

z

o 0.15

0

LL

Ld 0.10"z

0w

* THYROTOXICOSISo HYPOTHYROIDISM

0

0

0

00

50 100 150 200 25(TBPA (,ug/100ml)

'O 350

FIG. 5. RELATION BETWEENFREE THYROXINEFRACTIONAND TBPA IN THYROTOXICOSIS AND HYPOTHYROIDISM.The cross-hatched areas indicate the normal ranges.

A-R

z° 0.15-H

L-

LLJ 0.10-

uJLL

* THYROTOXICOSIS|o HYPOTHYROIDISM

.

S

S *e

0o00 W0 00 0

5 10 15 20 25 30 35TBG (,ug/IOOml)

FIG. 6. RELATION BETWEENFREE THYROXINEFRACTIONAND TBG IN THYROTOXICOSIS AND HYPOTHYROIDISM.The cross-hatched areas indicate the normal ranges.

cant role of TBGthan TBPA as a determinant ofthe free thyroxine fraction in thyroid disease (seediscussion).

For purposes of illustration, it appeared reason-able to divide the patients with thyrotoxicosis intotwo groups based upon the TBGcapacities: groupI with diminished TBGand group II with normalTBG. The difference between TBG capacity ingroup I (14.0 + 2.8 jug/100 ml) and that in groupII (21.7 ± 1.1 ug/100 ml) was appreciable,whereas the TBPA capacity in group I was notdifferent from that in group II (Table I andFig. 2). Despite the diminution of TBPA ca-pacities in all thyrotoxic sera, the free thyroxinefractions in group I with diminished TBG ca-pacities were much higher than those in group IIwith normal TBGcapacities (Table I and Fig. 2).

The patients with hypothyroidism could also bedivided into two groups based upon TBG capaci-ties: group I with elevated TBG capacities andgroup II with normal TBGcapacities. As shownin Table II and Fig. 3, group I with elevatedTBG capacities had markedly diminished freethyroxine fractions, while group II with normalTBG capacities had normal or nearly normal freethyroxine fractions. The difference between freethyroxine fractions in groups I and II was highlysignificant statistically (P < 0.001).

1447

MITSUO INADA AND KENNETHSTERLING

Z 25c0

I--F-

F- 20-zLLU

o 5-

IEELUD 0

0~~~~~~~~

LUJOf * THYROTOXICOSISLL ° HYPOTHYROIDISM

FIG. 7. FREE THYROXINE IODINE CONCENTRATIONINTHYROTOXICOSISAND HYPOTHYROIDISM. The cross-hatchedarea indicates the normal range. Group I has abnormalTBG capacity; group II has normal TBG.

The calculated free thyroxine iodine concentra-tions, representing the products of PBI values andfree thyroxine fractions, revealed marked devia-tions from the normal range in both thyrotoxicosisand hypothyroidism, regardless of subdivision intogroups I and II. The magnitudes of these ab-normalities (Fig. 7) are similar to differences pre-

viously observed between the normal range andthyroid diseases, when estimations of the absolutefree hormone concentration have been compared(5, 6). There was no statistically significant dif-ference in each diagnostic category between groups

I and II in this parameter (Tables I and II), as

anticipated from inspection of Fig. 7. The PBIvalues were significantly higher in thyrotoxicosisgroup II with normal TBG than in group I withdiminished TBG; evidently the higher PBI con-

centrations offset the lower free thyroxine frac-tions in yielding similarly elevated unbound hor-mone concentrations. In hypothyroidism a similaroffsetting difference in PBI concentrations was

present, although in this instance the differencewas not statistically significant (Table II).

DiscussionReverse flow paper electrophoresis (9) with

barbital buffer or ammonium bicarbonate buffer

at pH 8.6 has been used as a standard method todetermine the maximal binding capacity of TBG.However, since it became evident that the barbital(Veronal) buffer inhibited thyroxine binding toTBPA, Tris-maleate buffer, pH 8.6, was intro-duced by Ingbar to determine TBG and TBPAcapacities (11, 12). Recently, thyroxine-bindingprealbumin has been purified by Purdy, Woeber,Holloway, and Ingbar (13) and Oppenheimer,Surks, Smith, and Squef (14). Studies with theaddition of purified TBPA to serum revealedcorrelation of added protein with maximal bindingcapacity and also with stainable protein in pre-albumin-1 on starch gel (14, 15). Therefore, itwas concluded that the normal serum maximalbinding capacity was approximately 270 ug/100ml, as determined by glycine acetate electrophore-sis-a value considerably higher than that ob-tained with Tris-maleate buffer, under the condi-tions employed (11, 12).

Sterling and Tabachnick (10) first reported theuse of a glycine acetate system at pH 8.6 to sepa-rate thyroxine-binding proteins on paper electro-phoresis. Oppenheimer et al. (4) reported thatthe patients with thyrotoxicosis had normal TBGcapacities; although they employed the glycine ace-tate system, they used conventional paper electro-phoresis. Consequently, it is entirely possible thattrailing albumin and prealbumin were associatedwith TBG and obscured alterations which mighthave been evident with the reverse flow technique.Silverstein et al. (16) and Elzinga, Carr, andBeierwaltes (17) adapted the Durrum type cellfor reverse flow paper electrophoresis, a procedurewhich has distinct technical advantages. An im-portant advantage of the Durrum type cell is thatthe buffer reservoir contains four baffles to preventelectrode products from reaching the paper strips.Therefore, it was possible to minimize the effectof pH change, which is pronounced during elec-trophoresis with the glycine acetate system. Inimmediate proximity to the anode, pH may fallto 5.0, whereas at the cathode it may rise to 9.8.However, in the outer side of each reservoir, intowhich the paper strips were dipped with the paperwick, pH was almost constant (8.6 ± 0.1). Inthree successive runs without change of the solu-tion the same samples gave values of TBG andTBPA capacities within the error of simultaneous

1448

THYROXINETRANSPORTIN THYROTOXICOSISAND HYPOTHYROIDISM

replicates. Nevertheless, we always employedfresh buffer in the present study.

Thus, we reinvestigated the thyroxine-bindingproteins in thyrotoxic and hypothyroid sera, usingreverse flow paper electrophoresis with the glycineacetate system, pH 8.6, in the Durrum type cell.

The diminished binding capacity of TBPA insera of patients with thyrotoxicosis was first notedby Richards, Dowling, and Ingbar (1). Themarkedly elevated free thyroxine fractions in seraof patients with thyrotoxicosis have been ascribedto their diminished TBPAcapacities, as well as tothe increased total hormone concentration (4, 5).This tentative explanation did not appear suf-ficient to account for the wide variation in freethyroxine fractions in thyrotoxic sera, some ofwhich showed strikingly high values. Moreover,the TBPA was not found to be diminished in allof the thyrotoxic sera in the preliminary reportby Richards, Dowling, and Ingbar (1).

Diminution of TBG in thyrotoxicosis has beenreported by Silverstein and coworkers who usedVeronal buffer with reverse flow technique (16)and by Cuaron, who used conventional Tris-male-ate electrophoresis (18). Moreover, simultaneouswith the present work, Schussler, employing quitedifferent methodology, reported in an abstract(19) the finding of diminution of TBG in thyro-toxicosis.

In the present paper, the reduction of TBGcapacities present in 17 of 23 thyrotoxic sera andthe elevation of TBGcapacities in 11 of 16 hypo-thyroid sera appeared sufficient to explain thepronounced abnormalities of free thyroxine frac-tions. Indeed, in thyrotoxicosis and hypothyroid-ism, the inverse correlation between free thyroxinefractions and TBG capacities was much closerthan that with TBPA capacities. When diag-nostic categories were considered separately, onlyTBGbore a significant inverse relation to the freethyroxine fraction. These results suggested thatin thyroid diseases, TBG may sometimes play amore important role than TBPA in determiningthe free thyroxine fraction. The regulation of thefree thyroxine fraction by TBG was illustratedmost clearly by the division of the sera into twogroups, based upon their TBGcapacities.

In thyrotoxicosis, both the sera of group Iwith diminished TBG capacities and those ofgroup II with normal TBG capacities had simi-

larly diminished TBPA capacities. Nevertheless,the elevation of free thyroxine fractions was sig-nificantly greater where diminution of TBG oc-curred, that is, in group I. It appeared reason-able, therefore, to attribute the more markedelevation of free thyroxine fractions to the reduc-tion not only of TBPAbut also of TBGcapacitiesin the group I sera.

Furthermore, TBPA capacities showed no con-sistent deviations from normal in hypothyroidism.The sera of group I with slightly but significantlyelevated TBGcapacities had markedly diminishedfree thyroxine fractions, whereas those of groupII with normal TBG capacities had normal ornearly normal free thyroxine fractions.

Thus, these results suggested a significant clini-cal role of a variation in TBG.

Recent studies (20, 21) suggested that thediminished TBPA capacity in chronic illness andacute injury was due to reduction of actual con-centration of prealbumin due to diminished syn-thesis. Presumably, the diminished TBPA ca-pacities in thyrotoxicosis might also be due todiminished prealbumin synthesis; however, theycould be due to accelerated turnover of serumproteins (22, 23). The diminished TBG capaci-ties are not explained by the present findings, andit is not proven that diminished TBGcapacity re-flects reduction of the actual concentration ofTBG itself, although this is a plausible hypothesis.

Acknowledgments

The authors wish to express gratitude for the valuabletechnical assistance of Mr. Milton A. Brenner and Mr.Edward S. Newman.

References

1. Richards, J. B., J. T. Dowling, and S. H. Ingbar.1959. Alterations in the plasma transport of thy-roxine in sick patients and their relation to theabnormality in Graves' disease. J. Clin. Invest.38: 1035.

2. Tanaka, S., and P. Starr. 1959. Clinical observa-tions on serum globulin thyroxine-binding ca-pacity, using a simplified technique. J. Clin. En-docrinol Metab. 19: 84.

3 a. Robbins, J., and J. E. Rall. 1957. The interactionof thyroid hormones and protein in biologicalfluids. Recent Progr. Hormone Res. 13: 161.

3 b. Robbins, J., and J. E. Rall. 1960. Proteins associ-ated with the thyroid hormones. Physiol. Rev. 40(Suppl. 4): 415.

1449

MITSUO INADA AND KENNETHSTERLING

4. Oppenheimer, J. H., R. Squef, M. I. Surks, and H.Hauer. 1963. Binding of thyroxine by serumproteins evaluated by equilibrium dialysis andelectrophoretic techniques. Alterations in non-thyroidal illness. J. Clin. Invest. 42: 1769.

5. Ingbar, S. H., L. E. Braverman, N. A. Dawber, andG. Y. Lee. 1965. A new method for measuringthe free thyroid hormone in human serum and ananalysis of the factors that influence its concen-tration. J. Clin. Invest. 44: 1679.

6. Sterling, K., and M. A. Brenner. 1966. Free thy-roxine in human serum: Simplified measurementwith the aid of magnesium precipitation. J. Clin.Invest. 45: 153.

7. Bernstein, G., and J. H. Oppenheimer. 1966. Factorsinfluencing the concentration of free and total thy-roxine in patients with nonthyroidal disease. J.Clin. Endocrinol Metab. 26: 195.

8. Berson, S. A., R. S. Yalow, J. Sorrentino, and B.Roswit. 1952. The determination of thyroidaland renal plasma I"s clearance rates as a routinediagnostic test of thyroid dysfunction. J. Clin.Invest. 31: 141.

9. Robbins, J. 1956. Reverse flow zone electrophoresis.A method for determining the thyroxine-bindingcapacity of serum protein. Arch. Biochem Bio-phys. 63: 461.

10. Sterling, K., and M. Tabachnick. 1961. Paperelectrophoretic demonstration of thyroxine-bind-ing prealbumin fraction in serum. Endocrinology.68: 1073.

11. Ingbar, S. H. 1958. Pre-albumin: A thyroxine-binding protein of human plasma. Endocrinology.63: 256.

12. Ingbar, S. H. 1961. Clinical and physiologicalobservations in a patient with an idiopathic de-crease in the thyroxine-binding globulin of plasma.J. Clin. Invest. 40: 2053.

13. Purdy, R. H., K. A. Woeber, M. T. Holloway, andS. H. Ingbar. 1965. Preparation of crystallinethyroxine-binding prealbumin from human plasma.Biochemistry. 4: 1888.

14. Oppenheimer, J. H., M. I. Surks, J. C. Smith, and

R. Squef. 1965. Isolation and characterization ofhuman thyroxine-binding prealbumin. I. Biol.Chem. 240: 173.

15. Oppenheimer, J. H., M. Martinez, and G. Bernstein.1966. Determination of the maximal binding ca-

pacity and protein concentration of thyroxine-binding prealbumin in human serum. J. Lab. Clin.Med. 67: 500.

16. Silverstein, J. N., H. L. Schwartz, E. B. Feldman,D. M. Kydd, and A. C. Carter. 1962. Correla-tion of the red blood cell uptake of I-L-triiodo-thyronine and thyroxine binding globulin ca-pacity in man. J. Clin. Endocrinol Metab. 22:1002.

17. Elzinga, K. E., E. A. Carr, and W. H. Beierwaltes.1961 Adapation of the standard Durrum-typecell for reverse-flow paper electrophoresis. Am.J. Clin. Pathol. 36: 125.

18. Cuaron, A. 1966. Relationship between the invitro uptake of .1.I-triiodothyronine by erythro-cytes and its binding by serum proteins in thy-roid disease. J. Clin. Endocrinol Metab. 26: 53.

19. Schussler, G. C. 1966. Thyroxine binding globulin(TBG) in thyrotoxicosis and nonthyroidal illness.The Endocrine Society: Program of the 48thmeeting in Chicago. 113. (Abstr.)

20. Oppenheimer, J. H., M. I. Surks, G. Bernstein, andJ. C. Smith. 1965. Metabolism of iodine 131-labeled thyroxine binding prealbumin in man.Science. 149: 748.

21. Socolow, E. L., K. A. Woeber, R. H. Purdy, M. T.Holloway, and S. H. Ingbar. 1965. Preparationof 113 labeled human serum prealbumin and itsmetabolism in normal and sick patients. J. Clin.Invest. 44: 1600.

22. Rothschild, M. A., A Bauman, R. S. Yalow, andS. A. Berson. 1957. The effect of large doses ofdesiccated thyroid on the distribution and metabo-lism of albumin I' in euthyroid subjects. J. Clin.Invest. 36: 422.

23. Lewallen, C. G., J. E. Rall, and M. Berman. 1959.Studies of iodoalbumin metabolism. If. The effectof thyroid hormone. J. Clin. Invest. 38: 88.

1450