Vol. 175, No. 3, 1991 March 29, 1991
BIOCHEMICAL
NATURALLY C. Pena*,
J.H.
OCCURRING Medina,
AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 1042-l 050
BENZODIAZEPINES
M. Piva*, L.E. and A-C. Paladini*
IN HUMAN MILK Diaz+,
C. Danilowicz'
Instituto de Biologia Celular, Facultad de Medicina, *Institute de Quimica y Fisicoquimica Biologicas (UBA-CONICET) and +Departamento de Quimica Analitica, Facultad de Farmacia y Bioquimica, Junin 956, 1113, Buenos Aires, Argentina Received
February
22,
1991
The presence of benzodiazepine-like molecules was detected radioimmunologically in the plasma and milk of 12 women and in the plasma of 9 men. All subjects were non-users of benzodiazepines. The concentration of these biological materials expressed as diazepam equivalents per mL amounted to 2.54 f 0.74 ng in male to 2.20 f 0.35 ng in female plasma and to 1.91 + 0.54 ng plasma; active compounds in milk in milk. Further investigation of the permitted the unequivocal identification of diazepam, both free three and bound to a presumably protein carrier and, at least, molecules. Their origin either from more benzodiazepine-like endogenoue biosynthesis is dietary 8ources or as a result of 0 1991AcademicPress, Inc. still unclear.
Benzodiazepines are organic molecules whose anxiolytic, anticonvulsant and sedative actions were discovered serendipitously in 1957 (1). For many years, they were considered non-natural heterocycles until recent studies in several laboratories have demonstrated that mammalian brains contain small amounts of BZDs (2-7). Furthermore, trace amounts of BZDs and other incompletely characterized BZD-like molecules were found in drug-free human sera (6,8-10) human cerebroapinal fluid (7,11), rodent plasma (9, 12), and rat adrenal gland (4). Recently, BZD-like molecules were also detected in several plants, including vegetables, grains and flowers (13-15) and in our laboratory they were detected in cow tentatively identified as DZ (5,161. milk (5), one of them being . . Abbrenahma
ACN: acetonitrile; BZD: benzodiaiepine; DZ: diazepam; tritium labelled flunitrazepam; HPLC: high performance chromatography; PEG: polyethylene glycol; DZ E: diazepam lent; MAb: monoclonal antibody. 0006-291X/91 Copyright All rights
$1.50
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
1042
["HI-FNZ: liquid equiva-
Vol.
175,
No.
3, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
These latter findings provide a possible explanation for the occurrence of BZDs in various organs of animals and man. Denker and Johansson (17) have found BZD-like agents in mothers-milk but their assay noncould not exclude the presence of BZD receptor ligands in the samples. Here we report the detection of BZD-like substances in milk from women known to be non-users of BZDs and the identification of the major component as DZ. Also, we confirm the presence of small amounts of BZD-like substances in drug-free human plasma. MATERIALS
AND METHODS Prmedvx
Blood samples (3-6 mL) were obtained from twenty one normal healthy volunteers (nine men and twelve women) known not to be exposed to BZD drugs during the previous 10 years of their life. A similar criteria was used to obtain milk samples (5-20 ml) from 12 newly delivered women 7 of which also provided blood samples. Sample collection was performed under conditions of asepsia and avoiding any possible contamination with BZDs. The extraction, partial purification and characterization of BZD-like molecules were carried out in glassware free of BZDs. Samples were extracted twice in 5 vol of 2 M acetic acid in the cold (5). After centrifugation the resulting supernatants Associates, were filtered through Cl8 Sep Pak cartridges (Waters USA) previously conditioned with 2 mL of methanol and5mL of water. The material eluted from the cartridges with 5 mL of 80% ACN was recovered by evaporation to dryness using a centrifugal apparatus (Savant,USA) and the BZD-like activity of the residue was measured by radioimmunoassay or radioreceptor assay as indicated below. With this technique there was less than 10% intersample variability as determined by radioimmunoassay. Lyophilized 80% ACN alone had no effect in both assays.
In order to identify the chemical nature of the BZD-like substances detected in the milk samples by radioimmunoassay, 930 mL of human milk from a single donor, were lyophilized and the residue extracted with 1L of 0.1% trifluoroacetic acid in methanol during 4 hs at 4"C, with agitation. The suspension was then centrifuged at 5,000 rpm for 30 min. The pellet was reextracted overnight and centrifuged. The pooled supernatants were concenin a rotary evaporator and centrifuged trated, in vacua, at 18,000 rpm for 90 min to obtain a clear supernatant that was submitted to solid phase extraction in five cl8 Sep Pak cartridges, as described above. The eluates were evaporated and the residue was chromatographed in reversed phase HPLC columns. Fractions were collected every minute and aliquots of each fraction were dried and tested for activity in the radioreceptor and radioimmunoassays. Ultraviolet absorption was detected in a 2140 rapid spectral detector (LKB, Sweden). The HPLC steps were as follows: step? I: The material eluted from the Sep Pak cartridges was applied to column (Vydac, The Separation a SeIIIipreparatiVe cl8 Group, Hesperia, CA, USA, 1 cmx25 cm, 5 urn particle size), equilibrated with 6% ACN in water and developed with a lo%-70% linear gradient of 80% ACN over 40 min at a flow rate of 2mL/min. 1043
Vol.
175, No. 3, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Step II: Active fractions from Step I were evaporated and apto a Cd column (Vydac, 0.46 cm x 25 cm, 5 um). Fractions 3 plied isocratically with 20% and 4 from Step I were chromatographed methanol, 10% ACN in water (Solvent A) at a flow rate of 1 mL/min. Fraction 5 from Step I was chromatographed isocratically with solvent A for 20 min, followed by a 0%-30% gradient of a flow rate of 1 mL/min. To 80% ACN (solvent B) over 20 min at test the degree of recovery of the procedure, a parallel experiment was carried out with 30 ml of milk additioned with a tracer amount of sH-FNZ. A recovery of radioactivity higher than 65% was found. &ss spentrometrv: The spectra were obtained with a Shimadzu GC-MS-QP 1000 operated in the EI mode. The ionizing voltage was 70 eV and the ionizing current 60 uA. The scan rate was 1 set/decade and the mass range scanned was 50-500 atomic mass units. The data were analyzed and the intensity pattern of selected characteristic masses was compared with those of reference BZDs.
Binding of sH-FNZ (spec. act. 85 Ci/mmol, NEN) was accomplished as already described (21), using an extensively washed crude synaptosomal membrane fraction from bovine cerebral cortex. Briefly, for each assay, triplicate or quadruplicate membrane samples containing 0.2-0.4 mg protein as determined by the Lowry's method (22) were suspended in 1 mL Tris-HCl buffer, pH 7.3. The incubation was carried out at 4°C for 60 min with 0.6 nM was determined in parallel incubasH-FNZ. Non-specific binding 3 UM FNZ or clonazepam and represented tions in the presence of 515% of the total. Bound and free radioactivity were separated filters with three by rapid filtration through GF/B (Whatman) washes of 5 mL each of the incubation buffer. The filters were then dried and transferred to vials with scintillation cocktail (2.5 diphenyloxazole-xilene) and the radioactivity was measured with 40% efficiency. Aliquots of the reconstituted Sep Pak or HPLC fractions of either human plasma or milk extracts were tested for their ability to displace the binding of 0.6 nM sH-FNZ to the above described membranes.
An antibenzodiazepine monoclonal antibody (MAb 21-7F9, kindly provided by Dr. A. De Blas) directed against 3 hemisuccinyloxyclonazepam was utilized (2). This antibody specifically recognizes BZDs with affinities ranging from 0.4 nM to 0.8 uM (2, 5). For the radioimmunoassay 30-60 ng of MAb 21-7FS were incubated with 0.6 nM sH-FNZ in 1 mL of 25 mM Tris HCl buffer, pH 7.3, in the absence or in the presence of reconstituted aliquots of the Sep Pak or HPLC fractions. After 60 min incubation at 4"C, 100 UL of gamma-globulin (1.2 mg/mL) and 400 UL of PEG 8000 (7.5% w/v) were added to each assay tube. After mixing and standing at 4°C for 10 min, the precipitated complex was separated by vacuum filtration through GF/B glass fiber filters and washed twice with 5 mL cold buffer containing 5% PEG. Non specific binding to MAb 217FS was determined in the presence of 1 uM FNZ or clonazepam and represented 2-10% of the total binding. The filters were dried and transferred to vials with scintillation cocktail and the radioactivity was measured with 40% efficiency. Data are expressed as diazepam equivalents (DZ E) based on extrapolation from displacement curves generated using DZ. A DZ E 1044
Vol.
175,
No.
BIOCHEMICAL
3, 1991
is the amount MAb 21-7F9.
of
DZ which
AND
causes
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
50% displacement
of
sH-FNZ
from
RESULTS Partially
purified
binding
to
stances
active out
Plasma
7.9
ng
specific was
content
DZ
of
E/mL.
Human milk
No
also
present
+ 0.54
eluate
ng DZ E/mL,
order
to
from
materials (Fig.
were
these
Cm
eH-FNZ
from
detected
showed
shown)
and displacing
branes
and to
by HPLCs
MAb 21-7F9.
8
The
to sexes
+ 0.36
ng
as revealed by these compounds
of
separate
a
mean f
SE
and
C4 columns 2, sH-FNZ
retention
of
human further In the
BZD-like
radioimmunoassays peaks
c). to of
were
ob-
The material that
binding time
in
with
single
b and
similar
of
present extracted and and Methods. regions
radioreceptor
spectra
capacity
= 2.2
materials
4 and 5 (Fig. absorption
between
women
with
donor were in Materials
on
0.3
1).
active
five
BZDexam-
samples between
molecules,
The milk concentration 7.3 ng DZ E/mL,
column
fractions
peaks
BZD-like
the
plasma
observed vs
submethod,
MAb 21-7F9.
ranged
was n = 9,
from one described
On further
the
molecules
n = 12 (Fig.
identify
the
2 a).
tained
displace
using all
difference
to
milk, 930 mL of milk purified by HPLC, as
in
BZD-like
contains
the radioimmunoassays. ranged between 0.4 In
plasma
radioimmunoassays
(men = 2.54 f 0.74 ng DZ E/mL, DZ E/mL, n = 12; Fig. 1).
1.91
human
synaptosomal membranes. Since different on the BZD receptor are detected with this
immunoreactivity
ined.
of
crude
we carried like
extracts
of
to the
in
DZ
brain
(not mem-
material
0
A
0 0
----;---i
----;----
0 ”
WOMEN
MEN
PLASMA
MILK
Firmre 1. BZD-like activity in 9 male and 12 female plasma samples and in 12 samples of human milk. Value8 are expressed as DZ E, in ng per mL. Average values f SE (ng) are 2.54 + 0.74 for male plasma, 2.20 + 0.35 for female plasma and 1.91 + 0.54 for milk samples.
1045
from
Vol.
175,
No.
3, 1991
BIOCHEMICAL
b
AND
DZ
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
c
HFLC of milk extract. In a) the material 2. eluted from the Sep Fak cartriges (see Materials and Methods) was applied to a Vydac Cle column (1 cm x 25 cm, 5 pm). equilibrated-with 6 % ACN in water and developed with the gradient of ACN shown. Solvent B is 80% ACN in water. The flow-rate was 2 mL/min. The bars indicate fractions active in the radioreceptor and radioimmunoassays. Fractions 4 and 5 were rechromatographed in a Vydac Cd column (0.46 cm x 25 cm, 5 pm); b) shows the result obtained with fraction 4. The column was developed ieocratically with 20% methanol, 10% ACN in water (solvent A), at a flow-rate of 1 mL/min. c) ie the chromatogram obtained with fraction 5 when the column was developed isocratically for 20 min with Solvent A and then with a 0%-30% linear gradient of solvent B. The arrows indicate the retention time of DZ under the stated conditions.
-
fraction
5
was identical
chromatographic The analysis (Fig.
3).
Although
The
to
systems
(Fig.
of
mass
the BZD present
that 2,
shown by
standard
DZ in
a and b).
spectra in Fraction
identified 4 could
fraction 5 as not be identified.
WV absorption and mass spectra were similar it had different retention characteristics given by DZ, reversed phase Cls (Fig. 2 a) and C4 columns (Fig. 2 b). more, tern
its
its
mass spectrum
both
showed fragments
not
present
DZ
to those on the Further-
in the
of DZ (Fig. 3) or in any of the known BZDs examined. The active material that eluted as a single peak at 44 min
patin
the first chromatographic step (see Fig. 2 a, fraction 3) was found to be separated into two components on the C4 column eluted isocratically with methanol shown). 20%, ACN 10% in water (not One of them was an inactive compound, with a retention time of 9 min that a highly
liberated hydrophobic
amino acids component
by acid with
1046
hydrolysis. a
retention
The other time
of
35
was min
Vol.
175,
No.
3, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
256 100
FRACTION
t
5
T
22
263
100
FRACTION
4
t
I
T
222
200
250 MASS
Fiaure 3. Mass spectrometric identification of fraction 5 (see Fig. 2, a) with DZ. The mass analysis of fraction 4 (see Fig. 2,a and b) indicates the close similarity of its fragmentation pattern to that of DZ. that
was identified
tographic
behavior
In order
(not
to
purification from
detect
HPLCs
any
contamination
a solvent of
to MAb 21-7F9
and by its
chroma-
shown).
procedure,
eluates binding
as DZ by mass spectrometry
such
could
blank
blanks,
occurring was
no
during
processed.
the
In
displacement
of
is
from
the
3H-FNZ
be detected.
DISCUSSION The main finding
in the
present
women contains
several
BZDs. The major
free
man milk
was identified
zodiazepine time
monoclonal
in two
analysis. Aa laboratory
different
the the
as DZ by its antibody, HPLC
contamination
the recently delivered vious ten years, our pollution by synthetic
study
mothers
high
its systems,
milk
BZD component
affinity
uv spectrum, and
was rigorously had not
that
taken
to its
by mass excluded
of
BZDhu-
an antibenretention epectrometry and
any BZD in the
since pre-
data strongly support that environmental BZDs can be virtually ruled out. Moreover, in human milk does not correlate with variety of BZDs found prevalence of marketed BZDs, further suggesting that the in1047
Vol.
175, No. 3, 1991
BIOCHEMICAL
AND BIOPHYSICAL RESEARCH COMMUNICATIONS
dustrial synthesis of these compounds is unlikely to contribute Also due to the pharmacoto their occurrence in human milk. kinetics and metabolism of DZ, recent therapeutic intake of this BZD would have produced detectable amounts of metabolites such as n-desmethyl-diazepam, temazepam and oxazepam. The concentration of DZ observed in human milk is similar to that reported previously for cow milk (5); also, similar amounts of DZ and its metabolites were found to be present in brains of drug free rats (2, 4, 5). Since it has been reported that oral therapeutic doses of DZ may produce brain concentrations between 10-s and lo-7M (3 to 30 (22), and because the chronic intake of small w/g wet weight) amounts of the BZDs present in human milk might lead to their acthe pharmacological relevance of these cumulation in the brain, invesnatural BZDs from human milk is a possibility that merits tigation. It is important to stress that for a baby suckling 250 mL of milk daily, the total amount of ingested BZDs, per day, would be equivalent to 500 ng DZ. Fraction 3 contains DZ probably non-covalently bound to a protein carrier molecule as shown by its release when this that material is submitted to rechromatography. It is well known serum albumin binds BZDs. The amount of BZD-like molecules present in human milk of drug free women is similar to that observed in seven plasma samples from the same volunteers. Since it has been reported that milk: plasma ratio is about 0.2 when the subjects receive a dose of 10 mg DZ daily (21), our results are not consistent with the assumption that a possible intake of DZ is responsible for the values found. In discussing possible explanations for the occurrence of BZDs in human milk one can consider two main hypotheses: 1) BZDs are the result of biosynthesis in animal tissues; 2) BZDs in human milk are biosynthesized by microorganisms and/or plants and hence they become constituents of our diet. This latter hypothesis is supported by recent studies showing the presence of BZDs in grains, vegetables, plants, flowers and cow milk (5,13-15). In addition, it is known that cyclopeptine, a BZD alkaloid, is formed from L-phenyl-alanine, anthranilic acid and L-methionine by Penicilium cyclopium (23). Also, anthramycin, a bacterial product of Streptomyces refuineus, contains a BZD structure (24). However, up to now, the interaction of a micro-
Vol.
175,
No.
3, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
bial BZD with the central type BZD receptor has not been described. egg white, chicken and hen Based on the presence of BZDa in that BZDa appear very early (15), it has been recently claimed during ontogenesis suggesting that during maturation the organism is probably exposed to these compounds. The hypothesis that BZDs are bioaynthesized by animal tissues is supported by the fact that BZD-like molecules were detected in the neuroblaatoma X glioma hybrid cell line NG 108-15 even after these cells were grown for 3 months in a serum free medium (3). it has been argued that mammalian tissues may not be Although synthesize chlorine containing molecules like BZDs (2, able to of cyclic molecules has 16) it is noteworthy that chlorination been demonstrated in human cells (26). Furthermore, we have found that brain homogenates incubated in a medium supplemented with aminoacids show a 5 fold increase in the content of BZD-like comcontrol non-incubated samples (unpounds as compared with published observations). m: This work was supported by grants from the National Research Council of Argentina, the University of Buenos Aires, The Antorchaa Foundation and the International Foundation for Science. We are indebted to M. Rubio and S. Wikinaki for the provision of some of the samples analyzed, to Dr. A.L. De Blas for providing the MAb 21-7F9 and to Ma. M. Levi de Stein for her skillful technical assistance. REFERENCES
k 3. 4. 5. 6. 7. 8. 9. 10. 11.
Sternbach, L.H. (1978) Prog. Drug. Res. 22, 229-266. Sangameswaran, L., Falea, H.M., Friedrich, P. and De Bias, A. (1986). Proc. Natl. Acad. Sci. USA, 83, 9236-9240. De Blaa, A.L. and Sotelo, C. (1987). Brain Rea. 413, 285-296. Wildmann, J., Mohler, H. and Vetter, W., Ranalder, U., Schmidt, K. and Maurer, R. (1987). Neural Tranam. 70, 383388. Medina, J-H., Pena, C., Piva, M., Paladini, A.C. and De Robertis, E. (1988). Biochem. Biophys. Rea. Comm. 152, 534539. Unaeld, E. and Klotz, U. (1989). Pharmaceut. Rea. 6, 1-3. Unaeld, E., Fischer, C., Rothemund, E., Klotz, U. (1990). Biochem. Pharmacol. 39, 210-212. Manning, R-W., Callahan, A.M., Paik, Y-K., Hayman, A., Davis, L-G. and Morris, H.R. (1986). In: Receptor binding in Drug Research (O'Brien, R-A. Ed) PP. 393-406. Dekker, New York. Wildmann, J., Niemann, J., Matthai, H. (1986). J. Neural. Tranam. 66, 151-160. Medina, J-H., Pefia, C., Piva, M., Rubio, M-C., Wikinski, S. and Paladini, A. (1990). Lancet, 336, 1379. Olasma, M., Guidotti, A., Costa, E., Rothstein, J-D., Goldman, M-E., Weber, R.J. and Paul, S.M. (1989). Lancet A, 491492 _ 1049
Vol.
12. 13. 14. 15. 16. 17. ,18. 19. 20. 21. 22. 23. 24. 25. 26.
175,
No.
3, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Wildmann, J. and Ranalder, U. (1988). Life Sci. 43, 12571260. Maurer, Wildmann, J., Vetter, W., Ranalder, U., Schmidt, K., R. and Mohler, H. (1988). Biochem. Pharm. 37, 3549-3559. Medina, J.H., Pena, C., Levi de Stein, M., Wolfman, C. and Paladini, A.C. (1989). Biochem. Biophys. Res. Comm. 165, 547553. Unseld, E., Krishna, D-R., Fischer, C., Klotz, U. (1989). Biochem. Pharmacol. 38, 2473-2478. De Robertis, E., PeAa, C., Paladini, A.C. and Medina, J.H. (1988). Neurochem. Int. 13, l-11. Denker, S.J. and Johansson, G. (1990). Lancet 335, 413. Medina, J.H. and De Robertis, E. (1985). J. Neurochem., 44, 1340-1345. Lowry, O-H., Rosebrough, N.T., Farr, A-L., and Randall, R.J. (1951). J. Biol. Chem. 193, 265-275. De Bias, A., Sangameswaran, L., Haney, S.A., Park, D., Abraham, C.J. and Rayner, C.A. (1985). J. Neurochem. 45, 1748-1753. Dusci, I-J., Goods, S-M., Hall, R-W. and Ilett, K.F. (1990). Brit. J. Clin. Pharmacol. 29, 123-126. (1985). Haefely, W. T Kybura, E., Gerecke, M. and Mohler, H. Adv. Drug. Res. 14, 165-322. Gerlach, M., Schwelle, N., Lerbs, W. and Luckner, M. (1985). Phytochemistry, 24, 1935-1939. Leimgruber, W., Batcho, A-D. and Schenfer, F, (1965). J. A. Chem. Sot. 87, 5793-5795. De Blas, A-L., Park, D. and Friederich, P. (1987). Brain Res. 413, 275-284. StelmasZgliczynski, J.M., Olszowska, E., Olszowski, S., Int. J. Biochem. 17, zynska, T. and Kwasnowska, E. (1985). 393-397.
1050