www.elsevier.com / locate / bres
Research report
Influence of corticotrophin releasing factor on neuronal cell death in
vitro and in vivo
a ,
*
a a aMark W. Craighead
, Herve Boutin , Kelly M.L. Middlehurst , Stuart M. Allan ,
b b a
Nigel Brooks , Ian Kimber , Nancy J. Rothwell
a
School of Biological Sciences, 1.124 Stopford Building, University of Manchester, Oxford Road, Manchester M13 9PT, UK b
Zeneca Central Toxicology Laboratory, Alderley Park, Macclesfield, Cheshire SK10 4TJ, UK Accepted 25 July 2000
Abstract
Several studies have demonstrated that antagonists of the corticotrophin releasing factor (CRF) receptor markedly inhibit experimental-ly induced excitotoxic, ischaemic and traumatic brain injury in the rat, and that CRF expression is elevated in response to experimentalexperimental-ly induced stroke or traumatic brain injury. CRF is also induced by the pro-inflammatory cytokine interleukin 1 (IL-1), which participates in various forms of neurodegeneration. The aim of this study was to test the hypothesis that CRF is toxic directly in vivo or in vitro. In primary cultures of rat cortical neurons, exposure to CRF (10 pM–100 nM) for 24 h failed to cause cell death directly, or to modify the neurotoxic effects of N-methyl-D-aspartate (NMDA). Similarly, infusion of CRF (0.3–5mg) into specific brain regions of the rat did not
induce cell death and did not significantly alter the neuronal damage produced by infusion of excitatory amino acids. These data demonstrate that CRF is not directly neurotoxic, and suggest that either CRF mediates neuronal damage by indirect actions (e.g. on the vasculature) and / or that CRF is not the endogenous ligand which contributes to neurodegeneration through activation of CRF receptors.
2000 Elsevier Science B.V. All rights reserved.
Theme: Disorders of the nervous system
Topic: Neurotoxicity
Keywords: CRF; Neuronal toxicity; Cortical neurons; Ischaemia; NMDA; AMPA
1. Introduction cortical brain injury or focal cerebral ischaemia in the rat
[13,14,17]. These data suggest that endogenous CRF
Corticotrophin releasing factor (CRF) is a key mediator causes, or contributes to, several forms of neuronal cell
of the stress response, and induces release of ACTH from death, though its mechanisms of action are not known.
the pituitary [19], but also has a number of other direct The pro-inflammatory cytokine interleukin 1 (IL-1) also
actions on the brain [1,3,5]. mediates diverse forms of acute neurodegeneration, and
Several studies have implicated CRF in acute neurode- our recent observations indicate that IL-1 acts at specific
generation. CRF mRNA levels are elevated in the cortex, brain sites (particularly the striatum and hypothalamus) to
amygdala and hypothalamus of rats exposed to traumatic cause cortical injury [2]. IL-1 induces the release of CRF
or ischaemic brain injury [14,20]. Furthermore, central [18], its effects on fever are dependent on CRF [12,15],
administration of CRF receptor antagonists such as a- and CRF mRNA is expressed in several of the sites of IL-1
helical CRF( 9 – 41 )or D-phenyl CRF( 12 – 41 ), attenuates mark- action in injured brain [14,20]. Thus, CRF may mediate
edly neuronal damage caused by infusion of excitotoxins, the neurodegeneration induced by IL-1.
The primary aim of these investigations was to test the hypothesis that CRF causes neuronal death directly in vivo
*Corresponding author. Tel.:144-161-275-5545; fax: 1
44-161-275-or in primary c44-161-275-ortical neurons in vitro, and to determine
5948.
E-mail address: [email protected] (M.W. Craighead). whether it exacerbates excitotoxic neuronal damage, at
specific brain sites where the pro-inflammatory cytokine ferred to nitrocellulose and probed with antibodies which
IL-1 acts to enhance neuronal damage [2]. recognised either total ERK (1:30,000, Santa Cruz
Bio-technology, USA) or the phosphorylated form of the protein (1:5000, New England Biolabs, USA) and
de-2. Materials and methods veloped using the enhanced chemiluminescence kit (ECL) from Amersham.
2.1. Cell culture
2.3. AMPA and CRF injections Cortical neurons were prepared from 18 day gestation,
Sprague–Dawley rat pups (Charles River, UK). Isolated Male, Sprague–Dawley rats (body weight, 28467 g,
cortices had the meninges dissected and were resuspended mean6S.E.M., Charles River UK) were maintained in
in 4 ml Neurobasal medium (GibCo, UK), dissociated by thermoregulated, humidity controlled facilities under a 12
trituration through glass-fired pipettes, then centrifuged at h / 12 h light / dark (light on between 08.00–20.00) cycle
1000 g for 5 min, and resuspended in plating medium and allowed free access to food and water. Anaesthesia
(NBM supplemented with 100 U / ml penicillin, 100mg / ml was induced by inhalation of halothane (5% in oxygen),
streptomycin, 2 mM glutamine, B27 supplement with anti- and thereafter maintained at 1.5% in a mixture of 70%
oxidants (GibCo, UK) and 25 mM glutamate). Cells were NO / 30% O . Vehicle, S-AMPA (15 mM in PBS buffer,2 2
grown on poly-D-lysine-coated, 24-well plates at a density Tocris, UK) or human recombinant CRF (hrCRF, 2.1 mM
4 2
of 8310 cell / cm . Half of the medium was replaced in saline 0.9% / BSA 0.1%) were injected stereotaxically
every 3 to 4 days with fresh maintenance medium (NBM using a 0.5ml or 1ml Hamilton syringe (injection rate: 0.5
supplemented with 100 U / ml penicillin, 100mg / ml strep- ml / min, UltraMicroPump II and Micro4 Controller, WPI
tomycin, 2 mM glutamine and B27 supplement without Inc., USA). Vehicle (0.5 ml) or S-AMPA (0.5 ml) were
anti-oxidants). injected into the right striatum (coordinates:
anterior–pos-terior, 0.7 mm; lateral, 2.7 mm (from Bregma);
dorsoventr-2.2. Drug treatment al, 5.5 mm), followed immediately by injections of hrCRF
(0.5 ml) or vehicle, in the amygdala (anterior–posterior,
Recombinant human / rat CRF was obtained from either 22.6 mm; lateral, 4.8 mm (from Bregma); dorsoventral,
Bachem, UK, or the Salk Institute, USA. Both preparations 7.3 mm), or ventromedial nucleus of the hypothalamus
were physiologically active as demonstrated by their effect (anterior–posterior, 22.6 mm; lateral, 0.5 mm from
on food intake when injected ICV into rats (0.5 ml, 2.1 Bregma; dorsoventral, 9.5 mm).
mM) (data not shown). Animals were maintained normothermic (body
tempera-Mature cortical neurons (cultured for 12 to 14 days) ture: 37.060.48C) during the surgery through the use of a
were exposed in vitro to CRF (10 pM–100 nM) for 24 h in heating blanket (Harvard Apparatus Ltd., UK).
250 ml conditioned medium. For excitotoxicity
experi-ments, cells were pre-incubated with CRF (1 nM or 100 2.4. Histological analysis of damage
nM) for 1 h prior to addition of NMDA (0–100mM). All
experiments were conducted in conditioned medium, as Forty-eight hours after injection, rats were sacrificed by
this reduced the cell death induced by media change. anaesthetic overdose followed by cervical dislocation.
Cytotoxicity was determined by measuring the release of Brains were removed quickly and frozen on isopentane in
the stable cytosolic enzyme lactate dehydrogenase (LDH) dry-ice. Brain damage, delineated by the relative paleness
[7], using the CytoTox-96 kit (Promega, UK), following of histological staining in the damaged tissue, was assessed
the manufacturers instructions and with conditioned on Cresyl Violet stained, coronal sections (20 mm).
medium as a blank. Maximal LDH release (100%) was Volumes of ipsilateral healthy tissue, contralateral
hemi-deduced by the addition of 1% Triton X-100 (final sphere, cortical damage and striatal damage were
calcu-concentration) to control cells. The data were then ex- lated by the integration of areas on each section quantified
pressed as LDH release as a percentage of maximal with a computer-assisted image analyser [9].
release.
For immunoblotting experiments, media was replaced 2.5. Statistical analysis of data
with EBSS, and cells were allowed to recover for 30 min
prior to being treated with CRF (100 nM) for a further 30 All data are expressed as mean6S.E.M. For assessment
min. Cells were then harvested in 75ml of lysis buffer per of neuronal cytotoxicity in vitro, ANOVA was used to
well at 48C (50 mM Tris pH 7.4, 1% TX-100, 50 mM compare effects of several doses of CRF on cell death. For
NaF, 10 mM NaVO , 50 mM4 b-glycerophosphate). in vivo studies the data did not follow a normal
dis-Lysates from three wells were combined. 4 ml of cell tribution, and the variances were unequal between the
extract were then mixed with 16ml of Lammli buffer and groups. Therefore, a Kruskal–Wallis test was performed,
analyses, the significance level was accepted to be P,
0.05.
3. Results
3.1. CRF activates the extracellular signal regulated
kinases ERK1 and ERK2 in primary cortical neurons
In order to verify biological activity of CRF in primary cortical neurons. Cells were exposed to 100 nM CRF for 30 min. Cell extracts were immunoblotted with antibodies which recognised either total ERK or the activated phos-phorylated form of the protein. Fig. 1 is a typical example of the results obtained, CRF caused an increase in the level of phosphorylated ERK whilst the total amount of ERK remained unchanged. The average increase in phosphoryla-tion observed was 45% above control in four separate experiments.
3.2. CRF is not directly neurotoxic in vitro and does not
exacerbate excitotoxicity Fig. 2. Effect of CRF on LDH release (as a measure of cell death) in
primary cortical neurons. Cells were exposed to CRF for 24 h in conditioned medium. LDH release is expressed as a percentage of
Basal cell death in cortical cultures was low (,10% of
maximal LDH release elicited by cell lysis. Values are mean6S.E.M. for
maximal LDH release). Exposure of cortical neurons to
six separate experiments.
CRF (10 pM–100 nM) alone caused no significant in-crease in cell death, as measured by LDH release (Fig. 2).
These concentrations of CRF were chosen because as they nM) for 1 h failed to exacerbate the excitotoxic effects of
cover a range above and below the reported EC50 values NMDA (Fig. 3).
for all known CRF receptor subtypes [8,16].
NMDA caused a dose-dependent increase in release of 3.3. CRF is not neurotoxic in vivo and does not
LDH from neurons, but pre-treatment with CRF (1 or 100 exacerbate excitotoxic damage
Direct injection of CRF into either the amygdala or
hypothalamus (0.5ml, 2.1 mM) failed to induce any gross
Fig. 1. CRF activates ERK in primary cortical neurons. Cells were Fig. 3. Effect of CRF on NMDA-induced cell death. Cortical neurons exposed to CRF (100 nM) for 30 min in EBSS. Cell extracts were were pre-incubated for 1 h with CRF (1 or 100 nM) in conditioned immunoblotted for either total ERK (A) or the active phosphorylated form medium. NMDA was then added directly to the cells (0–100mM). LDH (B). CRF increased the level of active ERK immunoreactivity whilst total release was assayed 24 h later. Values are mean6S.E.M. for four separate
however, a range of concentrations of CRF failed to cause cell death, and did not modify significantly the toxic effect of NMDA (Figs. 2 and 3). Similar results were obtained with cerebellar granule neurons (data not shown). RT-PCR confirmed that cortical neurons did express the mRNA for type 1 CRF receptor (data not shown). Similarly infusions of CRF into either the hypothalamus or amygdala in vivo (Fig. 4), both known sites of action of CRF, caused no cell death and failed to modify the damage caused by striatal injections of the excitotoxin S-AMPA.
These results demonstrate that exogenous CRF is not directly neurotoxic and does not exacerbate cell death caused by excitotoxicity in vivo or in vitro. The present data are unexpected in view of our results, and those of other laboratories, which show that several CRF antago-nists markedly inhibit ischaemic, excitotoxic and traumatic brain injury [13,14,17], suggesting that endogenous CRF directly mediates neuronal cell death.
There are several possible explanations for this apparent inconsistency. The most straightforward is that while CRF is required for the cell death observed during acute neurodegeneration it is not in itself sufficient to cause this toxicity, and requires additional factors which were not present in the conditions reported here. However, this
appears unlikely as the CRF antagonista-helical CRF( 9 – 41 )
inhibits excitotoxicity [17], while exogenous CRF fails to
Fig. 4. Effect of CRF on excitotoxic damage in vivo. CRF (0.5ml, 2.1
exacerbate such damage.
mM) was injected into either the hypothalamus (h) or amygdala (a) with
In vivo endogenous CRF will be released locally and
either saline or S-AMPA (7.5 nmol) injected into the striatum. Values for
lesion volume are corrected for oedema. CRF did not significantly alter may, therefore, act at discrete sites inducing specific
lesion volume when co-injected with AMPA. Values are mean6S.E.M., effects. Such effects would be blocked by the addition of
n57.
exogenous receptor antagonists. In the experiments pre-sented here, CRF was added exogenously and will, there-fore, act at all CRF receptors, potentially evoking conflict-ing effects, which could therefore mask the neurotoxicity
morphological damage (Fig. 4). Similarly, no damage was evoked by activating CRF receptors at specific sites or on
observed when CRF was infused into either the striatum or specific cells.
cortex (data not shown). Another possibility is that CRF and its antagonists have
Injection of the glutamate agonist S-AMPA (7.5 nmol) no direct effect on neuronal viability, but rather exert their
into the striatum of anaesthetised rats induced substantial effect on a separate pathway which leads to the observed
striatal damage [9]. This excitotoxin was chosen because neuroprotection. CRF has been reported to elevate both
we have shown previously that co-administration of IL-1 heart rate and blood pressure by stimulating noradrenergic
exacerbates excitotoxic damage [2]. S-AMPA (7.5 nmol) sympathetic nervous outflow [6] and to cause vasodilation
induced a large striatal lesion (Fig. 4). However, the extent [10]. Recently it has been shown that CRF can
dramatical-of damage was not significantly affected by the co-ad- ly inhibit leukocyte recruitment to endothelia by blocking
ministration of CRF at any site tested. up-regulation of ICAM-1 expression [4] and that this effect
can be inhibited by astressin. Therefore neuroprotective effects of CRF antagonists may be due to an alteration in blood flow and / or recruitment of invading leukocytes to
4. Discussion the area of damage.
In conclusion we have demonstrated that CRF is not
The aim of this study was to investigate whether CRF directly toxic to neurons either in vitro or in vivo and does
causes neuronal cell death directly, in vivo or in vitro, or not exacerbate cell death resulting from exposure to
whether it modifies neurodegeneration caused by other excitatory amino acids. These results, in conjunction with
stimuli. previous reports, indicate that CRF plays a complex role in
In agreement with previously published data [11] CRF acute neurodegeneration and that further investigations will
[9] C.B. Lawrence, S.M. Allan, N.J. Rothwell, Interleukin-1beta and the Acknowledgements
interleukin-1 receptor antagonist act in the striatum to modify excitotoxic brain damage in the rat, Eur. J. Neurosci. 10 (1998)
This study was supported by a Strategic Research Fund 1188–1195.
grant from Astra-Zeneca (MWC) and Medical Research [10] S. Lei, R. Richter, M. Bienert, M.J. Mulvany, Relaxing actions of
Council (SA, HB, KMLM and NJR). The authors would corticotropin-releasing factor on rat resistance arteries, Br. J.
Phar-macol. 108 (1993) 941–947.
like to thank Prof. Jean Rivier at the Salk Institute for his
[11] H. Li, P.J. Robinson, S. Kawashima, J.W. Funder, J.P. Liu,
Differen-kind gift of CRF. We would also like to thank Dr Catherine
tial regulation of MAP kinase activity by corticotropin-releasing
Lawrence for her very generous help with the in vivo hormone in normal and neoplastic corticotropes, Int. J. Biochem.
feeding experiments involving CRF. Cell Biol. 30 (1998) 1389–1401.
[12] J. Lundkvist, Z. Chai, R. Teheranian, H. Hasanvan, T. Bartfai, F. Jenck, U. Widmer, J.L. Moreau, A non peptidic corticotropin releasing factor receptor antagonist attenuates fever and exhibits References
anxiolytic-like activity, Eur. J. Pharmacol. 309 (1996) 195–200. [13] M.K. Lyons, R.E. Anderson, F.B. Meyer, Corticotropin releasing [1] J.B. Aldenhoff, D.L. Gruol, J. Rivier, W. Vale, G.R. Siggins, factor antagonist reduces ischemic hippocampal neuronal injury,
Corticotropin releasing factor decreases postburst hyperpolarizations Brain Res. 545 (1991) 339–342.
and excites hippocampal neurons, Science 221 (1983) 875–877. [14] S.Y. Roe, E.M. McGowan, N.J. Rothwell, Evidence for the in-[2] S.M. Allan, L.C. Parker, B. Collins, R. Davies, G.N. Luheshi, N.J. volvement of corticotrophin-releasing hormone in the pathogenesis
Rothwell, Cortical cell death induced by interleukin-1 is mediated of traumatic brain injury, Eur. J. Neurosci. 10 (1998) 553–559. via actions in the hypothalamus of the rat, Proc. Natl. Acad. Sci. [15] N.J. Rothwell, CRF is involved in the pyrogenic and thermogenic USA 97 (2000) 5580–5585. effects of interleukin 1 beta in the rat, Am. J. Physiol. 256 (1989) [3] T.Z. Baram, E. Hirsch, O.C.D. Snead, L. Schultz, Corticotropin- E111–E115.
releasing hormone-induced seizures in infant rats originate in the [16] P. Schoeffter, D. Feuerbach, I. Bobirnac, L. Gazi, R. Longato, amygdala, Ann. Neurol. 31 (1992) 488–494. Functional, endogenously expressed corticotropin-releasing factor [4] M. Casadevall, E. Saperas, J. Panes, A. Salas, D.C. Anderson, J.R. receptor type 1 (CRF1) and CRF1 receptor mRNA expression in Malagelada, J.M. Pique, Mechanisms underlying the anti-inflamma- human neuroblastoma SH-SY5Y cells, Fundam. Clin. Pharmacol. 13 tory actions of central corticotropin-releasing factor, Am. J. Physiol. (1999) 484–489.
276 (1999) G1016–G1026. [17] P.J. Strijbos, J.K. Relton, N.J. Rothwell, Corticotrophin-releasing [5] C.L. Ehlers, S.J. Henriksen, M. Wang, J. Rivier, W. Vale, F.E. factor antagonist inhibits neuronal damage induced by focal cerebral Bloom, Corticotropin releasing factor produces increases in brain ischaemia or activation of NMDA receptors in the rat brain, Brain excitability and convulsive seizures in rats, Brain Res. 278 (1983) Res. 656 (1994) 405–408.
332–336. [18] A.V. Turnbull, C. Rivier, Regulation of the HPA axis by cytokines, [6] L.A. Fisher, G. Jessen, M.R. Brown, Corticotropin-releasing factor Brain Behav. Immun. 9 (1995) 253–275.
(CRF): mechanism to elevate mean arterial pressure and heart rate, [19] W. Vale, J. Spiess, C. Rivier, J. Rivier, Characterization of a Regul. Pept. 5 (1983) 153–161. 41-residue ovine hypothalamic peptide that stimulates secretion of [7] J.Y. Koh, D.W. Choi, Quantitative determination of glutamate corticotropin and beta-endorphin, Science 213 (1981) 1394–1397.
mediated cortical neuronal injury in cell culture by lactate dehydro- [20] M.L. Wong, S.A. Loddick, P.B. Bongiorno, P.W. Gold, N.J. Rot-genase efflux assay, J. Neurosci. Meth. 20 (1987) 83–90. hwell, J. Licinio, Focal cerebral ischemia induces CRH mRNA in rat [8] W.A. Kostich, A. Chen, K. Sperle, B.L. Largent, Molecular identifi- cerebral cortex and amygdala, NeuroReport 6 (1995) 1785–1788.
