news and views

did not appear to mediate the analgesic effects ­excitability and neurotransmitter release. It will side effects. However, targeting the anand- of this drug. Thus, the mechanism of drug be interesting to see whether the CB1 receptors amide degrading enzyme may provide more action is via anandamide actions on peripheral responsible for anandamide control of pain act specificity and fewer side effects in com- CB1 receptors. This raises questions regarding at peripheral nerve endings and, if so, what parison to cannabinoid receptor agonists. the relative roles of anandamide and 2-AG in molecular targets and other neurotransmitters Previously, it was noted that endocannabi- peripheral analgesic actions that will no doubt are implicated in these analgesic actions. noids were elevated locally by inflammatory/ be addressed in future experiments, possibly Other recent reports have also highlighted painful stimuli3. Thus, inhibiting FAAH may through the use of peripherally restricted the potential therapeutic usefulness of drugs predominantly affect endocannabinoid/CB1 inhibitors of the 2-AG degrading enzymes targeting the peripheral endocannabinoid sys- signaling in affected regions, avoiding wide- MAGL or ABDH6 (refs. 5,6). It is interest- tem7. For example, CB1 antagonists are known spread effects produced by activation of all ing that elimination of the FAAH enzyme in to reduce weight via effects on eating behavior peripheral cannabinoid receptors. gene-targeted mice did not produce an analge- and peripheral metabolism8,9. Indeed, the CB1 Ultimately, clinical studies of safety and effi- sic effect similar to FAAH inhibitors, whereas antagonist rimonabant that acts both peripher- cacy will be needed to assess the usefulness the antinociceptive actions of URB937 did not ally and centrally was developed for treatment of the peripherally targeted FAAH inhibitor. diminish after 7 d of drug treatment. These of obesity and metabolic syndrome, but its use For now, Clapper et al.4 have found that anan­ findings indicate that compensation for loss was quickly discontinued as a result of side damide participates in important antinocicep- of FAAH activity can occur, but perhaps only effects likely arising from CNS drug actions10, tive actions in the PNS. when the enzyme is out of commission for highlighting the need for peripherally re­stricted COMPETING FINANCIAL INTERESTS prolonged periods or early in development. endocannabinoid-targeted drugs. A recent The author declares no competing financial interests. This bodes well for use of peripheral FAAH study11 found that a peripherally restricted inhibitors in pain management, at least with CB1 antagonist reduced untoward metabolic 1. Calignano, A., La Rana, G., Giuffrida, A. & Piomelli, D. Nature 394, 277–281 (1998). relatively short-term treatment. However, effects of obesity, suggesting another use for 2. Walker, J.M. & Hohmann, A.G. Handb. Exp. Pharmacol. more extensive testing will be necessary to peripherally restricted drugs targeting the 168, 509–554 (2005). determine whether the inhibitor is effective endocannabinoid system. Particularly rel- 3. Agarwal, N. et al. Nat. Neurosci. 10, 870–879 4 (2007). with prolonged exposure. evant to Clapper et al. is a recent preliminary 4. Clapper, J.R. et al. Nat. Neurosci. 13, 1265–1270 Additional studies will also be needed to report that peripherally active cannabinoid (2010). identify the site and mechanisms of the periph- receptor agonists do not reduce acute pain in 5. Marrs, W.R. et al. Nat. Neurosci. 13, 951–957 (2010). eral URB937 analgesic actions. Although the humans with chronic lower back pain, while 6. Kinsey, S.G. et al. J. Pharmacol. Exp. Ther. 330, authors suggest that spinal mechanisms con- some weight gain and metabolic side effects 902–910 (2009). tribute to the antinociceptive effects, the initial were observed12. These findings suggest that 7. Kunos, G., Osei-Hyiaman, D., Bátkai, S., Sharkey, K.A. & Makriyannis, A. Trends Pharmacol. Sci. 30, site of drug action is likely on peripheral nerves widespread activation of peripheral cannabi- 1–7 (2009). and may be at nerve endings in peripheral noid receptors may not be efficacious for pain 8. Di Marzo, V. et al. Nature 410, 822–825 (2001). organs (Fig. 1). Indeed, it was previously found treatment and may have undesirable conse- 9. Engeli, S. et al. Diabetes 54, 2838–2843 (2005). 10. Christensen, R., Kristensen, P.K., Bartels, E.M., that CB1 receptors on the peripheral endings quences. Caution must therefore be exercised Bliddal, H. & Astrup, A. Lancet 370, 1706–1713 of nociceptive sensory neurons mediate anal- in judging the potential safety and efficacy (2007). gesia produced by local or systemic treatment of peripherally targeted FAAH inhibitors, as 11. Tam, J. et al. J. Clin. Invest. 120, 2953–2966 (2010). 3 with systemic CB1 agonists . Activation of prolonged anandamide activation of cannabi- 12. Groblewski, T. et al. Int. Cannabinoid Res. Soc. Abstr. CB1 receptors generally inhibits neuronal noid receptors could still produce unwanted 38 (2010). © 2010 Nature America, Inc. All rights reserved. All rights Inc. America, Nature © 2010 Prime time for stress

Richard Piet & Olivier J Manzoni Stress primes the hypothalamic-pituitary-adrenal axis response to subsequent stressors. A new study finds that acute stress modifies the properties of excitatory synapses impinging on parvocellular neurons of the paraventricular nucleus.

When faced with a perilous or unexpected fight or flight. Responses to stress are mediated, and the adaptation of future behaviors must situation, such as a close encounter with a in part, by the hypothalamic-pituitary-adrenal be managed. Indeed, chronic, as well as acute, hungry Ursus spelaeus or Panthera spelea, our (HPA) axis and a single close encounter with stress is known to induce long-term decreases ­ancestors responded to that stress with either a cave bear or lion would have had protracted (habituation) or increases (sensitization) in the consequences on our Flintstone HPA axes, HPA axis response to the subsequent expo- Richard Piet is in the Centre for Neuroendocrinology resulting in altered sensitivity to future stres- sure to a stressor1–3. Despite its importance and Department of Physiology, University of Otago, sors. Although cave bears and cave lions are in understanding the physiopathology of the Dunedin, New Zealand. Olivier J. Manzoni is at now long extinct, in the modern world, the stress system, the cellular mechanism under- the Institut de Neurobiologie de la Méditerranée HPA axis kicks in when we face challenges lying the ability of a previous stressor to alter INSERM U901, Université de la Méditerranée, Parc in everyday life, as well as more extraordi- the responsiveness to further stressors remains Scientifique de Luminy, Marseille, France. nary circumstances. Versatility in the stress mostly unresolved. e-mail: [email protected] or response of the HPA axis is necessary because A study by Kuzmiski et al.4 sheds new light [email protected] both the immediate response to the stressor on this important question and reveals how

1156 volume 13 | number 10 | October 2010 nature neuroscience news and views

Figure 1 Acute stress primes excitatory synapses NaiveAcute stress in the PVN. Top, proposed mechanism of synaptic priming in the PVN. Excitatory synapses on AMPA receptor parvocellular neurons are under the control of NMDA receptor an inhibitory feedback loop involving NMDA receptors, calcium influx and exocytotic release of an unknown retrograde messenger (messenger Multivesicular x, left). This tonic feedback loop prevents release – these synapses from expressing short-term potentiation of glutamate release in response Messenger x to high-frequency stimulation. Acute stress, ? – via intra-PVN release of CRH acting at CRHR1, Ca2+ ? 2+ may relieve glutamatergic synapses from this + x Ca inhibitory control by decreasing NMDAR function SNARE- CRH dependent (right). This allows glutamatergic synapses to exocytosis CRHR1 CRHR1 undergo synaptic potentiation (which likely involves multivesicular release) and therefore No short-term potentiation Short-term potentiation may increase the output of the PVN and augment Increased the release of CRH, ACTH and glucocorticoids. output Bottom, schematic representation of the response of the HPA axis to stress. Anterior pituitary Adrenal cortex

Brain stem Stress + hypothalamus + CRH synaptic plasticity in the output nucleus of and limbic structures PVN AVPACTH the HPA axis may contribute to this phenom- – Glucocorticoids enon. Kuzmiski et al.4 combined in vivo stress challenges and in vitro patch-clamp electro- physiology and found that acute stress, either a 1 nonsocial emotional stressor (immobilization) HPA axis or a psychogenic stress (predator odor, fox)1, modifies the properties of the excitatory syn- apses impinging on parvocellular neurons rats, the authors found that high-frequency either ­projecting from the bed nucleus of the of the paraventricular nucleus (PVN) of the stimulation of the glutamatergic input to stria terminalis5 or collaterals of parvocellular hypothalamus for several days. To the best parvocellular neurons resulted in the short- neurons’ axons. More work is needed to distin- of our knowledge, this is the first report that term potentiation (STP) of glutamate release guish between these possibilities. synaptic mechanisms in the HPA axis itself (Fig. 1). Notably, the stress-induced synaptic Are synaptic priming and downregulation may underlie the long-term effects of acute priming could still be seen up to 72 h after the of NMDARs independent consequences of stress. In particular, these results may account exposure to stress and vanished after 10 days, acute stress? In an elegant series of experi- for the stress-induced long-term sensitization a time course that is consistent with HPA axis ments conducted in slices from naive rats, of the HPA axis response to stress1. sensitization. This observation argues in favor Kuzmiski et al.4 discovered a causal relation- Activation of the HPA axis by stress results of synaptic priming as a neural substrate for ship between the two phenomena, finding that © 2010 Nature America, Inc. All rights reserved. All rights Inc. America, Nature © 2010 in the release of corticotrophin-releasing HPA axis sensitization. the intracellular blockade of NMDARs in a hormone (CRH) and vasopressin (AVP) Investigation of the induction mechanisms single parvocellular neuron unmasks STP and from parvocellular neurosecretory cells of stress-induced priming of synaptic trans- occludes additional CRH priming. Further of the PVN. CRH stimulates the release of mission in the PVN revealed that CRH, which experiments revealed that NMDARs are adrenocorticotropic hormone from ante- can be released in the PVN5 and robustly involved in an endogenous inhibitory feedback rior pituitary cells, which in turn stimulates enhances synaptic efficacy and plasticity in loop involving calcium influx and somato­ the production of glucocorticoids from the the hippocampus7,8, is important. First, direct dendritic exocytosis of an as yet unidentified adrenal gland. Glucocorticoids act both in treatment of naive slices with CRH was suffi- retrograde messenger (the usual suspects10, the periphery, to mobilize energy, and in the cient to depress NMDAR currents and induce endocannabinoids, adenosine and opioids, brain, where they mediate a negative feed- synaptic priming via corticotrophin type 1 were swiftly and convincingly excluded) that back that eventually shuts down the stress receptors (CRHR1). Moreover, the injection controls glutamate release in the PVN (Fig. 1). response (Fig. 1, see refs. 1–3,5 for in-depth of a CRHR1 antagonist before the exposure to Stress may therefore induce synaptic priming reviews). Kuzmiski et al.4 report that stress stress prevented synaptic priming. Although by inhibiting NMDARs, thereby unmasking was associated with two changes in gluta- these results suggest that the local release of the ability of excitatory synapses in the PVN matergic synaptic transmission in the PVN. CRH in the PVN is responsible for the stress- to undergo synaptic plasticity. Although ques- First, they observed a downregulation of induced effects on synaptic transmission in the tions regarding the identity of the retrograde NMDA receptors, a subtype of ionotropic PVN, the origin of CRH remains obscure. Two messenger, the cellular origin of CRH and how glutamate receptors that participate in syn- possible routes can be imagined. Parvocellular CRHR1 inhibits NMDARs remain unanswered, aptic learning and memory throughout the neurons could somatodendritically release these finding open new avenues of investigation CNS6. Second, the authors found that stress- CRH locally, as is the case for other neuro- into the potential modulation of this synaptic primed glutamatergic synapses express a peptides in the hypothalamus9. Alternatively, priming by multiple stressors, chronic stress, form of synaptic plasti­city that is not seen in but not exclusively, CRH could be released glucocorticoid feedback, aging and neuropsy- naive rats. In slices obtained from stressed from CRH-containing axons in the PVN, chiatric disorders3.

nature neuroscience volume 13 | number 10 | OCTOBER 2010 1157 news and views

The results of Kuzmiski et al.4 are impor- synaptic priming has the potential to affect the different synaptic adaptations. Indeed, deci- tant because they link basic synaptic plasticity entire functional repertoire of the HPA axis, phering the protracted adaptive regulation of mechanisms to whole-organism physiology assuming that the majority of the excitatory the stress response is crucial to understanding processes that we may experience in our daily synapses on parvocellular neurons are under the role of stress in the etiology of major stress- life. Although the idea is well accepted that the control of the mysterious retrograde mes- related neuropsychiatric diseases such as drug experience-­dependent plasticity of NMDAR senger (Fig. 1). Alternatively, if only a subset addiction14, depression and post-traumatic is central to the dynamic control of synaptic of glutamate afferents is sensitive to retrograde stress disorder1,3,15. Multiple neuronal circuits functions6, there still is a big gap between plasticity, then one expects stress-induced and stress mediators orchestrate the ‘neuro- the elucidation of the versatile mechanisms priming to displace the balance toward a par- symphony of stress’2, and by introducing new mediating synaptic plasticity in vitro and the ticular set of neuroendocrine, synaptic and players to the band, Kuzmiski et al.4 substan- ­realization that these mechanisms may partici- behavioral responses. Resolving these issues tially extends the repertoire of the orchestra. pate in a physiological behavioral response. For will first necessitate drawing a clear picture of 4 COMPETING FINANCIAL INTERESTS example, Kuzmiski et al. show that although the specific sources of the glutamatergic inner- The authors declare no competing financial interests. stress-induced priming involves the long-term vation of PVN parvocellular neurons (such as depression of postsynaptic NMDARs, the STP the dorso-medial hypothalamic nucleus and 1. Armario, A., Escorihuela, R.M. & Nadal, R. Neurosci. 1–3,5 Biobehav. Rev. 32, 1121–1135 (2008). that it unmasks is instead expressed presynapti- the bed nucleus of the stria terminalis) . 2. Joëls, M. & Baram, T.Z. Nat. Rev. Neurosci. 10, 459–466 cally and is mediated, at least in part, by multi­ A related issue in need of further investigation is (2009). vesicular glutamate release (Fig. 1). Although the modulation of synaptic priming in the PVN 3. Lupien, S.J., McEwen, B.S., Gunnar, M.R. & Heim, C. Nat. Rev. Neurosci. 10, 434–445 (2009). multivesicular release has been increasingly by other stress mediators, such as monoamines, 4. Kuzmiski, J.B., Marty, V., Baimhoukhametova, D.V. & observed in short- and long-term potentiation neuropeptides and steroids. These other media- Bains, J.S. Nat. Neurosci. 13, 1257–1267 (2010). (see ref. 11), this is the first report, to the best of tors can potentially modulate synaptic plasticity, 5. Ulrich-Lai, Y.M. & Herman, J.P. Nat. Rev. Neurosci. 10, 397–409 (2009). our knowledge, that this atypical phenomenon and precise interactions among them are neces- 6. Lau, C.G. & Zukin, R.S. Nat. Rev. Neurosci. 8, 413– 2 occurs in a physiological context. Similarly, this sary to achieve the appropriate stress response . 426 (2007). report provides a new physiological context for The advent of optogenetic approaches allowing 7. Schierloh, A., Deussing, J., Wurst, W., Zieglgansberger, W. & Rammes, G. Neurosci. Lett. 416, 82–86 (2007). postsynaptic vesicular release, a phenomenon targeted stimulation of precise neuronal net- 8. Sheng, H. et al. Endocrinology 149, 1389–1398 that was shown to participate in hippocampal works in specific brain areas may help clarify (2008). long-term plasticity over a decade ago12. By the exact circuitry at work. 9. Ludwig, M. & Leng, G. Nat. Rev. Neurosci. 7, 126–136 showing that ­NMDAR-dependent ­exocytosis Finally, it is important to remember that (2006). 10. Regehr, W.G., Carey, M.R. & Best, A.R. Neuron 63, represses synaptic gain ­independently of stress comes in two different colors. Hans Selye, 154–170 (2009). AMPAR trafficking and desensitization in naive who first put stress in a physiological context, 11. Bender, V.A., Pugh, J.R. & Jahr, C.E. J. Neurosci. 29, 4 10974–10978 (2009). PVN, Kuzmiski et al. expand the ­functions of coined the terms ‘distress’ for negative stress 12. Lledo, P.M., Zhang, X., Sudhof, T.C., Malenka, R.C. & activity-dependent vesicular release beyond (such as punishment, danger) and ‘eustress’ Nicoll, R.A. Science 279, 399–403 (1998). classical views and bring retrograde signaling for positive stress (reward)13. Kuzmiski et al.4 13. Selye, H. Stress Without Distress (New American back into the spotlight10. reveal that two different forms of distress can Library, New York, 1975). 14. Koob, G.F. Neuron 59, 11–34 (2008). Because it occurs in the PVN, the output trigger priming. It is now important to deter- 15. Feder, A., Nestler, E.J. & Charney, D.S. Nat. Rev. structure of the HPA, environment-regulated mine whether eustress triggers similar or Neurosci. 10, 446–457 (2009).

© 2010 Nature America, Inc. All rights reserved. All rights Inc. America, Nature © 2010 It takes all kinds to make a brain

Rachel I Wilson Variation in neuronal properties is often thought of as noise that interferes with information processing. A study now suggests that neuronal diversity may actually improve the coding capacity of neural ensembles.

As neuroscientists, we sometimes wish our tend to feel that variation is merely a result of In this issue of Nature Neuroscience, data looked a bit tidier than it actually does. For Mother Nature’s poor quality control. Padmanabhan and Urban5 show us another example, we tend to report our measurements However, variation in the nervous system reason why variation isn’t intrinsically bad. as a mean plus or minus error, but many of us isn’t necessarily a bad thing. In an evolving Specifically, they found that variation among secretly yearn for small error bars. When we population, variation among the nervous neurons of the same type increases the ­coding measure the same variable from many ­neurons systems of different organisms is part of the capacity of neural ensembles (we define neu- of the same type (even when our notion of a diversity that natural selection acts on1,2. In a rons of the same type as being neurons that ‘type’ is fuzzy), we suspect this variable should developing organism, variation among neurons carry approximately the same signal). To get really have a fixed value. In other words, we competing for territory and survival may help an intuition for why this should be so, consider to ensure that the winners are fit3. Finally, the following problem. You are trying to learn The author is in the Department of Neurobiology some variation may simply be neutral. If the plot of a movie you haven’t seen based on and the Howard Hughes Medical Institute, Harvard ­variable neurons can combine in many ways conversations with several friends. All of the Medical School, Boston, Massachusetts, USA. to produce adequately functional circuits, then friends saw the same movie (the same signal), e-mail: [email protected] there is no disadvantage to this variability4. but each friend is attuned to something

1158 volume 13 | number 10 | October 2010 nature neuroscience