10.1192/bjp.158.1.59Access the most recent version at doi:
1991 158: 59-63 The British Journal of Psychiatry
CB Nemeroff, G Bissette, H Akil and M Fink
Corticotrophin- releasing factor, beta-endorphin and somatostatin
depressed patients treated with electroconvulsive therapy.
Neuropeptide concentrations in the cerebrospinal fluid of
Article cited in:
To obtain reprints or permission to reproduce material from this paper, please write
to this article at
You can respond
top right corner of the article or
Receive free email alerts when new articles cite this article - sign up in the box at the
The Royal College of Psychiatrists Published by
on March 8, 2011 bjp.rcpsych.org
go to: The British Journal of PsychiatryTo subscribe to
British Journal of Psychiatry (1991),158, 59â€”63
In research on the physiological
(CNS) and their putative role in the pathophysiology
of affective disordersand schizophrenia,
factor (CRF), the 41-amino-acid hypothalamic peptide
that regulates the secretion of adrenocorticotrophin
(ACTH) and other pro-opiomelanocortin-derived
Bissette, 1986). Wheninjected
CNS, CRF decreases food consumption and sexual
locomotoractivity (Koob & Thatcher-Britton,
In conjunction with the hypercortisolaemia commonly
observed in depressed patients, these findings suggest
that CRF is hypersecretedin depressed patients (Gold
et a!, 1984; Nemeroff,1988). This hypothesis
supported by findings of: elevated CSF concentrations
comparedwith controls (Nemeroff eta!,
et a!, 1987; France et a!, 1988; Arato et a!, 1988); a
blunted ACFH response to intravenously administered
CRF (Gold et a!, 1984; Holsboer et a!, 1984); and
a decreased CRF receptor number in the frontal
cortex of suicide victims (Nemeroff
Somatostatinis a tetradecapeptide
inhibit the secretion of growth hormone
other anterior pituitary hormones. Somatostatin and
concentrations of somatostatin
free depressed patients (Rubmow eta!, 1983; Bissette
et a!, 1986; Davis et a!, 1988).
While there is no proof that $-endorphin
play a direct role in psychiatric
are knownto be responsive
participate in the expression
role of neuropeptide
nervous in the centralsystem
et a!, 1983), and alters
et a!, 1988a).
are reduced in drug
1986; Berger & Nemeroff, 1987).
We sought to determine whether clinical improve
ment in depressed patients
vulsive therapy (ECT) is associated
in CSF concentrations of CRF,
(Akil et a!, 1986; Nemeroff& Bissette,
Nine patients (two men, seven women) fulfilling DSMâ€”III
University Hospital, Stony Brook. The 30-bed in-patient
psychiatric unit is part of a large, suburban teaching
hospital complex, which serves as the main tertiary-care
centre for a population of two million. The in-patient service
is a main teaching unit of a psychiatric department, and
the study of new treatments is accepted as a departmental
mission: it is not unusual for more than half the in-patients
to be part of ongoing researchprotocols. In thisenvironment
we have come to know many affectively ill patients with
recurrent illness who view our professionals
psychiatrictherapists.It isfromthiscadrethat thepatients
for the present studies were selected.
The decision to use ECT is made by the treatment team
and is based on the severity of the illness, the patient's
treatment history, tolerance and response to medications
and ECT, and medical status, as well as the wishes of the
patient and the family. The ECT unit is well known, and
the patients see staff members as both research scientists
and clinicians. The ECT consent process follows the
guidelines of the American Psychiatric Association (1978).
Thebenefitsand risksof the treatmentare discussedwith
the patient and close family members, and they viewa
specific researchfor whichthe patient
The written signedconsent describesthe indications for
ECT,its risks and benefits, the experimental
(miiltifr@nd piptrnpnrpnhnlnaranhv nnnrnnndncrinn teats
1980) were studied
as their primary
Depressed Patients Treated with Electroconvulsive
Concentrations in the CerebrospinalFluid of
CHARLES B. NEMEROFF, GARTH BISSETTE, HUDA AKIL and MAX FINK
TheCSFconcentrationsof CRF,somatostatin andj3-endorphinweredeterminedin ninepatients
who fulfilled DSMâ€”lll criteria for major depressionwith psychotic features. CSFsampleswere
obtained at baseline in the depressed state, and again after a course of ECT.Concentrations
of both CRFand @5-endorphin decreased after ECT, while the concentration of somatostatin
increased, although the latter difference did not attain statistical significance. The increase
in CSF concentrations of CRF and @9-endorphin in depressed patients is therefore seen to be
60 NEMEROFF ET AL
CSF measures, neuropsychologicaltests), and detailed
statements of the patient's
obtained from both patientand close family members.
For thesestudies,consentwas obtained for two CSF
samples, one before treatment, and one when the patient
had improved. The consent process, both for ECT and for
the various experimental studies, was approved by the
InstitutionalReview Boards of the Health Sciences Center
and University Hospital, and by the University Committee
on Research in Human subjects.
Once consent was obtained,
were done. These included medical, neurological, dental,
and anaesthesia assessments; and complete blood count,
performed about four days after admission in all patients
admitted to our unit, and repeated in the four days before
the first ECT if there has been an intervening period longer
than on@week. Cortisol concentrations in plasma were
measured from blood samples obtained at 8.00 a.m., 4.00
p.m., and 11.00 p.m. on the day after dexamethasone
administration (1 mg at 11.00p.m.).
commercial kits obtained from Diagnostic Laboratories Inc.
Based on our experience with patients at University
Hospital, we selected a cut-off point of 5.5 @g/d1 as the
upper limit for â€˜¿?normal'values in the post-dexamethasone
assays,because this placed90Â°!. of
patients in the normal DST group (Fink et a!, 1987).
The Hamilton Rating Scale for Depression (HRSD;
Hamilton, 1967) was completed at the time of the DST.
Diagnoses were based on DSMâ€”IIIcriteria and the
patientswere classifiedas major
(296.34 or 236.54). All psychoactive
tapered on admission or when ECT was being considered.
Medications had generally been suspended for four days
at the time of first ECT. Benzodiazepines are routinely
suspended on admission for affectively illpatients, so that
the patients were free of such drugs for at least a week
before treatment began. Assessments of mood, behaviour,
weekly throughout the course of treatment, on a day
The CSF was obtained by lumbar puncture, with patients
and 7.30 a.m. with patient activity limited to toiletting. A
22-gaugeneedlewasused and all sampleswereobtained
by MF. Approximately 15ml of CSF was obtained, and
aliquots from the second and third tubes (6â€”15 ml) were
used in the analyses. The first CSF sample was obtained
from one to four days before the first ECT treatment.
Repeat CSF examinations weredone one (sixpatients), two,
three, and four days (one patient each) after an ECT
treatment. Thetime of sampling
change in mood and behaviour, and the second sample was
obtained after ECT sessions6and 10(7 patients), 12and
14(one patient each). After the second CSF sampling, four
patients had no further ECT; two patients had one, two
patients three, and one patient five additional treatments.
right to refuse tests or to
at any time.Consent is
the pre-treatment tests
test (DST) is usually
wasdetermined by the
Theseweredeterminedbythepatient's degreeof improvement
and our estimate of the need for continuing treatment to
maintain it. Thusthe meannumber
8.4 (s.d. 0.95, range 6â€”14).
Patients received modified ECT with glycopyrrolate
(0.2 mg), methohexital (0.5 mg/kg),
(0.5 mg/kg), and assisted ventilation with oxygen; electro
cardiogram, blood pressure,and blood oxygen saturation
werecontinuously monitored. Seizureduration ismonitored
by cuff and EEG methods (Gujavarty et a!, 1987).Seizures
areconsideredadequatewhen the duration by cuff monitoring
was25secondsor longer, and the EEG duration at least
equal (Fink & Johnson, 1982).Treatments were begun with
placementif the response was not considered
six or seven treatments (Abrams
used with intensitiesselected as slightly above threshold.
Treatmentswere given three times weekly.
$-endorphin were measured
radioimmunoassay procedures. The concentration of CRF
was measured, as previously described in detail (Nemeroff
et a!, 1984; Chappell et a!, 1986; Banki et a!, 1987)
using an antiserum to ovine CRF generously provided by
Dr WylieVale(SalkInstitute, LaJolla, CA).Themethod
has a sensitivity of 2.5 pg/tube, an IC@(mid-point) of
40 pg. and does not cross-react
endogenous peptides. The concentration of somatostatin
was measured as described in detail by Bissette et a! (1986)
using an antiserum also provided by Vale; it has a sensitivity
of 2.5 pg/tube,an lCn of 100pg, and shows cross
reactivity only with somatostatin
Concentrationsof @-endorphinwere determined by a
previously well characterised
a!, 1983)which recognises j3-endorphinand fl-lipotrophin
equimolarly.CSF (1.2 ml total) was extracted using Sep Pak
cartridges as detailed in Cahill et al(l983). The eluate was
dried down and resuspended in a radioimmunoassay buffer,
allowing the equivalent of 0.6 ml of CSF to be added to
each assay tube. Each sample was assayed in duplicate.
ArtificialCSFwas usedto verify
insensitive to salts in the CSF. The midpoint of the assay
is 15â€”20 fmol/tube or the equivalent of 50â€”70 pg of (3-
endorphin/tube.It can reliably detect 3â€”5 pg/tube of (3-
endorphin or its molar equivalent in (3-lipotrophin.
be noted that there is very little j3-lipotrophin-sized
immunoreactivity in mammalian brain or in CSF, and the
values read represent over 90Â°lo (3-endorphin-sizedmaterial.
Statistically significant differences were sought between
CSF peptide concentrationsat baseline and after ECT using
a paired Student's i-test.
of ECT sessionswas
& Fink, 1984). Brief-pulse
D or Thymatron)were
thatthe assay was
The HRSD scores and the post-dexamethasone
cortisol concentrations for each
score for the nine depressedpatients
nine patients studied, eight were DST non-suppressors.
patientsof the depressed
was 25.1 (2.6).Of the
include the DST, the blunted
response to intravenously administered thyrotrophin
concentrations of monoamine metabolites (Asberg
et a!, 1976; Loosen & Prange,
1981; Gold et a!, 1984;Holsboer et a!, 1984;Reynolds
& Shipley,1985; Doyleeta!,1985; Langer eta!,1986;
Nemeroff et a!, 1988b). Whether these biochemical
or are state-dependent
resolving this issue is the confounding
may be unrelated to their therapeutic actions. The
study of changes induced by ECT is an ideal paradigm
for determining whether one or another putative
marker,reflecting a patient's
is ideal because of its documented
because there are no confounding pharmacological
effects of drugs (Fink,
It is important
studying in the present report is not the acute effects
of ECT on CSF neuropeptide
first CSF sample was obtained before any ECT. The
second was obtained at least 24 hours after the last
ECT session. The changes observed are thought to
be associated with the changes in behaviour induced
by the treatment, and not an immediate post-seizure
effect. A parallel is seen in studies of plasma
prolactinafter seizures. There is an immediate
seizure, and a return
With successive treatments,
prolactin is obtunded, but the baseline level remains
the same when measured 24 hours after a seizure.
The alterations in CSF measures were in expected
evidenced by increased CSF concentrations.
improvementwith ECT is associated with reductions
in CRF and $-endorphin
believe that the concentrations
statistically insignificant,increasein CSFsomatostatin
NEUROPEPTIDECONCENTRATIONSIN CSF BEFORE AND AFTER ECT
in CSF somatostatin levels after treatment.
Five of eight patients exhibitedan increase
1980; Carroll et a!,
alterations represent trait markers
is unclear. An impediment to
is a trait marker,
or conversely is a state
to emphasise that what we are
with a peak20â€”30minutes
in two to four hours.
patients, of CRFas
of these two peptides
in CSF. We
Depressionratings and post-DSTcortisol concentrations
The CSF neuropeptide data, expressed as pg/ml, are
presented in Fig. 1.The mean (s.c.m.) CSF concentration
of CRFin the depressed patients
(8.8) pg/mI. This concentration isclearlyelevatedcompared
with normal controls studied in our previous investigations,
who had groupmean CSF CRF concentrations
40-60 pg/nil. After a course of ECT, there was a reduction
in mean CSF CRF concentrations
Six of nine patients exhibited a reduction in CSF CRF
concentrations after ECT.
(P<0.025, Fig. 1).
corticotrophin-releasing factor (CRF), @-endorphin and somatostatin
before (0) and after (@) a course of ECT. There was a
significant reduction in the CSF concentrations of CRF (P<0.025)
and fi endorphin (P<0.05).
CSF concentrations(pg/mI) (errorbars show s.c.m.) of
Similar results were obtained with (3-endorphin (Fig. 1).
Pre-treatmentCSF (3-endorphin concentrations
(s.c.m.)) in the depressed patients was 39.9 (5.8) pg/mI.
After ECT, CSF (3-endorphin concentrations were reduced
to 21.5 (8.7) pg/mI (P<0.05). Of the eight patients studied
(one sample was lost for (3-endorphinmeasurement), seven
patients exhibited a fall in CSF (3-endorphinconcentrations
The CSF somatostatin concentrations before and after
ECT exhibited a different pattern to that of CRF and
(3-endorphin (Fig. 1). After ECT there was a small,
Several groups have reported that depressed patients
have low CSF concentrations of this tetradecapeptide
(Rubinow et a!, 1983; Bissette et a!, 1986; Davis
et a!, 1988). AfterECT,
insignificant increase in somatostatin
in CSF. Of particular interest,
that the CSF somatostatin
show a patternsimilarto CRF and (3-endorphin,
rendering specificity to the findings observed.
Is the covariation between CRF and (3-endorphin
in the CSF due to a common regulator which controls
both peptides, is it caused by a direct interaction
between these two substances, or is it simply the
reflection of unrelated consequences of ECT? Basic
studies have begun to shed light on this issue. Both
neuropeptide systems are intimately
responsiveness to stress, not only at an endocrine
level, but also within the brain. In addition, results
on the relase of fi-endorphin from hypothalamic
slices show that CRF is a potent
13-endorphin(Young et a!, 1989). Thus,
elevation of CRF in depressed subjects may be
associated witha concomitant
endorphin release. Conversely, ECT treatment may
lower j3-endorphin levels through
mechanisms, including the decrease in CRF being
liberated. It is likelythat numerous other neurotrans
Do the changes in CSF peptides represent a non
specific stress response or a differential response of
different tissues to the stimulus or the seizure? While
many authors have argued that the release of peptides
into CSF and blood is non-specific,
Deakin eta! (1983), Whalley eta! (1982) and others,
cited by Fink (1987), argue for a differential
response. The findings in the present study support
the view that the release of peptides following
induced seizures is a specific tissue response.
Do these observations contribute to our understand
ing of the therapeutic chain in ECT? There are two
prevailing hypotheses of the antidepressant
of ECT (Fink,1979, 1984). The neurohumoral
receptor hypothesis accepts the changes reported by
pharmacologists as the basis for the antidepressant
efficacy of antidepressant drugs as the same for the
ECT process; that is, the increase in uptake and in
receptor activity following induced seizures reversing
a reduction in neurohumoral activity seen as the basis
for the altered mood (Grahame-Smith et a!, 1978;
An alternative view argues that altered mood states
are associated with a change in brain peptide levels,
of depression, particularly
findings are also of interest.
there was a small but
is the fact
releaser of brain
the reports by
and that ECT alters these by direct stimulation
of brain cells and by altering acetylcholine and
norepinephrine brain concentrations
enhance specific brain tissue responses (Fink, 1979;
Fink & Ottosson, 1980). The present observations
find differential effects of seizures on three brain
peptides to identify those which may directly affect
mood and behaviour.
furtherstudy of brain
Thestudy wassupported byNIMH MH-42088(CBN),MH-4225l
(HA),NIDADA-02265 (HA)andtheInternational Association for
ABRAMS,R. & FINK, M. (1984) The present
ECT: some recommendations. Journal of Affective Disorders,
AKIL, H.,YOUNG, E., WALKER, J. M., et al (1986) The many
possiblerolesof opioidsand relatedpeptidesin stress-induced
analgesia. Annals of the New York Academy of Sciences, 467,
Therapy (Task Force Report No.
(1980)Diagnosticand StatisticalManual of Mental
Disorders (3rd edn) (DSM-III).
ARATO, M., BANKI, C.M..BissErrE,
cerebrospinalfluid concentrationsof corticotrophin-releasing
ASBERG, M.,THOREN, P.,TRASKEN, L.,
depressionâ€”¿?a biochemical subgroup
BANJU, C. M., BISSETTE,0., Aizj@ro, M.,cia!(1987)Cerebrospinal
fluid corticotrophin-releasing factor-likeimmunoreactivityin
BERGER,P. A.&NEMEROFF, C.
in affective disorders. In Psychopharmacology, The Third
Generationof Progress(ed. H. Y. Meitzer). New York: Raven
Bisss-rrs, 0., WIDERLOV,E.,WALLEUS,
GeneralPsychiatry, 43, 1148â€”1154.
CAHILL, C. A.,MATrREWS, .1. D. & AKIL, H. (1983) Human
extraction technique and validation of radioimmunoassay.
Journal of Clinical Endocrinology and Metabolism, 56,
specific laboratory test for the diagnosis of melancholia.
of General Psychiatry, 38,15-22.
CHAPPELL, P.B.,SMITH, M. A.,
Alterations in corticotrophin-releasingfactor-ike immunoreactivity
Journal of Neuroscience,6, 2908â€”2914.
DAVIS,K. L., DAVIDSON,M., YANG,
somatostatin in Alzheimer's disease, depressed patients, and
Washington, DC: APA.
0. cial (1988)Elevated
et al (1976)
OREDEN,J. F.,ci al (1981)
KILTS, C. D.,ci al(1986)
R.-K.,ci al(1988) CSF
NEUROPEPTIDECONCENTRATIONSIN CSF BEFORE AND AFTERECT
DEAKIN, J. F., FERRIER, I. N.,
ECT on pituitary hormone release: relationship to seizure,
clinicalvariables andoutcome. BritishJournalof Psychiatry,
DOYLE, M. C., GEORGE, A. 3., RAVINDRAN,A. V., ci al (1985)
Platelet a2-adrenoreceptor binding in elderly depressed patients.
AmericanJournalof Psychiatry,142, 1489-1490.
FINK, M. (1979)Convulsive Therapy:
(1984)Theories of the antidepressant efficacy of convulsive
therapy(EC'r).InNeurobiology ofManicDepresive Illness(eds
R. M. Post&J. C. Ballenger).Baltimore:Williams&Wilkins.
(1987)Neuroendocrineaspects of convulsivetherapy:
review of recent developments.
Psychoneuroendocnnology(eds C. B. Neineroff& P.T. Loosen).
New York: Guilford Press.
&OrrossoN,J..O. (1980) Atheoryofconvulsive therapyin
PsychiatryResearch, 2, 49â€”61.
CROW, T. 3., ci al(1983)Effectsof
In Handbookof Clinical
of hypothalamic functions.
and EEG methods compared. Archives of General
GUJAVARTY, K. & GREENBERG, L. (1987)
sone suppression tests and clinical outcome in ECT. Convulsive
Therapy, 3. 111-120.
FRANca, R. D., URBAN, B., KRISHNAN, K. R. R., etal (1988)CSF
corticotrophin.releasing factor-ikeimmunoreactivity inchronic
pain patients with and withoutmajor depression.Biological
GOLD, P. W., CHRoUsos. 0., KELLNER,C., ci al(l984)
implicationsof basicand clinicalstudieswithcorticotrophin
GRAHAME-SMITH,D. G., GREEN,
HAMILTON,M. (1967) Development
depressive illness. British Journal of Social and Clinical
Psychology, 6, 278â€”296.
HOLSBOER,F., VONBARDELEBAN, U. V., GERKEN,A., ci al (1984)
Blunted corticotrophin and norma] cortisol response to human
corticotrophin-releasingfactor in depression.New England
Journalof Medicine, 311,1127.
anxiogeniceffects of corticotrophin-releasingfactor. In Endocoids
R. & COSTAIN, D. W.
of a rating scale for primary
(eds H. La], F. LaBella & J. Lane). New York: Alan R.
binding sites in brain and platelets. In Aniidepressants and
Receptor Function (CIBAFoundation Symposium 123).
receptors:implicationsforthemechanismof actionof electro
LOOSEN,P. T. & PRANGE,A. J., Jr (1980) Thyrotrophin-releasing
hormone (TRH): a useful tool for psychoneuroendocrine
MAURER, R. & REUBI, J.-C. (1985)Somatostatin
of theAmerican MedicalAssociation, 253,2741.
NEMEROFF,C. B. (1988) The role of corticotrophin.releasing
in the pathogenesis of major depression.Pharmacopsychiatry,
â€”¿?, WIDERLOV,E.,BI55Erre,G.,ci al (1984)Elevated
reactivity in depressedpatients. Science,226, 1342â€”1344.
of CSFcorticotrophin-releasing factor-like immuno.
â€”¿?& BIssErm,G.(1986) Neuropeptidesinpsychiatricdisorders.
InAmerican Handbook ofPsychiatry, vol.VIII (eds P.A.Berger
&H. K. H. Brodie).NewYork:BasicBooks.
â€”¿?,OwENS,M. J., BIsSETTE, G.,ci al (1988a)Reduced
cortex of suicides. Archives of General Psychiatry, 45, 577-579.
â€”¿? , KNIGHT, D. L.,K sHNAN, K. R.R., etal(1988b) Marked
reductionin the numberof platelet [3HJ.imipraminebinding
sites ingeriatricdepression. Archives ofGeneral Psychiatry,
REYNOLDS,C. F., III & SHIPLEY,J. E. (1985) Sleep in depressive
disorders. In Psychiatry Update, Vol.4. Washington, DC: APA.
RUBINOW,D. R.,Gou, P. W.,
somatostatin inaffective illness. Archives of General Psychiatry,
SIRINATHSINGHJI, D. J. S., REES, L. H.,
Corticotrophin-releasing factoris a potent inhibitorof sexual
receptivity in the female rat. Nature, 305, 232â€”235.
YOUNG, E. A., BRONSTEIN, D. M. & AKIL, H. (1989)
of ft-endorphin release from rat hypothalamus in vitro. Society
for NeuroscienceAbstract, 15, 185.
WHALLEY, L. J., ROSIE, R., DICK, H., ci a! (1982) Immediate
increasesin plasmaprolactinand neurophysin
hormones after electroconvulsive therapy. Lancet, ii, 1064-1068.
PosT, R.M.,ci al (1983)CSF
RIVIER, J., ci al (1983)
but not other
Department of Psychiatry, Duke University Medica! Center, Durham, North Carolina 27710; Huda Akil,
PhD, Menta! HeafthResearchInstitute, Department
Medicine, Ann Arbor, Michigan 48109; Max Fink, MD, Department of Psychiatry, State University of
New York at Stony Brook, Stony Brook, New York 11780, USA
B. Nemeroff,MD, PhD, Departmentof Psychiatry and Pharmacology;Garth Bissette,PhD,
of Psychiatry,University of MichiganSchoo!of
Correspondence:Box3859,Duke UniversityMedicalCenter, Durham, NC 27710, USA