Mass fragmentography of cortisol and cortisone: Preliminary studies on the development of a reference method

Article abstract of DOI:10.1135/cccc20020001

Cortisol and cortisone are the two main glucocorticoids found in human plasma. Cortisol is commonly measured as a means of assessing adrenal function, usually by immunoassay, but such procedures must be subject to proper quality control which ideally uses

Full text of DOI:10.1135/cccc20020001

Mass Fragmentography of Cortisol and Cortisone
1
MASS FRAGMENTOGRAPHY OF CORTISOL AND CORTISONE:
PRELIMINARY STUDIES ON THE DEVELOPMENT
OF A REFERENCE METHOD
a
b
Hugh L. J. MAKINa1,*, David J. H. TRAFFORD and Norman F. TAYLOR
a
Department of Clinical Biochemistry, St. Bartholomew’s & the Royal London School
1
of Medicine & Dentistry, London E1 2AD, U.K.; e-mail: h.l.j.makin@mds.qmw.ac.uk
Department of Clinical Biochemistry, King’s College School of Medicine, Denmark Hill,
London SE5 9RS, U.K.; e-mail: norman.taylor@kcl.ac.uk
b
Received October 3, 2001
Accepted Jan uary 4, 2002
Dedicated to the memory of Dr Václav Černý.
Cortisol an d cortison e are th e two m ain glucocorticoids foun d in h um an plasm a. Cortisol is
com m on ly m easured as a m ean s of assessin g adren al fun ction , usually by im m un oassay, but
such procedures m ust be subject to proper quality con trol wh ich ideally uses an alyte target
values m easured by isotope dilution m ass fragm en tograph y (GC-MS). Th ree derivatives of
th ese steroids h ave been in vestigated for th eir ch rom atograph ic an d m ass spectral ch aracter-
istics in order to develop a sim ple an d rapid GC-MS m eth od for th is purpose. Trim eth ylsilyl
eth er-O-m eth yloxim e derivatives gave poor sen sitivity an d th e syn- an d anti-isom ers were re-
solved in to two peaks. Cyclic m eth ylboron ate-O-m eth yloxim e derivatives gave excellen t sen -
sitivity but were difficult to prepare, occasion ally givin g rise, by som e un kn own m ech an ism ,
to 11-oxidation of cortisol. Th e un derivatised 11-h ydroxyl of th is derivative of cortisol also
gave rise to ch rom atograph ic problem s. En ol-trim eth ylsilyl eth ers h owever were easy to pre-
pare an d gave excellen t sen sitivity (10–17 tim es th at ach ieved with th e trim eth ylsilyl
eth er-O-m eth yloxim e derivatives). Straigh t lin e calibration graph s were obtain ed an d th e
m eth od gave clean traces allowin g h igh specificity an d sen sitive m easurem en t of cortisol
an d cortison e in h um an plasm a usin g tri-deuterated cortisol as an in tern al stan dard.
Keyw ord s: Steroids; Mass spectrom etry; Glucocorticoid h orm on es; An alytical m eth ods;
GC-MS.
Cortisol (1) is th e m ain glucocorticoid h orm on e circulatin g in h um an
plasm a. It is secreted from th e adren al cortex an d con verted in both periph -
eral an d target tissues to cortison e, wh ich is also foun d at low levels in h u-
m an plasm a. Th e m easurem en t of plasm a cortisol is widely used in clin ical
laboratories as a m ean s of assessin g adren al an d pituitary fun ction in con -
jun ction with suppression an d stim ulation tests1. Th ese routin e assays are
Collect. Czech. Chem. Commun. (Vol. 67) (2002)
doi:10.1135/cccc20020001

2
Makin, Trafford, Taylor:
n ow carried out usin g im m un oassay system s, wh ich are con ven ien t, sim ple
an d capable of autom ation . However, th e use of such system s, wh ich are
n ot always specific, requires careful con trol, preferably usin g extern al qual-
ity assuran ce sch em es. In th e U.K., Nation al Health Service Laboratories are
required, for th e purposes of accreditation , to belon g to th e Nation al Exter-
n al Quality Assuran ce Sch em e (NEQAS), wh ich in cludes QA of cortisol as-
says2. Sch em es such as th ese sh ould h ave target con cen tration s obtain ed on
sam ples circulated to participatin g laboratories usin g referen ce m eth ods,
usually agreed to utilise gas ch rom atograph y-m ass spectrom etry (GC-MS).
In th e EU, efforts h ave been m ade to produce certified referen ce m aterial,
usin g GC-MS for th e m easurem en t, by a n um ber of differen t laboratories3,
of cortisol con cen tration in th e m aterial provided. Con tin uin g m on itorin g
of perform an ce of routin e laboratories is essen tial for th e m ain ten an ce of
h igh stan dards an d th e use of agreed referen ce m eth odology, usually
GC-MS, is an im portan t part of proper surveillan ce of perform an ce.
COCH2OH
OH
COCH2OH
OH
O
HO
O
O
2
1
Cortisol is a polar m olecule, wh ich can n ot be subjected to gas ch rom a-
tograph y with out derivatisation , both to im prove ch rom atograph ic perfor-
m an ce an d to protect th e steroid side ch ain . In th e EU study3, all seven
participatin g laboratories used O-m eth yloxim e-per-trim eth ylsilyl eth er
derivatisation an d [2H4]-cortisol as an in tern al stan dard. In th is exercise
on ly cortisol was m easured, alth ough cortison e is also presen t, albeit in
lower con cen tration s, in h um an plasm a. An exam in ation of th e literature
over th e last ten years h as in dicated a m ove towards LC-MS as a m ean s of
an alysis for plasm a cortisol an d oth er steroids presen t in h um an plasm a. In
our view such system s, wh ile m ore con ven ien t in th at th ey do n ot require
derivatisation , are m ore expen sive th an today’s sim ple GC-MS system s an d
also sacrifice a con siderable degree of specificity (un less tan dem system s are
used) an d sen sitivity. Modern tech n ology often ign ores th e lesson s of th e
past an d in order to provide a sim ple, sen sitive an d, perh aps m ost im por-
tan t, specific referen ce GC-MS procedure for cortisol (1) an d cortison e (2),
we in vestigated th e ch rom atograph ic an d m ass spectral ch aracteristics of
en ol-trim eth ylsilyl an d cyclic boron ate derivatives an d com pared th em to
th ose of th e widely used O-m eth yloxim e-trim eth ylsilyl eth er derivatives.
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

Mass Fragmentography of Cortisol and Cortisone
3
EXPERIMENTAL
Materials
All ch em icals were obtain ed in th e purest form possible, usually an alytical reagen ts from
Merck Eurolab (Lutterworth , Leicestersh ire LE17 4XN, U.K.), Sigm a–Aldrich Com pan y (Poole,
Dorset BH12 4QH, U.K.) or Perbio Scien ce U.K. Ltd., Tatten h all, Ch esh ire CH3 9RJ, U.K.) and
were used without further purification. [2H4]-Cortisol was a kind gift from Dr C. H. L. Shackleton
(Oaklan d (CA), U.S.A.) an d cortisol an d cortison e were from Steraloids (U.K.) Ltd. (New Barn et,
Hertfordsh ire EN5 1PR, U.K.). GC-MS was carried out usin g an Hewlett–Packard HP5890 gas
ch rom atograph lin ked to an HP5970 m ass selective detector in positive ion electron im pact
m ode, as previously described4,5. Electron im pact spectra were recorded across each peak
an d stored. Mass spectra recorded h ere were obtain ed by averagin g th e spectra from each
peak an d subtractin g th e backgroun d. GC was carried out usin g an HP1 cross-lin ked silicon e
station ary ph ase (15–20 m × ca 0.22 m m i.d.) as previously described6. Th e in jection system
of th e GC was m odified to accom m odate a JADE septum less in jection system , wh ich elim i-
n ates septum bleed an d allows h igh -tem perature in jection (350 °C). Th e advan tages of th is
in jection system h ave been discussed elsewh ere7.
Meth ods
Formation of O-methyloximes. Steroids were in cubated at 40 °C for 1 h or altern atively
overn igh t (>12 h ) at room tem perature with a solution of 10% m eth oxylam in e h ydroch lo-
ride in pyridin e. Pyridin e was rem oved un der vacuum an d th e steroid derivatives dissolved
in toluen e.
Formation of per-trimethylsilyl ethers. Th is was ach ieved by in cubation of th e O-m eth yl-
oxim e derivative with (trim eth ylsilyl)im idazole at 50 °C for 1 h . Th is derivatises all h ydroxy
groups. Th e steroid derivatives are th en purified on a sm all Lipidex 5000 colum n as previ-
ously described8. Th is procedure form s th e trim eth ylsilyl (TMS) eth ers on all h ydroxy
groups, in cludin g th e sterically h in dered 17α-h ydroxyl.
Formation of enol-trimethylsilyl ethers. Th e un derivatised steroid was in cubated at 60 °C for
15 m in with N-m eth yl-N-(trim eth ylsilyl)trifluoroacetam ide : iodo(trim eth yl)silan e : dith io-
th reitol (1 000 : 2 : 2, v/v/v). Th ese en ol-TMS derivatives were very rapidly form ed. Th e re-
agen t m ixture could be in jected directly in to th e GC with out furth er purification .
Formation of cyclic boronate esters. Th ese were form ed by in cubatin g th e steroid, after for-
m ation of th e O-m eth yloxim e derivative as described above, with m eth ylboron ic acid as de-
scribed previously9.
RESULTS
O-Meth yloxim e-TMS eth er derivatives of cortisol an d cortison e gave rise to
two peaks durin g GC (see Fig. 1). Th ese peaks were resolved in th e case of
cortison e but n ot in th e case of cortisol. Mass spectrom etry of th e cortisol
derivative sh owed a low abun dan ce m olecular ion m/z 636 with th e usual
ion s at m/z 605 (M – 31)+, m/z 515 (M – 90 – 31)+ an d m/z 425 (M – 90 – 90 –
31)+ all of rough ly equal abun dan ce.
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

4
Makin, Trafford, Taylor:
En ol-TMS eth ers also gave two peaks durin g GC but th e late elutin g peak
was m in or. It will be seen from Fig. 2 th at th e m ajor peaks from cortisol
an d cortison e are resolved durin g GC an d th e use of sin gle ion m on itorin g
allows resolution of th e m ajor peak of th e cortisol derivative from th e m i-
n or peak of th e cortison e derivative. Th e m ass spectrum of th e cortisol de-
rivatives sh owed a h igh -abun dan ce ion (base peak) at m/z 632 (M – 90)+
with very few other m ajor ions. The m ass spectrum of the cortisone derivative
showed high-abundance ions at m/z 630 (M – 90)+ and m/z 615 (M – 90 – 15)+
(base peak) an d again very few oth er m ajor ion s.
Th e steroid side ch ain can be protected by th e form ation of a cyclic
boron ate esters10 across C17–C21. Th e procedure used for th e form ation of
th e cyclic m eth ylboron ate esters appeared to be very destructive in th e case
of cortisol an d cortison e. Protection of th e oxo groups as O-m eth yloxim es
prior to form ation of cyclic boron ates lim ited th is destruction but th ere
were still m ajor problem s with quan titative con version to th ese derivatives.
Th e ch rom atograph y of th e cortisol O-m eth yloxim e-cyclic boron ate was
very poor presum ably due to absorption by th e un derivatised 11β-h ydroxy
group. In addition th e reagen t gave rise to con version of th e cortisol deriva-
1 000
73
800
600
400
605
361
515
246
425
200
0
296
143
1 000
800
100
200
300
400
500
600
m/z
Cortisol-OMO-TMS
syn- and anti-
600
400
200
0
Cortisone-OMO-TMS
syn- and anti-
11.0
11.5
12.0
12.5
13.0
13.5
Time, min
FIG. 1
GC-MS of O-m eth yloxim e-TMS (OMO-TMS) derivatives of cortisol an d cortison e. Ion s m/z 531
(cortison e derivative) an d m/z 605 (cortisol derivative) were m on itored an d th e m ass spectrum
obtain ed from th e m ajor peak (un resolved) of th e cortisol derivative is sh own in th e upper
righ t h an d corn er
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

Mass Fragmentography of Cortisol and Cortisone
5
tive to th e cortison e derivative. Alth ough th is con version could be m in i-
m ised by th e addition of ascorbic acid to th e derivatisin g reagen t, it was n ot
possible to obtain quan titative con version in a reproducible fash ion an d
ch rom atograph ic beh aviour of th e cortisol derivative could n ot be im -
proved (see Fig. 3).
Th e sen sitivity of detection of both cortisol an d cortison e was exam in ed
usin g both O-m eth yloxim e-TMS an d en ol-TMS derivatives. Stan dard curves
for each derivative were set up over th e ran ge 0–120 n g for cortisol an d for
cortison e, usin g an in tern al stan dard of ch olestan e (40 n g). Figure 4 illus-
trates th ese curves sh owin g th e 10–17-fold in crease in sen sitivity from th e
O-m eth yloxim e-TMS eth er wh en usin g th e en ol-TMS derivatives.
A sim ple procedure usin g th e en ol-TMS derivatives for th e m easurem en t
of plasm a cortisol an d cortison e was set up an d th is is sum m arised in Fig. 5.
Stan dard curves for both cortisol an d cortison e were lin ear (correlation
coeficien ts 0.999) an d th e in tercept on th e y axes were n ot sign ifican tly dif-
feren t from zero (Fig. 6). Figure 7 sh ows th e results from th e use of th is
1 000
74
800
615
600
400
147
200
527
436
400
243
300
315
467
500
0
100
200
600
m/z
Cortisone-
enol-TMS
1 000
632
800
600
400
200
74
8
10
12
Time, min
147
347
316
219
437
542
600
Cortisol-enol-TMS
0
100
200
300
400
500
m/z
FIG. 2
GC-MS of en ol-TMS derivatives of cortisol an d cortison e. Ion s m/z 615 (cortison e derivative)
an d m/z 632 (cortisol derivative) were m on itored an d th e m ass spectra of each of th e m ajor
peaks are sh own on th e left. Th e ordin ate of th e m ass fragm en togram is tim e in m in an d th e
abscissa is a m easure of relative abun dan ce
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

6
Makin, Trafford, Taylor:
m eth od for th e an alysis of 1 m l of h um an serum . Wh ile furth er de-
velopment work is still required, it is estimated that normal levels of cortisol and
cortisone could be routinely m easured in 0.5 m l of serum with signal : noise
ratios of >10 for cortison e an d >100 for cortisol.
30 000
Ion 444
20 000
Cortisol
derivative
10 000
0
Ion 442
30 000
20 000
Cortisone derivative
10 000
0
5
10
15
Time, min
FIG. 3
GC-MS of O-m eth yloxim e-m eth ylboron ate derivative of cortisol. Ion s m/z 444 (cortisol deriva-
tive; poor peak sh ape due to adsorption durin g ch rom atograph y) an d m/z 442 (cortison e deriv-
ative; th is cortison e peak was form ed from cortisol durin g derivatisation ) were m on itored
8
6
8
6
b
a
4
4
2
0
2
0
0
20
40
60
80
100
0
20
40
60
80
100
Mass, ng
Mass, ng
FIG. 4
Com parison of GC-MS respon se between O-m eth yloxim e-TMS (ꢀ) an d en ol-TMS (ꢀ) deriva-
tives of cortisol (a) an d cortison e (b). Th e slope of th e lin e in dicates th e sen sitivity
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

Mass Fragmentography of Cortisol and Cortisone
7
1 ml plasma + deuteriated cortisol
Extract with acetonitrile, centrifuge, dilute
with phosphate buffer, pH 10.5
C
cartridge
18
Wash with water
Wash with 20% methanol
Elute with methanol, redissolve in
5% isopropylalcohol in hexane
SEP-PAK SIL
Wash with 10% isopropylalcohol
in hexane
Elute with 3 ml of methanol
GC-MS (after derivatisation)
FIG. 5
Flow ch art sh owin g proposed m eth odology for th e m easurem en t of cortisol an d cortison e in
h um an plasm a usin g th e en ol-TMS derivatisation procedure described in th e text
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0
10
20
30
40
50
Mass, ng
FIG. 6
Stan dard curves for cortisol an d cortison e usin g th e m eth odology outlin ed in Fig. 5. En ol-TMS
eth ers were m ade as described in th e text an d subjected to GC-MS. Mon itorin g ion s m/z 632
(cortisol), m/z 615 (cortison e) an d m/z 635 [2H3]-cortisol used as an in tern al stan dard. Th e area
ratios 632/635 (ꢀ; y = –2.2337 · 10–3 + 2.7104 · 10–2x, R2 = 0.999) and 615/635 (ꢀ; y = 1.2957 · 10–2
+ 1.7967 · 10–2x, R2 = 0.999) are plotted on th e abscissa again st m ass of cortisol an d cortison e
on th e ordin ate
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

8
Makin, Trafford, Taylor:
DISCUSSION
Th e use of O-m eth yloxim e-TMS derivatives for th e GC-MS of steroids is still
widely used. Th ese derivatives an d procedures based th ereon h ave proved
very useful, particularly wh en studyin g urin ary steroid profiles11. However
wh en applyin g GC-MS for th e m easurem en t of th e lower con cen tration s of
cortisol an d cortison e in h um an plasm a, sen sitivity assum es con siderable
im portan ce. Th e work described h ere h as sh own th at th e use of en ol-TMS
derivatives in creases th e sen sitivity of detection by 1–2 orders of m agn i-
tude. Use of th ese derivatives for th e detection an d m easurem en t of an abo-
lic steroids is widespread12 an d alth ough a paper was publish ed13 in 1972
advocatin g th e use of en ol-TMS derivatives, to th e best of our kn owledge it
h as n ot been applied to th e GC-MS of cortisol. Th e sim ple m eth od de-
scribed h ere n eeds furth er evaluation , in cludin g recovery experim en ts an d
assessm en t of with in - an d between -batch precision , before it could be
widely adopted. Th is in vestigation h as dem on strated th e n eed fully to eval-
1 200
Cortisol
800
Cortisone
400
2 000
2
[ H ]-Cortisol
3
1 600
1 200
800
400
0
7
8
9
10
Time, min
FIG. 7
GC-MS of extracts from 1 m l of h um an plasm a processed as outlin ed in Fig. 5. Th e ion s m on i-
tored durin g GC-MS are sh own on th e upper trace for cortisol (m/z 632) an d cortison e (m/z
615). Th e ion m/z 635 was m on itored to produce th e lower trace, sh owin g th e [2H3]-cortisol
Collect. Czech. Chem. Commun. (Vol. 67) (2002)

Mass Fragmentography of Cortisol and Cortisone
9
uate existin g m eth odology before use an d to con sider derivatives described
m an y years before wh ich m ay provide solution s to today’s problem s, espe-
cially wh en allied to m odern tech n ology.
REFERENCES
1. Grant J. K., Beastall G. H.: Clinical Chemistry of Steroid Hormones, p. 91. Croom Helm,
London 1983.
2. Middle J. in: Steroid Analysis (H. L. J. Makin, D. B. Gower and D. N. Kirk, Eds), p. 647.
Blackie Academic and Professional, Glasgow 1995.
3. Thienpont L., Siekmann L., Lawson A., Colinet E., De Leenheer A.: Clin. Chem.
(Washington, D. C.) 1991, 37, 540.
4. Makin H. L. J., Calverley M. J., Jones G. in: Steroid Analysis (H. L. J. Makin, D. B. Gower
and D. N. Kirk, Eds), p. 562. Blackie Academic and Professional, Glasgow 1995.
5. Jones G., Makin H. L. J. in: Modern Chromatographic Analysis of the Vitamins (A. P. De Leenheer,
W. E. Lambert and J. F. van Bocxlaer, Eds), 3rd ed., p. 75. Marcel Dekker, New York 2000.
6. Qaw F., Calverley M. J., Schroeder N. J., Trafford D. J. H., Makin H. L. J., Jones G.: J. Biol.
Chem. 1993, 269, 282.
7. Coldwell R. D., Trafford D. J. H., Makin H. L. J.: J. Mass Spectrom. 1995, 30, 348.
8. Street J. M., Trafford D. J. H., Makin H. L. J.: J. Lipid Res. 1986, 27, 208.
9. Coldwell R. D., Trafford D. J. H., Makin H. L. J., Varley M. J., Kirk D. N.: Clin. Chem.
(Washington, D. C.) 1984, 30, 1193.
10. Brooks C. J. W., Cole W. J.: J. Chromatogr. 1986, 362, 113.
11. Trainer P. J., Drake W. M., Perry L. A., Taylor N. F., Besser G. M., Monson J. P.: J. Clin.
Endocrinol. Metab. 2001, 86, 2989.
12. Gower D. B., Houghton E., Kicman A. T. in: Steroid Analysis (H. L. J. Makin, D. B. Gower
and D. N. Kirk, Eds), p. 468. Blackie Academic and Professional, Glasgow 1995.
13. Chambaz E. M., Madani A., Ros A.: J. Steroid Biochem. 1972, 3, 741.
Collect. Czech. Chem. Commun. (Vol. 67) (2002)