A 1111. clin. Biochem.

10 (1973) 115

Plasma Testosterone Using Sephadex LH-20 and Saturation

Analysis by Competitive Protein Binding

DAPHNE M. LAWRENCE and G. I. SWYER

Obstetric Research Laboratories, University Colleg« Hospital Medical School, London, WC IE 15 DH

A method for measuring plasma testosterone using separation on short columns of Sephadex
LH-20 and competitive protein binding is described. This is more reliable and practicable than
methods using paper or thin layer chromatography separation and Sephadex LH-20 can be re-used
after washing.

Methodological problems of variable blanks from used for making up the standards and rinsing the
0-0.4 ng equivalent of testosterone from paper used assay tubes was refluxed over pellets of sodium
in chromatography after refluxing with solvents for hydroxide A.R. and redistilled twice.
24 h and washing with methanol have been en- All the glassware was rinsed twice with solvents
countered in this laboratory. This made the assay before use.
unreliable but the problems were solved with the use Extraction: 20 fLl (50000 DPM) I, 2"H-testosterone
of Sephadex LH-20 by column chromatography in ethanol and 3 ml 0.01 rnol/l NaOH and 0.2--0.4
(Murphy, 1970). For the elution of the testosterone ml of male plasma or 2-3 ml of female plasma was
fraction a column of 31 x 0.9 ern and an elution extracted with 30 ml dichloromethane. After
volume of 26 ml was recommended by Murphy aspiration of the aqueous layer the solvent was
(1970). This was reduced, apparently without loss of washed twice with 2 ml volumes of water. After
specificity, and a routine method established, whereby freezing any remaining water at - 20"C the solvent
a large number of columns could easily be managed was poured into 50 ml flasks with round bottoms and
and 5 ml of the solvent system used for the elution dried on a water bath at 60-70°C. It has been
evaporated directly in the tube for the assay and assumed that losses by volatilisation at the operating
measured by competitive protein binding. temperatures of 60-70"C were minimal. In support
of this, the studies of Marshall-Jones (1966) with
MATERIALS AND METHODS
cortisol and aldosterone using temperatures varying
from 40-80°C and either an air jet or nitrogen,
I, 2"H-testosterone (specific activity: 40-45 curies/ showed no significant losses of either steroid.
rnmol) was purchased from New England Nuclear Column chromatograplty : the dried extracts were
Corporation, Frankfurt. It was used once without transferred to the columns in 0.2 ml of the solvent
purification and subsequently purified on Sephadex system and then a further wash with 0.2 ml solvent
LH-20 before each batch of test samples was assayed. system and the testosterone eluted in the 7-12 ml
Five fLl of I, 2"H-testosterone in benzene/ethanol, fraction. In this system testosterone is adequately
9: I was evaporated and added in 0.2 ml of the resolved from t>'-androstenedione, ""'-androstenediol
solvent system to a column of Sephadex LH-20 and and ~"-androstenedjol.There is some overlap of the
eluted in the 7-12 ml fraction into a glass-stoppered end of the dihydrotestosterone peak and the start of
tube. After evaporation on a Tecam 'Dri-Block ' the testosterone peak (Fig. I) but this was reduced
under nitrogen, this was taken up in 5 ml of absolute by discarding the first 0.5 ml of the testosterone
alcohol. eluate (6.5-7.0 ml). The 5 ml eluate was collected into
Solvent system: chloroform, heptane, ethanol test tubes (7.5 x 1.2 cm) and evaporated under
50:50:0.25, water to saturation (Murphy, 1970). nitrogen on a 'Dri-Block ' at 70°C.
Sephadex LH-20, Pharmacia Fine Chemicals AB. Competitive protein binding: the method was that
Equilibrated for at least 12 h in the solvent system of Mayes and Nugent (1968). Plasma from the last
and the fines removed by washing the columns with trimester of pregnancy was used as a source of
solvent and discarding the eluate. steroid hormone binding globulin (S.H.B.G.) and
Chromatography columns: columns 18 x 0.7 cru diluted I :30 in glass distilled water. S.H.B.G.
with a sintered glass disk and column height of 13 ern solution, 0.15 rnl, was added to each tube and to a
Sephadex LH-20 were used. The outlet could be series of tubes containing 0, 0.5, 1.0, 1.5 and 2.0 ng
stoppered with a plug of solvent-resistant Auto- testosterone with a 12 fLl I, 2"H-testosterone (30000
Analyzer tubing attached to a piece of glass rod. DPM). The tubes were mixed on a 'Rota-mixer' and
Glass distilled water was used for the solutions. either incubated for 2 h at room temperature, 10 min
Solvents were redistilled before use. The ethanol at 37"C, or overnight at 4°C. One third of the total
115
116

3 r.p.rn., a sample of supernatant, 0.1 ml (A2) was

transferred to a vial.
T Counting: 10 ml of Dioxan scintillant was added
and the vials counted in a Packard Tri-Carb model
3375 liquid scintillation spectrometer for 5 min or
10000 c.p.m.
Calculation: a standard curve was plotted so that
100 x A2/A I or 'x, free testosterone is plotted
against concentration of testosterone in ng where A I
is the CPM in the 0.05 ml aliquot for the determina-
tion of sample recovery and A2 is the CPM of the
0.1 ml of supernatant after centrifugation which
contains the free testosterone not bound to protein.
The plasma sample = amount calculated from the
standard curve x R/3(AI) ng where R/3 (AI) is
the correction for the recovery. R is the 20 ,.,.1
I, 2"H-testosterone in CPM added to the samples at
the start of the extraction.
Precision: the coefficient of variation for between-
assay precision was 5.4 ~{, for a male plasma of
average mean value 744 ng/ I 00 ml; standard devia-
tion ± 40.5, N = 18. For a female plasma the
5 6 7 B ~ 10 11 12 13 coefficient of variation was 11.0 % mean value
63 ng/lOO ml; standard deviation ± 6.9, N = 10.
El ua t e ml. Recovery: recoveries of 40-86 ~{, with a mean of
Fig. t.-Elution of I, 23H-testosterone (T) and dihydro- 61 %werefound(N = 300).lnthosesituationswhere
testosterone (DHT) using Sephadex LH-20. Solvent very low levels of endogenous testosterone has to be
system: heptane SO, chloroform SO, ethanol 0.25, water to measured, allowance for the variable contribution of
saturation. the labelled marker to the total mass present, can be
made according to the formulation of Mayes and
sample, 0.05 ml, (A I), was removed for determination Nugent (1968).
of sample recovery and 0.05 ml of saturated ammon- Specificity: the specificity of the method depends
ium sulphate was added before counting. Saturated on the isolation by Sephadex and the assay by
ammonium sulphate solution, 0.1 ml, was added to competitive protein binding. In practice evaluation
each tube, mixed, then covered with parafilm and was by comparison of the results with those derived
left for 5 min. After centrifuging for IS min at 3000 from other methods for normal adult ranges, Table I.

Table I. Plasma testosterone concentrations in normal subjects (ngiIOO ml)

~

Author Method Male subjects range Female subjects range

- ~~- --~~-

Mean S.D. Range Mean S.D. Range

~~- ~~- ~~--

Segre ('I al. (1964) Double isotope deriv. 650 140 440-960 54 14 34-101
(N 8) (N 50)
Collins et al. (1968) Gas chromatographic 528 261 238-1195 40 14 18-71
(N 41) (N 20)
Dufau et al. (1972) Saturation analysis by radio- 709 143 450-1100 54 22 25-80
immunoassay (N 30) (N 22)

Lawrence and Swyer- Saturation analysis by competi- 629 J60 408-1022 51 14 28-71
present method tive protein binding (N 27) (N II)
 117

ng.l100ml Fig. 2.-Normal male plasma testosterone concentrations

compared with those obtained in hypogonadal males and
normal female values compared with hirsute females.

1100

1000

•
ng / l11U ml
900

800 • 160 •
•
•
700 ·• 140
..I
600
-.--
•• 120
•
I
500 100
·
..• ••
400
•• -t-
80 •
•
•

·•·•
I

...
300
•
60
· •
•
•
200
------
40
··
• •
100 20
I
•••
Normal Hypogonadal Normal Hirsute
males males females females
(H 27) (N 27) (N 11) (Ho 16)
J 18

Table 3. Clinical detail» ofpatients with hypouonadism

Case Age Clinical data Plasma

testosterone
ng/lOO ml

I 39 Eunuchoidism, gynaecornastia. Sex chromatin-ve. 30

2 31 Right testicle atrophic, left not palpable. Scanty facial and pubic hair. 37
3 19 Delayed puberty. Voice not broken and not shaving. 45
4 36 Eunuchoidism. 49
5 17 Delayed puberty with no secondary sexual characteristics. 70
6 19 Eunuchoidism, gynaecomastia. Sex chrornatin-ve 79
7 32 Eunuchoidism. 84
8 36 Pituitary lesion. 87
9 19 Delayed puberty. Some development of secondary sexual characteristics
following HCG injection at 16. 89
10 23 Small, soft testicles. Voice not completely broken and not shaving. Downy
pubic hair. 98
II 45 Pre-pubertal-sized testicles. Adult male voice but not shaving. A little pubic
and axillary hair. Has had HCG injections at 15 and 16. 126
12 23 Cryptorchidism and testicles not identifiable. Voice unbroken but abundant
axillary and pubic hair. 181
13 34 Small, soft testicles. Voice not completely broken. Scanty facial and pubic
hair. 225
14 36 Small soft testicles. A little facial, axillary and pubic hair. Poor musculature. 255
15 29 Adult panhypopituitarism. 274
16 16 Delayed puberty, obesity, gynaecomastia. 290
17 39 Klinefelter's syndrome. Sex chromatin + ve. 305
18 33 Klinefelter's syndrome. Sex chromatin t- ve. 330
19 14 Gynaecornastia and he was reared as a girl. L. testicle undescended. Sex
chromatin - ve. 341
20 34 Right testicle small and in upper part of scrotum. Normal androgenic charac-
teristics. 349
21 30 Bilateral undescended testicles. 378

Table 4. Clinical details offemale patients with hirsutism,

Case Age Clinical condition and urinary 17 KS Plasma

testosterone
ngjlOO ml
----- - - - - ---------._-----.~
-------
I 18 Hirsutism, normal menstruation. Normal 17 KS. 18
2 20 Hirsutism, oligomenorrhoea. Normal 17 KS. 43
3 28 Mild adrenal hyperplasia with hirsutism. Raised 17 KS. 54
4 26 Hirsutism, oligomenorrhoea. Normal 17 KS. 59
5 23 Hirsutism, regular menstruation. Normal 17 KS. 71
6 19 Hirsutism, oligomcnorrhoea. S I. raised 17 KS. 7M
7 25 Hirsutism, regular menstrual cycles. Normal 17 KS. 80
8 31 Hirsute, obese. Normal 17 KS. M2
9 19 Hirsutism, secondary amenorrhoea. Upper normal 17 KS. 84
10 29 Hirsutism, secondary amenorrhoea. S l. raised 17 KS. 88
II 29 Hirsutism, secondary amenorrhoea. SI. raised 17 KS. 89
12 21 Hirsutism. Stein-Leventhal syndrome. S I. raised 17 KS. 92
13 18 Hirsutism, oligomenorrhoea. S I. raised 17 KS. 97
14 24 Hirsutism, Stein-Leventhal syndrome. S I. raised 17 KS. 104
15 28 Hirsutism. Stein-Leventhal syndrome. S I. raised 17 KS. 104
16 24 Extensive hirsutism, mild adrenal hyperplasia. Raised 17 KS. 162
\ I
 119

A further aspect of specificity was the low blank 177 ng/loo ml (S.D. 122, N = 21) and this was
values of the method. In agreement with Murphy statistically different (P = 0.001) from the mean for
(1970) the blanks most frequently approached the the normal group of 629 ngj I 00 ml (S. D. 160,
zero point of the curve. Recoveries of added testo- N = 27).
sterone to plasma stripped of steroids or to glass A group of hirsute women with and without
distilled water were as follows: Recoveries for 0.5 ng menstrual irregularities had a mean value of 81 ngj
were 0.53 ng (S.D. 0.092, N = 10), for 0.7 ng, 100 ml (S.D. 31, N = 16) which differed from the
0.72 ng (S.D. 0.089, N = 10) and for 1.0 ng, 1.04 ng normalmeanvalueof51 ng{looml(S.D.14, N = II)
(S.D. 0.13, N = 13). but five values in the 15 hirsute subjects were within
A batch of 32 samples can be processed within the normal range. (Table 4 and Fig. 2).
two days. A row of columns can easily be managed
and the advantage of the short columns was the Thanks are due to Mr. H. T. McGarrigle for helping
speed of elution and the reduction of solvent volume with the chromatographic columns. The Wellcome Trust
which can be directly evaporated in the assay tube. are thanked for a generous grant for the purchase of the
Sephadex LH-20 was washed with 35 ml of solvent Packard Tri-Carb model 3375 liquid scintillation spectro-
before re-using. The height of the column must be meter.
exactly 13 ern and checked before use.
REFERENCES

RESULTS Collins, W. P., Sisterson, 1. M., Koullapis, E. N.,

Mansfield, M. D., Sommerville, I. F. The Evaluation of
The normal ranges are given in Table I and the a gas-liquid chromatographic method for the determin-
individual results are plotted in Fig. 2. I n three of ation of plasma testosterone using nickel-63 electron
five male subjects examined a variation with higher capture detection. J. Chromatog. 37 (1968) 33.
results in the morning was observed (Table 2). De la Pena, A., Goldzieher, 1. W. Separation of free and
bound steroid in the competitive protein binding assay
for testosterone: accuracy and reproducibility. Steroids
18 (1971) 195.
Dufau, M. L., Catt, K. 1., Tsuruhara, T., Ryan, D.
Table 2. Plasma testosterone in jive male subjects Radioimmunoassay of plasma testosterone. Clin.
(n/;,/IDO ml) at three limes ofday, chim. Acta. 37 (1972) 109.
Marshall-Jones, P. Ph.D. Thesis, Steroids in human
12.DO-13.DO h 16.DO-17.DO h body fluids. University College Hospital Medical
9.DO-10.DO h
School, 1966.
Mayes, D., Nugent, C. A. Determination of plasma
660 655 545
588 testosterone by the use of competitive protein binding.
706 605
.I. clin. Endocr. 28 (1968) 1169.
818 775 580
Murphy, B. E. P. Methodological problems in competitive
995 990 994
960 1021 protein binding techniques: the use of Sephadex
1022
column chromatography to separate steroids in
Diczfalusy, E. (ed.) Steroid Assay by Protein Binding,
2nd Karolinska Symposium on Research Methods
in Reproductive Endocrinology, Copenhagen, 1970,
A series of males with clinical evidence of hypo- p.37.
gonadism were all found to have lower plasma Segre, E. J., Klaiber, E. L., Lobotsky, J., Lloyd, C. W.
testosterone than the normal subjects (Table 3). The Hirsutism and virilising syndromes. Ann. Rev. Med. 15
mean value in these males with hypogonadism was (1964)315.